release 9.2

also version 9.1

.github/workflows/all-ci.yml

@@ -10,13 +10,18 @@ jobs:
 
     steps:
       - name: Checkout code
-        uses: actions/checkout@v2
+        uses: actions/checkout@v3
 
-      - name: Install 64tass
-        run: sudo apt-get update -y && sudo apt-get install -y 64tass
+      - name: build and install recent 64tass
+        run: |
+          sudo apt-get install -y make build-essential
+          git clone --depth=1 https://github.com/irmen/64tass
+          cd 64tass
+          make -j4
+          sudo make install
 
       - name: Set up JDK 11
-        uses: actions/setup-java@v2
+        uses: actions/setup-java@v3
         with:
           java-version: 11
           distribution: adopt

.gitignore

@@ -15,6 +15,7 @@ out/
 parser/**/*.interp
 parser/**/*.tokens
 parser/**/*.java
+compiler/src/prog8/buildversion/*
 *.py[cod]
 *.egg
 *.egg-info
@@ -29,6 +30,8 @@ parsetab.py
 compiler/lib/
 
 .gradle
+**/BuildVersion.kt
 /prog8compiler.jar
 sd*.img
 *.d64
+

.idea/kotlinc.xml

@@ -4,6 +4,6 @@
     <option name="jvmTarget" value="11" />
   </component>
   <component name="KotlinJpsPluginSettings">
-    <option name="version" value="1.8.20" />
+    <option name="version" value="1.9.0-release-358" />
   </component>
 </project>

.idea/libraries/antlr_antlr4.xml

@@ -1,13 +1,13 @@
 <component name="libraryTable">
   <library name="antlr.antlr4" type="repository">
-    <properties maven-id="org.antlr:antlr4:4.12.0">
+    <properties maven-id="org.antlr:antlr4:4.13.0">
       <exclude>
         <dependency maven-id="com.ibm.icu:icu4j" />
       </exclude>
     </properties>
     <CLASSES>
-      <root url="jar://$MAVEN_REPOSITORY$/org/antlr/antlr4/4.12.0/antlr4-4.12.0.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/antlr/antlr4-runtime/4.12.0/antlr4-runtime-4.12.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/antlr/antlr4/4.13.0/antlr4-4.13.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/antlr/antlr4-runtime/4.13.0/antlr4-runtime-4.13.0.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/org/antlr/antlr-runtime/3.5.3/antlr-runtime-3.5.3.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/org/antlr/ST4/4.3.4/ST4-4.3.4.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/org/abego/treelayout/org.abego.treelayout.core/1.0.3/org.abego.treelayout.core-1.0.3.jar!/" />

.idea/libraries/io_kotest_assertions_core_jvm.xml

@@ -1,21 +1,21 @@
 <component name="libraryTable">
   <library name="io.kotest.assertions.core.jvm" type="repository">
-    <properties maven-id="io.kotest:kotest-assertions-core-jvm:5.5.5" />
+    <properties maven-id="io.kotest:kotest-assertions-core-jvm:5.6.2" />
     <CLASSES>
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-core-jvm/5.5.5/kotest-assertions-core-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.6.21/kotlin-stdlib-jdk8-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.6.21/kotlin-stdlib-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/annotations/13.0/annotations-13.0.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.6.21/kotlin-stdlib-jdk7-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-shared-jvm/5.5.5/kotest-assertions-shared-jvm-5.5.5.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-core-jvm/5.6.2/kotest-assertions-core-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.8.10/kotlin-stdlib-jdk8-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.8.10/kotlin-stdlib-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.8.10/kotlin-stdlib-jdk7-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-shared-jvm/5.6.2/kotest-assertions-shared-jvm-5.6.2.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/io/github/java-diff-utils/java-diff-utils/4.12/java-diff-utils-4.12.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/org/opentest4j/opentest4j/1.2.0/opentest4j-1.2.0.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-common/1.6.21/kotlin-stdlib-common-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-jdk8/1.6.4/kotlinx-coroutines-jdk8-1.6.4.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-reflect/1.6.21/kotlin-reflect-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-common-jvm/5.5.5/kotest-common-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-api-jvm/5.5.5/kotest-assertions-api-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-core-jvm/1.6.4/kotlinx-coroutines-core-jvm-1.6.4.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-common/1.8.10/kotlin-stdlib-common-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-jdk8/1.7.0/kotlinx-coroutines-jdk8-1.7.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-reflect/1.8.10/kotlin-reflect-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-common-jvm/5.6.2/kotest-common-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-api-jvm/5.6.2/kotest-assertions-api-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-core-jvm/1.7.0/kotlinx-coroutines-core-jvm-1.7.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/annotations/23.0.0/annotations-23.0.0.jar!/" />
     </CLASSES>
     <JAVADOC />
     <SOURCES />

.idea/libraries/io_kotest_property_jvm.xml

@@ -1,22 +1,22 @@
 <component name="libraryTable">
   <library name="io.kotest.property.jvm" type="repository">
-    <properties maven-id="io.kotest:kotest-property-jvm:5.5.5" />
+    <properties maven-id="io.kotest:kotest-property-jvm:5.6.2" />
     <CLASSES>
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-property-jvm/5.5.5/kotest-property-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-common-jvm/5.5.5/kotest-common-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-shared-jvm/5.5.5/kotest-assertions-shared-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-api-jvm/5.5.5/kotest-assertions-api-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-jdk8/1.6.4/kotlinx-coroutines-jdk8-1.6.4.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/opentest4j/opentest4j/1.2.0/opentest4j-1.2.0.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-common/1.6.21/kotlin-stdlib-common-1.6.21.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-property-jvm/5.6.2/kotest-property-jvm-5.6.2.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/com/github/curious-odd-man/rgxgen/1.4/rgxgen-1.4.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.6.21/kotlin-stdlib-jdk8-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.6.21/kotlin-stdlib-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/annotations/13.0/annotations-13.0.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.6.21/kotlin-stdlib-jdk7-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-reflect/1.6.21/kotlin-reflect-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-core-jvm/1.6.4/kotlinx-coroutines-core-jvm-1.6.4.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.8.10/kotlin-stdlib-jdk8-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.8.10/kotlin-stdlib-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.8.10/kotlin-stdlib-jdk7-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-common-jvm/5.6.2/kotest-common-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-shared-jvm/5.6.2/kotest-assertions-shared-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-api-jvm/5.6.2/kotest-assertions-api-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-jdk8/1.7.0/kotlinx-coroutines-jdk8-1.7.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/opentest4j/opentest4j/1.2.0/opentest4j-1.2.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-common/1.8.10/kotlin-stdlib-common-1.8.10.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/io/github/java-diff-utils/java-diff-utils/4.12/java-diff-utils-4.12.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-reflect/1.8.10/kotlin-reflect-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-core-jvm/1.7.0/kotlinx-coroutines-core-jvm-1.7.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/annotations/23.0.0/annotations-23.0.0.jar!/" />
     </CLASSES>
     <JAVADOC />
     <SOURCES />

.idea/libraries/io_kotest_runner_junit5_jvm.xml

@@ -1,55 +1,42 @@
 <component name="libraryTable">
   <library name="io.kotest.runner.junit5.jvm" type="repository">
-    <properties maven-id="io.kotest:kotest-runner-junit5-jvm:5.5.5" />
+    <properties maven-id="io.kotest:kotest-runner-junit5-jvm:5.6.2" />
     <CLASSES>
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-runner-junit5-jvm/5.5.5/kotest-runner-junit5-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-api-jvm/5.5.5/kotest-framework-api-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-test-jvm/1.6.4/kotlinx-coroutines-test-jvm-1.6.4.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-shared-jvm/5.5.5/kotest-assertions-shared-jvm-5.5.5.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-runner-junit5-jvm/5.6.2/kotest-runner-junit5-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-api-jvm/5.6.2/kotest-framework-api-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-shared-jvm/5.6.2/kotest-assertions-shared-jvm-5.6.2.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/io/github/java-diff-utils/java-diff-utils/4.12/java-diff-utils-4.12.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-common/1.6.21/kotlin-stdlib-common-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-common-jvm/5.5.5/kotest-common-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-engine-jvm/5.5.5/kotest-framework-engine-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/github/classgraph/classgraph/4.8.154/classgraph-4.8.154.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-test-jvm/1.7.0/kotlinx-coroutines-test-jvm-1.7.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-common-jvm/5.6.2/kotest-common-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-engine-jvm/5.6.2/kotest-framework-engine-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/github/classgraph/classgraph/4.8.157/classgraph-4.8.157.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/com/github/ajalt/mordant/1.2.1/mordant-1.2.1.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/com/github/ajalt/colormath/1.2.0/colormath-1.2.0.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/org/opentest4j/opentest4j/1.2.0/opentest4j-1.2.0.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-debug/1.6.4/kotlinx-coroutines-debug-1.6.4.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-debug/1.7.0/kotlinx-coroutines-debug-1.7.0.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/net/java/dev/jna/jna/5.9.0/jna-5.9.0.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/net/java/dev/jna/jna-platform/5.9.0/jna-platform-5.9.0.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/net/bytebuddy/byte-buddy/1.10.9/byte-buddy-1.10.9.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/net/bytebuddy/byte-buddy-agent/1.10.9/byte-buddy-agent-1.10.9.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-discovery-jvm/5.5.5/kotest-framework-discovery-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-core-jvm/5.5.5/kotest-assertions-core-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-jdk8/1.6.4/kotlinx-coroutines-jdk8-1.6.4.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-api-jvm/5.5.5/kotest-assertions-api-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-extensions-jvm/5.5.5/kotest-extensions-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/mockk/mockk-jvm/1.13.1/mockk-jvm-1.13.1.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/mockk/mockk-dsl-jvm/1.13.1/mockk-dsl-jvm-1.13.1.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/mockk/mockk-agent-jvm/1.13.1/mockk-agent-jvm-1.13.1.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/objenesis/objenesis/3.2/objenesis-3.2.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/mockk/mockk-agent-api-jvm/1.13.1/mockk-agent-api-jvm-1.13.1.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/mockk/mockk-core-jvm/1.13.1/mockk-core-jvm-1.13.1.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/slf4j/slf4j-api/1.7.36/slf4j-api-1.7.36.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/junit/junit/4.13.2/junit-4.13.2.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/hamcrest/hamcrest-core/1.3/hamcrest-core-1.3.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/junit/jupiter/junit-jupiter/5.8.2/junit-jupiter-5.8.2.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/junit/jupiter/junit-jupiter-params/5.8.2/junit-jupiter-params-5.8.2.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/junit/jupiter/junit-jupiter-engine/5.8.2/junit-jupiter-engine-5.8.2.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-core/1.6.4/kotlinx-coroutines-core-1.6.4.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-concurrency-jvm/5.5.5/kotest-framework-concurrency-jvm-5.5.5.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-core-jvm/1.6.4/kotlinx-coroutines-core-jvm-1.6.4.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/junit/platform/junit-platform-engine/1.7.2/junit-platform-engine-1.7.2.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/apiguardian/apiguardian-api/1.1.0/apiguardian-api-1.1.0.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/junit/platform/junit-platform-commons/1.7.2/junit-platform-commons-1.7.2.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/junit/platform/junit-platform-suite-api/1.7.2/junit-platform-suite-api-1.7.2.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/junit/platform/junit-platform-launcher/1.7.2/junit-platform-launcher-1.7.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-discovery-jvm/5.6.2/kotest-framework-discovery-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-core-jvm/5.6.2/kotest-assertions-core-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-jdk8/1.7.0/kotlinx-coroutines-jdk8-1.7.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-api-jvm/5.6.2/kotest-assertions-api-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-extensions-jvm/5.6.2/kotest-extensions-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-concurrency-jvm/5.6.2/kotest-framework-concurrency-jvm-5.6.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlinx/kotlinx-coroutines-core-jvm/1.7.0/kotlinx-coroutines-core-jvm-1.7.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/annotations/23.0.0/annotations-23.0.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/junit/platform/junit-platform-engine/1.8.2/junit-platform-engine-1.8.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/junit/platform/junit-platform-commons/1.8.2/junit-platform-commons-1.8.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/apiguardian/apiguardian-api/1.1.2/apiguardian-api-1.1.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/junit/platform/junit-platform-suite-api/1.8.2/junit-platform-suite-api-1.8.2.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/junit/platform/junit-platform-launcher/1.8.2/junit-platform-launcher-1.8.2.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/org/junit/jupiter/junit-jupiter-api/5.8.2/junit-jupiter-api-5.8.2.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.6.21/kotlin-stdlib-jdk8-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.6.21/kotlin-stdlib-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/annotations/13.0/annotations-13.0.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.6.21/kotlin-stdlib-jdk7-1.6.21.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-reflect/1.6.21/kotlin-reflect-1.6.21.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.8.10/kotlin-stdlib-jdk8-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.8.10/kotlin-stdlib-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.8.10/kotlin-stdlib-jdk7-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-common/1.8.10/kotlin-stdlib-common-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-reflect/1.8.10/kotlin-reflect-1.8.10.jar!/" />
     </CLASSES>
     <JAVADOC />
     <SOURCES />

.idea/libraries/michael_bull_kotlin_result_jvm.xml

@@ -1,13 +1,13 @@
 <component name="libraryTable">
   <library name="michael.bull.kotlin.result.jvm" type="repository">
-    <properties maven-id="com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.16" />
+    <properties maven-id="com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18" />
     <CLASSES>
-      <root url="jar://$MAVEN_REPOSITORY$/com/michael-bull/kotlin-result/kotlin-result-jvm/1.1.16/kotlin-result-jvm-1.1.16.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-common/1.6.20/kotlin-stdlib-common-1.6.20.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.6.20/kotlin-stdlib-jdk8-1.6.20.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.6.20/kotlin-stdlib-1.6.20.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/com/michael-bull/kotlin-result/kotlin-result-jvm/1.1.18/kotlin-result-jvm-1.1.18.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.8.10/kotlin-stdlib-jdk8-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.8.10/kotlin-stdlib-1.8.10.jar!/" />
       <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/annotations/13.0/annotations-13.0.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.6.20/kotlin-stdlib-jdk7-1.6.20.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.8.10/kotlin-stdlib-jdk7-1.8.10.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-common/1.8.10/kotlin-stdlib-common-1.8.10.jar!/" />
     </CLASSES>
     <JAVADOC />
     <SOURCES />

README.md

@@ -14,6 +14,21 @@ Documentation
 Full documentation (syntax reference, how to use the language and the compiler, etc.) can be found at:
 https://prog8.readthedocs.io/
 
+How to get it/build it
+----------------------
+
+- Download the latest [official release](https://github.com/irmen/prog8/releases) from github.
+- Or, if you want/need a bleeding edge development version, you can:
+  - download a build artifact zipfile from a recent [github action build](https://github.com/irmen/prog8/actions).
+  - you can also compile it yourself from source. [Instructions here](https://prog8.readthedocs.io/en/latest/compiling.html).
+
+Community
+---------
+Most of the development on Prog8 and the use of it is currently centered around 
+the [Commander X16](https://www.commanderx16.com/) retro computer. Their [discord server](https://discord.gg/nS2PqEC) contains a small channel
+dedicated to Prog8. Other than that, use the issue tracker on github.
+
+
 Software license
 ----------------
 GNU GPL 3.0 (see file LICENSE), with exception for generated code:
@@ -76,7 +91,9 @@ IntelliJ IDEA with the Kotlin plugin).
 
 It's handy to have an emulator (or a real machine perhaps!) to run the programs on. The compiler assumes the presence
 of the [Vice emulator](http://vice-emu.sourceforge.net/)  for the C64 target,
-and the [x16emu emulator](https://github.com/commanderx16/x16-emulator) for the CommanderX16 target.
+and a recent emulator version (R42 or newer) for the CommanderX16, such as [x16emu](https://cx16forum.com/forum/viewforum.php?f=30) 
+(preferred, this is the official emulator. If required, source code is [here](https://github.com/X16Community/x16-emulator/)).
+There is also [Box16](https://github.com/indigodarkwolf/box16) which has powerful debugging features.
 
 **Syntax highlighting:** for a few different editors, syntax highlighting definition files are provided.
 Look in the [syntax-files](https://github.com/irmen/prog8/tree/master/syntax-files) directory in the github repository to find them.

codeCore/build.gradle

@@ -26,7 +26,7 @@ compileTestKotlin {
 dependencies {
     // should have no dependencies to other modules
     implementation "org.jetbrains.kotlin:kotlin-stdlib-jdk8"
-    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.16"
+    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18"
 }
 
 sourceSets {

codeCore/src/prog8/code/SymbolTable.kt

@@ -80,6 +80,16 @@ class SymbolTable(astProgram: PtProgram) : StNode(astProgram.name, StNodeType.GL
     }
 
     override fun lookup(scopedName: String) = flat[scopedName]
+
+    fun getLength(name: String): Int? {
+        val node = flat[name]
+        return when(node) {
+            is StMemVar -> node.length
+            is StMemorySlab -> node.size.toInt()
+            is StStaticVariable -> node.length
+            else -> null
+        }
+    }
 }
 
 
@@ -235,7 +245,7 @@ class StSub(name: String, val parameters: List<StSubroutineParameter>, val retur
 
 
 class StRomSub(name: String,
-               val address: UInt,
+               val address: UInt?,      // null in case of asmsub, specified in case of romsub
                val parameters: List<StRomSubParameter>,
                val returns: List<StRomSubParameter>,
                astNode: PtNode) :

codeCore/src/prog8/code/SymbolTableMaker.kt

@@ -40,14 +40,9 @@ class SymbolTableMaker(private val program: PtProgram, private val options: Comp
     private fun addToSt(node: PtNode, scope: Stack<StNode>) {
         val stNode = when(node) {
             is PtAsmSub -> {
-                if(node.address==null) {
-                    val params = node.parameters.map { StSubroutineParameter(it.second.name, it.second.type) }
-                    StSub(node.name, params, node.returns.singleOrNull()?.second, node)
-                } else {
-                    val parameters = node.parameters.map { StRomSubParameter(it.first, it.second.type) }
-                    val returns = node.returns.map { StRomSubParameter(it.first, it.second) }
-                    StRomSub(node.name, node.address, parameters, returns, node)
-                }
+                val parameters = node.parameters.map { StRomSubParameter(it.first, it.second.type) }
+                val returns = node.returns.map { StRomSubParameter(it.first, it.second) }
+                StRomSub(node.name, node.address, parameters, returns, node)
             }
             is PtBlock -> {
                 StNode(node.name, StNodeType.BLOCK, node)
@@ -99,6 +94,9 @@ class SymbolTableMaker(private val program: PtProgram, private val options: Comp
                     initialString = null
                     numElements = node.arraySize?.toInt()
                 }
+//                if(node.type in SplitWordArrayTypes) {
+//                    ... split array also add _lsb and _msb to symboltable?
+//                }
                 StStaticVariable(node.name, node.type, initialNumeric, initialString, initialArray, numElements, node.zeropage, node)
             }
             is PtBuiltinFunctionCall -> {

codeCore/src/prog8/code/ast/AstBase.kt

@@ -37,16 +37,17 @@ class PtNodeGroup : PtNode(Position.DUMMY)
 sealed class PtNamedNode(var name: String, position: Position): PtNode(position) {
     // Note that as an exception, the 'name' is not read-only
     // but a var. This is to allow for cheap node renames.
-    val scopedName: String by lazy {
-        var namedParent: PtNode = this.parent
-        if(namedParent is PtProgram)
-            name
-        else {
-            while (namedParent !is PtNamedNode)
-                namedParent = namedParent.parent
-            namedParent.scopedName + "." + name
+    val scopedName: String
+        get() {
+            var namedParent: PtNode = this.parent
+            return if(namedParent is PtProgram)
+                name
+            else {
+                while (namedParent !is PtNamedNode)
+                    namedParent = namedParent.parent
+                namedParent.scopedName + "." + name
+            }
         }
-    }
 }
 
 
@@ -63,7 +64,9 @@ class PtProgram(
         children.asSequence().filterIsInstance<PtBlock>()
 
     fun entrypoint(): PtSub? =
-        allBlocks().firstOrNull { it.name == "main" }?.children?.firstOrNull { it is PtSub && it.name == "start" } as PtSub?
+        allBlocks().firstOrNull { it.name == "main" || it.name=="p8_main" }
+            ?.children
+            ?.firstOrNull { it is PtSub && (it.name == "start" || it.name=="main.start" || it.name=="p8_start" || it.name=="p8_main.p8_start") } as PtSub?
 }
 
 
@@ -71,6 +74,7 @@ class PtBlock(name: String,
               val address: UInt?,
               val library: Boolean,
               val forceOutput: Boolean,
+              val noSymbolPrefixing: Boolean,
               val alignment: BlockAlignment,
               val source: SourceCode,       // taken from the module the block is defined in.
               position: Position

codeCore/src/prog8/code/ast/AstExpressions.kt

@@ -25,7 +25,7 @@ sealed class PtExpression(val type: DataType, position: Position) : PtNode(posit
     infix fun isSameAs(other: PtExpression): Boolean {
         return when(this) {
             is PtAddressOf -> other is PtAddressOf && other.type==type && other.identifier isSameAs identifier
-            is PtArrayIndexer -> other is PtArrayIndexer && other.type==type && other.variable isSameAs variable && other.index isSameAs index
+            is PtArrayIndexer -> other is PtArrayIndexer && other.type==type && other.variable isSameAs variable && other.index isSameAs index && other.splitWords==splitWords
             is PtBinaryExpression -> other is PtBinaryExpression && other.left isSameAs left && other.right isSameAs right
             is PtContainmentCheck -> other is PtContainmentCheck && other.type==type && other.element isSameAs element && other.iterable isSameAs iterable
             is PtIdentifier -> other is PtIdentifier && other.type==type && other.name==name
@@ -48,7 +48,7 @@ sealed class PtExpression(val type: DataType, position: Position) : PtNode(posit
                 this.name == target.identifier!!.name
             }
             target.array != null && this is PtArrayIndexer -> {
-                this.variable.name == target.array!!.variable.name && this.index isSameAs target.array!!.index
+                this.variable.name == target.array!!.variable.name && this.index isSameAs target.array!!.index && this.splitWords==target.array!!.splitWords
             }
             else -> false
         }
@@ -115,6 +115,9 @@ class PtArrayIndexer(elementType: DataType, position: Position): PtExpression(el
     val index: PtExpression
         get() = children[1] as PtExpression
 
+    val splitWords: Boolean
+        get() = variable.type in SplitWordArrayTypes
+
     init {
         require(elementType in NumericDatatypes)
     }

codeCore/src/prog8/code/ast/AstPrinter.kt

@@ -17,7 +17,7 @@ fun printAst(root: PtNode, skipLibraries: Boolean, output: (text: String) -> Uni
             is PtConditionalBranch -> "if_${node.condition.name.lowercase()}"
             is PtAddressOf -> "&"
             is PtArray -> "array len=${node.children.size} ${type(node.type)}"
-            is PtArrayIndexer -> "<arrayindexer> ${type(node.type)}"
+            is PtArrayIndexer -> "<arrayindexer> ${type(node.type)} ${if(node.splitWords) "[splitwords]" else ""}"
             is PtBinaryExpression -> "<expr> ${node.operator} ${type(node.type)}"
             is PtBuiltinFunctionCall -> {
                 val str = if(node.void) "void " else ""
@@ -96,13 +96,14 @@ fun printAst(root: PtNode, skipLibraries: Boolean, output: (text: String) -> Uni
                 str
             }
             is PtVariable -> {
+                val split = if(node.type in SplitWordArrayTypes) "@split" else ""
                 val str = if(node.arraySize!=null) {
                     val eltType = ArrayToElementTypes.getValue(node.type)
-                    "${eltType.name.lowercase()}[${node.arraySize}] ${node.name}"
+                    "${eltType.name.lowercase()}[${node.arraySize}] $split ${node.name}"
                 }
                 else if(node.type in ArrayDatatypes) {
                     val eltType = ArrayToElementTypes.getValue(node.type)
-                    "${eltType.name.lowercase()}[] ${node.name}"
+                    "${eltType.name.lowercase()}[] $split ${node.name}"
                 }
                 else
                     "${node.type.name.lowercase()} ${node.name}"

codeCore/src/prog8/code/core/BuiltinFunctions.kt

@@ -68,7 +68,6 @@ class FSignature(val pure: Boolean,      // does it have side effects?
     }
 }
 
-
 val BuiltinFunctions: Map<String, FSignature> = mapOf(
     // this set of function have no return value and operate in-place:
     "rol"       to FSignature(false, listOf(FParam("item", arrayOf(DataType.UBYTE, DataType.UWORD))), null),
@@ -80,19 +79,41 @@ val BuiltinFunctions: Map<String, FSignature> = mapOf(
     // cmp returns a status in the carry flag, but not a proper return value
     "cmp"       to FSignature(false, listOf(FParam("value1", IntegerDatatypesNoBool), FParam("value2", NumericDatatypesNoBool)), null),
     "prog8_lib_stringcompare"       to FSignature(true, listOf(FParam("str1", arrayOf(DataType.STR)), FParam("str2", arrayOf(DataType.STR))), DataType.BYTE),
-    "abs"       to FSignature(true, listOf(FParam("value", IntegerDatatypesNoBool)), DataType.UWORD),
+    "abs"       to FSignature(true, listOf(FParam("value", NumericDatatypesNoBool)), null),
+    "abs__byte"  to FSignature(true, listOf(FParam("value", arrayOf(DataType.BYTE))), DataType.BYTE),
+    "abs__word"  to FSignature(true, listOf(FParam("value", arrayOf(DataType.WORD))), DataType.WORD),
+    "abs__float" to FSignature(true, listOf(FParam("value", arrayOf(DataType.FLOAT))), DataType.FLOAT),
     "len"       to FSignature(true, listOf(FParam("values", IterableDatatypes)), DataType.UWORD),
     // normal functions follow:
-    "sizeof"    to FSignature(true, listOf(FParam("object", DataType.values())), DataType.UBYTE),
+    "sizeof"    to FSignature(true, listOf(FParam("object", DataType.entries.toTypedArray())), DataType.UBYTE),
     "sgn"       to FSignature(true, listOf(FParam("value", NumericDatatypesNoBool)), DataType.BYTE),
-    "sqrt16"    to FSignature(true, listOf(FParam("value", arrayOf(DataType.UWORD))), DataType.UBYTE),
-    "divmod"    to FSignature(false, listOf(FParam("number", arrayOf(DataType.UBYTE)), FParam("divident", arrayOf(DataType.UBYTE)), FParam("division", arrayOf(DataType.UBYTE)), FParam("remainder", arrayOf(DataType.UBYTE))), null),
-    "divmodw"   to FSignature(false, listOf(FParam("number", arrayOf(DataType.UWORD)), FParam("divident", arrayOf(DataType.UWORD)), FParam("division", arrayOf(DataType.UWORD)), FParam("remainder", arrayOf(DataType.UWORD))), null),
+    "sqrt"        to FSignature(true, listOf(FParam("value", NumericDatatypesNoBool)), null),
+    "sqrt__ubyte" to FSignature(true, listOf(FParam("value", arrayOf(DataType.UBYTE))), DataType.UBYTE),
+    "sqrt__uword" to FSignature(true, listOf(FParam("value", arrayOf(DataType.UWORD))), DataType.UBYTE),
+    "sqrt__float" to FSignature(true, listOf(FParam("value", arrayOf(DataType.FLOAT))), DataType.FLOAT),
+    "divmod"           to FSignature(false, listOf(FParam("number", arrayOf(DataType.UBYTE, DataType.UWORD)), FParam("divident", arrayOf(DataType.UBYTE, DataType.UWORD)), FParam("division", arrayOf(DataType.UBYTE, DataType.UWORD)), FParam("remainder", arrayOf(DataType.UBYTE, DataType.UWORD))), null),
+    "divmod__ubyte"    to FSignature(false, listOf(FParam("number", arrayOf(DataType.UBYTE)), FParam("divident", arrayOf(DataType.UBYTE)), FParam("division", arrayOf(DataType.UBYTE)), FParam("remainder", arrayOf(DataType.UBYTE))), null),
+    "divmod__uword"    to FSignature(false, listOf(FParam("number", arrayOf(DataType.UWORD)), FParam("divident", arrayOf(DataType.UWORD)), FParam("division", arrayOf(DataType.UWORD)), FParam("remainder", arrayOf(DataType.UWORD))), null),
     "any"       to FSignature(true, listOf(FParam("values", ArrayDatatypes)), DataType.UBYTE),
     "all"       to FSignature(true, listOf(FParam("values", ArrayDatatypes)), DataType.UBYTE),
     "lsb"       to FSignature(true, listOf(FParam("value", arrayOf(DataType.UWORD, DataType.WORD))), DataType.UBYTE),
     "msb"       to FSignature(true, listOf(FParam("value", arrayOf(DataType.UWORD, DataType.WORD))), DataType.UBYTE),
     "mkword"    to FSignature(true, listOf(FParam("msb", arrayOf(DataType.UBYTE)), FParam("lsb", arrayOf(DataType.UBYTE))), DataType.UWORD),
+    "clamp"     to FSignature(true, listOf(FParam("value", arrayOf(DataType.BYTE)), FParam("minimum", arrayOf(DataType.BYTE)), FParam("maximum", arrayOf(DataType.BYTE))), null),
+    "clamp__byte"  to FSignature(true, listOf(FParam("value", arrayOf(DataType.BYTE)), FParam("minimum", arrayOf(DataType.BYTE)), FParam("maximum", arrayOf(DataType.BYTE))), DataType.BYTE),
+    "clamp__ubyte" to FSignature(true, listOf(FParam("value", arrayOf(DataType.UBYTE)), FParam("minimum", arrayOf(DataType.UBYTE)), FParam("maximum", arrayOf(DataType.UBYTE))), DataType.UBYTE),
+    "clamp__word"  to FSignature(true, listOf(FParam("value", arrayOf(DataType.WORD)), FParam("minimum", arrayOf(DataType.WORD)), FParam("maximum", arrayOf(DataType.WORD))), DataType.WORD),
+    "clamp__uword" to FSignature(true, listOf(FParam("value", arrayOf(DataType.UWORD)), FParam("minimum", arrayOf(DataType.UWORD)), FParam("maximum", arrayOf(DataType.UWORD))), DataType.UWORD),
+    "min"        to FSignature(true, listOf(FParam("val1", arrayOf(DataType.BYTE)), FParam("val2", arrayOf(DataType.BYTE))), null),
+    "min__byte"  to FSignature(true, listOf(FParam("val1", arrayOf(DataType.BYTE)), FParam("val2", arrayOf(DataType.BYTE))), DataType.BYTE),
+    "min__ubyte" to FSignature(true, listOf(FParam("val1", arrayOf(DataType.UBYTE)), FParam("val2", arrayOf(DataType.UBYTE))), DataType.UBYTE),
+    "min__word"  to FSignature(true, listOf(FParam("val1", arrayOf(DataType.WORD)), FParam("val2", arrayOf(DataType.WORD))), DataType.WORD),
+    "min__uword" to FSignature(true, listOf(FParam("val1", arrayOf(DataType.UWORD)), FParam("val2", arrayOf(DataType.UWORD))), DataType.UWORD),
+    "max"        to FSignature(true, listOf(FParam("val1", arrayOf(DataType.BYTE)), FParam("val2", arrayOf(DataType.BYTE))), null),
+    "max__byte"  to FSignature(true, listOf(FParam("val1", arrayOf(DataType.BYTE)), FParam("val2", arrayOf(DataType.BYTE))), DataType.BYTE),
+    "max__ubyte" to FSignature(true, listOf(FParam("val1", arrayOf(DataType.UBYTE)), FParam("val2", arrayOf(DataType.UBYTE))), DataType.UBYTE),
+    "max__word"  to FSignature(true, listOf(FParam("val1", arrayOf(DataType.WORD)), FParam("val2", arrayOf(DataType.WORD))), DataType.WORD),
+    "max__uword" to FSignature(true, listOf(FParam("val1", arrayOf(DataType.UWORD)), FParam("val2", arrayOf(DataType.UWORD))), DataType.UWORD),
     "peek"      to FSignature(true, listOf(FParam("address", arrayOf(DataType.UWORD))), DataType.UBYTE),
     "peekw"     to FSignature(true, listOf(FParam("address", arrayOf(DataType.UWORD))), DataType.UWORD),
     "poke"      to FSignature(false, listOf(FParam("address", arrayOf(DataType.UWORD)), FParam("value", arrayOf(DataType.UBYTE))), null),

codeCore/src/prog8/code/core/CompilationOptions.kt

@@ -19,8 +19,8 @@ class CompilationOptions(val output: OutputType,
                          var asmQuiet: Boolean = false,
                          var asmListfile: Boolean = false,
                          var experimentalCodegen: Boolean = false,
-                         var varsHigh: Boolean = false,
-                         var useNewExprCode: Boolean = false,
+                         var varsHighBank: Int? = null,
+                         var splitWordArrays: Boolean = false,
                          var evalStackBaseAddress: UInt? = null,
                          var outputDir: Path = Path(""),
                          var symbolDefs: Map<String, String> = emptyMap()

codeCore/src/prog8/code/core/Enumerations.kt

@@ -11,7 +11,9 @@ enum class DataType {
     ARRAY_UB,           // pass by reference
     ARRAY_B,            // pass by reference
     ARRAY_UW,           // pass by reference
+    ARRAY_UW_SPLIT,     // pass by reference, lo/hi byte split
     ARRAY_W,            // pass by reference
+    ARRAY_W_SPLIT,      // pass by reference, lo/hi byte split
     ARRAY_F,            // pass by reference
     ARRAY_BOOL,         // pass by reference
     UNDEFINED;
@@ -119,12 +121,14 @@ val IntegerDatatypesNoBool = arrayOf(DataType.UBYTE, DataType.BYTE, DataType.UWO
 val NumericDatatypes = arrayOf(DataType.UBYTE, DataType.BYTE, DataType.UWORD, DataType.WORD, DataType.FLOAT, DataType.BOOL)
 val NumericDatatypesNoBool = arrayOf(DataType.UBYTE, DataType.BYTE, DataType.UWORD, DataType.WORD, DataType.FLOAT)
 val SignedDatatypes =  arrayOf(DataType.BYTE, DataType.WORD, DataType.FLOAT)
-val ArrayDatatypes = arrayOf(DataType.ARRAY_UB, DataType.ARRAY_B, DataType.ARRAY_UW, DataType.ARRAY_W, DataType.ARRAY_F, DataType.ARRAY_BOOL)
+val ArrayDatatypes = arrayOf(DataType.ARRAY_UB, DataType.ARRAY_B, DataType.ARRAY_UW, DataType.ARRAY_UW_SPLIT, DataType.ARRAY_W, DataType.ARRAY_W_SPLIT, DataType.ARRAY_F, DataType.ARRAY_BOOL)
 val StringlyDatatypes = arrayOf(DataType.STR, DataType.ARRAY_UB, DataType.ARRAY_B, DataType.UWORD)
+val SplitWordArrayTypes = arrayOf(DataType.ARRAY_UW_SPLIT, DataType.ARRAY_W_SPLIT)
 val IterableDatatypes = arrayOf(
     DataType.STR,
     DataType.ARRAY_UB, DataType.ARRAY_B,
     DataType.ARRAY_UW, DataType.ARRAY_W,
+    DataType.ARRAY_UW_SPLIT, DataType.ARRAY_W_SPLIT,
     DataType.ARRAY_F, DataType.ARRAY_BOOL
 )
 val PassByValueDatatypes = NumericDatatypes
@@ -135,6 +139,8 @@ val ArrayToElementTypes = mapOf(
     DataType.ARRAY_UB to DataType.UBYTE,
     DataType.ARRAY_W to DataType.WORD,
     DataType.ARRAY_UW to DataType.UWORD,
+    DataType.ARRAY_W_SPLIT to DataType.WORD,
+    DataType.ARRAY_UW_SPLIT to DataType.UWORD,
     DataType.ARRAY_F to DataType.FLOAT,
     DataType.ARRAY_BOOL to DataType.BOOL
 )

codeCore/src/prog8/code/core/IErrorReporter.kt

@@ -3,6 +3,7 @@ package prog8.code.core
 interface IErrorReporter {
     fun err(msg: String, position: Position)
     fun warn(msg: String, position: Position)
+    fun undefined(symbol: List<String>, position: Position)
     fun noErrors(): Boolean
     fun report()
     fun finalizeNumErrors(numErrors: Int, numWarnings: Int) {

codeCore/src/prog8/code/core/MemoryRegions.kt

@@ -131,6 +131,7 @@ abstract class Zeropage(options: CompilationOptions): MemoryAllocator(options) {
 }
 
 
+// TODO: this class is not yet used
 class GoldenRam(options: CompilationOptions, val region: UIntRange): MemoryAllocator(options) {
     private var nextLocation: UInt = region.first
 

codeCore/src/prog8/code/target/atari/AtariZeropage.kt

@@ -13,6 +13,10 @@ class AtariZeropage(options: CompilationOptions) : Zeropage(options) {
     override val SCRATCH_W2 = 0xcfu      // temp storage 2 for a word  $cf+$d0        TODO is $d0 okay to use?
 
     init {
+        if (options.floats) {
+            throw InternalCompilerException("Atari target doesn't yet support floating point routines")
+        }
+
         if (options.floats && options.zeropage !in arrayOf(
                 ZeropageType.FLOATSAFE,
                 ZeropageType.BASICSAFE,
@@ -39,9 +43,9 @@ class AtariZeropage(options: CompilationOptions) : Zeropage(options) {
             }
         }
 
-        val distictFree = free.distinct()
+        val distinctFree = free.distinct()
         free.clear()
-        free.addAll(distictFree)
+        free.addAll(distinctFree)
 
         removeReservedFromFreePool()
     }

codeCore/src/prog8/code/target/c128/C128Zeropage.kt

@@ -6,15 +6,22 @@ import prog8.code.core.Zeropage
 import prog8.code.core.ZeropageType
 
 
+// reference: "Mapping the C128" zero page chapter.
+
 class C128Zeropage(options: CompilationOptions) : Zeropage(options) {
 
-    override val SCRATCH_B1 = 0x9bu      // temp storage for a single byte
-    override val SCRATCH_REG = 0x9cu     // temp storage for a register, must be B1+1
+    override val SCRATCH_B1 = 0x74u      // temp storage for a single byte
+    override val SCRATCH_REG = 0x75u     // temp storage for a register, must be B1+1
     override val SCRATCH_W1 = 0xfbu      // temp storage 1 for a word  $fb+$fc
     override val SCRATCH_W2 = 0xfdu      // temp storage 2 for a word  $fd+$fe
 
-
     init {
+        if (options.floats) {
+            throw InternalCompilerException("C128 target doesn't yet support floating point routines")
+            // note: in git commit labeled 'c128: remove floats module' the floats.p8 and floats.asm files are removed,
+            //       they could be retrieved again at a later time if the compiler somehow *does* store the fp variables in bank1.
+        }
+
         if (options.floats && options.zeropage !in arrayOf(
                 ZeropageType.FLOATSAFE,
                 ZeropageType.BASICSAFE,
@@ -24,23 +31,42 @@ class C128Zeropage(options: CompilationOptions) : Zeropage(options) {
 
         when (options.zeropage) {
             ZeropageType.FULL -> {
-                // TODO all c128 usable zero page locations, except the ones used by the system's IRQ routine
-                free.addAll(0x0au..0xffu)       // TODO c128 what about $02-$09?
-                // TODO c128  free.removeAll(setOf(0xa0u, 0xa1u, 0xa2u, 0x91u, 0xc0u, 0xc5u, 0xcbu, 0xf5u, 0xf6u))        // these are updated by IRQ
+                // $00/$01 are data port IO registers, // $02-$09 are storage locations for JSRFAR and such
+                free.addAll(0x0au..0xffu)
+                free.removeAll(listOf(0x90u, 0x91u, 0xa0u, 0xa1u, 0xa2u, 0xc0u, 0xccu, 0xcdu, 0xd0u, 0xd1u, 0xd2u, 0xd3u, 0xd4u, 0xd5u, 0xf7u))        // these are updated/used by IRQ
+            }
+            ZeropageType.KERNALSAFE -> {
+                free.addAll(0x0au..0x8fu)       // BASIC variables
+                free.addAll(listOf(0x92u, 0x96u, 0x9bu, 0x9cu, 0x9eu, 0x9fu, 0xa4u, 0xa7u, 0xa8u, 0xa9u, 0xaau, 0xabu,
+                    0xb0u, 0xb1u, 0xb4u, 0xb5u, 0xb6u))
             }
-            ZeropageType.KERNALSAFE,
             ZeropageType.FLOATSAFE,
             ZeropageType.BASICSAFE -> {
-                free.clear()   // TODO c128 usable zero page addresses
+                free.addAll(listOf(0x0bu, 0x0cu, 0x0du, 0x0eu, 0x0fu, 0x10u, 0x11u, 0x12u, 0x16u, 0x17u, 0x18u, 0x19u, 0x1au))
+                free.addAll(0x1bu..0x23u)
+                free.addAll(listOf(0x3fu, 0x40u, 0x41u, 0x42u, 0x43u, 0x44u, 0x47u, 0x48u, 0x49u, 0x4au, 0x4bu, 0x4cu, 0x4fu,
+                    0x55u, 0x56u, 0x57u, 0x58u,
+                    0x74u, 0x75u, 0x78u, 0x80u, 0x83u, 0x87u, 0x88u, 0x89u, 0x8au, 0x8bu, 0x8cu, 0x8du, 0x8eu, 0x8fu,
+                    0x92u, 0x96u, 0x9bu, 0x9cu, 0x9eu, 0x9fu, 0xa4u, 0xa7u, 0xa8u, 0xa9u, 0xaau, 0xabu,
+                    0xb0u, 0xb1u, 0xb4u, 0xb5u, 0xb6u
+                ))
+
+                // if(options.zeropage==ZeropageType.BASICSAFE) {
+                    // can also clobber the FP locations (unconditionally, because the C128 target doesn't support floating point calculations in prog8 at this time0
+                    free.addAll(listOf(0x14u, 0x28u, 0x29u, 0x2au, 0x2bu, 0x2cu,
+                        0x50u, 0x51u, 0x52u, 0x53u, 0x54u, 0x59u, 0x5au, 0x5bu, 0x5cu, 0x5du, 0x5eu, 0x5fu, 0x60u, 0x61u, 0x62u,
+                        0x63u, 0x64u, 0x65u, 0x66u, 0x67u, 0x68u,
+                        0x6au, 0x6bu, 0x6cu, 0x6du, 0x6eu, 0x6fu, 0x71u))
+                // }
             }
             ZeropageType.DONTUSE -> {
                 free.clear()  // don't use zeropage at all
             }
         }
 
-        val distictFree = free.distinct()
+        val distinctFree = free.distinct()
         free.clear()
-        free.addAll(distictFree)
+        free.addAll(distinctFree)
 
         removeReservedFromFreePool()
     }

codeCore/src/prog8/code/target/c64/C64Zeropage.kt

@@ -65,9 +65,9 @@ class C64Zeropage(options: CompilationOptions) : Zeropage(options) {
             }
         }
 
-        val distictFree = free.distinct()
+        val distinctFree = free.distinct()
         free.clear()
-        free.addAll(distictFree)
+        free.addAll(distinctFree)
 
         removeReservedFromFreePool()
 

codeCore/src/prog8/code/target/cx16/CX16MachineDefinition.kt

@@ -64,7 +64,7 @@ class CX16MachineDefinition: IMachineDefinition {
 
     override fun initializeMemoryAreas(compilerOptions: CompilationOptions) {
         zeropage = CX16Zeropage(compilerOptions)
-        golden = GoldenRam(compilerOptions, 0x0600u until 0x0800u)
+        golden = GoldenRam(compilerOptions, 0x0400u until ESTACK_LO)
     }
 
 }

codeCore/src/prog8/code/target/cx16/CX16Zeropage.kt

@@ -43,9 +43,9 @@ class CX16Zeropage(options: CompilationOptions) : Zeropage(options) {
                 else -> throw InternalCompilerException("for this machine target, zero page type 'floatsafe' is not available. ${options.zeropage}")
             }
 
-            val distictFree = free.distinct()
+            val distinctFree = free.distinct()
             free.clear()
-            free.addAll(distictFree)
+            free.addAll(distinctFree)
 
             removeReservedFromFreePool()
 

codeGenCpu6502/build.gradle

@@ -28,9 +28,9 @@ dependencies {
     implementation project(':codeCore')
     implementation "org.jetbrains.kotlin:kotlin-stdlib-jdk8"
     // implementation "org.jetbrains.kotlin:kotlin-reflect"
-    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.16"
+    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18"
 
-    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.5.5'
+    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.6.2'
 }
 
 sourceSets {

codeGenCpu6502/src/prog8/codegen/cpu6502/AsmGen.kt

@@ -3,6 +3,7 @@ package prog8.codegen.cpu6502
 import com.github.michaelbull.result.fold
 import prog8.code.StNodeType
 import prog8.code.SymbolTable
+import prog8.code.SymbolTableMaker
 import prog8.code.ast.*
 import prog8.code.core.*
 import prog8.codegen.cpu6502.assignment.*
@@ -14,18 +15,180 @@ import kotlin.io.path.writeLines
 internal const val subroutineFloatEvalResultVar1 = "prog8_float_eval_result1"
 internal const val subroutineFloatEvalResultVar2 = "prog8_float_eval_result2"
 
-class AsmGen6502: ICodeGeneratorBackend {
+class AsmGen6502(val prefixSymbols: Boolean): ICodeGeneratorBackend {
     override fun generate(
         program: PtProgram,
         symbolTable: SymbolTable,
         options: CompilationOptions,
         errors: IErrorReporter
     ): IAssemblyProgram? {
-        val asmgen = AsmGen6502Internal(program, symbolTable, options, errors)
+        val st = if(prefixSymbols) prefixSymbols(program, options, symbolTable) else symbolTable
+        val asmgen = AsmGen6502Internal(program, st, options, errors)
         return asmgen.compileToAssembly()
     }
+
+    private fun prefixSymbols(program: PtProgram, options: CompilationOptions, st: SymbolTable): SymbolTable {
+        val nodesToPrefix = mutableListOf<Pair<PtNode, Int>>()              // parent + index
+        val functionCallsToPrefix = mutableListOf<Pair<PtNode, Int>>()      // parent + index
+
+        fun prefixNamedNode(node: PtNamedNode) {
+            node.name = "p8_${node.name}"
+        }
+
+        fun prefixSymbols(node: PtNode) {
+            when(node) {
+                is PtAsmSub -> {
+                    prefixNamedNode(node)
+                    node.parameters.forEach { (_, param) -> prefixNamedNode(param) }
+                }
+                is PtSub -> {
+                    prefixNamedNode(node)
+                    node.parameters.forEach { prefixNamedNode(it) }
+                }
+                is PtFunctionCall -> {
+                    val stNode = st.lookup(node.name)!!
+                    if(stNode.astNode.definingBlock()?.noSymbolPrefixing!=true) {
+                        val index = node.parent.children.indexOf(node)
+                        functionCallsToPrefix += node.parent to index
+                    }
+                }
+                is PtIdentifier -> {
+                    var lookupName = node.name
+                    if(node.type in SplitWordArrayTypes && (lookupName.endsWith("_lsb") || lookupName.endsWith("_msb"))) {
+                        lookupName = lookupName.dropLast(4)
+                    }
+                    val stNode = st.lookup(lookupName)!!
+                    if(stNode.astNode.definingBlock()?.noSymbolPrefixing!=true) {
+                        val index = node.parent.children.indexOf(node)
+                        nodesToPrefix += node.parent to index
+                    }
+                }
+                is PtJump -> {
+                    if(node.identifier!=null) {
+                        val stNode = st.lookup(node.identifier!!.name)!!
+                        if(stNode.astNode.definingBlock()?.noSymbolPrefixing!=true) {
+                            val index = node.parent.children.indexOf(node)
+                            nodesToPrefix += node.parent to index
+                        }
+                    }
+                    else if(node.generatedLabel!=null) {
+                        val stNode = st.lookup(node.generatedLabel!!)!!
+                        if(stNode.astNode.definingBlock()?.noSymbolPrefixing!=true) {
+                            val index = node.parent.children.indexOf(node)
+                            nodesToPrefix += node.parent to index
+                        }
+                    }
+                }
+                is PtBlock -> prefixNamedNode(node)
+                is PtConstant -> prefixNamedNode(node)
+                is PtLabel -> prefixNamedNode(node)
+                is PtMemMapped -> prefixNamedNode(node)
+                is PtSubroutineParameter -> prefixNamedNode(node)
+                is PtVariable -> {
+                    val index = node.parent.children.indexOf(node)
+                    nodesToPrefix += node.parent to index
+                }
+                else -> { }
+            }
+            node.children.forEach { prefixSymbols(it) }
+        }
+
+        program.allBlocks().forEach { block ->
+            if (!block.noSymbolPrefixing) {
+                prefixSymbols(block)
+            }
+        }
+
+        nodesToPrefix.forEach { (parent, index) ->
+            val node = parent.children[index]
+            when(node) {
+                is PtIdentifier -> parent.children[index] = node.prefix(parent, st)
+                is PtFunctionCall ->  throw AssemblyError("PtFunctionCall should be processed in their own list, last")
+                is PtJump -> parent.children[index] = node.prefix(parent, st)
+                is PtVariable -> parent.children[index] = node.prefix(st)
+                else -> throw AssemblyError("weird node to prefix $node")
+            }
+        }
+
+        // reversed so inner calls (such as arguments to a function call) get processed before the actual function call itself
+        functionCallsToPrefix.reversed().forEach { (parent, index) ->
+            val node = parent.children[index]
+            if(node is PtFunctionCall) {
+                parent.children[index] = node.prefix(parent)
+            } else {
+                throw AssemblyError("expected PtFunctionCall")
+            }
+        }
+
+        return SymbolTableMaker(program, options).make()
+    }
 }
 
+private fun PtVariable.prefix(st: SymbolTable): PtVariable {
+    name = name.split('.').map {"p8_$it" }.joinToString(".")
+    if(value==null)
+        return this
+
+    val arrayValue = value as? PtArray
+    return if(arrayValue!=null && arrayValue.children.any { it !is PtNumber} ) {
+        val newValue = PtArray(arrayValue.type, arrayValue.position)
+        arrayValue.children.forEach { elt ->
+            when(elt) {
+                is PtIdentifier -> newValue.add(elt.prefix(arrayValue, st))
+                is PtNumber -> newValue.add(elt)
+                is PtAddressOf -> {
+                    if(elt.definingBlock()?.noSymbolPrefixing==true)
+                        newValue.add(elt)
+                    else {
+                        val newAddr = PtAddressOf(elt.position)
+                        newAddr.children.add(elt.identifier.prefix(newAddr, st))
+                        newAddr.parent = arrayValue
+                        newValue.add(newAddr)
+                    }
+                }
+                else -> throw AssemblyError("weird array value element $elt")
+            }
+        }
+        PtVariable(name, type, zeropage, newValue, arraySize, position)
+    }
+    else this
+}
+
+private fun PtJump.prefix(parent: PtNode, st: SymbolTable): PtJump {
+    val jump = if(identifier!=null) {
+        val prefixedIdent = identifier!!.prefix(this, st)
+        PtJump(prefixedIdent, address, generatedLabel, position)
+    } else {
+        val prefixedLabel = generatedLabel!!.split('.').map {"p8_$it" }.joinToString(".")
+        PtJump(null, address, prefixedLabel, position)
+    }
+    jump.parent = parent
+    return jump
+}
+
+private fun PtFunctionCall.prefix(parent: PtNode): PtFunctionCall {
+    val newName = name.split('.').map {"p8_$it" }.joinToString(".")
+    val call = PtFunctionCall(newName, void, type, position)
+    call.children.addAll(children)
+    call.children.forEach { it.parent = call }
+    call.parent = parent
+    if(name.endsWith("concat_string"))
+        println("CONCAT ${this.position}")
+    return call
+}
+
+private fun PtIdentifier.prefix(parent: PtNode, st: SymbolTable): PtIdentifier {
+    val target = st.lookup(name)
+    if(target?.astNode?.definingBlock()?.noSymbolPrefixing==true)
+        return this
+
+    val newName = name.split('.').map { "p8_$it" }.joinToString(".")
+    val node = PtIdentifier(newName, type, position)
+    node.parent = parent
+    return node
+}
+
+
 class AsmGen6502Internal (
     val program: PtProgram,
     internal val symbolTable: SymbolTable,
@@ -362,13 +525,15 @@ class AsmGen6502Internal (
     internal fun loadScaledArrayIndexIntoRegister(
         expr: PtArrayIndexer,
         elementDt: DataType,
-        register: CpuRegister,
-        addOneExtra: Boolean = false
+        register: CpuRegister
     ) {
         val reg = register.toString().lowercase()
         val indexnum = expr.index.asConstInteger()
         if (indexnum != null) {
-            val indexValue = indexnum * options.compTarget.memorySize(elementDt) + if (addOneExtra) 1 else 0
+            val indexValue = if(expr.splitWords)
+                indexnum
+            else
+                indexnum * options.compTarget.memorySize(elementDt)
             out("  ld$reg  #$indexValue")
             return
         }
@@ -377,77 +542,42 @@ class AsmGen6502Internal (
             ?: throw AssemblyError("array indexer should have been replaced with a temp var @ ${expr.index.position}")
 
         val indexName = asmVariableName(indexVar)
-        if (addOneExtra) {
-            // add 1 to the result
-            when (elementDt) {
-                in ByteDatatypes -> {
-                    out("  ldy  $indexName |  iny")
-                    when (register) {
-                        CpuRegister.A -> out(" tya")
-                        CpuRegister.X -> out(" tyx")
-                        CpuRegister.Y -> {
-                        }
-                    }
-                }
-                in WordDatatypes -> {
-                    out("  lda  $indexName |  sec |  rol  a")
-                    when (register) {
-                        CpuRegister.A -> {
-                        }
-                        CpuRegister.X -> out(" tax")
-                        CpuRegister.Y -> out(" tay")
-                    }
-                }
-                DataType.FLOAT -> {
-                    require(options.compTarget.memorySize(DataType.FLOAT) == 5) {"invalid float size ${expr.position}"}
-                    out(
-                        """
-                                lda  $indexName
-                                asl  a
-                                asl  a
-                                sec
-                                adc  $indexName"""
-                    )
-                    when (register) {
-                        CpuRegister.A -> {
-                        }
-                        CpuRegister.X -> out(" tax")
-                        CpuRegister.Y -> out(" tay")
-                    }
-                }
-                else -> throw AssemblyError("weird dt")
+
+        if(expr.splitWords) {
+            when (register) {
+                CpuRegister.A -> out("  lda  $indexName")
+                CpuRegister.X -> out("  ldx  $indexName")
+                CpuRegister.Y -> out("  ldy  $indexName")
             }
-        } else {
-            when (elementDt) {
-                in ByteDatatypes -> out("  ld$reg  $indexName")
-                in WordDatatypes -> {
-                    out("  lda  $indexName |  asl  a")
-                    when (register) {
-                        CpuRegister.A -> {
-                        }
-                        CpuRegister.X -> out(" tax")
-                        CpuRegister.Y -> out(" tay")
-                    }
+            return
+        }
+
+        when (elementDt) {
+            in ByteDatatypes -> out("  ld$reg  $indexName")
+            in WordDatatypes -> {
+                out("  lda  $indexName |  asl  a")
+                when (register) {
+                    CpuRegister.A -> {}
+                    CpuRegister.X -> out(" tax")
+                    CpuRegister.Y -> out(" tay")
                 }
-                DataType.FLOAT -> {
-                    require(options.compTarget.memorySize(DataType.FLOAT) == 5) {"invalid float size ${expr.position}"}
-                    out(
-                        """
-                                lda  $indexName
-                                asl  a
-                                asl  a
-                                clc
-                                adc  $indexName"""
-                    )
-                    when (register) {
-                        CpuRegister.A -> {
-                        }
-                        CpuRegister.X -> out(" tax")
-                        CpuRegister.Y -> out(" tay")
-                    }
-                }
-                else -> throw AssemblyError("weird dt")
             }
+            DataType.FLOAT -> {
+                require(options.compTarget.memorySize(DataType.FLOAT) == 5) {"invalid float size ${expr.position}"}
+                out("""
+                    lda  $indexName
+                    asl  a
+                    asl  a
+                    clc
+                    adc  $indexName"""
+                )
+                when (register) {
+                    CpuRegister.A -> {}
+                    CpuRegister.X -> out(" tax")
+                    CpuRegister.Y -> out(" tay")
+                }
+            }
+            else -> throw AssemblyError("weird dt")
         }
     }
 
@@ -483,7 +613,7 @@ class AsmGen6502Internal (
         when(reg) {
             RegisterOrPair.A,
             RegisterOrPair.X,
-            RegisterOrPair.Y -> assignmentAsmGen.assignRegisterByte(target, reg.asCpuRegister())
+            RegisterOrPair.Y -> assignmentAsmGen.assignRegisterByte(target, reg.asCpuRegister(), target.datatype in SignedDatatypes)
             RegisterOrPair.AX,
             RegisterOrPair.AY,
             RegisterOrPair.XY,
@@ -546,7 +676,6 @@ class AsmGen6502Internal (
     private fun translate(stmt: PtIfElse) {
         val condition = stmt.condition as? PtBinaryExpression
         if(condition!=null) {
-            require(!options.useNewExprCode)
             requireComparisonExpression(condition)  // IfStatement: condition must be of form  'x <comparison> <value>'
             if (stmt.elseScope.children.isEmpty()) {
                 val jump = stmt.ifScope.children.singleOrNull()
@@ -674,65 +803,48 @@ class AsmGen6502Internal (
     private fun repeatWordCount(count: Int, stmt: PtRepeatLoop) {
         require(count in 257..65535) { "invalid repeat count ${stmt.position}" }
         val repeatLabel = makeLabel("repeat")
-        if(isTargetCpu(CpuType.CPU65c02)) {
-            val counterVar = createRepeatCounterVar(DataType.UWORD, true, stmt)
-            out("""
-                lda  #<$count
-                ldy  #>$count
-                sta  $counterVar
-                sty  $counterVar+1
+        val counterVar = createRepeatCounterVar(DataType.UWORD, isTargetCpu(CpuType.CPU65c02), stmt)
+        // the iny + double dec is microoptimization of the 16 bit loop
+        out("""
+            ldy  #>$count
+            lda  #<$count
+            beq  +
+            iny
++           sta  $counterVar
+            sty  $counterVar+1
 $repeatLabel""")
-            translate(stmt.statements)
-            out("""
-                lda  $counterVar
-                bne  +
-                dec  $counterVar+1
-+               dec  $counterVar
-                lda  $counterVar
-                ora  $counterVar+1
-                bne  $repeatLabel""")
-        } else {
-            val counterVar = createRepeatCounterVar(DataType.UWORD, false, stmt)
-            out("""
-                lda  #<$count
-                ldy  #>$count
-                sta  $counterVar
-                sty  $counterVar+1
-$repeatLabel""")
-            translate(stmt.statements)
-            out("""
-                lda  $counterVar
-                bne  +
-                dec  $counterVar+1
-+               dec  $counterVar
-                lda  $counterVar
-                ora  $counterVar+1
-                bne  $repeatLabel""")
-        }
+        translate(stmt.statements)
+        out("""
+            dec  $counterVar
+            bne  $repeatLabel
+            dec  $counterVar+1
+            bne  $repeatLabel""")
     }
 
     private fun repeatWordCountInAY(endLabel: String, stmt: PtRepeatLoop) {
         // note: A/Y must have been loaded with the number of iterations!
-        // no need to explicitly test for 0 iterations as this is done in the countdown logic below
+        // the iny + double dec is microoptimization of the 16 bit loop
         val repeatLabel = makeLabel("repeat")
         val counterVar = createRepeatCounterVar(DataType.UWORD, false, stmt)
         out("""
-                sta  $counterVar
-                sty  $counterVar+1
-$repeatLabel    lda  $counterVar
-                bne  +
-                lda  $counterVar+1
-                beq  $endLabel
-                lda  $counterVar
-                bne  +
-                dec  $counterVar+1
-+               dec  $counterVar
-""")
+            cmp  #0
+            beq  +
+            iny
++           sta  $counterVar
+            sty  $counterVar+1
+            ora  $counterVar+1
+            beq  $endLabel
+$repeatLabel""")
         translate(stmt.statements)
-        jmp(repeatLabel)
+        out("""
+            dec  $counterVar
+            bne  $repeatLabel
+            dec  $counterVar+1
+            bne  $repeatLabel""")
         out(endLabel)
     }
 
+
     private fun repeatByteCount(count: Int, stmt: PtRepeatLoop) {
         require(count in 2..256) { "invalid repeat count ${stmt.position}" }
         val repeatLabel = makeLabel("repeat")
@@ -1023,18 +1135,10 @@ $repeatLabel    lda  $counterVar
     }
 
     internal fun pointerViaIndexRegisterPossible(pointerOffsetExpr: PtExpression): Pair<PtExpression, PtExpression>? {
-        val left: PtExpression
-        val right: PtExpression
-        val operator: String
-
-        if (pointerOffsetExpr is PtBinaryExpression) {
-            require(!options.useNewExprCode)
-            operator = pointerOffsetExpr.operator
-            left = pointerOffsetExpr.left
-            right = pointerOffsetExpr.right
-        }
-        else return null
-
+        if (pointerOffsetExpr !is PtBinaryExpression) return null
+        val operator = pointerOffsetExpr.operator
+        val left = pointerOffsetExpr.left
+        val right = pointerOffsetExpr.right
         if (operator != "+") return null
         val leftDt = left.type
         val rightDt = right.type
@@ -1146,10 +1250,7 @@ $repeatLabel    lda  $counterVar
     }
 
     internal fun findSubroutineParameter(name: String, asmgen: AsmGen6502Internal): PtSubroutineParameter? {
-        val stScope = asmgen.symbolTable.lookup(name)
-        require(stScope!=null) {
-            "invalid name lookup $name"
-        }
+        val stScope = asmgen.symbolTable.lookup(name) ?: return null
         val node = stScope.astNode
         if(node is PtSubroutineParameter)
             return node
@@ -1639,14 +1740,7 @@ $repeatLabel    lda  $counterVar
         if(byteJumpForSimpleRightOperand(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            assignExpressionToRegister(left, RegisterOrPair.A)
-        } else {
-            pushCpuStack(DataType.UBYTE, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            out("  pla")
-        }
+        assignByteOperandsToAAndVar(left, right, "P8ZP_SCRATCH_B1")
         return code("P8ZP_SCRATCH_B1")
     }
 
@@ -1680,14 +1774,7 @@ $repeatLabel    lda  $counterVar
         if(byteJumpForSimpleRightOperand(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            assignExpressionToRegister(left, RegisterOrPair.A)
-        } else {
-            pushCpuStack(DataType.UBYTE, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            out("  pla")
-        }
+        assignByteOperandsToAAndVar(left, right, "P8ZP_SCRATCH_B1")
         return code("P8ZP_SCRATCH_B1")
     }
 
@@ -1723,15 +1810,7 @@ $repeatLabel    lda  $counterVar
         if(wordJumpForSimpleRightOperands(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.UWORD)
-            assignExpressionToRegister(left, RegisterOrPair.AY)
-        } else {
-            pushCpuStack(DataType.UWORD, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.UWORD)
-            restoreRegisterStack(CpuRegister.Y, false)
-            restoreRegisterStack(CpuRegister.A, false)
-        }
+        assignWordOperandsToAYAndVar(left, right, "P8ZP_SCRATCH_W2")
         return out("  jsr  prog8_lib.reg_less_uw |  beq  $jumpIfFalseLabel")
     }
 
@@ -1769,15 +1848,7 @@ $repeatLabel    lda  $counterVar
         if(wordJumpForSimpleRightOperands(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.WORD)
-            assignExpressionToRegister(left, RegisterOrPair.AY)
-        } else {
-            pushCpuStack(DataType.WORD, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.WORD)
-            restoreRegisterStack(CpuRegister.Y, false)
-            restoreRegisterStack(CpuRegister.A, false)
-        }
+        assignWordOperandsToAYAndVar(left, right, "P8ZP_SCRATCH_W2")
         return out("  jsr  prog8_lib.reg_less_w |  beq  $jumpIfFalseLabel")
     }
 
@@ -1816,14 +1887,7 @@ $repeatLabel    lda  $counterVar
         if(byteJumpForSimpleRightOperand(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            assignExpressionToRegister(left, RegisterOrPair.A)
-        } else {
-            pushCpuStack(DataType.UBYTE, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            out("  pla")
-        }
+        assignByteOperandsToAAndVar(left, right, "P8ZP_SCRATCH_B1")
         return code("P8ZP_SCRATCH_B1")
     }
 
@@ -1859,14 +1923,7 @@ $repeatLabel    lda  $counterVar
         if(byteJumpForSimpleRightOperand(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.BYTE)
-            assignExpressionToRegister(left, RegisterOrPair.A)
-        } else {
-            pushCpuStack(DataType.BYTE, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.BYTE)
-            out("  pla")
-        }
+        assignByteOperandsToAAndVar(left, right, "P8ZP_SCRATCH_B1")
         return code("P8ZP_SCRATCH_B1")
     }
 
@@ -1908,15 +1965,7 @@ $repeatLabel    lda  $counterVar
         if(wordJumpForSimpleRightOperands(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.UWORD)
-            assignExpressionToRegister(left, RegisterOrPair.AY)
-        } else {
-            pushCpuStack(DataType.UWORD, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.UWORD)
-            restoreRegisterStack(CpuRegister.Y, false)
-            restoreRegisterStack(CpuRegister.A, false)
-        }
+        assignWordOperandsToAYAndVar(left, right, "P8ZP_SCRATCH_W2")
         return code("P8ZP_SCRATCH_W2+1", "P8ZP_SCRATCH_W2")
     }
 
@@ -1959,15 +2008,7 @@ $repeatLabel    lda  $counterVar
         if(wordJumpForSimpleLeftOperand(left, right, ::code))
             return
 
-        if(right.isSimple()) {
-            assignExpressionToVariable(left, "P8ZP_SCRATCH_W2", DataType.WORD)
-            assignExpressionToRegister(right, RegisterOrPair.AY)
-        }  else {
-            pushCpuStack(DataType.WORD, right)
-            assignExpressionToVariable(left, "P8ZP_SCRATCH_W2", DataType.WORD)
-            restoreRegisterStack(CpuRegister.Y, false)
-            restoreRegisterStack(CpuRegister.A, false)
-        }
+        assignWordOperandsToAYAndVar(right, left, "P8ZP_SCRATCH_W2")
         return out("  jsr  prog8_lib.reg_less_w |  beq  $jumpIfFalseLabel")
     }
 
@@ -2007,14 +2048,7 @@ $repeatLabel    lda  $counterVar
         if(byteJumpForSimpleRightOperand(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            assignExpressionToRegister(left, RegisterOrPair.A)
-        } else {
-            pushCpuStack(DataType.UBYTE, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            out("  pla")
-        }
+        assignByteOperandsToAAndVar(left, right, "P8ZP_SCRATCH_B1")
         return code("P8ZP_SCRATCH_B1")
     }
 
@@ -2050,14 +2084,7 @@ $repeatLabel    lda  $counterVar
         if(byteJumpForSimpleRightOperand(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.BYTE)
-            assignExpressionToRegister(left, RegisterOrPair.A)
-        } else {
-            pushCpuStack(DataType.BYTE, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.BYTE)
-            out("  pla")
-        }
+        assignByteOperandsToAAndVar(left, right, "P8ZP_SCRATCH_B1")
         return code("P8ZP_SCRATCH_B1")
     }
 
@@ -2101,15 +2128,7 @@ $repeatLabel    lda  $counterVar
         if(wordJumpForSimpleRightOperands(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.UWORD)
-            assignExpressionToRegister(left, RegisterOrPair.AY)
-        } else {
-            pushCpuStack(DataType.UWORD, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.UWORD)
-            restoreRegisterStack(CpuRegister.Y, false)
-            restoreRegisterStack(CpuRegister.A, false)
-        }
+        assignWordOperandsToAYAndVar(left, right, "P8ZP_SCRATCH_W2")
         return out("  jsr  prog8_lib.reg_lesseq_uw |  beq  $jumpIfFalseLabel")
     }
 
@@ -2156,15 +2175,7 @@ $repeatLabel    lda  $counterVar
             return code(asmVariableName(left))
         }
 
-        if(right.isSimple()) {
-            assignExpressionToVariable(left, "P8ZP_SCRATCH_W2", DataType.WORD)
-            assignExpressionToRegister(right, RegisterOrPair.AY)
-        } else {
-            pushCpuStack(DataType.WORD, right)
-            assignExpressionToVariable(left, "P8ZP_SCRATCH_W2", DataType.WORD)
-            restoreRegisterStack(CpuRegister.Y, false)
-            restoreRegisterStack(CpuRegister.A, false)
-        }
+        assignWordOperandsToAYAndVar(right, left, "P8ZP_SCRATCH_W2")
         return out("  jsr  prog8_lib.reg_lesseq_w |  beq  $jumpIfFalseLabel")
     }
 
@@ -2200,14 +2211,7 @@ $repeatLabel    lda  $counterVar
         if(byteJumpForSimpleRightOperand(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            assignExpressionToRegister(left, RegisterOrPair.A)
-        } else {
-            pushCpuStack(DataType.UBYTE, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-            out("  pla")
-        }
+        assignByteOperandsToAAndVar(left, right, "P8ZP_SCRATCH_B1")
         return code("P8ZP_SCRATCH_B1")
     }
 
@@ -2240,15 +2244,31 @@ $repeatLabel    lda  $counterVar
         if(byteJumpForSimpleRightOperand(left, right, ::code))
             return
 
+        assignByteOperandsToAAndVar(left, right, "P8ZP_SCRATCH_B1")
+        return code("P8ZP_SCRATCH_B1")
+    }
+
+    internal fun assignByteOperandsToAAndVar(left: PtExpression, right: PtExpression, rightVarName: String) {
         if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.BYTE)
+            assignExpressionToVariable(right, rightVarName, DataType.UBYTE)
             assignExpressionToRegister(left, RegisterOrPair.A)
         } else {
-            pushCpuStack(DataType.BYTE, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_B1", DataType.BYTE)
+            pushCpuStack(DataType.UBYTE, left)
+            assignExpressionToVariable(right, rightVarName, DataType.UBYTE)
             out("  pla")
         }
-        return code("P8ZP_SCRATCH_B1")
+    }
+
+    internal fun assignWordOperandsToAYAndVar(left: PtExpression, right: PtExpression, rightVarname: String) {
+        if(left.isSimple()) {
+            assignExpressionToVariable(right, rightVarname, DataType.UWORD)
+            assignExpressionToRegister(left, RegisterOrPair.AY)
+        }  else {
+            pushCpuStack(DataType.UWORD, left)
+            assignExpressionToVariable(right, rightVarname, DataType.UWORD)
+            restoreRegisterStack(CpuRegister.Y, false)
+            restoreRegisterStack(CpuRegister.A, false)
+        }
     }
 
     private fun translateUwordGreaterOrEqualOrJumpElsewhere(left: PtExpression, right: PtExpression, leftConstVal: PtNumber?, rightConstVal: PtNumber?, jumpIfFalseLabel: String) {
@@ -2282,15 +2302,7 @@ $repeatLabel    lda  $counterVar
         if(wordJumpForSimpleRightOperands(left, right, ::code))
             return
 
-        if(right.isSimple()) {
-            assignExpressionToVariable(left, "P8ZP_SCRATCH_W2", DataType.UWORD)
-            assignExpressionToRegister(right, RegisterOrPair.AY)
-        }  else {
-            pushCpuStack(DataType.UWORD, right)
-            assignExpressionToVariable(left, "P8ZP_SCRATCH_W2", DataType.UWORD)
-            restoreRegisterStack(CpuRegister.Y, false)
-            restoreRegisterStack(CpuRegister.A, false)
-        }
+        assignWordOperandsToAYAndVar(right, left, "P8ZP_SCRATCH_W2")
         return out("  jsr  prog8_lib.reg_lesseq_uw |  beq  $jumpIfFalseLabel")
     }
 
@@ -2328,15 +2340,7 @@ $repeatLabel    lda  $counterVar
         if(wordJumpForSimpleRightOperands(left, right, ::code))
             return
 
-        if(left.isSimple()) {
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.WORD)
-            assignExpressionToRegister(left, RegisterOrPair.AY)
-        } else {
-            pushCpuStack(DataType.WORD, left)
-            assignExpressionToVariable(right, "P8ZP_SCRATCH_W2", DataType.WORD)
-            restoreRegisterStack(CpuRegister.Y, false)
-            restoreRegisterStack(CpuRegister.A, false)
-        }
+        assignWordOperandsToAYAndVar(left, right, "P8ZP_SCRATCH_W2")
         return out("  jsr  prog8_lib.reg_lesseq_w |  beq  $jumpIfFalseLabel")
     }
 
@@ -2894,7 +2898,6 @@ $repeatLabel    lda  $counterVar
                     out("  sta  P8ESTACK_LO,x |  dex")
             }
             is PtBinaryExpression -> {
-                require(!options.useNewExprCode)
                 val addrExpr = expr.address as PtBinaryExpression
                 if(tryOptimizedPointerAccessWithA(addrExpr, addrExpr.operator, false)) {
                     if(pushResultOnEstack)

codeGenCpu6502/src/prog8/codegen/cpu6502/AsmOptimizer.kt

@@ -43,7 +43,14 @@ internal fun optimizeAssembly(lines: MutableList<String>, machine: IMachineDefin
         numberOfOptimizations++
     }
 
-    mods= optimizeJsrRtsAndOtherCombinations(linesByFour)
+    mods = optimizeJsrRtsAndOtherCombinations(linesByFour)
+    if(mods.isNotEmpty()) {
+        apply(mods, lines)
+        linesByFour = getLinesBy(lines, 4)
+        numberOfOptimizations++
+    }
+
+    mods = optimizeUselessPushPopStack(linesByFour)
     if(mods.isNotEmpty()) {
         apply(mods, lines)
         linesByFour = getLinesBy(lines, 4)
@@ -517,3 +524,40 @@ private fun optimizeJsrRtsAndOtherCombinations(linesByFour: List<List<IndexedVal
     }
     return mods
 }
+
+private fun optimizeUselessPushPopStack(linesByFour: List<List<IndexedValue<String>>>): List<Modification> {
+    val mods = mutableListOf<Modification>()
+
+    fun optimize(register: Char, lines: List<IndexedValue<String>>) {
+        if(lines[0].value.trimStart().startsWith("ph$register")) {
+            if(lines[2].value.trimStart().startsWith("pl$register")) {
+                val second = lines[1].value.trimStart().take(6).lowercase()
+                if(register!in second
+                    && !second.startsWith("jsr")
+                    && !second.startsWith("pl")
+                    && !second.startsWith("ph")) {
+                    mods.add(Modification(lines[0].index, true, null))
+                    mods.add(Modification(lines[2].index, true, null))
+                }
+            }
+            else if (lines[3].value.trimStart().startsWith("pl$register")) {
+                val second = lines[1].value.trimStart().take(6).lowercase()
+                val third = lines[2].value.trimStart().take(6).lowercase()
+                if(register !in second && register !in third
+                    && !second.startsWith("jsr") && !third.startsWith("jsr")
+                    && !second.startsWith("pl") && !third.startsWith("pl")
+                    && !second.startsWith("ph") && !third.startsWith("ph")) {
+                    mods.add(Modification(lines[0].index, true, null))
+                    mods.add(Modification(lines[3].index, true, null))
+                }
+            }
+        }
+    }
+
+    for (lines in linesByFour) {
+        optimize('a', lines)
+        optimize('x', lines)
+        optimize('y', lines)
+    }
+    return mods
+}

codeGenCpu6502/src/prog8/codegen/cpu6502/BuiltinFunctionsAsmGen.kt

@@ -1,5 +1,6 @@
 package prog8.codegen.cpu6502
 
+import prog8.code.StStaticVariable
 import prog8.code.ast.*
 import prog8.code.core.*
 import prog8.codegen.cpu6502.assignment.*
@@ -30,12 +31,15 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             "msb" -> funcMsb(fcall, resultToStack, resultRegister)
             "lsb" -> funcLsb(fcall, resultToStack, resultRegister)
             "mkword" -> funcMkword(fcall, resultToStack, resultRegister)
-            "abs" -> funcAbs(fcall, resultToStack, resultRegister, sscope)
+            "clamp__byte", "clamp__ubyte", "clamp__word", "clamp__uword" -> funcClamp(fcall, resultToStack, resultRegister)
+            "min__byte", "min__ubyte", "min__word", "min__uword" -> funcMin(fcall, resultToStack, resultRegister)
+            "max__byte", "max__ubyte", "max__word", "max__uword" -> funcMax(fcall, resultToStack, resultRegister)
+            "abs__byte", "abs__word", "abs__float" -> funcAbs(fcall, resultToStack, resultRegister, sscope)
             "any", "all" -> funcAnyAll(fcall, resultToStack, resultRegister, sscope)
             "sgn" -> funcSgn(fcall, resultToStack, resultRegister, sscope)
-            "sqrt16" -> funcSqrt16(fcall, resultToStack, resultRegister, sscope)
-            "divmod" -> funcDivmod(fcall)
-            "divmodw" -> funcDivmodW(fcall)
+            "sqrt__ubyte", "sqrt__uword", "sqrt__float" -> funcSqrt(fcall, resultToStack, resultRegister, sscope)
+            "divmod__ubyte" -> funcDivmod(fcall)
+            "divmod__uword" -> funcDivmodW(fcall)
             "rol" -> funcRol(fcall)
             "rol2" -> funcRol2(fcall)
             "ror" -> funcRor(fcall)
@@ -90,13 +94,12 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         val var3name = asmgen.asmVariableName(fcall.args[3] as PtIdentifier)
         val divisionTarget = AsmAssignTarget(TargetStorageKind.VARIABLE, asmgen, DataType.UBYTE, fcall.definingISub(), fcall.args[2].position, var2name)
         val remainderTarget = AsmAssignTarget(TargetStorageKind.VARIABLE, asmgen, DataType.UBYTE, fcall.definingISub(), fcall.args[3].position, var3name)
-        assignAsmGen.assignRegisterByte(remainderTarget, CpuRegister.A)
-        assignAsmGen.assignRegisterByte(divisionTarget, CpuRegister.Y)
+        assignAsmGen.assignRegisterByte(remainderTarget, CpuRegister.A, false)
+        assignAsmGen.assignRegisterByte(divisionTarget, CpuRegister.Y, false)
     }
 
     private fun funcDivmodW(fcall: PtBuiltinFunctionCall) {
-        assignAsmGen.assignExpressionToVariable(fcall.args[0], "P8ZP_SCRATCH_W1", DataType.UWORD)
-        assignAsmGen.assignExpressionToRegister(fcall.args[1], RegisterOrPair.AY, false)
+        asmgen.assignWordOperandsToAYAndVar(fcall.args[1], fcall.args[0], "P8ZP_SCRATCH_W1")
         // math.divmod_uw_asm: -- divide two unsigned words (16 bit each) into 16 bit results
         //    input:  P8ZP_SCRATCH_W1 in ZP: 16 bit number, A/Y: 16 bit divisor
         //    output: P8ZP_SCRATCH_W2 in ZP: 16 bit remainder, A/Y: 16 bit division result
@@ -113,8 +116,7 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
     }
 
     private fun funcStringCompare(fcall: PtBuiltinFunctionCall, resultToStack: Boolean) {
-        assignAsmGen.assignExpressionToVariable(fcall.args[1], "P8ZP_SCRATCH_W2", DataType.UWORD)
-        assignAsmGen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.AY, false)
+        asmgen.assignWordOperandsToAYAndVar(fcall.args[0], fcall.args[1], "P8ZP_SCRATCH_W2")
         asmgen.out("  jsr  prog8_lib.strcmp_mem")
         if(resultToStack)
             asmgen.out("  sta  P8ESTACK_LO,x |  dex")
@@ -206,27 +208,13 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                             asmgen.assignExpressionToRegister(arg1, RegisterOrPair.A)
                             asmgen.out("  cmp  ${arg2.address.asConstInteger()!!.toHex()}")
                         } else {
-                            if(arg1.isSimple()) {
-                                asmgen.assignExpressionToVariable(arg2, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-                                asmgen.assignExpressionToRegister(arg1, RegisterOrPair.A)
-                                asmgen.out("  cmp  P8ZP_SCRATCH_B1")
-                            } else {
-                                asmgen.pushCpuStack(DataType.UBYTE, arg1)
-                                asmgen.assignExpressionToVariable(arg2, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-                                asmgen.out("  pla |  cmp  P8ZP_SCRATCH_B1")
-                            }
+                            asmgen.assignByteOperandsToAAndVar(arg1, arg2, "P8ZP_SCRATCH_B1")
+                            asmgen.out("  cmp  P8ZP_SCRATCH_B1")
                         }
                     }
                     else -> {
-                        if(arg1.isSimple()) {
-                            asmgen.assignExpressionToVariable(arg2, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-                            asmgen.assignExpressionToRegister(arg1, RegisterOrPair.A)
-                            asmgen.out("  cmp  P8ZP_SCRATCH_B1")
-                        } else {
-                            asmgen.pushCpuStack(DataType.UBYTE, arg1)
-                            asmgen.assignExpressionToVariable(arg2, "P8ZP_SCRATCH_B1", DataType.UBYTE)
-                            asmgen.out("  pla |  cmp  P8ZP_SCRATCH_B1")
-                        }
+                        asmgen.assignByteOperandsToAAndVar(arg1, arg2, "P8ZP_SCRATCH_B1")
+                        asmgen.out("  cmp  P8ZP_SCRATCH_B1")
                     }
                 }
             } else
@@ -252,25 +240,12 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
 +""")
                     }
                     else -> {
-                        if(arg1.isSimple()) {
-                            asmgen.assignExpressionToVariable(arg2, "P8ZP_SCRATCH_W1", DataType.UWORD)
-                            asmgen.assignExpressionToRegister(arg1, RegisterOrPair.AY)
-                            asmgen.out("""
-                                cpy  P8ZP_SCRATCH_W1+1
-                                bne  +
-                                cmp  P8ZP_SCRATCH_W1
-    +""")
-                        } else {
-                            asmgen.pushCpuStack(DataType.UWORD, arg1)
-                            asmgen.assignExpressionToVariable(arg2, "P8ZP_SCRATCH_W1", DataType.UWORD)
-                            asmgen.restoreRegisterStack(CpuRegister.Y, false)
-                            asmgen.restoreRegisterStack(CpuRegister.A, false)
-                            asmgen.out("""
-                                cpy  P8ZP_SCRATCH_W1+1
-                                bne  +
-                                cmp  P8ZP_SCRATCH_W1
-    +""")
-                        }
+                        asmgen.assignWordOperandsToAYAndVar(arg1, arg2, "P8ZP_SCRATCH_W1")
+                        asmgen.out("""
+                            cpy  P8ZP_SCRATCH_W1+1
+                            bne  +
+                            cmp  P8ZP_SCRATCH_W1
++""")
                     }
                 }
             } else
@@ -297,13 +272,34 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         asmgen.translateNormalAssignment(assign, fcall.definingISub())
     }
 
-    private fun funcSqrt16(fcall: PtBuiltinFunctionCall, resultToStack: Boolean, resultRegister: RegisterOrPair?, scope: IPtSubroutine?) {
+    private fun funcSqrt(fcall: PtBuiltinFunctionCall, resultToStack: Boolean, resultRegister: RegisterOrPair?, scope: IPtSubroutine?) {
         translateArguments(fcall, scope)
-        if(resultToStack)
-            asmgen.out("  jsr  prog8_lib.func_sqrt16_stack")
-        else {
-            asmgen.out("  jsr  prog8_lib.func_sqrt16_into_A")
-            assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A)
+        when(fcall.args[0].type) {
+            DataType.UBYTE -> {
+                if(resultToStack)
+                    asmgen.out("  ldy  #0 |  jsr  prog8_lib.func_sqrt16_stack")
+                else {
+                    asmgen.out("  ldy  #0 |  jsr  prog8_lib.func_sqrt16_into_A")
+                    assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, false)
+                }
+            }
+            DataType.UWORD -> {
+                if(resultToStack)
+                    asmgen.out("  jsr  prog8_lib.func_sqrt16_stack")
+                else {
+                    asmgen.out("  jsr  prog8_lib.func_sqrt16_into_A")
+                    assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, false)
+                }
+            }
+            DataType.FLOAT -> {
+                asmgen.out("  jsr  floats.func_sqrt_into_FAC1")
+                if(resultToStack)
+                    assignAsmGen.assignFAC1float(AsmAssignTarget(TargetStorageKind.STACK, asmgen, DataType.FLOAT, scope, fcall.position))
+                else {
+                    assignAsmGen.assignFAC1float(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.FAC1, true, fcall.position, scope, asmgen))
+                }
+            }
+            else -> throw AssemblyError("weird dt")
         }
     }
 
@@ -337,6 +333,16 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                         lda  #$numElements
                         jsr  prog8_lib.func_reverse_w""")
                 }
+                DataType.STR -> {
+                    val stringLength = (symbol as StStaticVariable).length!!-1
+                    asmgen.out("""
+                        lda  #<$varName
+                        ldy  #>$varName
+                        sta  P8ZP_SCRATCH_W1
+                        sty  P8ZP_SCRATCH_W1+1
+                        lda  #$stringLength
+                        jsr  prog8_lib.func_reverse_b""")
+                }
                 DataType.ARRAY_F -> {
                     asmgen.out("""
                         lda  #<$varName
@@ -346,6 +352,7 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                         lda  #$numElements
                         jsr  floats.func_reverse_f""")
                 }
+                in SplitWordArrayTypes -> TODO("split word reverse")
                 else -> throw AssemblyError("weird type")
             }
         }
@@ -381,7 +388,18 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                         lda  #$numElements""")
                     asmgen.out(if (decl.type == DataType.ARRAY_UW) "  jsr  prog8_lib.func_sort_uw" else "  jsr  prog8_lib.func_sort_w")
                 }
+                DataType.STR -> {
+                    val stringLength = (symbol as StStaticVariable).length!!-1
+                    asmgen.out("""
+                        lda  #<$varName
+                        ldy  #>$varName
+                        sta  P8ZP_SCRATCH_W1
+                        sty  P8ZP_SCRATCH_W1+1
+                        lda  #$stringLength
+                        jsr  prog8_lib.func_sort_ub""")
+                }
                 DataType.ARRAY_F -> throw AssemblyError("sorting of floating point array is not supported")
+                in SplitWordArrayTypes -> TODO("split word sort")
                 else -> throw AssemblyError("weird type")
             }
         } else
@@ -416,6 +434,8 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UWORD -> {
                 when (what) {
                     is PtArrayIndexer -> {
+                        if(what.splitWords)
+                            TODO("ror2 split words ${what.position}")
                         translateRolRorArrayArgs(what.variable, what, "ror2", 'w')
                         asmgen.out("  jsr  prog8_lib.ror2_array_uw")
                     }
@@ -475,6 +495,8 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UWORD -> {
                 when (what) {
                     is PtArrayIndexer -> {
+                        if(what.splitWords)
+                            TODO("ror split words ${what.position}")
                         translateRolRorArrayArgs(what.variable, what, "ror", 'w')
                         asmgen.out("  jsr  prog8_lib.ror_array_uw")
                     }
@@ -517,6 +539,8 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UWORD -> {
                 when (what) {
                     is PtArrayIndexer -> {
+                        if(what.splitWords)
+                            TODO("rol2 split words ${what.position}")
                         translateRolRorArrayArgs(what.variable, what, "rol2", 'w')
                         asmgen.out("  jsr  prog8_lib.rol2_array_uw")
                     }
@@ -576,6 +600,8 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UWORD -> {
                 when (what) {
                     is PtArrayIndexer -> {
+                        if(what.splitWords)
+                            TODO("rol split words ${what.position}")
                         translateRolRorArrayArgs(what.variable, what, "rol", 'w')
                         asmgen.out("  jsr  prog8_lib.rol_array_uw")
                     }
@@ -591,14 +617,17 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
     }
 
     private fun translateRolRorArrayArgs(arrayvar: PtIdentifier, indexer: PtArrayIndexer, operation: String, dt: Char) {
+        if(indexer.splitWords)
+            TODO("rol/ror split words access ${indexer.position}")
         if(arrayvar.type==DataType.UWORD) {
             if(dt!='b')
                 throw AssemblyError("non-array var indexing requires bytes dt")
             asmgen.assignExpressionToVariable(arrayvar, "prog8_lib.${operation}_array_u${dt}._arg_target", DataType.UWORD)
         } else {
+            val p = arrayvar.parent
             val addressOf = PtAddressOf(arrayvar.position)
             addressOf.add(arrayvar)
-            addressOf.parent = arrayvar.parent.parent
+            addressOf.parent = p
             asmgen.assignExpressionToVariable(addressOf, "prog8_lib.${operation}_array_u${dt}._arg_target", DataType.UWORD)
         }
         asmgen.assignExpressionToVariable(indexer.index, "prog8_lib.${operation}_array_u${dt}._arg_index", DataType.UBYTE)
@@ -625,7 +654,7 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                 DataType.FLOAT -> asmgen.out("  jsr  floats.func_sign_f_into_A")
                 else -> throw AssemblyError("weird type $dt")
             }
-            assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A)
+            assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, true)
         }
     }
 
@@ -637,6 +666,7 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                 DataType.ARRAY_B, DataType.ARRAY_UB, DataType.STR -> asmgen.out("  jsr  prog8_lib.func_${fcall.name}_b_stack")
                 DataType.ARRAY_UW, DataType.ARRAY_W -> asmgen.out("  jsr  prog8_lib.func_${fcall.name}_w_stack")
                 DataType.ARRAY_F -> asmgen.out("  jsr  floats.func_${fcall.name}_f_stack")
+                in SplitWordArrayTypes -> TODO("split word any/all")
                 else -> throw AssemblyError("weird type $dt")
             }
         } else {
@@ -644,9 +674,10 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                 DataType.ARRAY_B, DataType.ARRAY_UB, DataType.STR -> asmgen.out("  jsr  prog8_lib.func_${fcall.name}_b_into_A |  ldy  #0")
                 DataType.ARRAY_UW, DataType.ARRAY_W -> asmgen.out("  jsr  prog8_lib.func_${fcall.name}_w_into_A |  ldy  #0")
                 DataType.ARRAY_F -> asmgen.out("  jsr  floats.func_${fcall.name}_f_into_A |  ldy  #0")
+                in SplitWordArrayTypes -> TODO("split word any/all")
                 else -> throw AssemblyError("weird type $dt")
             }
-            assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A)
+            assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, dt in SignedDatatypes)
         }
     }
 
@@ -655,21 +686,35 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         val dt = fcall.args.single().type
         if(resultToStack) {
             when (dt) {
-                DataType.UBYTE -> asmgen.out("  ldy  #0")
                 DataType.BYTE -> asmgen.out("  jsr  prog8_lib.abs_b_stack")
-                DataType.UWORD -> {}
                 DataType.WORD -> asmgen.out("  jsr  prog8_lib.abs_w_stack")
-                else -> throw AssemblyError("weird type")
+                else -> {
+                    asmgen.out("  jsr  floats.func_abs_f_into_FAC1")
+                    assignAsmGen.assignFAC1float(AsmAssignTarget(TargetStorageKind.STACK, asmgen, DataType.FLOAT, scope, fcall.position))
+                }
             }
         } else {
             when (dt) {
-                DataType.UBYTE -> asmgen.out("  ldy  #0")
-                DataType.BYTE -> asmgen.out("  jsr  prog8_lib.abs_b_into_AY")
-                DataType.UWORD -> {}
-                DataType.WORD -> asmgen.out("  jsr  prog8_lib.abs_w_into_AY")
+                DataType.BYTE -> {
+                    asmgen.out("  jsr  prog8_lib.abs_b_into_A")
+                    assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A,false)
+                }
+                DataType.WORD -> {
+                    asmgen.out("  jsr  prog8_lib.abs_w_into_AY")
+                    assignAsmGen.assignRegisterpairWord(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.AY, false, fcall.position, scope, asmgen), RegisterOrPair.AY)
+                }
+                DataType.FLOAT -> {
+                    asmgen.out("  jsr  floats.func_abs_f_into_FAC1")
+                    assignAsmGen.assignFAC1float(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.FAC1, true, fcall.position, scope, asmgen))
+                }
+                DataType.UBYTE -> {
+                    asmgen.assignRegister(RegisterOrPair.A, AsmAssignTarget.fromRegisters(resultRegister?:RegisterOrPair.A, false, fcall.position, scope, asmgen))
+                }
+                DataType.UWORD -> {
+                    asmgen.assignRegister(RegisterOrPair.AY, AsmAssignTarget.fromRegisters(resultRegister?:RegisterOrPair.AY, false, fcall.position, scope, asmgen))
+                }
                 else -> throw AssemblyError("weird type")
             }
-            assignAsmGen.assignRegisterpairWord(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.AY, false, fcall.position, scope, asmgen), RegisterOrPair.AY)
         }
     }
 
@@ -706,7 +751,6 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                 }
             }
             is PtBinaryExpression -> {
-                require(!asmgen.options.useNewExprCode)
                 val result = asmgen.pointerViaIndexRegisterPossible(addrExpr)
                 val pointer = result?.first as? PtIdentifier
                 if(result!=null && pointer!=null && asmgen.isZpVar(pointer)) {
@@ -731,8 +775,7 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         }
 
         // fall through method:
-        asmgen.assignExpressionToVariable(fcall.args[0], "P8ZP_SCRATCH_W1", DataType.UWORD)
-        asmgen.assignExpressionToRegister(fcall.args[1], RegisterOrPair.AY)
+        asmgen.assignWordOperandsToAYAndVar(fcall.args[1], fcall.args[0], "P8ZP_SCRATCH_W1")
         asmgen.out("  jsr  prog8_lib.func_pokew")
     }
 
@@ -769,7 +812,6 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                 } else fallback()
             }
             is PtBinaryExpression -> {
-                require(!asmgen.options.useNewExprCode)
                 val result = asmgen.pointerViaIndexRegisterPossible(addrExpr)
                 val pointer = result?.first as? PtIdentifier
                 if(result!=null && pointer!=null && asmgen.isZpVar(pointer)) {
@@ -804,11 +846,167 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         }
     }
 
+    private fun funcClamp(fcall: PtBuiltinFunctionCall, resultToStack: Boolean, resultRegister: RegisterOrPair?) {
+        val signed = fcall.type in SignedDatatypes
+        when(fcall.type) {
+            in ByteDatatypes -> {
+                assignAsmGen.assignExpressionToVariable(fcall.args[1], "P8ZP_SCRATCH_W1", fcall.args[1].type)  // minimum
+                assignAsmGen.assignExpressionToVariable(fcall.args[2], "P8ZP_SCRATCH_W1+1", fcall.args[2].type)  // maximum
+                assignAsmGen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.A, signed)    // value
+                asmgen.out("  jsr  prog8_lib.func_clamp_${fcall.type.toString().lowercase()}")
+                if(resultToStack) {
+                    asmgen.out("  sta  P8ESTACK_LO,x |  dex")
+                } else {
+                    val targetReg = AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, signed, fcall.position, fcall.definingISub(), asmgen)
+                    assignAsmGen.assignRegisterByte(targetReg, CpuRegister.A, signed)
+                }
+            }
+            in WordDatatypes -> {
+                assignAsmGen.assignExpressionToVariable(fcall.args[1], "P8ZP_SCRATCH_W1", fcall.args[1].type)  // minimum
+                assignAsmGen.assignExpressionToVariable(fcall.args[2], "P8ZP_SCRATCH_W2", fcall.args[2].type)  // maximum
+                assignAsmGen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.AY, signed)    // value
+                asmgen.out("  jsr  prog8_lib.func_clamp_${fcall.type.toString().lowercase()}")
+                if(resultToStack) {
+                    asmgen.out("  sta  P8ESTACK_LO,x |  sty  P8ESTACK_HI,x |  dex")
+                } else {
+                    val targetReg = AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.AY, signed, fcall.position, fcall.definingISub(), asmgen)
+                    assignAsmGen.assignRegisterpairWord(targetReg, RegisterOrPair.AY)
+                }
+            }
+            else -> throw AssemblyError("invalid dt")
+        }
+    }
+
+    private fun funcMin(fcall: PtBuiltinFunctionCall, resultToStack: Boolean, resultRegister: RegisterOrPair?) {
+        val signed = fcall.type in SignedDatatypes
+        if(fcall.type in ByteDatatypes) {
+            asmgen.assignExpressionToVariable(fcall.args[1], "P8ZP_SCRATCH_B1", fcall.type)     // right
+            asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.A)          // left
+            asmgen.out("  cmp  P8ZP_SCRATCH_B1")
+            if(signed) asmgen.out("  bmi  +") else asmgen.out("  bcc  +")
+            asmgen.out("""
+                lda  P8ZP_SCRATCH_B1
++""")
+            if(resultToStack) {
+                asmgen.out("  sta  P8ESTACK_LO,x |  dex")
+            } else {
+                val targetReg = AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, signed, fcall.position, fcall.definingISub(), asmgen)
+                asmgen.assignRegister(RegisterOrPair.A, targetReg)
+            }
+        } else if(fcall.type in WordDatatypes) {
+            asmgen.assignExpressionToVariable(fcall.args[0], "P8ZP_SCRATCH_W1", fcall.type)     // left
+            asmgen.assignExpressionToVariable(fcall.args[1], "P8ZP_SCRATCH_W2", fcall.type)     // right
+            if(signed) {
+                asmgen.out("""
+                    lda  P8ZP_SCRATCH_W1
+                    ldy  P8ZP_SCRATCH_W1+1
+                    cmp  P8ZP_SCRATCH_W2
+                    tya
+                    sbc  P8ZP_SCRATCH_W2+1
+                    bvc  +
+                    eor  #$80
++                   bpl  +
+                    lda  P8ZP_SCRATCH_W1                   
+                    ldy  P8ZP_SCRATCH_W1+1
+                    jmp  ++
++                   lda  P8ZP_SCRATCH_W2
+                    ldy  P8ZP_SCRATCH_W2+1
++""")
+            } else {
+                asmgen.out("""
+                    lda  P8ZP_SCRATCH_W1+1
+                    cmp  P8ZP_SCRATCH_W2+1
+                    bcc  ++
+                    bne  +
+                    lda  P8ZP_SCRATCH_W1
+                    cmp  P8ZP_SCRATCH_W2
+                    bcc  ++
++                   lda  P8ZP_SCRATCH_W2
+                    ldy  P8ZP_SCRATCH_W2+1
+                    jmp  ++
++                   lda  P8ZP_SCRATCH_W1
+                    ldy  P8ZP_SCRATCH_W1+1
++""")
+            }
+            if(resultToStack) {
+                asmgen.out("  sta  P8ESTACK_LO,x |  sty  P8ESTACK_HI,x |  dex")
+            } else {
+                val targetReg = AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.AY, signed, fcall.position, fcall.definingISub(), asmgen)
+                asmgen.assignRegister(RegisterOrPair.AY, targetReg)
+            }
+        } else {
+            throw AssemblyError("min float not supported")
+        }
+    }
+
+    private fun funcMax(fcall: PtBuiltinFunctionCall, resultToStack: Boolean, resultRegister: RegisterOrPair?) {
+        val signed = fcall.type in SignedDatatypes
+        if(fcall.type in ByteDatatypes) {
+            asmgen.assignExpressionToVariable(fcall.args[0], "P8ZP_SCRATCH_B1", fcall.type)     // left
+            asmgen.assignExpressionToRegister(fcall.args[1], RegisterOrPair.A)          // right
+            asmgen.out("  cmp  P8ZP_SCRATCH_B1")
+            if(signed) asmgen.out("  bpl  +") else asmgen.out("  bcs  +")
+            asmgen.out("""
+                lda  P8ZP_SCRATCH_B1
++""")
+            if(resultToStack) {
+                asmgen.out("  sta  P8ESTACK_LO,x |  dex")
+            } else {
+                val targetReg = AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, signed, fcall.position, fcall.definingISub(), asmgen)
+                asmgen.assignRegister(RegisterOrPair.A, targetReg)
+            }
+        } else if(fcall.type in WordDatatypes) {
+            asmgen.assignExpressionToVariable(fcall.args[0], "P8ZP_SCRATCH_W1", fcall.type)     // left
+            asmgen.assignExpressionToVariable(fcall.args[1], "P8ZP_SCRATCH_W2", fcall.type)     // right
+            if(signed) {
+                asmgen.out("""
+                    lda  P8ZP_SCRATCH_W1
+                    ldy  P8ZP_SCRATCH_W1+1
+                    cmp  P8ZP_SCRATCH_W2
+                    tya
+                    sbc  P8ZP_SCRATCH_W2+1
+                    bvc  +
+                    eor  #$80
++                   bmi  +
+                    lda  P8ZP_SCRATCH_W1                   
+                    ldy  P8ZP_SCRATCH_W1+1
+                    jmp  ++
++                   lda  P8ZP_SCRATCH_W2
+                    ldy  P8ZP_SCRATCH_W2+1
++""")
+            } else {
+                asmgen.out("""
+                    lda  P8ZP_SCRATCH_W1+1
+                    cmp  P8ZP_SCRATCH_W2+1
+                    bcc  ++
+                    bne  +
+                    lda  P8ZP_SCRATCH_W1
+                    cmp  P8ZP_SCRATCH_W2
+                    bcc  ++
++                   lda  P8ZP_SCRATCH_W1
+                    ldy  P8ZP_SCRATCH_W1+1
+                    jmp  ++
++                   lda  P8ZP_SCRATCH_W2
+                    ldy  P8ZP_SCRATCH_W2+1
++""")
+            }
+            if(resultToStack) {
+                asmgen.out("  sta  P8ESTACK_LO,x |  sty  P8ESTACK_HI,x |  dex")
+            } else {
+                val targetReg = AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.AY, signed, fcall.position, fcall.definingISub(), asmgen)
+                asmgen.assignRegister(RegisterOrPair.AY, targetReg)
+            }
+        } else {
+            throw AssemblyError("max float not supported")
+        }
+    }
+
     private fun funcMkword(fcall: PtBuiltinFunctionCall, resultToStack: Boolean, resultRegister: RegisterOrPair?) {
         if(resultToStack) {
-            asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.Y)      // msb
+            asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.A)      // msb
+            asmgen.out("  pha")
             asmgen.assignExpressionToRegister(fcall.args[1], RegisterOrPair.A)      // lsb
-            asmgen.out("  sta  P8ESTACK_LO,x |  tya |  sta  P8ESTACK_HI,x |  dex")
+            asmgen.out("  sta  P8ESTACK_LO,x |  pla |  sta  P8ESTACK_HI,x |  dex")
         } else {
             val reg = resultRegister ?: RegisterOrPair.AY
             var needAsave = asmgen.needAsaveForExpr(fcall.args[0])
@@ -962,9 +1160,27 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                         //       and will not generate another cmp when lsb() is directly used inside a comparison expression.
                     }
                     RegisterOrPair.Y -> {
-                        asmgen.out("  pha")
                         asmgen.assignExpressionToRegister(fcall.args.single(), RegisterOrPair.AY)
-                        asmgen.out("  tay |  pla |  cpy  #0")
+                        asmgen.out("  tay |  cpy  #0")
+                    }
+                    RegisterOrPair.AY -> {
+                        asmgen.assignExpressionToRegister(fcall.args.single(), RegisterOrPair.AY)
+                        asmgen.out("  ldy  #0 |  cmp  #0")
+                    }
+                    RegisterOrPair.AX -> {
+                        asmgen.assignExpressionToRegister(fcall.args.single(), RegisterOrPair.AX)
+                        asmgen.out("  ldx  #0 |  cmp  #0")
+                    }
+                    RegisterOrPair.XY -> {
+                        asmgen.assignExpressionToRegister(fcall.args.single(), RegisterOrPair.XY)
+                        asmgen.out("  ldy  #0 |  cpx  #0")
+                    }
+                    in Cx16VirtualRegisters -> {
+                        asmgen.assignExpressionToRegister(fcall.args.single(), resultRegister)
+                        val zero = PtNumber(DataType.UBYTE, 0.0, Position.DUMMY)
+                        zero.parent=fcall
+                        assignAsmGen.assignExpressionToVariable(zero, "cx16.${resultRegister.toString().lowercase()}H", DataType.UBYTE)
+                        asmgen.out("  lda  cx16.r0L")
                     }
                     else -> throw AssemblyError("invalid reg")
                 }

codeGenCpu6502/src/prog8/codegen/cpu6502/ExpressionsAsmGen.kt

@@ -237,6 +237,9 @@ internal class ExpressionsAsmGen(private val program: PtProgram,
             DataType.FLOAT -> {
                 asmgen.out(" lda  #<$varname |  ldy  #>$varname|  jsr  floats.push_float")
             }
+            in SplitWordArrayTypes -> {
+                throw AssemblyError("can't push address of split-word array ${expr.position}")
+            }
             in IterableDatatypes -> {
                 asmgen.out("  lda  #<$varname  |  sta  P8ESTACK_LO,x  |  lda  #>$varname |  sta  P8ESTACK_HI,x |  dex")
             }
@@ -245,8 +248,6 @@ internal class ExpressionsAsmGen(private val program: PtProgram,
     }
 
     private fun translateExpression(expr: PtBinaryExpression) {
-        require(!asmgen.options.useNewExprCode)
-
         // Uses evalstack to evaluate the given expression.  THIS IS SLOW AND SHOULD BE AVOIDED!
         if(translateSomewhatOptimized(expr.left, expr.operator, expr.right))
             return
@@ -761,6 +762,9 @@ internal class ExpressionsAsmGen(private val program: PtProgram,
             return
         }
 
+        if(arrayExpr.splitWords)
+            TODO("split words expression ${arrayExpr.position}")
+
         val constIndexNum = arrayExpr.index.asConstInteger()
         if(constIndexNum!=null) {
             val indexValue = constIndexNum * program.memsizer.memorySize(elementDt)

codeGenCpu6502/src/prog8/codegen/cpu6502/ForLoopsAsmGen.kt

@@ -51,7 +51,35 @@ internal class ForLoopsAsmGen(private val program: PtProgram,
                     // loop over byte range via loopvar
                     val varname = asmgen.asmVariableName(stmt.variable)
                     asmgen.assignExpressionToVariable(range.from, varname, ArrayToElementTypes.getValue(iterableDt))
-                    asmgen.assignExpressionToVariable(range.to, "$modifiedLabel+1", ArrayToElementTypes.getValue(iterableDt))
+                    asmgen.assignExpressionToRegister(range.to, RegisterOrPair.A, false)
+                    // pre-check for end already reached
+                    if(iterableDt==DataType.ARRAY_B) {
+                        asmgen.out("  sta  $modifiedLabel+1")
+                        if(stepsize<0)
+                            asmgen.out("""
+                                clc
+                                sbc  $varname
+                                bvc  +
+                                eor  #${'$'}80
++                               bpl  $endLabel""")
+                        else
+                            asmgen.out("""
+                                sec
+                                sbc  $varname
+                                bvc  +
+                                eor  #${'$'}80
++                               bmi  $endLabel""")
+                    } else {
+                        if(stepsize<0)
+                            asmgen.out("""
+                                cmp  $varname
+                                beq  +
+                                bcs  $endLabel
++""")
+                        else
+                            asmgen.out("  cmp  $varname  |  bcc  $endLabel")
+                        asmgen.out("  sta  $modifiedLabel+1")
+                    }
                     asmgen.out(loopLabel)
                     asmgen.translate(stmt.statements)
                     asmgen.out("""
@@ -69,7 +97,35 @@ $modifiedLabel          cmp  #0         ; modified
                     // loop over byte range via loopvar
                     val varname = asmgen.asmVariableName(stmt.variable)
                     asmgen.assignExpressionToVariable(range.from, varname, ArrayToElementTypes.getValue(iterableDt))
-                    asmgen.assignExpressionToVariable(range.to, "$modifiedLabel+1", ArrayToElementTypes.getValue(iterableDt))
+                    asmgen.assignExpressionToRegister(range.to, RegisterOrPair.A, false)
+                    // pre-check for end already reached
+                    if(iterableDt==DataType.ARRAY_B) {
+                        asmgen.out("  sta  $modifiedLabel+1")
+                        if(stepsize<0)
+                            asmgen.out("""
+                                clc
+                                sbc  $varname
+                                bvc  +
+                                eor  #${'$'}80
++                               bpl  $endLabel""")
+                        else
+                            asmgen.out("""
+                                sec
+                                sbc  $varname
+                                bvc  +
+                                eor  #${'$'}80
++                               bmi  $endLabel""")
+                    } else {
+                        if(stepsize<0)
+                            asmgen.out("""
+                                cmp  $varname
+                                beq  +
+                                bcs  $endLabel
++""")
+                        else
+                            asmgen.out("  cmp  $varname  |  bcc  $endLabel")
+                        asmgen.out("  sta  $modifiedLabel+1")
+                    }
                     asmgen.out(loopLabel)
                     asmgen.translate(stmt.statements)
                     if(stepsize>0) {
@@ -100,8 +156,9 @@ $modifiedLabel              cmp  #0     ; modified
 
                     stepsize == 1 || stepsize == -1 -> {
                         val varname = asmgen.asmVariableName(stmt.variable)
-                        assignLoopvar(stmt, range)
+                        assignLoopvarWord(stmt, range)
                         asmgen.assignExpressionToRegister(range.to, RegisterOrPair.AY)
+                        precheckFromToWord(iterableDt, stepsize, varname, endLabel)
                         asmgen.out("""
                             sty  $modifiedLabel+1
                             sta  $modifiedLabel2+1
@@ -134,8 +191,9 @@ $modifiedLabel2 cmp  #0    ; modified
 
                         // (u)words, step >= 2
                         val varname = asmgen.asmVariableName(stmt.variable)
-                        assignLoopvar(stmt, range)
+                        assignLoopvarWord(stmt, range)
                         asmgen.assignExpressionToRegister(range.to, RegisterOrPair.AY)
+                        precheckFromToWord(iterableDt, stepsize, varname, endLabel)
                         asmgen.out("""
                             sty  $modifiedLabel+1
                             sta  $modifiedLabel2+1
@@ -182,31 +240,15 @@ $endLabel""")
 
                         // (u)words, step <= -2
                         val varname = asmgen.asmVariableName(stmt.variable)
-                        assignLoopvar(stmt, range)
+                        assignLoopvarWord(stmt, range)
                         asmgen.assignExpressionToRegister(range.to, RegisterOrPair.AY)
+                        precheckFromToWord(iterableDt, stepsize, varname, endLabel)
                         asmgen.out("""
                             sty  $modifiedLabel+1
                             sta  $modifiedLabel2+1
 $loopLabel""")
                         asmgen.translate(stmt.statements)
 
-                        if(iterableDt==DataType.ARRAY_UW) {
-                            asmgen.out("""
-                lda  $varname
-                sec
-                sbc  #<${stepsize.absoluteValue}
-                sta  $varname
-                lda  $varname+1
-                sbc  #>${stepsize.absoluteValue}
-                sta  $varname+1
-$modifiedLabel  cmp  #0    ; modified                
-                bcc  $endLabel
-                bne  $loopLabel
-                lda  $varname
-$modifiedLabel2 cmp  #0    ; modified 
-                bcs  $loopLabel
-$endLabel""")
-                        } else {
                             asmgen.out("""
                 lda  $varname
                 sec
@@ -224,7 +266,6 @@ $modifiedLabel  sbc  #0    ; modified
                 eor  #$80
 +               bpl  $loopLabel                
 $endLabel""")
-                        }
                     }
                 }
             }
@@ -234,6 +275,56 @@ $endLabel""")
         asmgen.loopEndLabels.pop()
     }
 
+    private fun precheckFromToWord(iterableDt: DataType, stepsize: Int, fromVar: String, endLabel: String) {
+        // pre-check for end already reached.
+        // 'to' is in AY, do NOT clobber this!
+        if(iterableDt==DataType.ARRAY_W) {
+            if(stepsize<0)
+                asmgen.out("""
+                    sta  P8ZP_SCRATCH_W2        ; to
+                    sty  P8ZP_SCRATCH_W2+1      ; to
+                    lda  $fromVar
+	                cmp  P8ZP_SCRATCH_W2
+                    lda  $fromVar+1
+	                sbc  P8ZP_SCRATCH_W2+1
+	                bvc  +
+	                eor  #${'$'}80
++           		bmi  $endLabel
+                    lda  P8ZP_SCRATCH_W2
+                    ldy  P8ZP_SCRATCH_W2+1""")
+            else
+                asmgen.out("""
+                    sta  P8ZP_SCRATCH_REG
+	                cmp  $fromVar
+	                tya
+	                sbc  $fromVar+1
+	                bvc  +
+	                eor  #${'$'}80
++           		bmi  $endLabel
+                    lda  P8ZP_SCRATCH_REG""")
+        } else {
+            if(stepsize<0)
+                asmgen.out("""
+                    cpy  $fromVar+1
+                    beq  +
+                    bcc  ++
+                    bcs  $endLabel
++                   cmp  $fromVar
+                    bcc  +
+                    beq  +
+                    bne  $endLabel
++""")
+            else
+                asmgen.out("""
+                	cpy  $fromVar+1
+	                bcc  $endLabel
+	                bne  +
+	                cmp  $fromVar
+	                bcc  $endLabel
++""")
+        }
+    }
+
     private fun translateForOverIterableVar(stmt: PtForLoop, iterableDt: DataType, ident: PtIdentifier) {
         val loopLabel = asmgen.makeLabel("for_loop")
         val endLabel = asmgen.makeLabel("for_end")
@@ -332,7 +423,46 @@ $loopLabel          sty  $indexVar
                     // allocate index var on ZP if possible
                     val result = zeropage.allocate(indexVar, DataType.UBYTE, null, stmt.position, asmgen.errors)
                     result.fold(
-                        success = { (address,_)-> asmgen.out("""$indexVar = $address  ; auto zp UBYTE""") },
+                        success = { (address,_,_)-> asmgen.out("""$indexVar = $address  ; auto zp UBYTE""") },
+                        failure = { asmgen.out("$indexVar    .byte  0") }
+                    )
+                } else {
+                    asmgen.out("$indexVar    .byte  0")
+                }
+                asmgen.out(endLabel)
+            }
+            DataType.ARRAY_UW_SPLIT, DataType.ARRAY_W_SPLIT -> {
+                numElements!!
+                val indexVar = asmgen.makeLabel("for_index")
+                val loopvarName = asmgen.asmVariableName(stmt.variable)
+                asmgen.out("""
+                    ldy  #0
+$loopLabel          sty  $indexVar
+                    lda  ${iterableName}_lsb,y
+                    sta  $loopvarName
+                    lda  ${iterableName}_msb,y
+                    sta  $loopvarName+1""")
+                asmgen.translate(stmt.statements)
+                if(numElements<=255u) {
+                    asmgen.out("""
+                        ldy  $indexVar
+                        iny
+                        cpy  #$numElements
+                        beq  $endLabel
+                        bne  $loopLabel""")
+                } else {
+                    // length is 256
+                    asmgen.out("""
+                        ldy  $indexVar
+                        iny
+                        bne  $loopLabel
+                        beq  $endLabel""")
+                }
+                if(numElements>=16u) {
+                    // allocate index var on ZP if possible
+                    val result = zeropage.allocate(indexVar, DataType.UBYTE, null, stmt.position, asmgen.errors)
+                    result.fold(
+                        success = { (address,_,_)-> asmgen.out("""$indexVar = $address  ; auto zp UBYTE""") },
                         failure = { asmgen.out("$indexVar    .byte  0") }
                     )
                 } else {
@@ -590,7 +720,7 @@ $loopLabel""")
         asmgen.loopEndLabels.pop()
     }
 
-    private fun assignLoopvar(stmt: PtForLoop, range: PtRange) =
+    private fun assignLoopvarWord(stmt: PtForLoop, range: PtRange) =
         asmgen.assignExpressionToVariable(
             range.from,
             asmgen.asmVariableName(stmt.variable),

codeGenCpu6502/src/prog8/codegen/cpu6502/FunctionCallAsmGen.kt

@@ -164,7 +164,7 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
             if(requiredDt!=value.type)
                 throw AssemblyError("for statusflag, byte value is required")
             if (statusflag == Statusflag.Pc) {
-                // this param needs to be set last, right before the jsr
+                // this boolean param needs to be set last, right before the jsr
                 // for now, this is already enforced on the subroutine definition by the Ast Checker
                 when(value) {
                     is PtNumber -> {
@@ -183,12 +183,7 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
                     }
                     else -> {
                         asmgen.assignExpressionToRegister(value, RegisterOrPair.A)
-                        asmgen.out("""
-                            beq  +
-                            sec
-                            bcs  ++
-+                           clc
-+""")
+                        asmgen.out("  ror  a")
                     }
                 }
             } else throw AssemblyError("can only use Carry as status flag parameter")

codeGenCpu6502/src/prog8/codegen/cpu6502/PostIncrDecrAsmGen.kt

@@ -65,6 +65,31 @@ internal class PostIncrDecrAsmGen(private val program: PtProgram, private val as
                 val asmArrayvarname = asmgen.asmVariableName(targetArrayIdx.variable)
                 val elementDt = targetArrayIdx.type
                 val constIndex = targetArrayIdx.index.asConstInteger()
+                if(targetArrayIdx.splitWords) {
+                    if(constIndex!=null) {
+                        if(incr)
+                            asmgen.out(" inc  ${asmArrayvarname}_lsb+$constIndex |  bne  + |  inc  ${asmArrayvarname}_msb+$constIndex |+")
+                        else
+                            asmgen.out("""
+        lda  ${asmArrayvarname}_lsb+$constIndex
+        bne  +
+        dec  ${asmArrayvarname}_msb+$constIndex
++       dec  ${asmArrayvarname}_lsb+$constIndex""")
+                    } else {
+                        asmgen.saveRegisterLocal(CpuRegister.X, scope!!)
+                        asmgen.loadScaledArrayIndexIntoRegister(targetArrayIdx, elementDt, CpuRegister.X)
+                        if(incr)
+                            asmgen.out(" inc  ${asmArrayvarname}_lsb,x |  bne  + |  inc  ${asmArrayvarname}_msb,x |+")
+                        else
+                            asmgen.out("""
+        lda  ${asmArrayvarname}_lsb,x
+        bne  +
+        dec  ${asmArrayvarname}_msb,x
++       dec  ${asmArrayvarname}_lsb,x""")
+                        asmgen.restoreRegisterLocal(CpuRegister.X)
+                    }
+                    return
+                }
                 if(constIndex!=null) {
                     val indexValue = constIndex * program.memsizer.memorySize(elementDt)
                     when(elementDt) {
@@ -74,11 +99,10 @@ internal class PostIncrDecrAsmGen(private val program: PtProgram, private val as
                                 asmgen.out(" inc  $asmArrayvarname+$indexValue |  bne  + |  inc  $asmArrayvarname+$indexValue+1 |+")
                             else
                                 asmgen.out("""
-    lda  $asmArrayvarname+$indexValue
-    bne  +
-    dec  $asmArrayvarname+$indexValue+1
-+       dec  $asmArrayvarname+$indexValue 
-""")
+        lda  $asmArrayvarname+$indexValue
+        bne  +
+        dec  $asmArrayvarname+$indexValue+1
++       dec  $asmArrayvarname+$indexValue""")
                         }
                         DataType.FLOAT -> {
                             asmgen.out("  lda  #<($asmArrayvarname+$indexValue) |  ldy  #>($asmArrayvarname+$indexValue)")
@@ -89,9 +113,8 @@ internal class PostIncrDecrAsmGen(private val program: PtProgram, private val as
                 }
                 else
                 {
-                    asmgen.loadScaledArrayIndexIntoRegister(targetArrayIdx, elementDt, CpuRegister.A)
                     asmgen.saveRegisterLocal(CpuRegister.X, scope!!)
-                    asmgen.out("  tax")
+                    asmgen.loadScaledArrayIndexIntoRegister(targetArrayIdx, elementDt, CpuRegister.X)
                     when(elementDt) {
                         in ByteDatatypes -> {
                             asmgen.out(if(incr) "  inc  $asmArrayvarname,x" else "  dec  $asmArrayvarname,x")
@@ -101,20 +124,20 @@ internal class PostIncrDecrAsmGen(private val program: PtProgram, private val as
                                 asmgen.out(" inc  $asmArrayvarname,x |  bne  + |  inc  $asmArrayvarname+1,x |+")
                             else
                                 asmgen.out("""
-    lda  $asmArrayvarname,x
-    bne  +
-    dec  $asmArrayvarname+1,x
+        lda  $asmArrayvarname,x
+        bne  +
+        dec  $asmArrayvarname+1,x
 +       dec  $asmArrayvarname,x 
 """)
                         }
                         DataType.FLOAT -> {
                             asmgen.out("""
-                    ldy  #>$asmArrayvarname
-                    clc
-                    adc  #<$asmArrayvarname
-                    bcc  +
-                    iny
-+                       jsr  floats.inc_var_f""")
+        ldy  #>$asmArrayvarname
+        clc
+        adc  #<$asmArrayvarname
+        bcc  +
+        iny
++       jsr  floats.inc_var_f""")
                         }
                         else -> throw AssemblyError("weird array elt dt")
                     }

codeGenCpu6502/src/prog8/codegen/cpu6502/ProgramAndVarsGen.kt

@@ -33,8 +33,9 @@ internal class ProgramAndVarsGen(
     internal fun generate() {
         header()
         val allBlocks = program.allBlocks()
-        if(allBlocks.first().name != "main")
-            throw AssemblyError("first block should be 'main'")
+
+        if(allBlocks.first().name != "p8_main" && allBlocks.first().name != "main")
+            throw AssemblyError("first block should be 'main' or 'p8_main'")
 
         if(errors.noErrors())  {
             program.allBlocks().forEach { block2asm(it) }
@@ -104,15 +105,15 @@ internal class ProgramAndVarsGen(
                         asmgen.out("+\t.word  0")
                         asmgen.out("prog8_entrypoint\t; assembly code starts here")
                         if(!options.noSysInit)
-                            asmgen.out("  jsr  ${compTarget.name}.init_system")
-                        asmgen.out("  jsr  ${compTarget.name}.init_system_phase2")
+                            asmgen.out("  jsr  sys.init_system")
+                        asmgen.out("  jsr  sys.init_system_phase2")
                     }
                     CbmPrgLauncherType.NONE -> {
                         asmgen.out("; ---- program without basic sys call ----")
                         asmgen.out("* = ${options.loadAddress.toHex()}")
                         if(!options.noSysInit)
-                            asmgen.out("  jsr  ${compTarget.name}.init_system")
-                        asmgen.out("  jsr  ${compTarget.name}.init_system_phase2")
+                            asmgen.out("  jsr  sys.init_system")
+                        asmgen.out("  jsr  sys.init_system_phase2")
                     }
                 }
             }
@@ -120,8 +121,8 @@ internal class ProgramAndVarsGen(
                 asmgen.out("; ---- atari xex program ----")
                 asmgen.out("* = ${options.loadAddress.toHex()}")
                 if(!options.noSysInit)
-                    asmgen.out("  jsr  ${compTarget.name}.init_system")
-                asmgen.out("  jsr  ${compTarget.name}.init_system_phase2")
+                    asmgen.out("  jsr  sys.init_system")
+                asmgen.out("  jsr  sys.init_system_phase2")
             }
         }
 
@@ -139,25 +140,24 @@ internal class ProgramAndVarsGen(
             "cx16" -> {
                 if(options.floats)
                     asmgen.out("  lda  #4 |  sta  $01")    // to use floats, make sure Basic rom is banked in
-                asmgen.out("  jsr  main.start")
-                asmgen.out("  jmp  ${compTarget.name}.cleanup_at_exit")
+                asmgen.out("  jsr  p8_main.p8_start")
+                asmgen.out("  jmp  sys.cleanup_at_exit")
             }
             "c64" -> {
-                asmgen.out("  jsr  main.start |  lda  #31 |  sta  $01")
+                asmgen.out("  jsr  p8_main.p8_start |  lda  #31 |  sta  $01")
                 if(!options.noSysInit)
-                    asmgen.out("  jmp  ${compTarget.name}.cleanup_at_exit")
+                    asmgen.out("  jmp  sys.cleanup_at_exit")
                 else
                     asmgen.out("  rts")
             }
             "c128" -> {
-                asmgen.out("  jsr  main.start")
-                // TODO c128: how to bank basic+kernal back in?
+                asmgen.out("  jsr  p8_main.p8_start |  lda  #0 |  sta ${"$"}ff00")
                 if(!options.noSysInit)
-                    asmgen.out("  jmp  ${compTarget.name}.cleanup_at_exit")
+                    asmgen.out("  jmp  sys.cleanup_at_exit")
                 else
                     asmgen.out("  rts")
             }
-            else -> asmgen.jmp("main.start")
+            else -> asmgen.jmp("p8_main.p8_start")
         }
     }
 
@@ -176,7 +176,8 @@ internal class ProgramAndVarsGen(
 
     private fun footer() {
         asmgen.out("; bss sections")
-        if(options.varsHigh) {
+        asmgen.out("PROG8_VARSHIGH_RAMBANK = ${options.varsHighBank ?: 1}")
+        if(options.varsHighBank!=null) {
             if(options.compTarget.machine.BSSHIGHRAM_START == 0u || options.compTarget.machine.BSSHIGHRAM_END==0u) {
                 throw AssemblyError("current compilation target hasn't got the high ram area properly defined")
             }
@@ -312,7 +313,9 @@ internal class ProgramAndVarsGen(
 
         asmgen.out("${sub.name}\t$asmStartScope")
 
-        val scope = symboltable.lookupOrElse(sub.scopedName) { throw AssemblyError("lookup") }
+        val scope = symboltable.lookupOrElse(sub.scopedName) {
+            throw AssemblyError("lookup ${sub.scopedName}")
+        }
         require(scope.type==StNodeType.SUBROUTINE)
         val varsInSubroutine = getVars(scope)
 
@@ -332,7 +335,7 @@ internal class ProgramAndVarsGen(
         asmsubs2asm(sub.children)
 
         // the main.start subroutine is the program's entrypoint and should perform some initialization logic
-        if(sub.name=="start" && sub.definingBlock()!!.name=="main")
+        if((sub.name=="start" || sub.name=="p8_start") && (sub.definingBlock()!!.name=="main" || sub.definingBlock()!!.name=="p8_main"))
             entrypointInitialization()
 
         if(functioncallAsmGen.optimizeIntArgsViaRegisters(sub)) {
@@ -481,8 +484,8 @@ internal class ProgramAndVarsGen(
         val vars = allocator.zeropageVars.filter { it.value.dt==DataType.STR }
         for (variable in vars) {
             val scopedName = variable.key
-            val svar = symboltable.flat.getValue(scopedName) as StStaticVariable
-            if(svar.onetimeInitializationStringValue!=null)
+            val svar = symboltable.lookup(scopedName) as? StStaticVariable
+            if(svar?.onetimeInitializationStringValue!=null)
                 result.add(ZpStringWithInitial(scopedName, variable.value, svar.onetimeInitializationStringValue!!))
         }
         return result
@@ -493,8 +496,8 @@ internal class ProgramAndVarsGen(
         val vars = allocator.zeropageVars.filter { it.value.dt in ArrayDatatypes }
         for (variable in vars) {
             val scopedName = variable.key
-            val svar = symboltable.flat.getValue(scopedName) as StStaticVariable
-            if(svar.onetimeInitializationArrayValue!=null)
+            val svar = symboltable.lookup(scopedName) as? StStaticVariable
+            if(svar?.onetimeInitializationArrayValue!=null)
                 result.add(ZpArrayWithInitial(scopedName, variable.value, svar.onetimeInitializationArrayValue!!))
         }
         return result
@@ -544,6 +547,11 @@ internal class ProgramAndVarsGen(
             DataType.UWORD -> asmgen.out("${variable.name}\t.word  ?")
             DataType.WORD -> asmgen.out("${variable.name}\t.sint  ?")
             DataType.FLOAT -> asmgen.out("${variable.name}\t.fill  ${compTarget.machine.FLOAT_MEM_SIZE}")
+            in SplitWordArrayTypes -> {
+                val numbytesPerHalf = compTarget.memorySize(variable.dt, variable.length!!) / 2
+                asmgen.out("${variable.name}_lsb\t.fill  $numbytesPerHalf")
+                asmgen.out("${variable.name}_msb\t.fill  $numbytesPerHalf")
+            }
             in ArrayDatatypes -> {
                 val numbytes = compTarget.memorySize(variable.dt, variable.length!!)
                 asmgen.out("${variable.name}\t.fill  $numbytes")
@@ -628,6 +636,18 @@ internal class ProgramAndVarsGen(
                         asmgen.out("  .sint  " + chunk.joinToString())
                 }
             }
+            DataType.ARRAY_UW_SPLIT -> {
+                val data = makeArrayFillDataUnsigned(dt, value, orNumberOfZeros)
+                asmgen.out("_array_$varname := ${data.joinToString()}")
+                asmgen.out("${varname}_lsb\t.byte <_array_$varname")
+                asmgen.out("${varname}_msb\t.byte >_array_$varname")
+            }
+            DataType.ARRAY_W_SPLIT -> {
+                val data = makeArrayFillDataSigned(dt, value, orNumberOfZeros)
+                asmgen.out("_array_$varname := ${data.joinToString()}")
+                asmgen.out("${varname}_lsb\t.byte <_array_$varname")
+                asmgen.out("${varname}_msb\t.byte >_array_$varname")
+            }
             DataType.ARRAY_F -> {
                 val array = value ?: zeroFilledArray(orNumberOfZeros!!)
                 val floatFills = array.map {
@@ -687,7 +707,7 @@ internal class ProgramAndVarsGen(
                     val number = it.number!!.toInt()
                     "$"+number.toString(16).padStart(2, '0')
                 }
-            DataType.ARRAY_UW -> array.map {
+            DataType.ARRAY_UW, DataType.ARRAY_UW_SPLIT -> array.map {
                 if(it.number!=null) {
                     "$" + it.number!!.toInt().toString(16).padStart(4, '0')
                 }
@@ -719,11 +739,11 @@ internal class ProgramAndVarsGen(
                     else
                         "-$$hexnum"
                 }
-            DataType.ARRAY_UW -> array.map {
+            DataType.ARRAY_UW, DataType.ARRAY_UW_SPLIT -> array.map {
                 val number = it.number!!.toInt()
                 "$" + number.toString(16).padStart(4, '0')
             }
-            DataType.ARRAY_W -> array.map {
+            DataType.ARRAY_W, DataType.ARRAY_W_SPLIT -> array.map {
                 val number = it.number!!.toInt()
                 val hexnum = number.absoluteValue.toString(16).padStart(4, '0')
                 if(number>=0)

codeGenCpu6502/src/prog8/codegen/cpu6502/assignment/AsmAssignment.kt

@@ -115,7 +115,7 @@ internal class AsmAssignTarget(val kind: TargetStorageKind,
                 left is PtIdentifier && left.name==scopedName
             }
             TargetStorageKind.ARRAY -> {
-                left is PtArrayIndexer && left isSameAs array!!
+                left is PtArrayIndexer && left isSameAs array!! && left.splitWords==array.splitWords
             }
             TargetStorageKind.MEMORY -> {
                 left isSameAs memory!!
@@ -178,8 +178,8 @@ internal class AsmAssignSource(val kind: SourceStorageKind,
                     AsmAssignSource(SourceStorageKind.EXPRESSION, program, asmgen, value.type, expression = value)
                 }
                 is PtFunctionCall -> {
-                    val symbol = asmgen.symbolTable.lookup(value.name)
-                    val sub = symbol!!.astNode as IPtSubroutine
+                    val symbol = asmgen.symbolTable.lookup(value.name) ?: throw AssemblyError("lookup error ${value.name}")
+                    val sub = symbol.astNode as IPtSubroutine
                     val returnType = sub.returnsWhatWhere().firstOrNull { rr -> rr.first.registerOrPair != null || rr.first.statusflag!=null }?.second
                             ?: throw AssemblyError("can't translate zero return values in assignment")
 

codeGenCpu6502/src/prog8/codegen/cpu6502/assignment/AugmentableAssignmentAsmGen.kt

@@ -93,8 +93,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                             SourceStorageKind.ARRAY -> inplaceModification_word_value_to_variable(target.asmVarname, target.datatype, operator, value.array!!)
                             SourceStorageKind.EXPRESSION -> {
                                 if(value.expression is PtTypeCast) {
-                                    if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator))
-                                        return
+                                    if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator)) return
                                     inplaceModification_word_value_to_variable(target.asmVarname, target.datatype, operator, value.expression)
                                 }
                                 else {
@@ -178,7 +177,6 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                             SourceStorageKind.ARRAY -> inplaceModification_byte_value_to_variable("P8ZP_SCRATCH_B1", DataType.UBYTE, operator, value.array!!)
                             SourceStorageKind.EXPRESSION -> {
                                 if(value.expression is PtTypeCast) {
-                                    if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator)) return
                                     inplaceModification_byte_value_to_variable("P8ZP_SCRATCH_B1", DataType.UBYTE, operator, value.expression)
                                 } else {
                                     inplaceModification_byte_value_to_variable("P8ZP_SCRATCH_B1", DataType.UBYTE, operator, value.expression!!)
@@ -195,7 +193,10 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                 val indexVar = target.array.index as? PtIdentifier
                 when {
                     indexNum!=null -> {
-                        val targetVarName = "${target.asmVarname} + ${indexNum.number.toInt()*program.memsizer.memorySize(target.datatype)}"
+                        val targetVarName = if(target.array.splitWords)
+                            "${target.asmVarname} + ${indexNum.number.toInt()}"
+                        else
+                            "${target.asmVarname} + ${indexNum.number.toInt()*program.memsizer.memorySize(target.datatype)}"
                         when (target.datatype) {
                             in ByteDatatypes -> {
                                 when(value.kind) {
@@ -257,6 +258,10 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                     indexVar!=null -> {
                         when (target.datatype) {
                             in ByteDatatypes -> {
+                                if(value.kind==SourceStorageKind.EXPRESSION
+                                    && value.expression is PtTypeCast
+                                    && tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator))
+                                    return
                                 asmgen.loadScaledArrayIndexIntoRegister(target.array, DataType.UBYTE, CpuRegister.Y)
                                 asmgen.out("  lda  ${target.array.variable.name},y |  sta  P8ZP_SCRATCH_B1")
                                 asmgen.saveRegisterLocal(CpuRegister.Y, target.scope!!)
@@ -268,8 +273,6 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                                     SourceStorageKind.ARRAY -> inplaceModification_byte_value_to_variable("P8ZP_SCRATCH_B1", target.datatype, operator, value.array!!)
                                     SourceStorageKind.EXPRESSION -> {
                                         if(value.expression is PtTypeCast) {
-                                            if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator))
-                                                return
                                             inplaceModification_byte_value_to_variable("P8ZP_SCRATCH_B1", target.datatype, operator, value.expression)
                                         } else {
                                             inplaceModification_byte_value_to_variable("P8ZP_SCRATCH_B1", target.datatype, operator, value.expression!!)
@@ -281,10 +284,19 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                                 asmgen.out("  lda  P8ZP_SCRATCH_B1 |  sta  ${target.array.variable.name},y")
                             }
                             in WordDatatypes -> {
+                                if(value.kind==SourceStorageKind.EXPRESSION
+                                    && value.expression is PtTypeCast
+                                    && tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator))
+                                    return
                                 asmgen.loadScaledArrayIndexIntoRegister(target.array, DataType.UWORD, CpuRegister.Y)
-                                asmgen.out("  lda  ${target.array.variable.name},y |  sta  P8ZP_SCRATCH_W1")
-                                asmgen.out("  iny |  lda  ${target.array.variable.name},y |  sta  P8ZP_SCRATCH_W1+1")
-                                asmgen.saveRegisterLocal(CpuRegister.Y, target.scope!!)
+                                if(target.array.splitWords) {
+                                    asmgen.out("  lda  ${target.array.variable.name}_lsb,y |  sta  P8ZP_SCRATCH_W1")
+                                    asmgen.out("  lda  ${target.array.variable.name}_msb,y |  sta  P8ZP_SCRATCH_W1+1")
+                                } else {
+                                    asmgen.out("  lda  ${target.array.variable.name},y |  sta  P8ZP_SCRATCH_W1")
+                                    asmgen.out("  lda  ${target.array.variable.name}+1,y |  sta  P8ZP_SCRATCH_W1+1")
+                                }
+                                asmgen.saveRegisterStack(CpuRegister.Y, false)
                                 when(value.kind) {
                                     SourceStorageKind.LITERALNUMBER -> inplaceModification_word_litval_to_variable("P8ZP_SCRATCH_W1", target.datatype, operator, value.number!!.number.toInt())
                                     SourceStorageKind.VARIABLE -> inplaceModification_word_variable_to_variable("P8ZP_SCRATCH_W1", target.datatype, operator, value.asmVarname, value.datatype)
@@ -293,8 +305,6 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                                     SourceStorageKind.ARRAY -> inplaceModification_word_value_to_variable("P8ZP_SCRATCH_W1", target.datatype, operator, value.array!!)
                                     SourceStorageKind.EXPRESSION -> {
                                         if(value.expression is PtTypeCast) {
-                                            if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator))
-                                                return
                                             inplaceModification_word_value_to_variable("P8ZP_SCRATCH_W1", target.datatype, operator, value.expression)
                                         } else {
                                             inplaceModification_word_value_to_variable("P8ZP_SCRATCH_W1", target.datatype, operator, value.expression!!)
@@ -302,9 +312,14 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                                     }
                                     else -> throw AssemblyError("weird source type ${value.kind}")
                                 }
-                                asmgen.restoreRegisterLocal(CpuRegister.Y)
-                                asmgen.out("  lda  P8ZP_SCRATCH_W1+1 |  sta  ${target.array.variable.name},y")
-                                asmgen.out("  lda  P8ZP_SCRATCH_W1 |  dey |  sta  ${target.array.variable.name},y")
+                                asmgen.restoreRegisterStack(CpuRegister.Y, false)
+                                if(target.array.splitWords) {
+                                    asmgen.out("  lda  P8ZP_SCRATCH_W1 |  sta  ${target.array.variable.name}_lsb,y")
+                                    asmgen.out("  lda  P8ZP_SCRATCH_W1+1 |  sta  ${target.array.variable.name}_msb,y")
+                                } else {
+                                    asmgen.out("  lda  P8ZP_SCRATCH_W1 |  sta  ${target.array.variable.name},y")
+                                    asmgen.out("  lda  P8ZP_SCRATCH_W1+1 |  sta  ${target.array.variable.name}+1,y")
+                                }
                             }
                             DataType.FLOAT -> {
                                 asmgen.loadScaledArrayIndexIntoRegister(target.array, DataType.FLOAT, CpuRegister.A)
@@ -585,8 +600,9 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                 asmgen.out("  clc |  adc  $name |  sta  $name")
             }
             "-" -> {
-                asmgen.assignExpressionToVariable(value, "P8ZP_SCRATCH_B1", dt)
-                asmgen.out(" lda  $name |  sec |  sbc  P8ZP_SCRATCH_B1 |  sta  $name")
+                val tmpByte = if(name!="P8ZP_SCRATCH_B1") "P8ZP_SCRATCH_B1" else "P8ZP_SCRATCH_REG"
+                asmgen.assignExpressionToVariable(value, tmpByte, dt)
+                asmgen.out(" lda  $name |  sec |  sbc  $tmpByte |  sta  $name")
             }
             "*" -> {
                 asmgen.assignExpressionToRegister(value, RegisterOrPair.A)
@@ -772,13 +788,12 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             "<" -> {
                 if(dt==DataType.UBYTE) {
                     asmgen.out("""
-                        lda  $name
-                        cmp  $otherName
-                        bcc  +
                         lda  #0
-                        beq  ++
-+                       lda  #1
-+                       sta  $name""")
+                        ldy  $name
+                        cpy  $otherName
+                        rol  a
+                        eor  #1
+                        sta  $name""")
                 }
                 else {
                     // see http://www.6502.org/tutorials/compare_beyond.html
@@ -798,13 +813,11 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             "<=" -> {
                 if(dt==DataType.UBYTE) {
                     asmgen.out("""
-                        lda  $otherName
-                        cmp  $name
-                        bcs  +
                         lda  #0
-                        beq  ++
-+                       lda  #1
-+                       sta  $name""")
+                        ldy  $otherName
+                        cpy  $name
+                        rol  a
+                        sta  $name""")
                 } else {
                     // see http://www.6502.org/tutorials/compare_beyond.html
                     asmgen.out("""
@@ -823,13 +836,11 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             ">" -> {
                 if(dt==DataType.UBYTE) {
                     asmgen.out("""
-                        lda  $name
-                        cmp  $otherName
+                        lda  #0
+                        ldy  $name
+                        cpy  $otherName
                         beq  +
-                        bcs  ++
-+                       lda  #0
-                        beq  ++
-+                       lda  #1
+                        rol  a
 +                       sta  $name""")
                 } else {
                     // see http://www.6502.org/tutorials/compare_beyond.html
@@ -849,13 +860,11 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             ">=" -> {
                 if(dt==DataType.UBYTE) {
                     asmgen.out("""
-                        lda  $name
-                        cmp  $otherName
-                        bcs  +
                         lda  #0
-                        beq  ++
-+                       lda  #1
-+                       sta  $name""")
+                        ldy  $name
+                        cpy  $otherName
+                        rol  a
+                        sta  $name""")
                 } else {
                     // see http://www.6502.org/tutorials/compare_beyond.html
                     asmgen.out("""
@@ -963,13 +972,12 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             "<" -> {
                 if(dt==DataType.UBYTE) {
                     asmgen.out("""
-                        lda  $name
-                        cmp  #$value
-                        bcc  +
                         lda  #0
-                        beq  ++
-+                       lda  #1
-+                       sta  $name""")
+                        ldy  $name
+                        cpy  #$value
+                        rol  a
+                        eor  #1
+                        sta  $name""")
                 }
                 else {
                     // see http://www.6502.org/tutorials/compare_beyond.html
@@ -989,13 +997,11 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             "<=" -> {
                 if(dt==DataType.UBYTE) {
                     asmgen.out("""
-                        lda  #$value
-                        cmp  $name
-                        bcs  +
                         lda  #0
-                        beq  ++
-+                       lda  #1
-+                       sta  $name""")
+                        ldy  #$value
+                        cpy  $name
+                        rol  a
+                        sta  $name""")
                 } else {
                     // see http://www.6502.org/tutorials/compare_beyond.html
                     asmgen.out("""
@@ -1014,13 +1020,11 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             ">" -> {
                 if(dt==DataType.UBYTE) {
                     asmgen.out("""
-                        lda  $name
-                        cmp  #$value
+                        lda  #0
+                        ldy  $name
+                        cpy  #$value
                         beq  +
-                        bcs  ++
-+                       lda  #0
-                        beq  ++
-+                       lda  #1
+                        rol  a
 +                       sta  $name""")
                 } else {
                     // see http://www.6502.org/tutorials/compare_beyond.html
@@ -1040,13 +1044,11 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             ">=" -> {
                 if(dt==DataType.UBYTE) {
                     asmgen.out("""
-                        lda  $name
-                        cmp  #$value
-                        bcs  +
                         lda  #0
-                        beq  ++
-+                       lda  #1
-+                       sta  $name""")
+                        ldy  $name
+                        cpy  #$value
+                        rol  a
+                        sta  $name""")
                 } else {
                     // see http://www.6502.org/tutorials/compare_beyond.html
                     asmgen.out("""
@@ -1077,11 +1079,12 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             }
             "-" -> {
                 asmgen.translateDirectMemReadExpressionToRegAorStack(memread, false)
+                val tmpByte = if(name!="P8ZP_SCRATCH_B1") "P8ZP_SCRATCH_B1" else "P8ZP_SCRATCH_REG"
                 asmgen.out("""
-                    sta  P8ZP_SCRATCH_B1
+                    sta  $tmpByte
                     lda  $name
                     sec
-                    sbc  P8ZP_SCRATCH_B1
+                    sbc  $tmpByte
                     sta  $name""")
             }
             "|" -> {
@@ -1116,11 +1119,12 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             }
             "-" -> {
                 asmgen.translateDirectMemReadExpressionToRegAorStack(memread, false)
+                val tmpByte = if(name!="P8ZP_SCRATCH_B1") "P8ZP_SCRATCH_B1" else "P8ZP_SCRATCH_REG"
                 asmgen.out("""
-                    sta  P8ZP_SCRATCH_B1
+                    sta  $tmpByte
                     lda  $name
                     sec
-                    sbc  P8ZP_SCRATCH_B1
+                    sbc  $tmpByte
                     sta  $name
                     bcc  +
                     dec  $name+1
@@ -1217,10 +1221,8 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                         lda  #<$value
                         ldy  #>$value
                         jsr  math.multiply_words
-                        lda  math.multiply_words.result
                         sta  $name
-                        lda  math.multiply_words.result+1
-                        sta  $name+1""")
+                        sty  $name+1""")
                 }
             }
             "/" -> {
@@ -1678,10 +1680,8 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                                 lda  $name
                                 ldy  $name+1
                                 jsr  math.multiply_words
-                                lda  math.multiply_words.result
                                 sta  $name
-                                lda  math.multiply_words.result+1
-                                sta  $name+1""")
+                                sty  $name+1""")
                     }
                     "/" -> {
                         if(dt==DataType.UWORD) {
@@ -1818,11 +1818,8 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                             lda  $name
                             ldy  $name+1
                             jsr  math.multiply_words
-                            lda  math.multiply_words.result
                             sta  $name
-                            lda  math.multiply_words.result+1
-                            sta  $name+1
-                        """)
+                            sty  $name+1""")
                     }
                     "/" -> {
                         if(dt==DataType.WORD) {
@@ -1835,8 +1832,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                                 ldy  $otherName+1
                                 jsr  math.divmod_w_asm
                                 sta  $name
-                                sty  $name+1
-                            """)
+                                sty  $name+1""")
                         }
                         else {
                             asmgen.out("""
@@ -1848,8 +1844,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                                 ldy  $otherName+1
                                 jsr  math.divmod_uw_asm
                                 sta  $name
-                                sty  $name+1
-                            """)
+                                sty  $name+1""")
                         }
                     }
                     "%" -> {
@@ -2005,8 +2000,6 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
     private fun inplaceModification_word_value_to_variable(name: String, dt: DataType, operator: String, value: PtExpression) {
         // this should be the last resort for code generation for this,
         // because the value is evaluated onto the eval stack (=slow).
-
-
         fun multiplyVarByWordInAY() {
             asmgen.out("""
                 sta  P8ZP_SCRATCH_W1
@@ -2014,10 +2007,8 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                 lda  $name
                 ldy  $name+1
                 jsr  math.multiply_words
-                lda  math.multiply_words.result
                 sta  $name
-                lda  math.multiply_words.result+1
-                sta  $name+1
+                sty  $name+1
             """)
         }
 
@@ -2218,8 +2209,10 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                         asmgen.out("  clc |  adc  $name |  sta  $name |  tya |  adc  $name+1 |  sta  $name+1")
                     }
                     "-" -> {
-                        asmgen.assignExpressionToVariable(value, "P8ZP_SCRATCH_W1", valueDt)
-                        asmgen.out(" lda  $name |  sec |  sbc  P8ZP_SCRATCH_W1 |  sta  $name |  lda  $name+1 |  sbc  P8ZP_SCRATCH_W1+1 |  sta  $name+1")
+                        val tmpWord = if(name!="P8ZP_SCRATCH_W1") "P8ZP_SCRATCH_W1" else "P8ZP_SCRATCH_W2"
+                        asmgen.assignExpressionToVariable(value, tmpWord, valueDt)
+                        asmgen.out(" lda  $name |  sec |  sbc  $tmpWord |  sta  $name |  lda  $name+1 |  sbc  $tmpWord+1 |  sta  $name+1")
+                        // TODO wrong signed word subtraction code??? ^^^
                     }
                     "*" -> {
                         asmgen.assignExpressionToRegister(value, RegisterOrPair.AY)

codeGenCpu6502/test/Dummies.kt

@@ -45,6 +45,10 @@ internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors:
             warnings.add(text)
     }
 
+    override fun undefined(symbol: List<String>, position: Position) {
+        err("undefined symbol: ${symbol.joinToString(".")}", position)
+    }
+
     override fun noErrors(): Boolean  = errors.isEmpty()
 
     override fun report() {

codeGenCpu6502/test/TestCodegen.kt

@@ -1,6 +1,9 @@
 package prog8tests.codegencpu6502
 
+import io.kotest.assertions.throwables.shouldNotThrowAny
+import io.kotest.assertions.withClue
 import io.kotest.core.spec.style.FunSpec
+import io.kotest.matchers.ints.shouldBeGreaterThanOrEqual
 import io.kotest.matchers.shouldBe
 import prog8.code.SymbolTableMaker
 import prog8.code.ast.*
@@ -39,9 +42,9 @@ class TestCodegen: FunSpec({
 //        xx += cx16.r0
 //    }
 //}
-        val codegen = AsmGen6502()
+        val codegen = AsmGen6502(prefixSymbols = false)
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         val sub = PtSub("start", emptyList(), null, Position.DUMMY)
         sub.add(PtVariable("pi", DataType.UBYTE, ZeropageWish.DONTCARE, PtNumber(DataType.UBYTE, 0.0, Position.DUMMY), null, Position.DUMMY))
         sub.add(PtVariable("particleX", DataType.ARRAY_UB, ZeropageWish.DONTCARE, null, 3u, Position.DUMMY))
@@ -91,7 +94,7 @@ class TestCodegen: FunSpec({
         program.add(block)
 
         // define the "cx16.r0" virtual register
-        val cx16block = PtBlock("cx16", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val cx16block = PtBlock("cx16", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         cx16block.add(PtMemMapped("r0", DataType.UWORD, 100u, null, Position.DUMMY))
         program.add(cx16block)
 
@@ -102,5 +105,24 @@ class TestCodegen: FunSpec({
         result.name shouldBe "test"
         Files.deleteIfExists(Path("${result.name}.asm"))
     }
+
+    test("64tass assembler available? - if this fails you need to install 64tass in the path") {
+        val command = mutableListOf("64tass", "--version")
+        shouldNotThrowAny {
+            val proc = ProcessBuilder(command).start()
+            val output = String(proc.inputStream.readBytes())
+            val result = proc.waitFor()
+            result.shouldBe(0)
+            val (_, version) = output.split('V')
+            val (major, minor, _) = version.split('.')
+            val majorNum = major.toInt()
+            val minorNum = minor.toInt()
+            withClue("64tass version should be 1.58 or newer") {
+                majorNum shouldBeGreaterThanOrEqual 1
+                if (majorNum == 1)
+                    minorNum shouldBeGreaterThanOrEqual 58
+            }
+        }
+    }
 })
 

codeGenExperimental/build.gradle

@@ -29,7 +29,7 @@ dependencies {
     implementation project(':codeGenIntermediate')
     implementation "org.jetbrains.kotlin:kotlin-stdlib-jdk8"
     // implementation "org.jetbrains.kotlin:kotlin-reflect"
-    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.16"
+    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18"
 
 }
 

codeGenExperimental/src/prog8/codegen/experimental/ExperiCodeGen.kt

@@ -2,7 +2,10 @@ package prog8.codegen.experimental
 
 import prog8.code.SymbolTable
 import prog8.code.ast.PtProgram
-import prog8.code.core.*
+import prog8.code.core.CompilationOptions
+import prog8.code.core.IAssemblyProgram
+import prog8.code.core.ICodeGeneratorBackend
+import prog8.code.core.IErrorReporter
 import prog8.codegen.intermediate.IRCodeGen
 import prog8.intermediate.IRFileWriter
 
@@ -23,6 +26,15 @@ class ExperiCodeGen: ICodeGeneratorBackend {
         IRFileWriter(irProgram, null).write()
 
         println("** experimental codegen stub: no assembly generated **")
-        return null
+        return EmptyProgram
     }
-}
+}
+
+private object EmptyProgram : IAssemblyProgram {
+    override val name = "<Empty Program>"
+    override fun assemble(options: CompilationOptions, errors: IErrorReporter): Boolean {
+        println("** nothing assembled **")
+        return true
+    }
+
+}

codeGenIntermediate/build.gradle

@@ -28,9 +28,9 @@ dependencies {
     implementation project(':intermediate')
     implementation "org.jetbrains.kotlin:kotlin-stdlib-jdk8"
     // implementation "org.jetbrains.kotlin:kotlin-reflect"
-    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.16"
+    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18"
 
-    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.5.5'
+    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.6.2'
 }
 
 sourceSets {

codeGenIntermediate/src/prog8/codegen/intermediate/AssignmentGen.kt

@@ -46,7 +46,7 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
         assignment: PtAugmentedAssign
     ): IRCodeChunks {
         val value = assignment.value
-        val vmDt = codeGen.irType(value.type)
+        val vmDt = irType(value.type)
         return when(assignment.operator) {
             "+" -> expressionEval.operatorPlusInplace(address, null, vmDt, value)
             "-" -> expressionEval.operatorMinusInplace(address, null, vmDt, value)
@@ -72,7 +72,7 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
 
     private fun assignVarAugmented(symbol: String, assignment: PtAugmentedAssign): IRCodeChunks {
         val value = assignment.value
-        val targetDt = codeGen.irType(assignment.target.type)
+        val targetDt = irType(assignment.target.type)
         return when (assignment.operator) {
             "+=" -> expressionEval.operatorPlusInplace(null, symbol, targetDt, value)
             "-=" -> expressionEval.operatorMinusInplace(null, symbol, targetDt, value)
@@ -104,27 +104,10 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
             value.add(origAssign.value)
         } else {
             require(origAssign.operator.endsWith('='))
-            if(codeGen.options.useNewExprCode) {
-                // X += Y  ->   temp = X,  temp += Y,  X = temp
-                val tempvar = codeGen.getReusableTempvar(origAssign.definingSub()!!, origAssign.target.type)
-                val assign = PtAssignment(origAssign.position)
-                val target = PtAssignTarget(origAssign.position)
-                target.add(tempvar)
-                assign.add(target)
-                assign.add(origAssign.target.children.single())
-                val augAssign = PtAugmentedAssign(origAssign.operator, origAssign.position)
-                augAssign.add(target)
-                augAssign.add(origAssign.value)
-                val assignBack = PtAssignment(origAssign.position)
-                assignBack.add(origAssign.target)
-                assignBack.add(tempvar)
-                return translateRegularAssign(assign) + translate(augAssign) + translateRegularAssign(assignBack)
-            } else {
-                value = PtBinaryExpression(origAssign.operator.dropLast(1), origAssign.value.type, origAssign.value.position)
-                val left: PtExpression = origAssign.target.children.single() as PtExpression
-                value.add(left)
-                value.add(origAssign.value)
-            }
+            value = PtBinaryExpression(origAssign.operator.dropLast(1), origAssign.value.type, origAssign.value.position)
+            val left: PtExpression = origAssign.target.children.single() as PtExpression
+            value.add(left)
+            value.add(origAssign.value)
         }
         val normalAssign = PtAssignment(origAssign.position)
         normalAssign.add(origAssign.target)
@@ -138,16 +121,16 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
             "+" -> { }
             "-" -> {
                 code += if(address!=null)
-                    IRInstruction(Opcode.NEGM, vmDt, value = address)
+                    IRInstruction(Opcode.NEGM, vmDt, address = address)
                 else
                     IRInstruction(Opcode.NEGM, vmDt, labelSymbol = symbol)
             }
             "~" -> {
                 val regMask = codeGen.registers.nextFree()
                 val mask = if(vmDt==IRDataType.BYTE) 0x00ff else 0xffff
-                code += IRInstruction(Opcode.LOAD, vmDt, reg1=regMask, value = mask)
+                code += IRInstruction(Opcode.LOAD, vmDt, reg1=regMask, immediate = mask)
                 code += if(address!=null)
-                    IRInstruction(Opcode.XORM, vmDt, reg1=regMask, value = address)
+                    IRInstruction(Opcode.XORM, vmDt, reg1=regMask, address = address)
                 else
                     IRInstruction(Opcode.XORM, vmDt, reg1=regMask, labelSymbol = symbol)
             }
@@ -161,7 +144,8 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
         val targetIdent = assignment.target.identifier
         val targetMemory = assignment.target.memory
         val targetArray = assignment.target.array
-        val vmDt = codeGen.irType(assignment.value.type)
+        val valueDt = irType(assignment.value.type)
+        val targetDt = irType(assignment.target.type)
         val result = mutableListOf<IRCodeChunkBase>()
 
         var valueRegister = -1
@@ -169,30 +153,45 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
         val zero = codeGen.isZero(assignment.value)
         if(!zero) {
             // calculate the assignment value
-            if (vmDt == IRDataType.FLOAT) {
+            if (valueDt == IRDataType.FLOAT) {
                 val tr = expressionEval.translateExpression(assignment.value)
                 valueFpRegister = tr.resultFpReg
                 addToResult(result, tr, -1, valueFpRegister)
             } else {
+                val extendByteToWord = if(targetDt != valueDt) {
+                    // usually an error EXCEPT when a byte is assigned to a word.
+                    if(targetDt==IRDataType.WORD && valueDt==IRDataType.BYTE)
+                        true
+                    else
+                        throw AssemblyError("assignment value and target dt mismatch")
+                } else false
                 if (assignment.value is PtMachineRegister) {
                     valueRegister = (assignment.value as PtMachineRegister).register
+                    if(extendByteToWord) {
+                        valueRegister = codeGen.registers.nextFree()
+                        addInstr(result, IRInstruction(Opcode.EXT, IRDataType.BYTE, reg1=valueRegister, reg2=(assignment.value as PtMachineRegister).register), null)
+                    }
                 } else {
                     val tr = expressionEval.translateExpression(assignment.value)
                     valueRegister = tr.resultReg
                     addToResult(result, tr, valueRegister, -1)
+                    if(extendByteToWord) {
+                        valueRegister = codeGen.registers.nextFree()
+                        val opcode = if(assignment.value.type in SignedDatatypes) Opcode.EXTS else Opcode.EXT
+                        addInstr(result, IRInstruction(opcode, IRDataType.BYTE, reg1=valueRegister, reg2=tr.resultReg), null)
+                    }
                 }
             }
         }
 
         if(targetIdent!=null) {
             val instruction = if(zero) {
-                IRInstruction(Opcode.STOREZM, vmDt, labelSymbol = targetIdent.name)
+                IRInstruction(Opcode.STOREZM, targetDt, labelSymbol = targetIdent.name)
             } else {
-                if (vmDt == IRDataType.FLOAT) {
-                    IRInstruction(Opcode.STOREM, vmDt, fpReg1 = valueFpRegister, labelSymbol = targetIdent.name)
-                }
+                if (targetDt == IRDataType.FLOAT)
+                    IRInstruction(Opcode.STOREM, targetDt, fpReg1 = valueFpRegister, labelSymbol = targetIdent.name)
                 else
-                    IRInstruction(Opcode.STOREM, vmDt, reg1 = valueRegister, labelSymbol = targetIdent.name)
+                    IRInstruction(Opcode.STOREM, targetDt, reg1 = valueRegister, labelSymbol = targetIdent.name)
             }
             result += IRCodeChunk(null, null).also { it += instruction }
             return result
@@ -214,70 +213,105 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
                 if(zero) {
                     // there's no STOREZIX instruction
                     valueRegister = codeGen.registers.nextFree()
-                    code += IRInstruction(Opcode.LOAD, vmDt, reg1=valueRegister, value=0)
+                    code += IRInstruction(Opcode.LOAD, targetDt, reg1=valueRegister, immediate = 0)
                 }
-                code += IRInstruction(Opcode.STOREIX, vmDt, reg1=valueRegister, reg2=idxReg, labelSymbol = variable)
+                code += IRInstruction(Opcode.STOREIX, targetDt, reg1=valueRegister, reg2=idxReg, labelSymbol = variable)
                 result += code
                 return result
             }
 
             val fixedIndex = constIntValue(targetArray.index)
+            val arrayLength = codeGen.symbolTable.getLength(targetArray.variable.name)
             if(zero) {
                 if(fixedIndex!=null) {
-                    val offset = fixedIndex*itemsize
-                    val chunk = IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREZM, vmDt, labelSymbol = "$variable+$offset") }
+                    val chunk = IRCodeChunk(null, null).also {
+                        if(targetArray.splitWords) {
+                            it += IRInstruction(Opcode.STOREZM, IRDataType.BYTE, immediate = arrayLength, labelSymbol = "${variable}_lsb+$fixedIndex")
+                            it += IRInstruction(Opcode.STOREZM, IRDataType.BYTE, immediate = arrayLength, labelSymbol = "${variable}_msb+$fixedIndex")
+                        }
+                        else
+                            it += IRInstruction(Opcode.STOREZM, targetDt, labelSymbol = "$variable+${fixedIndex*itemsize}")
+                    }
                     result += chunk
                 } else {
                     val (code, indexReg) = loadIndexReg(targetArray, itemsize)
                     result += code
-                    result += IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREZX, vmDt, reg1=indexReg, labelSymbol = variable) }
+                    result += IRCodeChunk(null, null).also {
+                        if(targetArray.splitWords) {
+                            it += IRInstruction(Opcode.STOREZX, IRDataType.BYTE, reg1 = indexReg, immediate = arrayLength, labelSymbol = variable+"_lsb")
+                            it += IRInstruction(Opcode.STOREZX, IRDataType.BYTE, reg1 = indexReg, immediate = arrayLength, labelSymbol = variable+"_msb")
+                        }
+                        else
+                            it += IRInstruction(Opcode.STOREZX, targetDt, reg1=indexReg, labelSymbol = variable)
+                    }
                 }
             } else {
-                if(vmDt== IRDataType.FLOAT) {
+                if(targetDt== IRDataType.FLOAT) {
                     if(fixedIndex!=null) {
                         val offset = fixedIndex*itemsize
-                        val chunk = IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREM, vmDt, fpReg1 = valueFpRegister, labelSymbol = "$variable+$offset") }
+                        val chunk = IRCodeChunk(null, null).also {
+                            it += IRInstruction(Opcode.STOREM, targetDt, fpReg1 = valueFpRegister, labelSymbol = "$variable+$offset")
+                        }
                         result += chunk
                     } else {
                         val (code, indexReg) = loadIndexReg(targetArray, itemsize)
                         result += code
-                        result += IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREX, vmDt, reg1 = indexReg, fpReg1 = valueFpRegister, labelSymbol = variable) }
+                        result += IRCodeChunk(null, null).also {
+                            it += IRInstruction(Opcode.STOREX, targetDt, reg1 = indexReg, fpReg1 = valueFpRegister, labelSymbol = variable)
+                        }
                     }
                 } else {
                     if(fixedIndex!=null) {
-                        val offset = fixedIndex*itemsize
-                        val chunk = IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREM, vmDt, reg1 = valueRegister, labelSymbol = "$variable+$offset") }
+                        val chunk = IRCodeChunk(null, null).also {
+                            if(targetArray.splitWords) {
+                                val msbReg = codeGen.registers.nextFree()
+                                it += IRInstruction(Opcode.STOREM, IRDataType.BYTE, reg1 = valueRegister, immediate = arrayLength, labelSymbol = "${variable}_lsb+$fixedIndex")
+                                it += IRInstruction(Opcode.MSIG, IRDataType.BYTE, reg1 = msbReg, reg2 = valueRegister)
+                                it += IRInstruction(Opcode.STOREM, IRDataType.BYTE, reg1 = msbReg, immediate = arrayLength, labelSymbol = "${variable}_msb+$fixedIndex")
+                            }
+                            else
+                                it += IRInstruction(Opcode.STOREM, targetDt, reg1 = valueRegister, labelSymbol = "$variable+${fixedIndex*itemsize}")
+                        }
                         result += chunk
                     } else {
                         val (code, indexReg) = loadIndexReg(targetArray, itemsize)
                         result += code
-                        result += IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREX, vmDt, reg1 = valueRegister, reg2=indexReg, labelSymbol = variable) }
+                        result += IRCodeChunk(null, null).also {
+                            if(targetArray.splitWords) {
+                                val msbReg = codeGen.registers.nextFree()
+                                it += IRInstruction(Opcode.STOREX, IRDataType.BYTE, reg1 = valueRegister, reg2=indexReg, immediate = arrayLength, labelSymbol = "${variable}_lsb")
+                                it += IRInstruction(Opcode.MSIG, IRDataType.BYTE, reg1 = msbReg, reg2 = valueRegister)
+                                it += IRInstruction(Opcode.STOREX, IRDataType.BYTE, reg1 = msbReg, reg2=indexReg, immediate = arrayLength, labelSymbol = "${variable}_msb")
+                            }
+                            else
+                                it += IRInstruction(Opcode.STOREX, targetDt, reg1 = valueRegister, reg2=indexReg, labelSymbol = variable)
+                        }
                     }
                 }
             }
             return result
         }
         else if(targetMemory!=null) {
-            require(vmDt== IRDataType.BYTE) { "must be byte type ${targetMemory.position}"}
+            require(targetDt == IRDataType.BYTE) { "must be byte type ${targetMemory.position}"}
             if(zero) {
                 if(targetMemory.address is PtNumber) {
-                    val chunk = IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREZM, vmDt, value=(targetMemory.address as PtNumber).number.toInt()) }
+                    val chunk = IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREZM, targetDt, address = (targetMemory.address as PtNumber).number.toInt()) }
                     result += chunk
                 } else {
                     val tr = expressionEval.translateExpression(targetMemory.address)
                     val addressReg = tr.resultReg
                     addToResult(result, tr, tr.resultReg, -1)
-                    result += IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREZI, vmDt, reg1=addressReg) }
+                    result += IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREZI, targetDt, reg1=addressReg) }
                 }
             } else {
                 if(targetMemory.address is PtNumber) {
-                    val chunk = IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREM, vmDt, reg1=valueRegister, value=(targetMemory.address as PtNumber).number.toInt()) }
+                    val chunk = IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREM, targetDt, reg1=valueRegister, address=(targetMemory.address as PtNumber).number.toInt()) }
                     result += chunk
                 } else {
                     val tr = expressionEval.translateExpression(targetMemory.address)
                     val addressReg = tr.resultReg
                     addToResult(result, tr, tr.resultReg, -1)
-                    result += IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREI, vmDt, reg1=valueRegister, reg2=addressReg) }
+                    result += IRCodeChunk(null, null).also { it += IRInstruction(Opcode.STOREI, targetDt, reg1=valueRegister, reg2=addressReg) }
                 }
             }
 
@@ -288,28 +322,21 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
     }
 
     private fun loadIndexReg(array: PtArrayIndexer, itemsize: Int): Pair<IRCodeChunks, Int> {
-        // returns the code to load the Index into the register, which is also return\ed.
+        // returns the code to load the Index into the register, which is also returned.
 
         val result = mutableListOf<IRCodeChunkBase>()
-        if(itemsize==1) {
+        if(itemsize==1 || array.splitWords) {
             val tr = expressionEval.translateExpression(array.index)
             addToResult(result, tr, tr.resultReg, -1)
             return Pair(result, tr.resultReg)
         }
 
-        if(codeGen.options.useNewExprCode) {
-            val tr = expressionEval.translateExpression(array.index)
-            result += tr.chunks
-            addInstr(result, IRInstruction(Opcode.MUL, tr.dt, reg1=tr.resultReg, value = itemsize), null)
-            return Pair(result, tr.resultReg)
-        } else {
-            val mult: PtExpression
-            mult = PtBinaryExpression("*", DataType.UBYTE, array.position)
-            mult.children += array.index
-            mult.children += PtNumber(DataType.UBYTE, itemsize.toDouble(), array.position)
-            val tr = expressionEval.translateExpression(mult)
-            addToResult(result, tr, tr.resultReg, -1)
-            return Pair(result, tr.resultReg)
-        }
+        val mult: PtExpression
+        mult = PtBinaryExpression("*", DataType.UBYTE, array.position)
+        mult.children += array.index
+        mult.children += PtNumber(DataType.UBYTE, itemsize.toDouble(), array.position)
+        val tr = expressionEval.translateExpression(mult)
+        addToResult(result, tr, tr.resultReg, -1)
+        return Pair(result, tr.resultReg)
     }
 }

codeGenIntermediate/src/prog8/codegen/intermediate/BuiltinFuncGen.kt

@@ -1,9 +1,10 @@
 package prog8.codegen.intermediate
 
-import prog8.code.StStaticVariable
 import prog8.code.ast.*
 import prog8.code.core.AssemblyError
 import prog8.code.core.DataType
+import prog8.code.core.SignedDatatypes
+import prog8.code.core.SplitWordArrayTypes
 import prog8.intermediate.*
 
 
@@ -13,12 +14,12 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         return when(call.name) {
             "any" -> funcAny(call)
             "all" -> funcAll(call)
-            "abs" -> funcAbs(call)
+            "abs__byte", "abs__word", "abs__float" -> funcAbs(call)
             "cmp" -> funcCmp(call)
             "sgn" -> funcSgn(call)
-            "sqrt16" -> funcSqrt16(call)
-            "divmod" -> funcDivmod(call, IRDataType.BYTE)
-            "divmodw" -> funcDivmod(call, IRDataType.WORD)
+            "sqrt__ubyte", "sqrt__uword", "sqrt__float" -> funcSqrt(call)
+            "divmod__ubyte" -> funcDivmod(call, IRDataType.BYTE)
+            "divmod__uword" -> funcDivmod(call, IRDataType.WORD)
             "pop" -> funcPop(call)
             "popw" -> funcPopw(call)
             "push" -> funcPush(call)
@@ -27,7 +28,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             "rsavex",
             "rrestore",
             "rrestorex" -> ExpressionCodeResult.EMPTY  // vm doesn't have registers to save/restore
-            "callfar" -> throw AssemblyError("callfar() is for cx16 target only")
+            "callfar" -> funcCallfar(call)
             "msb" -> funcMsb(call)
             "lsb" -> funcLsb(call)
             "memory" -> funcMemory(call)
@@ -37,6 +38,9 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             "pokew" -> funcPokeW(call)
             "pokemon" -> ExpressionCodeResult.EMPTY     // easter egg function
             "mkword" -> funcMkword(call)
+            "clamp__byte", "clamp__ubyte", "clamp__word", "clamp__uword" -> funcClamp(call)
+            "min__byte", "min__ubyte", "min__word", "min__uword" -> funcMin(call)
+            "max__byte", "max__ubyte", "max__word", "max__uword" -> funcMax(call)
             "sort" -> funcSort(call)
             "reverse" -> funcReverse(call)
             "rol" -> funcRolRor(Opcode.ROXL, call)
@@ -48,24 +52,44 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         }
     }
 
+    private fun funcCallfar(call: PtBuiltinFunctionCall): ExpressionCodeResult {
+        val result = mutableListOf<IRCodeChunkBase>()
+        val bankTr = exprGen.translateExpression(call.args[0])
+        val addressTr = exprGen.translateExpression(call.args[1])
+        val argumentwordTr = exprGen.translateExpression(call.args[2])
+        addToResult(result, bankTr, bankTr.resultReg, -1)
+        addToResult(result, addressTr, addressTr.resultReg, -1)
+        addToResult(result, argumentwordTr, argumentwordTr.resultReg, -1)
+        result += codeGen.makeSyscall(IMSyscall.CALLFAR, listOf(IRDataType.BYTE to bankTr.resultReg, IRDataType.WORD to addressTr.resultReg, IRDataType.WORD to argumentwordTr.resultReg), IRDataType.WORD to argumentwordTr.resultReg)
+        return ExpressionCodeResult(result, IRDataType.WORD, argumentwordTr.resultReg, -1)
+    }
+
     private fun funcDivmod(call: PtBuiltinFunctionCall, type: IRDataType): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
         val number = call.args[0]
         val divident = call.args[1]
+        val divisionReg: Int
+        val remainderReg: Int
         if(divident is PtNumber) {
             val tr = exprGen.translateExpression(number)
             addToResult(result, tr, tr.resultReg, -1)
-            addInstr(result, IRInstruction(Opcode.DIVMOD, type, reg1 = tr.resultReg, value=divident.number.toInt()), null)
+            addInstr(result, IRInstruction(Opcode.DIVMOD, type, reg1 = tr.resultReg, immediate = divident.number.toInt()), null)
+            divisionReg = tr.resultReg
+            remainderReg = codeGen.registers.nextFree()
         } else {
             val numTr = exprGen.translateExpression(number)
             addToResult(result, numTr, numTr.resultReg, -1)
             val dividentTr = exprGen.translateExpression(divident)
             addToResult(result, dividentTr, dividentTr.resultReg, -1)
             addInstr(result, IRInstruction(Opcode.DIVMODR, type, reg1 = numTr.resultReg, reg2=dividentTr.resultReg), null)
+            divisionReg = numTr.resultReg
+            remainderReg = dividentTr.resultReg
         }
-        // DIVMOD result convention: division in r0, remainder in r1
-        result += assignRegisterTo(call.args[2], 0)
-        result += assignRegisterTo(call.args[3], 1)
+        // DIVMOD result convention: on value stack, division and remainder on top.
+        addInstr(result, IRInstruction(Opcode.POP, type, reg1=remainderReg), null)
+        addInstr(result, IRInstruction(Opcode.POP, type, reg1=divisionReg), null)
+        result += assignRegisterTo(call.args[2], divisionReg)
+        result += assignRegisterTo(call.args[3], remainderReg)
         return ExpressionCodeResult(result, type, -1, -1)
     }
 
@@ -73,12 +97,10 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         val result = mutableListOf<IRCodeChunkBase>()
         val left  = exprGen.translateExpression(call.args[0])
         val right = exprGen.translateExpression(call.args[1])
-        val targetReg = codeGen.registers.nextFree()
-        addToResult(result, left, 65500, -1)
-        addToResult(result, right, 65501, -1)
-        addInstr(result, IRInstruction(Opcode.SYSCALL, value=IMSyscall.COMPARE_STRINGS.number), null)
-        addInstr(result, IRInstruction(Opcode.LOADR, IRDataType.BYTE, reg1=targetReg, reg2=0), null)
-        return ExpressionCodeResult(result, IRDataType.BYTE, targetReg, -1)
+        addToResult(result, left, left.resultReg, -1)
+        addToResult(result, right, right.resultReg, -1)
+        result += codeGen.makeSyscall(IMSyscall.COMPARE_STRINGS, listOf(IRDataType.WORD to left.resultReg, IRDataType.WORD to right.resultReg), IRDataType.BYTE to left.resultReg)
+        return ExpressionCodeResult(result, IRDataType.BYTE, left.resultReg, -1)
     }
 
     private fun funcCmp(call: PtBuiltinFunctionCall): ExpressionCodeResult {
@@ -87,7 +109,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         addToResult(result, leftTr, leftTr.resultReg, -1)
         val rightTr = exprGen.translateExpression(call.args[1])
         addToResult(result, rightTr, rightTr.resultReg, -1)
-        val dt = codeGen.irType(call.args[0].type)
+        val dt = irType(call.args[0].type)
         result += IRCodeChunk(null, null).also {
             it += IRInstruction(Opcode.CMP, dt, reg1=leftTr.resultReg, reg2=rightTr.resultReg)
         }
@@ -96,9 +118,9 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
 
     private fun funcAny(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val arrayName = call.args[0] as PtIdentifier
-        val array = codeGen.symbolTable.flat.getValue(arrayName.name) as StStaticVariable
+        val arrayLength = codeGen.symbolTable.getLength(arrayName.name)
         val syscall =
-            when (array.dt) {
+            when (arrayName.type) {
                 DataType.ARRAY_UB,
                 DataType.ARRAY_B -> IMSyscall.ANY_BYTE
                 DataType.ARRAY_UW,
@@ -108,22 +130,18 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             }
         val result = mutableListOf<IRCodeChunkBase>()
         val tr = exprGen.translateExpression(call.args[0])
-        addToResult(result, tr, SyscallRegisterBase, -1)
-        result += IRCodeChunk(null, null).also {
-            it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = SyscallRegisterBase+1, value = array.length)
-            it += IRInstruction(Opcode.SYSCALL, value = syscall.number)
-            // SysCall call convention: return value in register r0
-            if(tr.resultReg!=0)
-                it += IRInstruction(Opcode.LOADR, IRDataType.BYTE, reg1 = tr.resultReg, reg2 = 0)
-        }
+        addToResult(result, tr, tr.resultReg, -1)
+        val lengthReg = codeGen.registers.nextFree()
+        addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = lengthReg, immediate = arrayLength), null)
+        result += codeGen.makeSyscall(syscall, listOf(IRDataType.WORD to tr.resultReg, IRDataType.BYTE to lengthReg), IRDataType.BYTE to tr.resultReg)
         return ExpressionCodeResult(result, IRDataType.BYTE, tr.resultReg, -1)
     }
 
     private fun funcAll(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val arrayName = call.args[0] as PtIdentifier
-        val array = codeGen.symbolTable.flat.getValue(arrayName.name) as StStaticVariable
+        val arrayLength = codeGen.symbolTable.getLength(arrayName.name)
         val syscall =
-            when(array.dt) {
+            when(arrayName.type) {
                 DataType.ARRAY_UB,
                 DataType.ARRAY_B -> IMSyscall.ALL_BYTE
                 DataType.ARRAY_UW,
@@ -133,14 +151,10 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             }
         val result = mutableListOf<IRCodeChunkBase>()
         val tr = exprGen.translateExpression(call.args[0])
-        addToResult(result, tr, SyscallRegisterBase, -1)
-        result += IRCodeChunk(null, null).also {
-            it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = SyscallRegisterBase+1, value = array.length)
-            it += IRInstruction(Opcode.SYSCALL, value = syscall.number)
-            // SysCall call convention: return value in register r0
-            if(tr.resultReg!=0)
-                it += IRInstruction(Opcode.LOADR, IRDataType.BYTE, reg1 = tr.resultReg, reg2 = 0)
-        }
+        addToResult(result, tr, tr.resultReg, -1)
+        val lengthReg = codeGen.registers.nextFree()
+        addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = lengthReg, immediate = arrayLength), null)
+        result += codeGen.makeSyscall(syscall, listOf(IRDataType.WORD to tr.resultReg, IRDataType.BYTE to lengthReg), IRDataType.BYTE to tr.resultReg)
         return ExpressionCodeResult(result, IRDataType.BYTE, tr.resultReg, -1)
     }
 
@@ -153,21 +167,13 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         val tr = exprGen.translateExpression(call.args[0])
         addToResult(result, tr, tr.resultReg, -1)
         when (sourceDt) {
-            DataType.UBYTE -> {
-                result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.EXT, IRDataType.BYTE, reg1 = tr.resultReg)
-                }
-                return ExpressionCodeResult(result, IRDataType.BYTE, tr.resultReg, -1)
-            }
             DataType.BYTE -> {
                 val notNegativeLabel = codeGen.createLabelName()
                 val compareReg = codeGen.registers.nextFree()
                 result += IRCodeChunk(null, null).also {
                     it += IRInstruction(Opcode.LOADR, IRDataType.BYTE, reg1=compareReg, reg2=tr.resultReg)
-                    it += IRInstruction(Opcode.AND, IRDataType.BYTE, reg1=compareReg, value=0x80)
-                    it += IRInstruction(Opcode.BZ, IRDataType.BYTE, reg1=compareReg, labelSymbol = notNegativeLabel)
+                    it += IRInstruction(Opcode.BSTPOS, labelSymbol = notNegativeLabel)
                     it += IRInstruction(Opcode.NEG, IRDataType.BYTE, reg1=tr.resultReg)
-                    it += IRInstruction(Opcode.EXT, IRDataType.BYTE, reg1=tr.resultReg)
                 }
                 result += IRCodeChunk(notNegativeLabel, null)
                 return ExpressionCodeResult(result, IRDataType.BYTE, tr.resultReg, -1)
@@ -177,20 +183,24 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
                 val compareReg = codeGen.registers.nextFree()
                 result += IRCodeChunk(null, null).also {
                     it += IRInstruction(Opcode.LOADR, IRDataType.WORD, reg1=compareReg, reg2=tr.resultReg)
-                    it += IRInstruction(Opcode.AND, IRDataType.WORD, reg1=compareReg, value=0x8000)
-                    it += IRInstruction(Opcode.BZ, IRDataType.WORD, reg1=compareReg, labelSymbol = notNegativeLabel)
+                    it += IRInstruction(Opcode.BSTPOS, labelSymbol = notNegativeLabel)
                     it += IRInstruction(Opcode.NEG, IRDataType.WORD, reg1=tr.resultReg)
                 }
                 result += IRCodeChunk(notNegativeLabel, null)
                 return ExpressionCodeResult(result, IRDataType.WORD, tr.resultReg, -1)
             }
-            else -> throw AssemblyError("weird type")
+            DataType.FLOAT -> {
+                val resultFpReg = codeGen.registers.nextFreeFloat()
+                addInstr(result, IRInstruction(Opcode.FABS, IRDataType.FLOAT, fpReg1 = resultFpReg, fpReg2 = tr.resultFpReg), null)
+                return ExpressionCodeResult(result, IRDataType.FLOAT, -1, resultFpReg)
+            }
+            else -> throw AssemblyError("weird dt")
         }
     }
 
     private fun funcSgn(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
-        val vmDt = codeGen.irType(call.type)
+        val vmDt = irType(call.type)
         val tr = exprGen.translateExpression(call.args.single())
         addToResult(result, tr, tr.resultReg, -1)
         val resultReg = codeGen.registers.nextFree()
@@ -200,15 +210,37 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         return ExpressionCodeResult(result, vmDt, resultReg, -1)
     }
 
-    private fun funcSqrt16(call: PtBuiltinFunctionCall): ExpressionCodeResult {
+    private fun funcSqrt(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
         val tr = exprGen.translateExpression(call.args.single())
-        addToResult(result, tr, tr.resultReg, -1)
-        val resultReg = codeGen.registers.nextFree()
-        result += IRCodeChunk(null, null).also {
-            it += IRInstruction(Opcode.SQRT, IRDataType.WORD, reg1=resultReg, reg2=tr.resultReg)
+        val dt = call.args[0].type
+        when(dt) {
+            DataType.UBYTE -> {
+                addToResult(result, tr, tr.resultReg, -1)
+                val resultReg = codeGen.registers.nextFree()
+                result += IRCodeChunk(null, null).also {
+                    it += IRInstruction(Opcode.SQRT, IRDataType.BYTE, reg1=resultReg, reg2=tr.resultReg)
+                }
+                return ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
+            }
+            DataType.UWORD -> {
+                addToResult(result, tr, tr.resultReg, -1)
+                val resultReg = codeGen.registers.nextFree()
+                result += IRCodeChunk(null, null).also {
+                    it += IRInstruction(Opcode.SQRT, IRDataType.WORD, reg1=resultReg, reg2=tr.resultReg)
+                }
+                return ExpressionCodeResult(result, IRDataType.WORD, resultReg, -1)
+            }
+            DataType.FLOAT -> {
+                addToResult(result, tr, -1, tr.resultFpReg)
+                val resultFpReg = codeGen.registers.nextFreeFloat()
+                result += IRCodeChunk(null, null).also {
+                    it += IRInstruction(Opcode.SQRT, IRDataType.FLOAT, fpReg1 = resultFpReg, fpReg2 = tr.resultFpReg)
+                }
+                return ExpressionCodeResult(result, IRDataType.FLOAT, -1, resultFpReg)
+            }
+            else -> throw AssemblyError("invalid dt for sqrt")
         }
-        return ExpressionCodeResult(result, IRDataType.WORD, resultReg, -1)
     }
 
     private fun funcPop(call: PtBuiltinFunctionCall): ExpressionCodeResult {
@@ -251,44 +283,44 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
 
     private fun funcReverse(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val arrayName = call.args[0] as PtIdentifier
-        val array = codeGen.symbolTable.flat.getValue(arrayName.name) as StStaticVariable
+        val arrayLength = codeGen.symbolTable.getLength(arrayName.name)
         val syscall =
-            when(array.dt) {
+            when(arrayName.type) {
                 DataType.ARRAY_UB, DataType.ARRAY_B, DataType.STR -> IMSyscall.REVERSE_BYTES
                 DataType.ARRAY_UW, DataType.ARRAY_W -> IMSyscall.REVERSE_WORDS
                 DataType.ARRAY_F -> IMSyscall.REVERSE_FLOATS
+                in SplitWordArrayTypes -> TODO("split word reverse")
                 else -> throw IllegalArgumentException("weird type to reverse")
             }
         val result = mutableListOf<IRCodeChunkBase>()
         val tr = exprGen.translateExpression(call.args[0])
-        addToResult(result, tr, SyscallRegisterBase, -1)
-        result += IRCodeChunk(null, null).also {
-            it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = SyscallRegisterBase+1, value = array.length)
-            it += IRInstruction(Opcode.SYSCALL, value = syscall.number)
-        }
+        addToResult(result, tr, tr.resultReg, -1)
+        val lengthReg = codeGen.registers.nextFree()
+        addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = lengthReg, immediate = if(arrayName.type==DataType.STR) arrayLength!!-1 else arrayLength), null)
+        result += codeGen.makeSyscall(syscall, listOf(IRDataType.WORD to tr.resultReg, IRDataType.BYTE to lengthReg), null)
         return ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
     }
 
     private fun funcSort(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val arrayName = call.args[0] as PtIdentifier
-        val array = codeGen.symbolTable.flat.getValue(arrayName.name) as StStaticVariable
+        val arrayLength = codeGen.symbolTable.getLength(arrayName.name)
         val syscall =
-            when(array.dt) {
+            when(arrayName.type) {
                 DataType.ARRAY_UB -> IMSyscall.SORT_UBYTE
                 DataType.ARRAY_B -> IMSyscall.SORT_BYTE
                 DataType.ARRAY_UW -> IMSyscall.SORT_UWORD
                 DataType.ARRAY_W -> IMSyscall.SORT_WORD
                 DataType.STR -> IMSyscall.SORT_UBYTE
                 DataType.ARRAY_F -> throw IllegalArgumentException("sorting a floating point array is not supported")
+                in SplitWordArrayTypes -> TODO("split word sort")
                 else -> throw IllegalArgumentException("weird type to sort")
             }
         val result = mutableListOf<IRCodeChunkBase>()
         val tr = exprGen.translateExpression(call.args[0])
-        addToResult(result, tr, SyscallRegisterBase, -1)
-        result += IRCodeChunk(null, null).also {
-            it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = SyscallRegisterBase+1, value = array.length)
-            it += IRInstruction(Opcode.SYSCALL, value = syscall.number)
-        }
+        addToResult(result, tr, tr.resultReg, -1)
+        val lengthReg = codeGen.registers.nextFree()
+        addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = lengthReg, immediate = if(arrayName.type==DataType.STR) arrayLength!!-1 else arrayLength), null)
+        result += codeGen.makeSyscall(syscall, listOf(IRDataType.WORD to tr.resultReg, IRDataType.BYTE to lengthReg), null)
         return ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
     }
 
@@ -298,10 +330,85 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         addToResult(result, msbTr, msbTr.resultReg, -1)
         val lsbTr = exprGen.translateExpression(call.args[1])
         addToResult(result, lsbTr, lsbTr.resultReg, -1)
+        val resultReg = codeGen.registers.nextFree()
         result += IRCodeChunk(null, null).also {
-            it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1 = lsbTr.resultReg, reg2 = msbTr.resultReg)
+            it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=resultReg, reg2 = lsbTr.resultReg, reg3 = msbTr.resultReg)
         }
-        return ExpressionCodeResult(result, IRDataType.WORD, lsbTr.resultReg, -1)
+        return ExpressionCodeResult(result, IRDataType.WORD, resultReg, -1)
+    }
+
+    private fun funcClamp(call: PtBuiltinFunctionCall): ExpressionCodeResult {
+        val result = mutableListOf<IRCodeChunkBase>()
+        val type = irType(call.type)
+        val valueTr = exprGen.translateExpression(call.args[0])
+        val minimumTr = exprGen.translateExpression(call.args[1])
+        val maximumTr = exprGen.translateExpression(call.args[2])
+        result += valueTr.chunks
+        result += minimumTr.chunks
+        result += maximumTr.chunks
+        if(type==IRDataType.FLOAT) {
+            result += codeGen.makeSyscall(
+                IMSyscall.CLAMP_FLOAT, listOf(
+                    valueTr.dt to valueTr.resultFpReg,
+                    minimumTr.dt to minimumTr.resultFpReg,
+                    maximumTr.dt to maximumTr.resultFpReg,
+                ), type to valueTr.resultFpReg
+            )
+            return ExpressionCodeResult(result, type, -1, valueTr.resultFpReg)
+        } else {
+            val syscall = when(call.type) {
+                DataType.UBYTE -> IMSyscall.CLAMP_UBYTE
+                DataType.BYTE -> IMSyscall.CLAMP_BYTE
+                DataType.UWORD -> IMSyscall.CLAMP_UWORD
+                DataType.WORD -> IMSyscall.CLAMP_WORD
+                else -> throw AssemblyError("invalid dt")
+            }
+            result += codeGen.makeSyscall(syscall, listOf(
+                    valueTr.dt to valueTr.resultReg,
+                    minimumTr.dt to minimumTr.resultReg,
+                    maximumTr.dt to maximumTr.resultReg,
+                ), type to valueTr.resultReg
+            )
+            return ExpressionCodeResult(result, type, valueTr.resultReg, -1)
+        }
+    }
+
+    private fun funcMin(call: PtBuiltinFunctionCall): ExpressionCodeResult {
+        val type = irType(call.type)
+        val result = mutableListOf<IRCodeChunkBase>()
+        val leftTr = exprGen.translateExpression(call.args[0])
+        addToResult(result, leftTr, leftTr.resultReg, -1)
+        val rightTr = exprGen.translateExpression(call.args[1])
+        addToResult(result, rightTr, rightTr.resultReg, -1)
+        val comparisonOpcode = if(call.type in SignedDatatypes) Opcode.BGTSR else Opcode.BGTR
+        val after = codeGen.createLabelName()
+        result += IRCodeChunk(null, null).also {
+            it += IRInstruction(comparisonOpcode, type, reg1 = rightTr.resultReg, reg2 = leftTr.resultReg, labelSymbol = after)
+            // right <= left, take right
+            it += IRInstruction(Opcode.LOADR, type, reg1=leftTr.resultReg, reg2=rightTr.resultReg)
+            it += IRInstruction(Opcode.JUMP, labelSymbol = after)
+        }
+        result += IRCodeChunk(after, null)
+        return ExpressionCodeResult(result, type, leftTr.resultReg, -1)
+    }
+
+    private fun funcMax(call: PtBuiltinFunctionCall): ExpressionCodeResult {
+        val type = irType(call.type)
+        val result = mutableListOf<IRCodeChunkBase>()
+        val leftTr = exprGen.translateExpression(call.args[0])
+        addToResult(result, leftTr, leftTr.resultReg, -1)
+        val rightTr = exprGen.translateExpression(call.args[1])
+        addToResult(result, rightTr, rightTr.resultReg, -1)
+        val comparisonOpcode = if(call.type in SignedDatatypes) Opcode.BGTSR else Opcode.BGTR
+        val after = codeGen.createLabelName()
+        result += IRCodeChunk(null, null).also {
+            it += IRInstruction(comparisonOpcode, type, reg1 = leftTr.resultReg, reg2 = rightTr.resultReg, labelSymbol = after)
+            // right >= left, take right
+            it += IRInstruction(Opcode.LOADR, type, reg1=leftTr.resultReg, reg2=rightTr.resultReg)
+            it += IRInstruction(Opcode.JUMP, labelSymbol = after)
+        }
+        result += IRCodeChunk(after, null)
+        return ExpressionCodeResult(result, type, leftTr.resultReg, -1)
     }
 
     private fun funcPokeW(call: PtBuiltinFunctionCall): ExpressionCodeResult {
@@ -310,7 +417,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             if (call.args[0] is PtNumber) {
                 val address = (call.args[0] as PtNumber).number.toInt()
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.STOREZM, IRDataType.WORD, value = address)
+                    it += IRInstruction(Opcode.STOREZM, IRDataType.WORD, address = address)
                 }
             } else {
                 val tr = exprGen.translateExpression(call.args[0])
@@ -325,7 +432,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
                 val tr = exprGen.translateExpression(call.args[1])
                 addToResult(result, tr, tr.resultReg, -1)
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.STOREM, IRDataType.WORD, reg1 = tr.resultReg, value = address)
+                    it += IRInstruction(Opcode.STOREM, IRDataType.WORD, reg1 = tr.resultReg, address = address)
                 }
             } else {
                 val addressTr = exprGen.translateExpression(call.args[0])
@@ -346,7 +453,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             if (call.args[0] is PtNumber) {
                 val address = (call.args[0] as PtNumber).number.toInt()
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.STOREZM, IRDataType.BYTE, value = address)
+                    it += IRInstruction(Opcode.STOREZM, IRDataType.BYTE, address = address)
                 }
             } else {
                 val tr = exprGen.translateExpression(call.args[0])
@@ -361,7 +468,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
                 val tr = exprGen.translateExpression(call.args[1])
                 addToResult(result, tr, tr.resultReg, -1)
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.STOREM, IRDataType.BYTE, reg1 = tr.resultReg, value = address)
+                    it += IRInstruction(Opcode.STOREM, IRDataType.BYTE, reg1 = tr.resultReg, address = address)
                 }
             } else {
                 val addressTr = exprGen.translateExpression(call.args[0])
@@ -382,7 +489,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             val resultRegister = codeGen.registers.nextFree()
             val address = (call.args[0] as PtNumber).number.toInt()
             result += IRCodeChunk(null, null).also {
-                it += IRInstruction(Opcode.LOADM, IRDataType.WORD, reg1 = resultRegister, value = address)
+                it += IRInstruction(Opcode.LOADM, IRDataType.WORD, reg1 = resultRegister, address = address)
             }
             ExpressionCodeResult(result, IRDataType.BYTE, resultRegister, -1)
         } else {
@@ -402,7 +509,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             val resultRegister = codeGen.registers.nextFree()
             val address = (call.args[0] as PtNumber).number.toInt()
             result += IRCodeChunk(null, null).also {
-                it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1 = resultRegister, value = address)
+                it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1 = resultRegister, address = address)
             }
             ExpressionCodeResult(result, IRDataType.BYTE, resultRegister, -1)
         } else {
@@ -442,7 +549,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
     }
 
     private fun funcRolRor(opcode: Opcode, call: PtBuiltinFunctionCall): ExpressionCodeResult {
-        val vmDt = codeGen.irType(call.args[0].type)
+        val vmDt = irType(call.args[0].type)
         val result = mutableListOf<IRCodeChunkBase>()
         val tr = exprGen.translateExpression(call.args[0])
         addToResult(result, tr, tr.resultReg, -1)

codeGenIntermediate/src/prog8/codegen/intermediate/ExpressionGen.kt

@@ -1,7 +1,6 @@
 package prog8.codegen.intermediate
 
 import prog8.code.StRomSub
-import prog8.code.StStaticVariable
 import prog8.code.StSub
 import prog8.code.ast.*
 import prog8.code.core.AssemblyError
@@ -28,26 +27,26 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
     fun translateExpression(expr: PtExpression): ExpressionCodeResult {
         return when (expr) {
             is PtMachineRegister -> {
-                ExpressionCodeResult(emptyList(), codeGen.irType(expr.type), expr.register, -1)
+                ExpressionCodeResult(emptyList(), irType(expr.type), expr.register, -1)
             }
             is PtNumber -> {
-                val vmDt = codeGen.irType(expr.type)
+                val vmDt = irType(expr.type)
                 val code = IRCodeChunk(null, null)
                 if(vmDt==IRDataType.FLOAT) {
                     val resultFpRegister = codeGen.registers.nextFreeFloat()
-                    code += IRInstruction(Opcode.LOAD, vmDt, fpReg1 = resultFpRegister, fpValue = expr.number.toFloat())
+                    code += IRInstruction(Opcode.LOAD, vmDt, fpReg1 = resultFpRegister, immediateFp = expr.number.toFloat())
                     ExpressionCodeResult(code, vmDt,-1, resultFpRegister)
                 }
                 else {
                     val resultRegister = codeGen.registers.nextFree()
-                    code += IRInstruction(Opcode.LOAD, vmDt, reg1 = resultRegister, value = expr.number.toInt())
+                    code += IRInstruction(Opcode.LOAD, vmDt, reg1 = resultRegister, immediate = expr.number.toInt())
                     ExpressionCodeResult(code, vmDt, resultRegister, -1)
                 }
             }
             is PtIdentifier -> {
-                val vmDt = codeGen.irType(expr.type)
                 val code = IRCodeChunk(null, null)
                 if (expr.type in PassByValueDatatypes) {
+                    val vmDt = irType(expr.type)
                     if(vmDt==IRDataType.FLOAT) {
                         val resultFpRegister = codeGen.registers.nextFreeFloat()
                         code += IRInstruction(Opcode.LOADM, vmDt, fpReg1 = resultFpRegister, labelSymbol = expr.name)
@@ -60,13 +59,14 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                     }
                 } else {
                     // for strings and arrays etc., load the *address* of the value instead
+                    val vmDt = if(expr.type==DataType.UNDEFINED) IRDataType.WORD else irType(expr.type)
                     val resultRegister = codeGen.registers.nextFree()
                     code += IRInstruction(Opcode.LOAD, vmDt, reg1 = resultRegister, labelSymbol = expr.name)
                     ExpressionCodeResult(code, vmDt, resultRegister, -1)
                 }
             }
             is PtAddressOf -> {
-                val vmDt = codeGen.irType(expr.type)
+                val vmDt = irType(expr.type)
                 val symbol = expr.identifier.name
                 // note: LOAD <symbol>  gets you the address of the symbol, whereas LOADM <symbol> would get you the value stored at that location
                 val code = IRCodeChunk(null, null)
@@ -79,7 +79,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 if(expr.address is PtNumber) {
                     val address = (expr.address as PtNumber).number.toInt()
                     val resultRegister = codeGen.registers.nextFree()
-                    addInstr(result, IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=resultRegister, value = address), null)
+                    addInstr(result, IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=resultRegister, address = address), null)
                     ExpressionCodeResult(result, IRDataType.BYTE, resultRegister, -1)
                 } else {
                     val tr = translateExpression(expr.address)
@@ -105,57 +105,45 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
 
     private fun translate(check: PtContainmentCheck): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
-        var tr = translateExpression(check.element)
-        addToResult(result, tr, tr.resultReg, -1)
-        val iterable = codeGen.symbolTable.flat.getValue(check.iterable.name) as StStaticVariable
-        when(iterable.dt) {
+        when(check.iterable.type) {
             DataType.STR -> {
-                tr = translateExpression(check.element)
-                addToResult(result, tr, SyscallRegisterBase, -1)
-                tr = translateExpression(check.iterable)
-                addToResult(result, tr, SyscallRegisterBase+1, -1)
-                val resultReg = codeGen.registers.nextFree()
-                result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.SYSCALL, value = IMSyscall.STRING_CONTAINS.number)
-                    it += IRInstruction(Opcode.LOADR, IRDataType.BYTE, reg1=resultReg, reg2=0)
-                }
-                return ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
+                val elementTr = translateExpression(check.element)
+                addToResult(result, elementTr, elementTr.resultReg, -1)
+                val iterableTr = translateExpression(check.iterable)
+                addToResult(result, iterableTr, iterableTr.resultReg, -1)
+                result += codeGen.makeSyscall(IMSyscall.STRING_CONTAINS, listOf(IRDataType.BYTE to elementTr.resultReg, IRDataType.WORD to iterableTr.resultReg), IRDataType.BYTE to elementTr.resultReg)
+                return ExpressionCodeResult(result, IRDataType.BYTE, elementTr.resultReg, -1)
             }
             DataType.ARRAY_UB, DataType.ARRAY_B -> {
-                tr = translateExpression(check.element)
-                addToResult(result, tr, SyscallRegisterBase, -1)
-                tr = translateExpression(check.iterable)
-                addToResult(result, tr, SyscallRegisterBase+1, -1)
-                val resultReg = codeGen.registers.nextFree()
-                result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=SyscallRegisterBase+2, value = iterable.length!!)
-                    it += IRInstruction(Opcode.SYSCALL, value = IMSyscall.BYTEARRAY_CONTAINS.number)
-                    it += IRInstruction(Opcode.LOADR, IRDataType.BYTE, reg1=resultReg, reg2=0)
-                }
-                // SysCall call convention: return value in register r0
-                return ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
+                val elementTr = translateExpression(check.element)
+                addToResult(result, elementTr, elementTr.resultReg, -1)
+                val iterableTr = translateExpression(check.iterable)
+                addToResult(result, iterableTr, iterableTr.resultReg, -1)
+                val lengthReg = codeGen.registers.nextFree()
+                val iterableLength = codeGen.symbolTable.getLength(check.iterable.name)
+                addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=lengthReg, immediate = iterableLength!!), null)
+                result += codeGen.makeSyscall(IMSyscall.BYTEARRAY_CONTAINS, listOf(IRDataType.BYTE to elementTr.resultReg, IRDataType.WORD to iterableTr.resultReg, IRDataType.BYTE to lengthReg), IRDataType.BYTE to elementTr.resultReg)
+                return ExpressionCodeResult(result, IRDataType.BYTE, elementTr.resultReg, -1)
             }
             DataType.ARRAY_UW, DataType.ARRAY_W -> {
-                tr = translateExpression(check.element)
-                addToResult(result, tr, SyscallRegisterBase, -1)
-                tr = translateExpression(check.iterable)
-                addToResult(result, tr, SyscallRegisterBase+1, -1)
-                val resultReg = codeGen.registers.nextFree()
-                result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=SyscallRegisterBase+2, value = iterable.length!!)
-                    it += IRInstruction(Opcode.SYSCALL, value = IMSyscall.WORDARRAY_CONTAINS.number)
-                    it += IRInstruction(Opcode.LOADR, IRDataType.BYTE, reg1=resultReg, reg2=0)
-                }
-                return ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
+                val elementTr = translateExpression(check.element)
+                addToResult(result, elementTr, elementTr.resultReg, -1)
+                val iterableTr = translateExpression(check.iterable)
+                addToResult(result, iterableTr, iterableTr.resultReg, -1)
+                val lengthReg = codeGen.registers.nextFree()
+                val iterableLength = codeGen.symbolTable.getLength(check.iterable.name)
+                addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=lengthReg, immediate = iterableLength!!), null)
+                result += codeGen.makeSyscall(IMSyscall.WORDARRAY_CONTAINS, listOf(IRDataType.WORD to elementTr.resultReg, IRDataType.WORD to iterableTr.resultReg, IRDataType.BYTE to lengthReg), IRDataType.BYTE to elementTr.resultReg)
+                return ExpressionCodeResult(result, IRDataType.BYTE, elementTr.resultReg, -1)
             }
             DataType.ARRAY_F -> throw AssemblyError("containment check in float-array not supported")
-            else -> throw AssemblyError("weird iterable dt ${iterable.dt} for ${check.iterable.name}")
+            else -> throw AssemblyError("weird iterable dt ${check.iterable.type} for ${check.iterable.name}")
         }
     }
 
     private fun translate(arrayIx: PtArrayIndexer): ExpressionCodeResult {
         val eltSize = codeGen.program.memsizer.memorySize(arrayIx.type)
-        val vmDt = codeGen.irType(arrayIx.type)
+        val vmDt = irType(arrayIx.type)
         val result = mutableListOf<IRCodeChunkBase>()
         val arrayVarSymbol = arrayIx.variable.name
 
@@ -173,6 +161,33 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         }
 
         var resultRegister = -1
+
+        if(arrayIx.splitWords) {
+            require(vmDt==IRDataType.WORD)
+            val arrayLength = codeGen.symbolTable.getLength(arrayIx.variable.name)
+            resultRegister = codeGen.registers.nextFree()
+            val finalResultReg = codeGen.registers.nextFree()
+            if(arrayIx.index is PtNumber) {
+                val memOffset = (arrayIx.index as PtNumber).number.toInt()
+                result += IRCodeChunk(null, null).also {
+                    val tmpRegMsb = codeGen.registers.nextFree()
+                    it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=resultRegister, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_lsb+$memOffset")
+                    it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=tmpRegMsb, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_msb+$memOffset")
+                    it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=finalResultReg, reg2=resultRegister, reg3=tmpRegMsb)
+                }
+            } else {
+                val tr = translateExpression(arrayIx.index)
+                addToResult(result, tr, tr.resultReg, -1)
+                result += IRCodeChunk(null, null).also {
+                    val tmpRegMsb = codeGen.registers.nextFree()
+                    it += IRInstruction(Opcode.LOADX, IRDataType.BYTE, reg1=resultRegister, reg2 = tr.resultReg, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_lsb")
+                    it += IRInstruction(Opcode.LOADX, IRDataType.BYTE, reg1=tmpRegMsb, reg2 = tr.resultReg, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_msb")
+                    it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=finalResultReg, reg2=resultRegister, reg3=tmpRegMsb)
+                }
+            }
+            return ExpressionCodeResult(result, vmDt, finalResultReg, -1)
+        }
+
         var resultFpRegister = -1
         if(arrayIx.index is PtNumber) {
             val memOffset = ((arrayIx.index as PtNumber).number.toInt() * eltSize).toString()
@@ -205,7 +220,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         val result = mutableListOf<IRCodeChunkBase>()
         val tr = translateExpression(expr.value)
         addToResult(result, tr, tr.resultReg, tr.resultFpReg)
-        val vmDt = codeGen.irType(expr.type)
+        val vmDt = irType(expr.type)
         when(expr.operator) {
             "+" -> { }
             "-" -> {
@@ -216,7 +231,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             }
             "~" -> {
                 val mask = if(vmDt==IRDataType.BYTE) 0x00ff else 0xffff
-                addInstr(result, IRInstruction(Opcode.XOR, vmDt, reg1 = tr.resultReg, value = mask), null)
+                addInstr(result, IRInstruction(Opcode.XOR, vmDt, reg1 = tr.resultReg, immediate = mask), null)
             }
             else -> throw AssemblyError("weird prefix operator")
         }
@@ -260,13 +275,13 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 when(cast.value.type) {
                     DataType.BYTE -> {
                         // byte -> uword:   sign extend
-                        actualResultReg2 = tr.resultReg
-                        addInstr(result, IRInstruction(Opcode.EXTS, type = IRDataType.BYTE, reg1 = actualResultReg2), null)
+                        actualResultReg2 = codeGen.registers.nextFree()
+                        addInstr(result, IRInstruction(Opcode.EXTS, type = IRDataType.BYTE, reg1 = actualResultReg2, reg2 = tr.resultReg), null)
                     }
                     DataType.UBYTE -> {
                         // ubyte -> uword:   sign extend
-                        actualResultReg2 = tr.resultReg
-                        addInstr(result, IRInstruction(Opcode.EXT, type = IRDataType.BYTE, reg1 = actualResultReg2), null)
+                        actualResultReg2 = codeGen.registers.nextFree()
+                        addInstr(result, IRInstruction(Opcode.EXT, type = IRDataType.BYTE, reg1 = actualResultReg2, reg2 = tr.resultReg), null)
                     }
                     DataType.WORD -> {
                         actualResultReg2 = tr.resultReg
@@ -282,13 +297,13 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 when(cast.value.type) {
                     DataType.BYTE -> {
                         // byte -> word:   sign extend
-                        actualResultReg2 = tr.resultReg
-                        addInstr(result, IRInstruction(Opcode.EXTS, type = IRDataType.BYTE, reg1 = actualResultReg2), null)
+                        actualResultReg2 = codeGen.registers.nextFree()
+                        addInstr(result, IRInstruction(Opcode.EXTS, type = IRDataType.BYTE, reg1 = actualResultReg2, reg2=tr.resultReg), null)
                     }
                     DataType.UBYTE -> {
                         // byte -> word:   sign extend
-                        actualResultReg2 = tr.resultReg
-                        addInstr(result, IRInstruction(Opcode.EXT, type = IRDataType.BYTE, reg1 = actualResultReg2), null)
+                        actualResultReg2 = codeGen.registers.nextFree()
+                        addInstr(result, IRInstruction(Opcode.EXT, type = IRDataType.BYTE, reg1 = actualResultReg2, reg2=tr.resultReg), null)
                     }
                     DataType.UWORD -> {
                         actualResultReg2 = tr.resultReg
@@ -321,12 +336,11 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             else -> throw AssemblyError("weird cast type")
         }
 
-        return ExpressionCodeResult(result, codeGen.irType(cast.type), actualResultReg2, actualResultFpReg2)
+        return ExpressionCodeResult(result, irType(cast.type), actualResultReg2, actualResultFpReg2)
     }
 
     private fun translate(binExpr: PtBinaryExpression): ExpressionCodeResult {
-        require(!codeGen.options.useNewExprCode)
-        val vmDt = codeGen.irType(binExpr.left.type)
+        val vmDt = irType(binExpr.left.type)
         val signed = binExpr.left.type in SignedDatatypes
         return when(binExpr.operator) {
             "+" -> operatorPlus(binExpr, vmDt)
@@ -353,87 +367,82 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         when (val callTarget = codeGen.symbolTable.flat.getValue(fcall.name)) {
             is StSub -> {
                 val result = mutableListOf<IRCodeChunkBase>()
+                // assign the arguments
+                val argRegisters = mutableListOf<FunctionCallArgs.ArgumentSpec>()
                 for ((arg, parameter) in fcall.args.zip(callTarget.parameters)) {
-                    val paramDt = codeGen.irType(parameter.type)
-                    val symbol = "${fcall.name}.${parameter.name}"
-                    if(codeGen.isZero(arg)) {
-                        addInstr(result, IRInstruction(Opcode.STOREZM, paramDt, labelSymbol = symbol), null)
-                    } else {
-                        if (paramDt == IRDataType.FLOAT) {
-                            val tr = translateExpression(arg)
-                            addToResult(result, tr, -1, tr.resultFpReg)
-                            addInstr(result, IRInstruction(Opcode.STOREM, paramDt, fpReg1 = tr.resultFpReg, labelSymbol = symbol), null)
-                        } else {
-                            val tr = translateExpression(arg)
-                            addToResult(result, tr, tr.resultReg, -1)
-                            addInstr(result, IRInstruction(Opcode.STOREM, paramDt, reg1 = tr.resultReg, labelSymbol = symbol), null)
-                        }
-                    }
+                    val paramDt = irType(parameter.type)
+                    val tr = translateExpression(arg)
+                    if(paramDt==IRDataType.FLOAT)
+                        argRegisters.add(FunctionCallArgs.ArgumentSpec(parameter.name, null, FunctionCallArgs.RegSpec(IRDataType.FLOAT, tr.resultFpReg, null)))
+                    else
+                        argRegisters.add(FunctionCallArgs.ArgumentSpec(parameter.name, null, FunctionCallArgs.RegSpec(paramDt, tr.resultReg, null)))
+                    result += tr.chunks
                 }
-                return if(fcall.void) {
-                    addInstr(result, IRInstruction(Opcode.CALL, labelSymbol = fcall.name), null)
+                // return value
+                val returnRegSpec = if(fcall.void) null else {
+                    val returnIrType = irType(callTarget.returnType!!)
+                    if(returnIrType==IRDataType.FLOAT)
+                        FunctionCallArgs.RegSpec(returnIrType, codeGen.registers.nextFreeFloat(), null)
+                    else
+                        FunctionCallArgs.RegSpec(returnIrType, codeGen.registers.nextFree(), null)
+                }
+                // create the call
+                val call = IRInstruction(Opcode.CALL, labelSymbol = fcall.name, fcallArgs = FunctionCallArgs(argRegisters, returnRegSpec))
+                addInstr(result, call, null)
+                return if(fcall.void)
                     ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
-                } else {
-                    var resultReg = -1
-                    var resultFpReg = -1
-                    if(fcall.type==DataType.FLOAT) {
-                        resultFpReg = codeGen.registers.nextFreeFloat()
-                        addInstr(result, IRInstruction(Opcode.CALLRVAL, IRDataType.FLOAT, fpReg1=resultFpReg, labelSymbol=fcall.name), null)
-                    } else {
-                        resultReg = codeGen.registers.nextFree()
-                        addInstr(result, IRInstruction(Opcode.CALLRVAL, codeGen.irType(fcall.type), reg1=resultReg, labelSymbol=fcall.name), null)
-                    }
-                    ExpressionCodeResult(result, codeGen.irType(fcall.type), resultReg, resultFpReg)
-                }
+                else if(fcall.type==DataType.FLOAT)
+                    ExpressionCodeResult(result, returnRegSpec!!.dt, -1, returnRegSpec.registerNum)
+                else
+                    ExpressionCodeResult(result, returnRegSpec!!.dt, returnRegSpec.registerNum, -1)
             }
             is StRomSub -> {
                 val result = mutableListOf<IRCodeChunkBase>()
+                // assign the arguments
+                val argRegisters = mutableListOf<FunctionCallArgs.ArgumentSpec>()
                 for ((arg, parameter) in fcall.args.zip(callTarget.parameters)) {
-                    val paramDt = codeGen.irType(parameter.type)
-                    val paramRegStr = if(parameter.register.registerOrPair!=null) parameter.register.registerOrPair.toString() else parameter.register.statusflag.toString()
-                    if(codeGen.isZero(arg)) {
-                        addInstr(result, IRInstruction(Opcode.STOREZCPU, paramDt, labelSymbol = paramRegStr), null)
+                    val paramDt = irType(parameter.type)
+                    val tr = translateExpression(arg)
+                    if(paramDt==IRDataType.FLOAT)
+                        argRegisters.add(FunctionCallArgs.ArgumentSpec("", null, FunctionCallArgs.RegSpec(IRDataType.FLOAT, tr.resultFpReg, parameter.register)))
+                    else
+                        argRegisters.add(FunctionCallArgs.ArgumentSpec("", null, FunctionCallArgs.RegSpec(paramDt, tr.resultReg, parameter.register)))
+                    result += tr.chunks
+                }
+                // return value
+                val returnRegSpec = if(fcall.void) null else {
+                    if(callTarget.returns.isEmpty())
+                        null
+                    else if(callTarget.returns.size==1) {
+                        val returns = callTarget.returns[0]
+                        val returnIrType = irType(returns.type)
+                        if(returnIrType==IRDataType.FLOAT)
+                            FunctionCallArgs.RegSpec(returnIrType, codeGen.registers.nextFreeFloat(), returns.register)
+                        else
+                            FunctionCallArgs.RegSpec(returnIrType, codeGen.registers.nextFree(), returns.register)
                     } else {
-                        if (paramDt == IRDataType.FLOAT)
-                            throw AssemblyError("doesn't support float register argument in asm romsub")
-                        val tr = translateExpression(arg)
-                        addToResult(result, tr, tr.resultReg, -1)
-                        addInstr(result, IRInstruction(Opcode.STORECPU, paramDt, reg1 = tr.resultReg, labelSymbol = paramRegStr), null)
+                        // multiple return values: take the first *register* (not status flag) return value and ignore the rest.
+                        val returns = callTarget.returns.first { it.register.registerOrPair!=null }
+                        val returnIrType = irType(returns.type)
+                        if(returnIrType==IRDataType.FLOAT)
+                            FunctionCallArgs.RegSpec(returnIrType, codeGen.registers.nextFreeFloat(), returns.register)
+                        else
+                            FunctionCallArgs.RegSpec(returnIrType, codeGen.registers.nextFree(), returns.register)
                     }
                 }
-                // just a regular call without using Vm register call convention: the value is returned in CPU registers!
-                addInstr(result, IRInstruction(Opcode.CALL, value=callTarget.address.toInt()), null)
-                val resultReg = codeGen.registers.nextFree()
-                if(!fcall.void) {
-                    when(callTarget.returns.size) {
-                        0 -> throw AssemblyError("expect a return value")
-                        1 -> {
-                            if(fcall.type==DataType.FLOAT)
-                                throw AssemblyError("doesn't support float register result in asm romsub")
-                            val returns = callTarget.returns.single()
-                            val regStr = if(returns.register.registerOrPair!=null) returns.register.registerOrPair.toString() else returns.register.statusflag.toString()
-                            addInstr(result, IRInstruction(Opcode.LOADCPU, codeGen.irType(fcall.type), reg1=resultReg, labelSymbol = regStr), null)
-                        }
-                        else -> {
-                            val returnRegister = callTarget.returns.singleOrNull{ it.register.registerOrPair!=null }
-                            if(returnRegister!=null) {
-                                // we skip the other values returned in the status flags.
-                                val regStr = returnRegister.register.registerOrPair.toString()
-                                addInstr(result, IRInstruction(Opcode.LOADCPU, codeGen.irType(fcall.type), reg1=resultReg, labelSymbol = regStr), null)
-                            } else {
-                                val firstReturnRegister = callTarget.returns.firstOrNull{ it.register.registerOrPair!=null }
-                                if(firstReturnRegister!=null) {
-                                    // we just take the first register return value and ignore the rest.
-                                    val regStr = firstReturnRegister.register.registerOrPair.toString()
-                                    addInstr(result, IRInstruction(Opcode.LOADCPU, codeGen.irType(fcall.type), reg1=resultReg, labelSymbol = regStr), null)
-                                } else {
-                                    throw AssemblyError("invalid number of return values from call")
-                                }
-                            }
-                        }
-                    }
-                }
-                return ExpressionCodeResult(result, if(fcall.void) IRDataType.BYTE else codeGen.irType(fcall.type), resultReg, -1)
+                // create the call
+                val call =
+                    if(callTarget.address==null)
+                        IRInstruction(Opcode.CALL, labelSymbol = fcall.name, fcallArgs = FunctionCallArgs(argRegisters, returnRegSpec))
+                    else
+                        IRInstruction(Opcode.CALL, address = callTarget.address!!.toInt(), fcallArgs = FunctionCallArgs(argRegisters, returnRegSpec))
+                addInstr(result, call, null)
+                return if(fcall.void)
+                    ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
+                else if(fcall.type==DataType.FLOAT)
+                    ExpressionCodeResult(result, returnRegSpec!!.dt, -1, returnRegSpec.registerNum)
+                else
+                    ExpressionCodeResult(result, returnRegSpec!!.dt, returnRegSpec.registerNum, -1)
             }
             else -> throw AssemblyError("invalid node type")
         }
@@ -446,6 +455,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         greaterEquals: Boolean
     ): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
+        val cmpResultReg = codeGen.registers.nextFree()
         if(vmDt==IRDataType.FLOAT) {
             val leftTr = translateExpression(binExpr.left)
             addToResult(result, leftTr, -1, leftTr.resultFpReg)
@@ -454,31 +464,17 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             val resultRegister = codeGen.registers.nextFree()
             addInstr(result, IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=resultRegister, fpReg1 = leftTr.resultFpReg, fpReg2 = rightTr.resultFpReg), null)
             val zeroRegister = codeGen.registers.nextFree()
-            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroRegister, value=0), null)
+            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroRegister, immediate = 0), null)
             val ins = if (signed) {
                 if (greaterEquals) Opcode.SGES else Opcode.SGTS
             } else {
                 if (greaterEquals) Opcode.SGE else Opcode.SGT
             }
-            addInstr(result, IRInstruction(ins, IRDataType.BYTE, reg1 = resultRegister, reg2 = zeroRegister), null)
-            return ExpressionCodeResult(result, IRDataType.BYTE, resultRegister, -1)
+            addInstr(result, IRInstruction(ins, IRDataType.BYTE, reg1=cmpResultReg, reg2 = resultRegister, reg3 = zeroRegister), null)
+            return ExpressionCodeResult(result, IRDataType.BYTE, cmpResultReg, -1)
         } else {
-            if(binExpr.left.type==DataType.STR && binExpr.right.type==DataType.STR) {
-                val leftTr = translateExpression(binExpr.left)
-                addToResult(result, leftTr, SyscallRegisterBase, -1)
-                val rightTr = translateExpression(binExpr.right)
-                addToResult(result, rightTr, SyscallRegisterBase+1, -1)
-                result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.SYSCALL, value = IMSyscall.COMPARE_STRINGS.number)
-                    // SysCall call convention: return value in register r0
-                    val zeroReg = codeGen.registers.nextFree()
-                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = zeroReg, value = 0)
-                    it += if (greaterEquals)
-                        IRInstruction(Opcode.SGES, IRDataType.BYTE, reg1 = leftTr.resultReg, reg2 = zeroReg)
-                    else
-                        IRInstruction(Opcode.SGTS, IRDataType.BYTE, reg1 = leftTr.resultReg, reg2 = zeroReg)
-                }
-                return ExpressionCodeResult(result, IRDataType.BYTE, leftTr.resultReg, -1)
+            if(binExpr.left.type==DataType.STR || binExpr.right.type==DataType.STR) {
+                throw AssemblyError("str compares should have been replaced with builtin function call to do the compare")
             } else {
                 val leftTr = translateExpression(binExpr.left)
                 addToResult(result, leftTr, leftTr.resultReg, -1)
@@ -489,8 +485,8 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 } else {
                     if (greaterEquals) Opcode.SGE else Opcode.SGT
                 }
-                addInstr(result, IRInstruction(ins, vmDt, reg1 = leftTr.resultReg, reg2 = rightTr.resultReg), null)
-                return ExpressionCodeResult(result, IRDataType.BYTE, leftTr.resultReg, -1)
+                addInstr(result, IRInstruction(ins, vmDt, reg1=cmpResultReg, reg2 = leftTr.resultReg, reg3 = rightTr.resultReg), null)
+                return ExpressionCodeResult(result, IRDataType.BYTE, cmpResultReg, -1)
             }
         }
     }
@@ -502,6 +498,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         lessEquals: Boolean
     ): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
+        val cmpResultRegister = codeGen.registers.nextFree()
         if(vmDt==IRDataType.FLOAT) {
             val leftTr = translateExpression(binExpr.left)
             addToResult(result, leftTr, -1, leftTr.resultFpReg)
@@ -510,33 +507,17 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             val resultRegister = codeGen.registers.nextFree()
             addInstr(result, IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=resultRegister, fpReg1 = leftTr.resultFpReg, fpReg2 = rightTr.resultFpReg), null)
             val zeroRegister = codeGen.registers.nextFree()
-            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroRegister, value=0), null)
+            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroRegister, immediate = 0), null)
             val ins = if (signed) {
                 if (lessEquals) Opcode.SLES else Opcode.SLTS
             } else {
                 if (lessEquals) Opcode.SLE else Opcode.SLT
             }
-            addInstr(result, IRInstruction(ins, IRDataType.BYTE, reg1 = resultRegister, reg2 = zeroRegister), null)
-            return ExpressionCodeResult(result, IRDataType.BYTE, resultRegister, -1)
+            addInstr(result, IRInstruction(ins, IRDataType.BYTE, reg1=cmpResultRegister, reg2 = resultRegister, reg3 = zeroRegister), null)
+            return ExpressionCodeResult(result, IRDataType.BYTE, cmpResultRegister, -1)
         } else {
-            if(binExpr.left.type==DataType.STR && binExpr.right.type==DataType.STR) {
-                val leftTr = translateExpression(binExpr.left)
-                addToResult(result, leftTr, SyscallRegisterBase, -1)
-                val rightTr = translateExpression(binExpr.right)
-                addToResult(result, rightTr, SyscallRegisterBase+1, -1)
-                val resultReg = codeGen.registers.nextFree()
-                result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.SYSCALL, value = IMSyscall.COMPARE_STRINGS.number)
-                    // SysCall call convention: return value in register r0
-                    val zeroReg = codeGen.registers.nextFree()
-                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = zeroReg, value = 0)
-                    it += if (lessEquals)
-                        IRInstruction(Opcode.SLES, IRDataType.BYTE, reg1 = 0, reg2 = zeroReg)
-                    else
-                        IRInstruction(Opcode.SLTS, IRDataType.BYTE, reg1 = 0, reg2 = zeroReg)
-                    it += IRInstruction(Opcode.LOADR, IRDataType.BYTE, reg1=resultReg, reg2=0)
-                }
-                return ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
+            if(binExpr.left.type==DataType.STR || binExpr.right.type==DataType.STR) {
+                throw AssemblyError("str compares should have been replaced with builtin function call to do the compare")
             } else {
                 val leftTr = translateExpression(binExpr.left)
                 addToResult(result, leftTr, leftTr.resultReg, -1)
@@ -547,8 +528,8 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 } else {
                     if (lessEquals) Opcode.SLE else Opcode.SLT
                 }
-                addInstr(result, IRInstruction(ins, vmDt, reg1 = leftTr.resultReg, reg2 = rightTr.resultReg), null)
-                return ExpressionCodeResult(result, IRDataType.BYTE, leftTr.resultReg, -1)
+                addInstr(result, IRInstruction(ins, vmDt, reg1=cmpResultRegister, reg2 = leftTr.resultReg, reg3 = rightTr.resultReg), null)
+                return ExpressionCodeResult(result, IRDataType.BYTE, cmpResultRegister, -1)
             }
         }
     }
@@ -563,32 +544,19 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             val resultRegister = codeGen.registers.nextFree()
             val valueReg = codeGen.registers.nextFree()
             val label = codeGen.createLabelName()
-            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=resultRegister, value=1), null)
+            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=resultRegister, immediate = 1), null)
             addInstr(result, IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=valueReg, fpReg1 = leftTr.resultFpReg, fpReg2 = rightTr.resultFpReg), null)
             if (notEquals) {
-                addInstr(result, IRInstruction(Opcode.BNZ, IRDataType.BYTE, reg1=valueReg, labelSymbol = label), null)
+                addInstr(result, IRInstruction(Opcode.BNE, IRDataType.BYTE, reg1=valueReg, immediate = 0, labelSymbol = label), null)
             } else {
-                addInstr(result, IRInstruction(Opcode.BZ, IRDataType.BYTE, reg1=valueReg, labelSymbol = label), null)
+                addInstr(result, IRInstruction(Opcode.BEQ, IRDataType.BYTE, reg1=valueReg, immediate = 0, labelSymbol = label), null)
             }
-            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=resultRegister, value=0), null)
+            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=resultRegister, immediate = 0), null)
             result += IRCodeChunk(label, null)
             return ExpressionCodeResult(result, IRDataType.BYTE, resultRegister, -1)
         } else {
-            if(binExpr.left.type==DataType.STR && binExpr.right.type==DataType.STR) {
-                val leftTr = translateExpression(binExpr.left)
-                addToResult(result, leftTr, SyscallRegisterBase, -1)
-                val rightTr = translateExpression(binExpr.right)
-                addToResult(result, rightTr, SyscallRegisterBase+1, -1)
-                val resultRegister = codeGen.registers.nextFree()
-                result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.SYSCALL, value = IMSyscall.COMPARE_STRINGS.number)
-                    // SysCall call convention: return value in register r0
-                    if (!notEquals)
-                        it += IRInstruction(Opcode.INV, vmDt, reg1 = 0)
-                    it += IRInstruction(Opcode.LOADR, vmDt, reg1=resultRegister, reg2=0)
-                    it += IRInstruction(Opcode.AND, vmDt, reg1 = resultRegister, value = 1)
-                }
-                return ExpressionCodeResult(result, IRDataType.BYTE, resultRegister, -1)
+            if(binExpr.left.type==DataType.STR || binExpr.right.type==DataType.STR) {
+                throw AssemblyError("str compares should have been replaced with builtin function call to do the compare")
             } else {
                 return if(constValue(binExpr.right)==0.0) {
                     val tr = translateExpression(binExpr.left)
@@ -603,8 +571,9 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                     val rightTr = translateExpression(binExpr.right)
                     addToResult(result, rightTr, rightTr.resultReg, -1)
                     val opcode = if (notEquals) Opcode.SNE else Opcode.SEQ
-                    addInstr(result, IRInstruction(opcode, vmDt, reg1 = leftTr.resultReg, reg2 = rightTr.resultReg), null)
-                    ExpressionCodeResult(result, IRDataType.BYTE, leftTr.resultReg, -1)
+                    val resultReg = codeGen.registers.nextFree()
+                    addInstr(result, IRInstruction(opcode, vmDt, reg1 = resultReg, reg2 = leftTr.resultReg, reg3 = rightTr.resultReg), null)
+                    ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
                 }
             }
         }
@@ -651,7 +620,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         return if(binExpr.right is PtNumber) {
             val tr = translateExpression(binExpr.left)
             addToResult(result, tr, tr.resultReg, -1)
-            addInstr(result, IRInstruction(Opcode.XOR, vmDt, reg1 = tr.resultReg, value=(binExpr.right as PtNumber).number.toInt()), null)
+            addInstr(result, IRInstruction(Opcode.XOR, vmDt, reg1 = tr.resultReg, immediate = (binExpr.right as PtNumber).number.toInt()), null)
             ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
         } else {
             val leftTr = translateExpression(binExpr.left)
@@ -668,7 +637,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         return if(binExpr.right is PtNumber) {
             val tr = translateExpression(binExpr.left)
             addToResult(result, tr, tr.resultReg, -1)
-            addInstr(result, IRInstruction(Opcode.AND, vmDt, reg1 = tr.resultReg, value=(binExpr.right as PtNumber).number.toInt()), null)
+            addInstr(result, IRInstruction(Opcode.AND, vmDt, reg1 = tr.resultReg, immediate = (binExpr.right as PtNumber).number.toInt()), null)
             ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
         } else {
             val leftTr = translateExpression(binExpr.left)
@@ -685,7 +654,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         return if(binExpr.right is PtNumber) {
             val tr = translateExpression(binExpr.left)
             addToResult(result, tr, tr.resultReg, -1)
-            addInstr(result, IRInstruction(Opcode.OR, vmDt, reg1 = tr.resultReg, value=(binExpr.right as PtNumber).number.toInt()), null)
+            addInstr(result, IRInstruction(Opcode.OR, vmDt, reg1 = tr.resultReg, immediate = (binExpr.right as PtNumber).number.toInt()), null)
             ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
         } else {
             val leftTr = translateExpression(binExpr.left)
@@ -703,7 +672,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         return if(binExpr.right is PtNumber) {
             val tr = translateExpression(binExpr.left)
             addToResult(result, tr, tr.resultReg, -1)
-            addInstr(result, IRInstruction(Opcode.MOD, vmDt, reg1 = tr.resultReg, value=(binExpr.right as PtNumber).number.toInt()), null)
+            addInstr(result, IRInstruction(Opcode.MOD, vmDt, reg1 = tr.resultReg, immediate = (binExpr.right as PtNumber).number.toInt()), null)
             ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
         } else {
             val leftTr = translateExpression(binExpr.left)
@@ -749,9 +718,9 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                     val leftTr = translateExpression(binExpr.left)
                     addToResult(result, leftTr, leftTr.resultReg, -1)
                     addInstr(result, if (signed)
-                            IRInstruction(Opcode.DIVS, vmDt, reg1 = leftTr.resultReg, value=(binExpr.right as PtNumber).number.toInt())
+                            IRInstruction(Opcode.DIVS, vmDt, reg1 = leftTr.resultReg, immediate = (binExpr.right as PtNumber).number.toInt())
                         else
-                            IRInstruction(Opcode.DIV, vmDt, reg1 = leftTr.resultReg, value=(binExpr.right as PtNumber).number.toInt())
+                            IRInstruction(Opcode.DIV, vmDt, reg1 = leftTr.resultReg, immediate = (binExpr.right as PtNumber).number.toInt())
                             , null)
                     ExpressionCodeResult(result, vmDt, leftTr.resultReg, -1)
                 } else {
@@ -832,7 +801,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 return if(binExpr.right is PtNumber) {
                     val tr = translateExpression(binExpr.left)
                     addToResult(result, tr, -1, tr.resultFpReg)
-                    addInstr(result, IRInstruction(Opcode.SUB, vmDt, fpReg1 = tr.resultFpReg, fpValue = (binExpr.right as PtNumber).number.toFloat()), null)
+                    addInstr(result, IRInstruction(Opcode.SUB, vmDt, fpReg1 = tr.resultFpReg, immediateFp = (binExpr.right as PtNumber).number.toFloat()), null)
                     ExpressionCodeResult(result, vmDt, -1, tr.resultFpReg)
                 } else {
                     val leftTr = translateExpression(binExpr.left)
@@ -852,9 +821,9 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             }
             else {
                 return if(binExpr.right is PtNumber) {
-                    val tr = translateExpression(binExpr.left,)
+                    val tr = translateExpression(binExpr.left)
                     addToResult(result, tr, tr.resultReg, -1)
-                    addInstr(result, IRInstruction(Opcode.SUB, vmDt, reg1 = tr.resultReg, value = (binExpr.right as PtNumber).number.toInt()), null)
+                    addInstr(result, IRInstruction(Opcode.SUB, vmDt, reg1 = tr.resultReg, immediate = (binExpr.right as PtNumber).number.toInt()), null)
                     ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
                 } else {
                     val leftTr = translateExpression(binExpr.left)
@@ -887,7 +856,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 return if(binExpr.right is PtNumber) {
                     val tr = translateExpression(binExpr.left)
                     addToResult(result, tr, -1, tr.resultFpReg)
-                    addInstr(result, IRInstruction(Opcode.ADD, vmDt, fpReg1 = tr.resultFpReg, fpValue = (binExpr.right as PtNumber).number.toFloat()), null)
+                    addInstr(result, IRInstruction(Opcode.ADD, vmDt, fpReg1 = tr.resultFpReg, immediateFp = (binExpr.right as PtNumber).number.toFloat()), null)
                     ExpressionCodeResult(result, vmDt, -1, tr.resultFpReg)
                 } else {
                     val leftTr = translateExpression(binExpr.left)
@@ -915,7 +884,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 return if(binExpr.right is PtNumber) {
                     val tr = translateExpression(binExpr.left)
                     addToResult(result, tr, tr.resultReg, -1)
-                    addInstr(result, IRInstruction(Opcode.ADD, vmDt, reg1 = tr.resultReg, value=(binExpr.right as PtNumber).number.toInt()), null)
+                    addInstr(result, IRInstruction(Opcode.ADD, vmDt, reg1 = tr.resultReg, immediate = (binExpr.right as PtNumber).number.toInt()), null)
                     ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
                 } else {
                     val leftTr = translateExpression(binExpr.left)
@@ -935,7 +904,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         val tr = translateExpression(operand)
         addToResult(result, tr, tr.resultReg, -1)
         addInstr(result, if(knownAddress!=null)
-            IRInstruction(Opcode.ANDM, vmDt, reg1=tr.resultReg, value=knownAddress)
+            IRInstruction(Opcode.ANDM, vmDt, reg1=tr.resultReg, address = knownAddress)
         else
             IRInstruction(Opcode.ANDM, vmDt, reg1=tr.resultReg, labelSymbol = symbol)
             ,null)
@@ -947,7 +916,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         val tr = translateExpression(operand)
         addToResult(result, tr, tr.resultReg, -1)
         addInstr(result, if(knownAddress!=null)
-            IRInstruction(Opcode.ORM, vmDt, reg1=tr.resultReg, value = knownAddress)
+            IRInstruction(Opcode.ORM, vmDt, reg1=tr.resultReg, address = knownAddress)
         else
             IRInstruction(Opcode.ORM, vmDt, reg1=tr.resultReg, labelSymbol = symbol)
             , null)
@@ -966,13 +935,13 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 addToResult(result, tr, -1, tr.resultFpReg)
                 val ins = if(signed) {
                     if(knownAddress!=null)
-                        IRInstruction(Opcode.DIVSM, vmDt, fpReg1 = tr.resultFpReg, value = knownAddress)
+                        IRInstruction(Opcode.DIVSM, vmDt, fpReg1 = tr.resultFpReg, address = knownAddress)
                     else
                         IRInstruction(Opcode.DIVSM, vmDt, fpReg1 = tr.resultFpReg, labelSymbol = symbol)
                 }
                 else {
                     if(knownAddress!=null)
-                        IRInstruction(Opcode.DIVM, vmDt, fpReg1 = tr.resultFpReg, value = knownAddress)
+                        IRInstruction(Opcode.DIVM, vmDt, fpReg1 = tr.resultFpReg, address = knownAddress)
                     else
                         IRInstruction(Opcode.DIVM, vmDt, fpReg1 = tr.resultFpReg, labelSymbol = symbol)
                 }
@@ -987,13 +956,13 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 addToResult(result, tr, tr.resultReg, -1)
                 val ins = if(signed) {
                     if(knownAddress!=null)
-                        IRInstruction(Opcode.DIVSM, vmDt, reg1 = tr.resultReg, value = knownAddress)
+                        IRInstruction(Opcode.DIVSM, vmDt, reg1 = tr.resultReg, address = knownAddress)
                     else
                         IRInstruction(Opcode.DIVSM, vmDt, reg1 = tr.resultReg, labelSymbol = symbol)
                 }
                 else {
                     if(knownAddress!=null)
-                        IRInstruction(Opcode.DIVM, vmDt, reg1 = tr.resultReg, value = knownAddress)
+                        IRInstruction(Opcode.DIVM, vmDt, reg1 = tr.resultReg, address = knownAddress)
                     else
                         IRInstruction(Opcode.DIVM, vmDt, reg1 = tr.resultReg, labelSymbol = symbol)
                 }
@@ -1014,7 +983,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 val tr = translateExpression(operand)
                 addToResult(result, tr, -1, tr.resultFpReg)
                 addInstr(result, if(knownAddress!=null)
-                    IRInstruction(Opcode.MULM, vmDt, fpReg1 = tr.resultFpReg, value = knownAddress)
+                    IRInstruction(Opcode.MULM, vmDt, fpReg1 = tr.resultFpReg, address = knownAddress)
                 else
                     IRInstruction(Opcode.MULM, vmDt, fpReg1 = tr.resultFpReg, labelSymbol = symbol)
                     , null)
@@ -1027,7 +996,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 val tr = translateExpression(operand)
                 addToResult(result, tr, tr.resultReg, -1)
                 addInstr(result, if(knownAddress!=null)
-                    IRInstruction(Opcode.MULM, vmDt, reg1=tr.resultReg, value = knownAddress)
+                    IRInstruction(Opcode.MULM, vmDt, reg1=tr.resultReg, address = knownAddress)
                 else
                     IRInstruction(Opcode.MULM, vmDt, reg1=tr.resultReg, labelSymbol = symbol)
                     , null)
@@ -1041,7 +1010,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         if(vmDt==IRDataType.FLOAT) {
             if((operand as? PtNumber)?.number==1.0) {
                 addInstr(result, if(knownAddress!=null)
-                    IRInstruction(Opcode.DECM, vmDt, value=knownAddress)
+                    IRInstruction(Opcode.DECM, vmDt, address = knownAddress)
                 else
                     IRInstruction(Opcode.DECM, vmDt, labelSymbol = symbol)
                     , null)
@@ -1050,7 +1019,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 val tr = translateExpression(operand)
                 addToResult(result, tr, -1, tr.resultFpReg)
                 addInstr(result, if(knownAddress!=null)
-                    IRInstruction(Opcode.SUBM, vmDt, fpReg1=tr.resultFpReg, value=knownAddress)
+                    IRInstruction(Opcode.SUBM, vmDt, fpReg1=tr.resultFpReg, address = knownAddress)
                 else
                     IRInstruction(Opcode.SUBM, vmDt, fpReg1=tr.resultFpReg, labelSymbol = symbol)
                     , null)
@@ -1058,7 +1027,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         } else {
             if((operand as? PtNumber)?.number==1.0) {
                 addInstr(result, if(knownAddress!=null)
-                    IRInstruction(Opcode.DECM, vmDt, value=knownAddress)
+                    IRInstruction(Opcode.DECM, vmDt, address = knownAddress)
                 else
                     IRInstruction(Opcode.DECM, vmDt, labelSymbol = symbol)
                     , null)
@@ -1067,7 +1036,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 val tr = translateExpression(operand)
                 addToResult(result, tr, tr.resultReg, -1)
                 addInstr(result, if(knownAddress!=null)
-                    IRInstruction(Opcode.SUBM, vmDt, reg1=tr.resultReg, value = knownAddress)
+                    IRInstruction(Opcode.SUBM, vmDt, reg1=tr.resultReg, address = knownAddress)
                 else
                     IRInstruction(Opcode.SUBM, vmDt, reg1=tr.resultReg, labelSymbol = symbol)
                     , null)
@@ -1081,7 +1050,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         if(vmDt==IRDataType.FLOAT) {
             if((operand as? PtNumber)?.number==1.0) {
                 addInstr(result, if(knownAddress!=null)
-                    IRInstruction(Opcode.INCM, vmDt, value = knownAddress)
+                    IRInstruction(Opcode.INCM, vmDt, address = knownAddress)
                 else
                     IRInstruction(Opcode.INCM, vmDt, labelSymbol = symbol)
                     , null)
@@ -1090,7 +1059,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 val tr = translateExpression(operand)
                 addToResult(result, tr, -1, tr.resultFpReg)
                 addInstr(result, if(knownAddress!=null)
-                        IRInstruction(Opcode.ADDM, vmDt, fpReg1=tr.resultFpReg, value = knownAddress)
+                        IRInstruction(Opcode.ADDM, vmDt, fpReg1=tr.resultFpReg, address = knownAddress)
                     else
                         IRInstruction(Opcode.ADDM, vmDt, fpReg1=tr.resultFpReg, labelSymbol = symbol)
                         , null)
@@ -1098,7 +1067,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         } else {
             if((operand as? PtNumber)?.number==1.0) {
                 addInstr(result, if(knownAddress!=null)
-                    IRInstruction(Opcode.INCM, vmDt, value = knownAddress)
+                    IRInstruction(Opcode.INCM, vmDt, address = knownAddress)
                 else
                     IRInstruction(Opcode.INCM, vmDt, labelSymbol = symbol)
                     , null)
@@ -1107,7 +1076,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 val tr = translateExpression(operand)
                 addToResult(result, tr, tr.resultReg, -1)
                 addInstr(result, if(knownAddress!=null)
-                        IRInstruction(Opcode.ADDM, vmDt, reg1=tr.resultReg, value=knownAddress)
+                        IRInstruction(Opcode.ADDM, vmDt, reg1=tr.resultReg, address = knownAddress)
                     else
                         IRInstruction(Opcode.ADDM, vmDt, reg1=tr.resultReg, labelSymbol = symbol)
                     , null)
@@ -1121,7 +1090,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         if(codeGen.isOne(operand)) {
             val opc = if (signed) Opcode.ASRM else Opcode.LSRM
             val ins = if(knownAddress!=null)
-                IRInstruction(opc, vmDt, value=knownAddress)
+                IRInstruction(opc, vmDt, address = knownAddress)
             else
                 IRInstruction(opc, vmDt, labelSymbol = symbol)
             addInstr(result, ins, null)
@@ -1130,7 +1099,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             addToResult(result, tr, tr.resultReg, -1)
             val opc = if (signed) Opcode.ASRNM else Opcode.LSRNM
             val ins = if(knownAddress!=null)
-                IRInstruction(opc, vmDt, reg1 = tr.resultReg, value=knownAddress)
+                IRInstruction(opc, vmDt, reg1 = tr.resultReg, address = knownAddress)
             else
                 IRInstruction(opc, vmDt, reg1 = tr.resultReg, labelSymbol = symbol)
             addInstr(result, ins, null)
@@ -1142,7 +1111,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         val result = mutableListOf<IRCodeChunkBase>()
         if(codeGen.isOne(operand)){
             addInstr(result, if(knownAddress!=null)
-                IRInstruction(Opcode.LSLM, vmDt, value=knownAddress)
+                IRInstruction(Opcode.LSLM, vmDt, address = knownAddress)
             else
                 IRInstruction(Opcode.LSLM, vmDt, labelSymbol = symbol)
                 , null)
@@ -1150,7 +1119,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             val tr = translateExpression(operand)
             addToResult(result, tr, tr.resultReg, -1)
             addInstr(result, if(knownAddress!=null)
-                IRInstruction(Opcode.LSLNM, vmDt, reg1=tr.resultReg, value=knownAddress)
+                IRInstruction(Opcode.LSLNM, vmDt, reg1=tr.resultReg, address = knownAddress)
             else
                 IRInstruction(Opcode.LSLNM, vmDt, reg1=tr.resultReg, labelSymbol = symbol)
                 ,null)
@@ -1163,7 +1132,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         val tr = translateExpression(operand)
         addToResult(result, tr, tr.resultReg, -1)
         addInstr(result, if(knownAddress!=null)
-            IRInstruction(Opcode.XORM, vmDt, reg1=tr.resultReg, value = knownAddress)
+            IRInstruction(Opcode.XORM, vmDt, reg1=tr.resultReg, address = knownAddress)
         else
             IRInstruction(Opcode.XORM, vmDt, reg1=tr.resultReg, labelSymbol = symbol)
             ,null)
@@ -1178,15 +1147,15 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             if (knownAddress != null) {
                 // @(address) = @(address) %= operand
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.LOADM, vmDt, reg1 = resultReg, value = knownAddress)
-                    it += IRInstruction(Opcode.MOD, vmDt, reg1 = resultReg, value = number)
-                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = resultReg, value = knownAddress)
+                    it += IRInstruction(Opcode.LOADM, vmDt, reg1 = resultReg, address = knownAddress)
+                    it += IRInstruction(Opcode.MOD, vmDt, reg1 = resultReg, immediate = number)
+                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = resultReg, address = knownAddress)
                 }
             } else {
                 // symbol = symbol %= operand
                 result += IRCodeChunk(null, null).also {
                     it += IRInstruction(Opcode.LOADM, vmDt, reg1 = resultReg, labelSymbol = symbol)
-                    it += IRInstruction(Opcode.MOD, vmDt, reg1 = resultReg, value = number)
+                    it += IRInstruction(Opcode.MOD, vmDt, reg1 = resultReg, immediate = number)
                     it += IRInstruction(Opcode.STOREM, vmDt, reg1 = resultReg, labelSymbol = symbol)
                 }
             }
@@ -1196,9 +1165,9 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             if (knownAddress != null) {
                 // @(address) = @(address) %= operand
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.LOADM, vmDt, reg1 = resultReg, value = knownAddress)
+                    it += IRInstruction(Opcode.LOADM, vmDt, reg1 = resultReg, address = knownAddress)
                     it += IRInstruction(Opcode.MODR, vmDt, reg1 = resultReg, reg2 = tr.resultReg)
-                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = resultReg, value = knownAddress)
+                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = resultReg, address = knownAddress)
                 }
             } else {
                 // symbol = symbol %= operand
@@ -1274,23 +1243,24 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
     ): MutableList<IRCodeChunkBase> {
         val result = mutableListOf<IRCodeChunkBase>()
         val valueReg = codeGen.registers.nextFree()
+        val cmpResultReg = codeGen.registers.nextFree()
         if(operand is PtNumber) {
             val numberReg = codeGen.registers.nextFree()
             if (knownAddress != null) {
                 // in-place modify a memory location
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.LOADM, vmDt, reg1 = valueReg, value = knownAddress)
-                    it += IRInstruction(Opcode.LOAD, vmDt, reg1=numberReg, value=operand.number.toInt())
-                    it += IRInstruction(compareAndSetOpcode, vmDt, reg1 = valueReg, reg2 = numberReg)
-                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = valueReg, value = knownAddress)
+                    it += IRInstruction(Opcode.LOADM, vmDt, reg1 = valueReg, address = knownAddress)
+                    it += IRInstruction(Opcode.LOAD, vmDt, reg1=numberReg, immediate = operand.number.toInt())
+                    it += IRInstruction(compareAndSetOpcode, vmDt, reg1 = cmpResultReg, reg2 = valueReg, reg3 = numberReg)
+                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = cmpResultReg, address = knownAddress)
                 }
             } else {
                 // in-place modify a symbol (variable)
                 result += IRCodeChunk(null, null).also {
                     it += IRInstruction(Opcode.LOADM, vmDt, reg1 = valueReg, labelSymbol = symbol)
-                    it += IRInstruction(Opcode.LOAD, vmDt, reg1=numberReg, value=operand.number.toInt())
-                    it += IRInstruction(compareAndSetOpcode, vmDt, reg1 = valueReg, reg2 = numberReg)
-                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = valueReg, labelSymbol = symbol)
+                    it += IRInstruction(Opcode.LOAD, vmDt, reg1=numberReg, immediate = operand.number.toInt())
+                    it += IRInstruction(compareAndSetOpcode, vmDt, reg1=cmpResultReg, reg2 = valueReg, reg3 = numberReg)
+                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = cmpResultReg, labelSymbol = symbol)
                 }
             }
         } else {
@@ -1299,16 +1269,16 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             if (knownAddress != null) {
                 // in-place modify a memory location
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.LOADM, vmDt, reg1 = valueReg, value = knownAddress)
-                    it += IRInstruction(compareAndSetOpcode, vmDt, reg1 = valueReg, reg2 = tr.resultReg)
-                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = valueReg, value = knownAddress)
+                    it += IRInstruction(Opcode.LOADM, vmDt, reg1 = valueReg, address = knownAddress)
+                    it += IRInstruction(compareAndSetOpcode, vmDt, reg1=cmpResultReg, reg2 = valueReg, reg3 = tr.resultReg)
+                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = cmpResultReg, address = knownAddress)
                 }
             } else {
                 // in-place modify a symbol (variable)
                 result += IRCodeChunk(null, null).also {
                     it += IRInstruction(Opcode.LOADM, vmDt, reg1 = valueReg, labelSymbol = symbol)
-                    it += IRInstruction(compareAndSetOpcode, vmDt, reg1 = valueReg, reg2 = tr.resultReg)
-                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = valueReg, labelSymbol = symbol)
+                    it += IRInstruction(compareAndSetOpcode, vmDt, reg1=cmpResultReg, reg2 = valueReg, reg3 = tr.resultReg)
+                    it += IRInstruction(Opcode.STOREM, vmDt, reg1 = cmpResultReg, labelSymbol = symbol)
                 }
             }
         }
@@ -1327,50 +1297,52 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         val zeroReg = codeGen.registers.nextFree()
         if(operand is PtNumber) {
             val numberReg = codeGen.registers.nextFreeFloat()
+            val cmpResultReg = codeGen.registers.nextFree()
             if (knownAddress != null) {
                 // in-place modify a memory location
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.LOADM, IRDataType.FLOAT, fpReg1 = valueReg, value = knownAddress)
-                    it += IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = numberReg, fpValue = operand.number.toFloat())
+                    it += IRInstruction(Opcode.LOADM, IRDataType.FLOAT, fpReg1 = valueReg, address = knownAddress)
+                    it += IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = numberReg, immediateFp = operand.number.toFloat())
                     it += IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg, fpReg2 = numberReg)
-                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, value=0)
-                    it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpReg, reg2=zeroReg)
-                    it += IRInstruction(Opcode.FFROMUB, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg)
-                    it += IRInstruction(Opcode.STOREM, IRDataType.FLOAT, fpReg1 = valueReg, value=knownAddress)
+                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, immediate = 0)
+                    it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpResultReg, reg2=cmpReg, reg3=zeroReg)
+                    it += IRInstruction(Opcode.FFROMUB, IRDataType.FLOAT, reg1=cmpResultReg, fpReg1 = valueReg)
+                    it += IRInstruction(Opcode.STOREM, IRDataType.FLOAT, fpReg1 = valueReg, address = knownAddress)
                 }
             } else {
                 // in-place modify a symbol (variable)
                 result += IRCodeChunk(null, null).also {
                     it += IRInstruction(Opcode.LOADM, IRDataType.FLOAT, fpReg1 = valueReg, labelSymbol = symbol)
-                    it += IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = numberReg, fpValue = operand.number.toFloat())
+                    it += IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = numberReg, immediateFp = operand.number.toFloat())
                     it += IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg, fpReg2 = numberReg)
-                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, value=0)
-                    it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpReg, reg2=zeroReg)
-                    it += IRInstruction(Opcode.FFROMUB, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg)
+                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, immediate = 0)
+                    it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpResultReg, reg2=cmpReg, reg3=zeroReg)
+                    it += IRInstruction(Opcode.FFROMUB, IRDataType.FLOAT, reg1=cmpResultReg, fpReg1 = valueReg)
                     it += IRInstruction(Opcode.STOREM, IRDataType.FLOAT, fpReg1 = valueReg, labelSymbol = symbol)
                 }
             }
         } else {
             val tr = translateExpression(operand)
+            val cmpResultReg = codeGen.registers.nextFree()
             addToResult(result, tr, -1, tr.resultFpReg)
             if (knownAddress != null) {
                 // in-place modify a memory location
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.LOADM, IRDataType.FLOAT, fpReg1 = valueReg, value = knownAddress)
+                    it += IRInstruction(Opcode.LOADM, IRDataType.FLOAT, fpReg1 = valueReg, address = knownAddress)
                     it += IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg, fpReg2 = tr.resultFpReg)
-                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, value=0)
-                    it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpReg, reg2=zeroReg)
-                    it += IRInstruction(Opcode.FFROMUB, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg)
-                    it += IRInstruction(Opcode.STOREM, IRDataType.FLOAT, fpReg1 = valueReg, value=knownAddress)
+                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, immediate = 0)
+                    it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpResultReg, reg2=cmpReg, reg3=zeroReg)
+                    it += IRInstruction(Opcode.FFROMUB, IRDataType.FLOAT, reg1=cmpResultReg, fpReg1 = valueReg)
+                    it += IRInstruction(Opcode.STOREM, IRDataType.FLOAT, fpReg1 = valueReg, address = knownAddress)
                 }
             } else {
                 // in-place modify a symbol (variable)
                 result += IRCodeChunk(null, null).also {
                     it += IRInstruction(Opcode.LOADM, IRDataType.FLOAT, fpReg1 = valueReg, labelSymbol = symbol)
                     it += IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg, fpReg2 = tr.resultFpReg)
-                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, value=0)
-                    it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpReg, reg2=zeroReg)
-                    it += IRInstruction(Opcode.FFROMUB, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg)
+                    it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, immediate = 0)
+                    it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpResultReg, reg2=cmpReg, reg3=zeroReg)
+                    it += IRInstruction(Opcode.FFROMUB, IRDataType.FLOAT, reg1=cmpResultReg, fpReg1 = valueReg)
                     it += IRInstruction(Opcode.STOREM, IRDataType.FLOAT, fpReg1 = valueReg, labelSymbol = symbol)
                 }
             }

codeGenIntermediate/src/prog8/codegen/intermediate/IRPeepholeOptimizer.kt

@@ -1,12 +1,42 @@
 package prog8.codegen.intermediate
 
+import prog8.code.core.IErrorReporter
 import prog8.intermediate.*
 
-internal class IRPeepholeOptimizer(private val irprog: IRProgram) {
-    fun optimize() {
-        irprog.blocks.asSequence().flatMap { it.children.filterIsInstance<IRSubroutine>() }.forEach { sub ->
+class IRPeepholeOptimizer(private val irprog: IRProgram) {
+    fun optimize(optimizationsEnabled: Boolean, errors: IErrorReporter) {
+        if(!optimizationsEnabled)
+            return optimizeOnlyJoinChunks()
+
+        peepholeOptimize()
+        val remover = IRUnusedCodeRemover(irprog, errors)
+        var totalRemovals = 0
+        do {
+            val numRemoved = remover.optimize()
+            totalRemovals += numRemoved
+        } while(numRemoved>0 && errors.noErrors())
+        errors.report()
+
+        if(totalRemovals>0) {
+            irprog.linkChunks()  // re-link again.
+        }
+    }
+
+    private fun optimizeOnlyJoinChunks() {
+        irprog.foreachSub { sub ->
+            joinChunks(sub)
             removeEmptyChunks(sub)
             joinChunks(sub)
+        }
+        irprog.linkChunks() // re-link
+    }
+
+    private fun peepholeOptimize() {
+        irprog.foreachSub { sub ->
+            joinChunks(sub)
+            removeEmptyChunks(sub)
+            joinChunks(sub)
+
             sub.chunks.withIndex().forEach { (index, chunk1) ->
                 // we don't optimize Inline Asm chunks here.
                 val chunk2 = if(index<sub.chunks.size-1) sub.chunks[index+1] else null
@@ -17,18 +47,20 @@ internal class IRPeepholeOptimizer(private val irprog: IRProgram) {
                         val changed = removeNops(chunk1, indexedInstructions)
                                 || removeDoubleLoadsAndStores(chunk1, indexedInstructions)       // TODO not yet implemented
                                 || removeUselessArithmetic(chunk1, indexedInstructions)
+                                || removeNeedlessCompares(chunk1, indexedInstructions)
                                 || removeWeirdBranches(chunk1, chunk2, indexedInstructions)
                                 || removeDoubleSecClc(chunk1, indexedInstructions)
                                 || cleanupPushPop(chunk1, indexedInstructions)
-                        // TODO other optimizations:
-                        //  more complex optimizations such as unused registers
+                        // TODO other optimizations
                     } while (changed)
                 }
             }
             removeEmptyChunks(sub)
         }
 
-        irprog.linkChunks() // re-link
+        // TODO also do register optimization step here at the end?
+
+        irprog.linkChunks()  // re-link
     }
 
     private fun removeEmptyChunks(sub: IRSubroutine) {
@@ -38,7 +70,7 @@ internal class IRPeepholeOptimizer(private val irprog: IRProgram) {
         /*
         Empty Code chunk with label ->
             If next chunk has no label -> move label to next chunk, remove original
-            If next chunk has label -> label name should be the same, remove original. Otherwise FOR NOW leave it in place. (TODO: consolidate labels into 1)
+            If next chunk has label -> label name should be the same, remove original. Otherwise FOR NOW leave it in place. (TODO: merge both labels into 1)
             If is last chunk -> keep chunk in place because of the label.
         Empty Code chunk without label ->
             should not have been generated! ERROR.
@@ -63,7 +95,7 @@ internal class IRPeepholeOptimizer(private val irprog: IRProgram) {
                             if (chunk.label == nextchunk.label)
                                 removeChunks += index
                             else {
-                                // TODO: consolidate labels on same chunk
+                                // TODO: merge labels on same chunk
                             }
                         }
                     }
@@ -85,7 +117,7 @@ internal class IRPeepholeOptimizer(private val irprog: IRProgram) {
         if(sub.chunks.isEmpty())
             return
 
-        fun mayJoin(previous: IRCodeChunkBase, chunk: IRCodeChunkBase): Boolean {
+        fun mayJoinCodeChunks(previous: IRCodeChunkBase, chunk: IRCodeChunkBase): Boolean {
             if(chunk.label!=null)
                 return false
             if(previous is IRCodeChunk && chunk is IRCodeChunk) {
@@ -102,12 +134,39 @@ internal class IRPeepholeOptimizer(private val irprog: IRProgram) {
         chunks += sub.chunks[0]
         for(ix in 1 until sub.chunks.size) {
             val lastChunk = chunks.last()
-            if(mayJoin(lastChunk, sub.chunks[ix])) {
-                lastChunk.instructions += sub.chunks[ix].instructions
-                lastChunk.next = sub.chunks[ix].next
+            val candidate = sub.chunks[ix]
+            when(candidate) {
+                is IRCodeChunk -> {
+                    if(mayJoinCodeChunks(lastChunk, candidate)) {
+                        lastChunk.instructions += candidate.instructions
+                        lastChunk.next = candidate.next
+                    }
+                    else
+                        chunks += candidate
+                }
+                is IRInlineAsmChunk -> {
+                    if(candidate.label!=null)
+                        chunks += candidate
+                    else if(lastChunk.isEmpty()) {
+                        val label = lastChunk.label
+                        if(label!=null)
+                            chunks += IRInlineAsmChunk(label, candidate.assembly, candidate.isIR, candidate.next)
+                        else
+                            chunks += candidate
+                    }
+                }
+                is IRInlineBinaryChunk -> {
+                    if(candidate.label!=null)
+                        chunks += candidate
+                    else if(lastChunk.isEmpty()) {
+                        val label = lastChunk.label
+                        if(label!=null)
+                            chunks += IRInlineBinaryChunk(label, candidate.data, candidate.next)
+                        else
+                            chunks += candidate
+                    }
+                }
             }
-            else
-                chunks += sub.chunks[ix]
         }
         sub.chunks.clear()
         sub.chunks += chunks
@@ -176,7 +235,7 @@ internal class IRPeepholeOptimizer(private val irprog: IRProgram) {
             }
 
             // remove useless RETURN
-            if(idx>0 && (ins.opcode == Opcode.RETURN || ins.opcode==Opcode.RETURNREG)) {
+            if(idx>0 && (ins.opcode == Opcode.RETURN || ins.opcode==Opcode.RETURNR)) {
                 val previous = chunk.instructions[idx-1]
                 if(previous.opcode in OpcodesThatJump) {
                     chunk.instructions.removeAt(idx)
@@ -194,12 +253,36 @@ internal class IRPeepholeOptimizer(private val irprog: IRProgram) {
             }
 
             // replace call + return --> jump
-            if(idx>0 && ins.opcode==Opcode.RETURN) {
-                val previous = chunk.instructions[idx-1]
-                if(previous.opcode==Opcode.CALL || previous.opcode==Opcode.CALLRVAL) {
-                    chunk.instructions[idx-1] = IRInstruction(Opcode.JUMP, value=previous.value, labelSymbol = previous.labelSymbol, branchTarget = previous.branchTarget)
+            // This can no longer be done here on the IR level, with the current CALL opcode that encodes the full subroutine call setup.
+            // If machine code is ever generated from this IR, *that* should possibly optimize the JSR + RTS into a JMP.
+//            if(idx>0 && ins.opcode==Opcode.RETURN) {
+//                val previous = chunk.instructions[idx-1]
+//                if(previous.opcode==Opcode.CALL) {
+//                    chunk.instructions[idx-1] = IRInstruction(Opcode.JUMP, address = previous.address, labelSymbol = previous.labelSymbol, branchTarget = previous.branchTarget)
+//                    chunk.instructions.removeAt(idx)
+//                    changed = true
+//                }
+//            }
+        }
+        return changed
+    }
+
+    private fun removeNeedlessCompares(chunk: IRCodeChunk, indexedInstructions: List<IndexedValue<IRInstruction>>): Boolean {
+        // a CMPI with 0, after an instruction like LOAD that already sets the status bits, can be removed.
+        // but only if the instruction after it is not using the Carry bit because that won't be set by a LOAD instruction etc.
+        var changed = false
+        indexedInstructions.reversed().forEach { (idx, ins) ->
+            if(idx>0 && idx<(indexedInstructions.size-1) && ins.opcode==Opcode.CMPI && ins.immediate==0) {
+                val previous = indexedInstructions[idx-1].value
+                if(previous.opcode in OpcodesThatSetStatusbitsIncludingCarry) {
                     chunk.instructions.removeAt(idx)
                     changed = true
+                } else if(previous.opcode in OpcodesThatSetStatusbitsButNotCarry) {
+                    val next = indexedInstructions[idx+1].value
+                    if(next.opcode !in arrayOf(Opcode.BSTCC, Opcode.BSTCS, Opcode.BSTPOS, Opcode.BSTNEG)) {
+                        chunk.instructions.removeAt(idx)
+                        changed = true
+                    }
                 }
             }
         }
@@ -212,47 +295,47 @@ internal class IRPeepholeOptimizer(private val irprog: IRProgram) {
         indexedInstructions.reversed().forEach { (idx, ins) ->
             when (ins.opcode) {
                 Opcode.DIV, Opcode.DIVS, Opcode.MUL, Opcode.MOD -> {
-                    if (ins.value == 1) {
+                    if (ins.immediate == 1) {
                         chunk.instructions.removeAt(idx)
                         changed = true
                     }
                 }
                 Opcode.ADD, Opcode.SUB -> {
-                    if (ins.value == 1) {
+                    if (ins.immediate == 1) {
                         chunk.instructions[idx] = IRInstruction(
                             if (ins.opcode == Opcode.ADD) Opcode.INC else Opcode.DEC,
                             ins.type,
                             ins.reg1
                         )
                         changed = true
-                    } else if (ins.value == 0) {
+                    } else if (ins.immediate == 0) {
                         chunk.instructions.removeAt(idx)
                         changed = true
                     }
                 }
                 Opcode.AND -> {
-                    if (ins.value == 0) {
-                        chunk.instructions[idx] = IRInstruction(Opcode.LOAD, ins.type, reg1 = ins.reg1, value = 0)
+                    if (ins.immediate == 0) {
+                        chunk.instructions[idx] = IRInstruction(Opcode.LOAD, ins.type, reg1 = ins.reg1, immediate = 0)
                         changed = true
-                    } else if (ins.value == 255 && ins.type == IRDataType.BYTE) {
+                    } else if (ins.immediate == 255 && ins.type == IRDataType.BYTE) {
                         chunk.instructions.removeAt(idx)
                         changed = true
-                    } else if (ins.value == 65535 && ins.type == IRDataType.WORD) {
+                    } else if (ins.immediate == 65535 && ins.type == IRDataType.WORD) {
                         chunk.instructions.removeAt(idx)
                         changed = true
                     }
                 }
                 Opcode.OR -> {
-                    if (ins.value == 0) {
+                    if (ins.immediate == 0) {
                         chunk.instructions.removeAt(idx)
                         changed = true
-                    } else if ((ins.value == 255 && ins.type == IRDataType.BYTE) || (ins.value == 65535 && ins.type == IRDataType.WORD)) {
-                        chunk.instructions[idx] = IRInstruction(Opcode.LOAD, ins.type, reg1 = ins.reg1, value = ins.value)
+                    } else if ((ins.immediate == 255 && ins.type == IRDataType.BYTE) || (ins.immediate == 65535 && ins.type == IRDataType.WORD)) {
+                        chunk.instructions[idx] = IRInstruction(Opcode.LOAD, ins.type, reg1 = ins.reg1, immediate = ins.immediate)
                         changed = true
                     }
                 }
                 Opcode.XOR -> {
-                    if (ins.value == 0) {
+                    if (ins.immediate == 0) {
                         chunk.instructions.removeAt(idx)
                         changed = true
                     }

codeGenIntermediate/src/prog8/codegen/intermediate/IRRegisterOptimizer.kt

@@ -0,0 +1,103 @@
+package prog8.codegen.intermediate
+
+import prog8.intermediate.IRProgram
+
+
+class IRRegisterOptimizer(private val irProg: IRProgram) {
+    fun optimize() {
+        // reuseRegisters()
+    }
+
+/*
+    TODO: this register re-use renumbering isn't going to work like this,
+    because subroutines will be clobbering the registers that the subroutine
+    which is calling them might be using...
+
+
+    private fun reuseRegisters() {
+
+        fun addToUsage(usage: MutableMap<Pair<Int, IRDataType>, MutableSet<IRCodeChunkBase>>,
+                       regnum: Int,
+                       dt: IRDataType,
+                       chunk: IRCodeChunkBase) {
+            val key = regnum to dt
+            val chunks = usage[key] ?: mutableSetOf()
+            chunks.add(chunk)
+            usage[key] = chunks
+        }
+
+        val usage: MutableMap<Pair<Int, IRDataType>, MutableSet<IRCodeChunkBase>> = mutableMapOf()
+
+        irProg.foreachCodeChunk { chunk ->
+            chunk.usedRegisters().regsTypes.forEach { (regNum, types) ->
+                types.forEach { dt ->
+                    addToUsage(usage, regNum, dt, chunk)
+                }
+            }
+        }
+
+        val registerReplacements = usage.asSequence()
+            .filter { it.value.size==1 }
+            .map { it.key to it.value.iterator().next() }
+            .groupBy({ it.second }, {it.first})
+            .asSequence()
+            .associate { (chunk, registers) ->
+                chunk to registers.withIndex().associate { (index, reg) -> reg to 50000+index }
+            }
+
+        registerReplacements.forEach { replaceRegisters(it.key, it.value) }
+    }
+
+    private fun replaceRegisters(chunk: IRCodeChunkBase, replacements: Map<Pair<Int, IRDataType>, Int>) {
+        val (rF, rI) = replacements.asSequence().partition { it.key.second==IRDataType.FLOAT }
+        val replacementsInt = rI.associate { it.key.first to it.value }
+        val replacementsFloat = rF.associate { it.key.first to it.value }
+
+        fun replaceRegs(fcallArgs: FunctionCallArgs?): FunctionCallArgs? {
+            if(fcallArgs==null)
+                return null
+            val args = if(fcallArgs.arguments.isEmpty()) fcallArgs.arguments else {
+                fcallArgs.arguments.map {
+                    FunctionCallArgs.ArgumentSpec(
+                        it.name,
+                        it.address,
+                        FunctionCallArgs.RegSpec(
+                            it.reg.dt,
+                            if(it.reg.dt==IRDataType.FLOAT)
+                                replacementsFloat.getOrDefault(it.reg.registerNum, it.reg.registerNum)
+                            else
+                                replacementsInt.getOrDefault(it.reg.registerNum, it.reg.registerNum),
+                            it.reg.cpuRegister
+                        )
+                    )
+                }
+            }
+            val rt = fcallArgs.returns
+            val returns = if(rt==null) null else {
+                FunctionCallArgs.RegSpec(
+                    rt.dt,
+                    if(rt.dt==IRDataType.FLOAT)
+                        replacementsFloat.getOrDefault(rt.registerNum, rt.registerNum)
+                    else
+                        replacementsInt.getOrDefault(rt.registerNum, rt.registerNum),
+                    rt.cpuRegister
+                )
+            }
+            return FunctionCallArgs(args, returns)
+        }
+
+        fun replaceRegs(instruction: IRInstruction): IRInstruction {
+            val reg1 = replacementsInt.getOrDefault(instruction.reg1, instruction.reg1)
+            val reg2 = replacementsInt.getOrDefault(instruction.reg2, instruction.reg2)
+            val fpReg1 = replacementsFloat.getOrDefault(instruction.fpReg1, instruction.fpReg1)
+            val fpReg2 = replacementsFloat.getOrDefault(instruction.fpReg2, instruction.fpReg2)
+            return instruction.copy(reg1 = reg1, reg2 = reg2, fpReg1 = fpReg1, fpReg2 = fpReg2, fcallArgs = replaceRegs(instruction.fcallArgs))
+        }
+        val newInstructions = chunk.instructions.map {
+            replaceRegs(it)
+        }
+        chunk.instructions.clear()
+        chunk.instructions.addAll(newInstructions)
+    }
+*/
+}

codeGenIntermediate/src/prog8/codegen/intermediate/IRUnusedCodeRemover.kt

@@ -5,35 +5,18 @@ import prog8.code.core.SourceCode.Companion.libraryFilePrefix
 import prog8.intermediate.*
 
 
-internal class IRUnusedCodeRemover(private val irprog: IRProgram, private val errors: IErrorReporter) {
+class IRUnusedCodeRemover(
+    private val irprog: IRProgram,
+    private val errors: IErrorReporter
+) {
     fun optimize(): Int {
-        val allLabeledChunks = mutableMapOf<String, IRCodeChunkBase>()
-
-        irprog.blocks.asSequence().flatMap { it.children.filterIsInstance<IRSubroutine>() }.forEach { sub ->
-            sub.chunks.forEach { chunk ->
-                chunk.label?.let { allLabeledChunks[it] = chunk }
-            }
-        }
-
-        var numRemoved = removeSimpleUnlinked(allLabeledChunks) + removeUnreachable(allLabeledChunks)
-
-        // remove empty subs
-        irprog.blocks.forEach { block ->
-            block.children.filterIsInstance<IRSubroutine>().reversed().forEach { sub ->
-                if(sub.isEmpty()) {
-                    if(!sub.position.file.startsWith(libraryFilePrefix)) {
-                        errors.warn("unused subroutine ${sub.label}", sub.position)
-                    }
-                    block.children.remove(sub)
-                    numRemoved++
-                }
-            }
-        }
+        var numRemoved = removeUnusedSubroutines() + removeUnusedAsmSubroutines()
 
         // remove empty blocks
         irprog.blocks.reversed().forEach { block ->
             if(block.isEmpty()) {
                 irprog.blocks.remove(block)
+                pruneSymboltable(block.label)
                 numRemoved++
             }
         }
@@ -41,10 +24,133 @@ internal class IRUnusedCodeRemover(private val irprog: IRProgram, private val er
         return numRemoved
     }
 
+    private fun pruneSymboltable(blockLabel: String) {
+        // we could clean up the SymbolTable as well, but ONLY if these symbols aren't referenced somewhere still in an instruction
+        val prefix = "$blockLabel."
+        val blockVars = irprog.st.allVariables().filter { it.name.startsWith(prefix) }
+        blockVars.forEach { stVar ->
+            irprog. allSubs().flatMap { it.chunks }.forEach { chunk ->
+                chunk.instructions.forEach { ins ->
+                    if(ins.labelSymbol == stVar.name) {
+                        return
+                    }
+                }
+            }
+        }
+        irprog.st.removeTree(blockLabel)
+    }
+
+    private fun removeUnusedSubroutines(): Int {
+        val allLabeledChunks = mutableMapOf<String, IRCodeChunkBase>()
+        irprog.foreachCodeChunk { chunk ->
+            chunk.label?.let { allLabeledChunks[it] = chunk }
+        }
+
+        var numRemoved = removeSimpleUnlinked(allLabeledChunks) + removeUnreachable(allLabeledChunks)
+        irprog.blocks.forEach { block ->
+            block.children.filterIsInstance<IRSubroutine>().reversed().forEach { sub ->
+                if(sub.isEmpty()) {
+                    if(!sub.position.file.startsWith(libraryFilePrefix)) {
+                        errors.warn("unused subroutine ${sub.label}", sub.position)
+                    }
+                    block.children.remove(sub)
+                    irprog.st.removeTree(sub.label)
+                    numRemoved++
+                }
+            }
+        }
+
+        return numRemoved
+    }
+
+    private fun removeUnusedAsmSubroutines(): Int {
+        val allLabeledAsmsubs = irprog.blocks.asSequence().flatMap { it.children.filterIsInstance<IRAsmSubroutine>() }
+            .associateBy { it.label }
+
+        var numRemoved = removeSimpleUnlinkedAsmsubs(allLabeledAsmsubs)
+        irprog.blocks.forEach { block ->
+            block.children.filterIsInstance<IRAsmSubroutine>().reversed().forEach { sub ->
+                if(sub.isEmpty()) {
+                    if(!sub.position.file.startsWith(libraryFilePrefix)) {
+                        errors.warn("unused asmsubroutine ${sub.label}", sub.position)
+                    }
+                    block.children.remove(sub)
+                    irprog.st.removeTree(sub.label)
+                    numRemoved++
+                }
+            }
+        }
+
+        return numRemoved
+    }
+
+    private fun removeSimpleUnlinkedAsmsubs(allSubs: Map<String, IRAsmSubroutine>): Int {
+        val linkedAsmSubs = mutableSetOf<IRAsmSubroutine>()
+
+        // TODO: asmsubs in library modules are never removed, we can't really tell here if they're actually being called or not...
+
+        // check if asmsub is called from another asmsub
+        irprog.blocks.asSequence().forEach { block ->
+            block.children.filterIsInstance<IRAsmSubroutine>().forEach { sub ->
+                if (block.forceOutput || block.library)
+                    linkedAsmSubs += sub
+                if (sub.asmChunk.isNotEmpty()) {
+                    allSubs.forEach { (label, asmsub) ->
+                        if (sub.asmChunk.assembly.contains(label))
+                            linkedAsmSubs += asmsub
+                    }
+                }
+                val inlineAsm = sub.asmChunk.next as? IRInlineAsmChunk
+                if(inlineAsm!=null) {
+                    allSubs.forEach { (label, asmsub) ->
+                        if (inlineAsm.assembly.contains(label))
+                            linkedAsmSubs += asmsub
+                    }
+                }
+            }
+        }
+
+        // check if asmsub is linked or called from another regular subroutine
+        irprog.foreachCodeChunk { chunk ->
+            chunk.instructions.forEach {
+                it.labelSymbol?.let { label -> allSubs[label]?.let { cc -> linkedAsmSubs += cc } }
+                // note: branchTarget can't yet point to another IRAsmSubroutine, so do nothing when it's set
+            }
+        }
+
+        return removeUnlinkedAsmsubs(linkedAsmSubs)
+    }
+
+    private fun removeUnlinkedAsmsubs(linkedAsmSubs: Set<IRAsmSubroutine>): Int {
+        var numRemoved = 0
+        irprog.blocks.asSequence().forEach { block ->
+            block.children.withIndex().reversed().forEach { (index, child) ->
+                if(child is IRAsmSubroutine && child !in linkedAsmSubs) {
+                    block.children.removeAt(index)
+                    numRemoved++
+                }
+            }
+        }
+        return numRemoved
+    }
+
     private fun removeUnreachable(allLabeledChunks: MutableMap<String, IRCodeChunkBase>): Int {
-        val entrypointSub = irprog.blocks.single { it.name=="main" }.children.single { it is IRSubroutine && it.label=="main.start" }
+        val entrypointSub = irprog.blocks.single { it.label=="main" }
+            .children.single { it is IRSubroutine && it.label=="main.start" }
         val reachable = mutableSetOf((entrypointSub as IRSubroutine).chunks.first())
 
+        // all chunks referenced in array initializer values are also 'reachable':
+        irprog.st.allVariables()
+            .filter { !it.uninitialized }
+            .forEach {
+                it.onetimeInitializationArrayValue?.let { array ->
+                    array.forEach {elt ->
+                        if(elt.addressOfSymbol!=null && irprog.st.lookup(elt.addressOfSymbol!!)==null)
+                            reachable.add(irprog.getChunkWithLabel(elt.addressOfSymbol!!))
+                    }
+                }
+            }
+
         fun grow() {
             val new = mutableSetOf<IRCodeChunkBase>()
             reachable.forEach {
@@ -73,29 +179,39 @@ internal class IRUnusedCodeRemover(private val irprog: IRProgram, private val er
     private fun removeSimpleUnlinked(allLabeledChunks: Map<String, IRCodeChunkBase>): Int {
         val linkedChunks = mutableSetOf<IRCodeChunkBase>()
 
-        irprog.blocks.asSequence().flatMap { it.children.filterIsInstance<IRSubroutine>() }.forEach { sub ->
-            sub.chunks.forEach { chunk ->
-                chunk.next?.let { next -> linkedChunks += next }
-                chunk.instructions.forEach {
-                    if(it.branchTarget==null) {
-                        it.labelSymbol?.let { label -> allLabeledChunks[label]?.let { cc -> linkedChunks += cc } }
-                    } else {
-                        linkedChunks += it.branchTarget!!
+        // all chunks referenced in array initializer values are linked as well!:
+        irprog.st.allVariables()
+            .filter { !it.uninitialized }
+            .forEach {
+                it.onetimeInitializationArrayValue?.let { array ->
+                    array.forEach {elt ->
+                        if(elt.addressOfSymbol!=null && irprog.st.lookup(elt.addressOfSymbol!!)==null)
+                            linkedChunks += irprog.getChunkWithLabel(elt.addressOfSymbol!!)
                     }
                 }
-                if (chunk.label == "main.start")
-                    linkedChunks += chunk
             }
+
+        irprog.foreachCodeChunk { chunk ->
+            chunk.next?.let { next -> linkedChunks += next }
+            chunk.instructions.forEach {
+                if(it.branchTarget==null) {
+                    it.labelSymbol?.let { label -> allLabeledChunks[label]?.let { cc -> linkedChunks += cc } }
+                } else {
+                    linkedChunks += it.branchTarget!!
+                }
+            }
+            if (chunk.label == "main.start")
+                linkedChunks += chunk
         }
 
         return removeUnlinkedChunks(linkedChunks)
     }
 
     private fun removeUnlinkedChunks(
-        linkedChunks: MutableSet<IRCodeChunkBase>
+        linkedChunks: Set<IRCodeChunkBase>
     ): Int {
         var numRemoved = 0
-        irprog.blocks.asSequence().flatMap { it.children.filterIsInstance<IRSubroutine>() }.forEach { sub ->
+        irprog.foreachSub { sub ->
             sub.chunks.withIndex().reversed().forEach { (index, chunk) ->
                 if (chunk !in linkedChunks) {
                     if (chunk === sub.chunks[0]) {

codeGenIntermediate/src/prog8/codegen/intermediate/RegisterPool.kt

@@ -1,11 +1,7 @@
 package prog8.codegen.intermediate
 
-import prog8.code.core.AssemblyError
-import prog8.intermediate.SyscallRegisterBase
-
 internal class RegisterPool {
-    // reserve 0,1,2 for return values of subroutine calls and syscalls
-    // TODO set this back to 0 once 'resultRegister' has been removed everywhere and SYSCALL/DIVMOD fixed?
+    // reserve 0,1,2 for return values of subroutine calls and syscalls in IR assembly code
     private var firstFree: Int=3
     private var firstFreeFloat: Int=3
 
@@ -15,16 +11,12 @@ internal class RegisterPool {
     fun nextFree(): Int {
         val result = firstFree
         firstFree++
-        if(firstFree >= SyscallRegisterBase)
-            throw AssemblyError("out of virtual registers (int)")
         return result
     }
 
     fun nextFreeFloat(): Int {
         val result = firstFreeFloat
         firstFreeFloat++
-        if(firstFreeFloat >= SyscallRegisterBase)
-            throw AssemblyError("out of virtual registers (fp)")
         return result
     }
 }

codeGenIntermediate/src/prog8/codegen/vm/VmCodeGen.kt

@@ -2,7 +2,10 @@ package prog8.codegen.vm
 
 import prog8.code.SymbolTable
 import prog8.code.ast.PtProgram
-import prog8.code.core.*
+import prog8.code.core.CompilationOptions
+import prog8.code.core.IAssemblyProgram
+import prog8.code.core.ICodeGeneratorBackend
+import prog8.code.core.IErrorReporter
 import prog8.codegen.intermediate.IRCodeGen
 import prog8.intermediate.IRFileWriter
 import prog8.intermediate.IRProgram

codeGenIntermediate/test/Dummies.kt

@@ -43,6 +43,10 @@ internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors:
             warnings.add(text)
     }
 
+    override fun undefined(symbol: List<String>, position: Position) {
+        err("undefined symbol: ${symbol.joinToString(".")}", position)
+    }
+
     override fun noErrors(): Boolean  = errors.isEmpty()
 
     override fun report() {

codeGenIntermediate/test/TestIRPeepholeOpt.kt

@@ -8,7 +8,7 @@ import prog8.intermediate.*
 class TestIRPeepholeOpt: FunSpec({
     fun makeIRProgram(chunks: List<IRCodeChunkBase>): IRProgram {
         require(chunks.first().label=="main.start")
-        val block = IRBlock("main", null, IRBlock.BlockAlignment.NONE, Position.DUMMY)
+        val block = IRBlock("main", null, false, false, IRBlock.BlockAlignment.NONE, Position.DUMMY)
         val sub = IRSubroutine("main.start", emptyList(), null, Position.DUMMY)
         chunks.forEach { sub += it }
         block += sub
@@ -36,17 +36,17 @@ class TestIRPeepholeOpt: FunSpec({
         return makeIRProgram(listOf(chunk))
     }
 
-    fun IRProgram.chunks(): List<IRCodeChunkBase> = this.blocks.flatMap { it.children.filterIsInstance<IRSubroutine>() }.flatMap { it.chunks }
+    fun IRProgram.chunks(): List<IRCodeChunkBase> = this.allSubs().flatMap { it.chunks }.toList()
 
     test("remove nops") {
         val irProg = makeIRProgram(listOf(
-            IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=1, value=42),
+            IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=1, immediate=42),
             IRInstruction(Opcode.NOP),
             IRInstruction(Opcode.NOP)
         ))
         irProg.chunks().single().instructions.size shouldBe 3
         val opt = IRPeepholeOptimizer(irProg)
-        opt.optimize()
+        opt.optimize(true, ErrorReporterForTests())
         irProg.chunks().single().instructions.size shouldBe 1
     }
 
@@ -65,7 +65,7 @@ class TestIRPeepholeOpt: FunSpec({
         irProg.chunks().size shouldBe 4
         irProg.chunks().flatMap { it.instructions }.size shouldBe 5
         val opt = IRPeepholeOptimizer(irProg)
-        opt.optimize()
+        opt.optimize(true, ErrorReporterForTests())
         irProg.chunks().size shouldBe 4
         irProg.chunks()[0].label shouldBe "main.start"
         irProg.chunks()[1].label shouldBe "label"
@@ -92,7 +92,7 @@ class TestIRPeepholeOpt: FunSpec({
         ))
         irProg.chunks().single().instructions.size shouldBe 6
         val opt = IRPeepholeOptimizer(irProg)
-        opt.optimize()
+        opt.optimize(true, ErrorReporterForTests())
         val instr = irProg.chunks().single().instructions
         instr.size shouldBe 1
         instr[0].opcode shouldBe Opcode.CLC
@@ -107,7 +107,7 @@ class TestIRPeepholeOpt: FunSpec({
         ))
         irProg.chunks().single().instructions.size shouldBe 4
         val opt = IRPeepholeOptimizer(irProg)
-        opt.optimize()
+        opt.optimize(true, ErrorReporterForTests())
         val instr = irProg.chunks().single().instructions
         instr.size shouldBe 1
         instr[0].opcode shouldBe Opcode.LOADR
@@ -117,31 +117,31 @@ class TestIRPeepholeOpt: FunSpec({
 
     test("remove useless div/mul, add/sub") {
         val irProg = makeIRProgram(listOf(
-            IRInstruction(Opcode.DIV, IRDataType.BYTE, reg1=42, value = 1),
-            IRInstruction(Opcode.DIVS, IRDataType.BYTE, reg1=42, value = 1),
-            IRInstruction(Opcode.MUL, IRDataType.BYTE, reg1=42, value = 1),
-            IRInstruction(Opcode.MOD, IRDataType.BYTE, reg1=42, value = 1),
-            IRInstruction(Opcode.DIV, IRDataType.BYTE, reg1=42, value = 2),
-            IRInstruction(Opcode.DIVS, IRDataType.BYTE, reg1=42, value = 2),
-            IRInstruction(Opcode.MUL, IRDataType.BYTE, reg1=42, value = 2),
-            IRInstruction(Opcode.MOD, IRDataType.BYTE, reg1=42, value = 2),
-            IRInstruction(Opcode.ADD, IRDataType.BYTE, reg1=42, value = 0),
-            IRInstruction(Opcode.SUB, IRDataType.BYTE, reg1=42, value = 0)
+            IRInstruction(Opcode.DIV, IRDataType.BYTE, reg1=42, immediate = 1),
+            IRInstruction(Opcode.DIVS, IRDataType.BYTE, reg1=42, immediate = 1),
+            IRInstruction(Opcode.MUL, IRDataType.BYTE, reg1=42, immediate = 1),
+            IRInstruction(Opcode.MOD, IRDataType.BYTE, reg1=42, immediate = 1),
+            IRInstruction(Opcode.DIV, IRDataType.BYTE, reg1=42, immediate = 2),
+            IRInstruction(Opcode.DIVS, IRDataType.BYTE, reg1=42, immediate = 2),
+            IRInstruction(Opcode.MUL, IRDataType.BYTE, reg1=42, immediate = 2),
+            IRInstruction(Opcode.MOD, IRDataType.BYTE, reg1=42, immediate = 2),
+            IRInstruction(Opcode.ADD, IRDataType.BYTE, reg1=42, immediate = 0),
+            IRInstruction(Opcode.SUB, IRDataType.BYTE, reg1=42, immediate = 0)
         ))
         irProg.chunks().single().instructions.size shouldBe 10
         val opt = IRPeepholeOptimizer(irProg)
-        opt.optimize()
+        opt.optimize(true, ErrorReporterForTests())
         irProg.chunks().single().instructions.size shouldBe 4
     }
 
     test("replace add/sub 1 by inc/dec") {
         val irProg = makeIRProgram(listOf(
-            IRInstruction(Opcode.ADD, IRDataType.BYTE, reg1=42, value = 1),
-            IRInstruction(Opcode.SUB, IRDataType.BYTE, reg1=42, value = 1)
+            IRInstruction(Opcode.ADD, IRDataType.BYTE, reg1=42, immediate = 1),
+            IRInstruction(Opcode.SUB, IRDataType.BYTE, reg1=42, immediate = 1)
         ))
         irProg.chunks().single().instructions.size shouldBe 2
         val opt = IRPeepholeOptimizer(irProg)
-        opt.optimize()
+        opt.optimize(true, ErrorReporterForTests())
         val instr = irProg.chunks().single().instructions
         instr.size shouldBe 2
         instr[0].opcode shouldBe Opcode.INC
@@ -150,40 +150,40 @@ class TestIRPeepholeOpt: FunSpec({
 
     test("remove useless and/or/xor") {
         val irProg = makeIRProgram(listOf(
-            IRInstruction(Opcode.AND, IRDataType.BYTE, reg1=42, value = 255),
-            IRInstruction(Opcode.AND, IRDataType.WORD, reg1=42, value = 65535),
-            IRInstruction(Opcode.OR, IRDataType.BYTE, reg1=42, value = 0),
-            IRInstruction(Opcode.XOR, IRDataType.BYTE, reg1=42, value = 0),
-            IRInstruction(Opcode.AND, IRDataType.BYTE, reg1=42, value = 200),
-            IRInstruction(Opcode.AND, IRDataType.WORD, reg1=42, value = 60000),
-            IRInstruction(Opcode.OR, IRDataType.BYTE, reg1=42, value = 1),
-            IRInstruction(Opcode.XOR, IRDataType.BYTE, reg1=42, value = 1)
+            IRInstruction(Opcode.AND, IRDataType.BYTE, reg1=42, immediate = 255),
+            IRInstruction(Opcode.AND, IRDataType.WORD, reg1=42, immediate = 65535),
+            IRInstruction(Opcode.OR, IRDataType.BYTE, reg1=42, immediate = 0),
+            IRInstruction(Opcode.XOR, IRDataType.BYTE, reg1=42, immediate = 0),
+            IRInstruction(Opcode.AND, IRDataType.BYTE, reg1=42, immediate = 200),
+            IRInstruction(Opcode.AND, IRDataType.WORD, reg1=42, immediate = 60000),
+            IRInstruction(Opcode.OR, IRDataType.BYTE, reg1=42, immediate = 1),
+            IRInstruction(Opcode.XOR, IRDataType.BYTE, reg1=42, immediate = 1)
         ))
         irProg.chunks().single().instructions.size shouldBe 8
         val opt = IRPeepholeOptimizer(irProg)
-        opt.optimize()
+        opt.optimize(true, ErrorReporterForTests())
         irProg.chunks().single().instructions.size shouldBe 4
     }
 
     test("replace and/or/xor by constant number") {
         val irProg = makeIRProgram(listOf(
-            IRInstruction(Opcode.AND, IRDataType.BYTE, reg1=42, value = 0),
-            IRInstruction(Opcode.AND, IRDataType.WORD, reg1=42, value = 0),
-            IRInstruction(Opcode.OR, IRDataType.BYTE, reg1=42, value = 255),
-            IRInstruction(Opcode.OR, IRDataType.WORD, reg1=42, value = 65535)
+            IRInstruction(Opcode.AND, IRDataType.BYTE, reg1=42, immediate = 0),
+            IRInstruction(Opcode.AND, IRDataType.WORD, reg1=42, immediate = 0),
+            IRInstruction(Opcode.OR, IRDataType.BYTE, reg1=42, immediate = 255),
+            IRInstruction(Opcode.OR, IRDataType.WORD, reg1=42, immediate = 65535)
         ))
         irProg.chunks().single().instructions.size shouldBe 4
         val opt = IRPeepholeOptimizer(irProg)
-        opt.optimize()
+        opt.optimize(true, ErrorReporterForTests())
         val instr = irProg.chunks().single().instructions
         instr.size shouldBe 4
         instr[0].opcode shouldBe Opcode.LOAD
         instr[1].opcode shouldBe Opcode.LOAD
         instr[2].opcode shouldBe Opcode.LOAD
         instr[3].opcode shouldBe Opcode.LOAD
-        instr[0].value shouldBe 0
-        instr[1].value shouldBe 0
-        instr[2].value shouldBe 255
-        instr[3].value shouldBe 65535
+        instr[0].immediate shouldBe 0
+        instr[1].immediate shouldBe 0
+        instr[2].immediate shouldBe 255
+        instr[3].immediate shouldBe 65535
     }
 })

codeGenIntermediate/test/TestVmCodeGen.kt

@@ -1,5 +1,6 @@
 import io.kotest.core.spec.style.FunSpec
 import io.kotest.matchers.ints.shouldBeGreaterThan
+import io.kotest.matchers.shouldBe
 import prog8.code.SymbolTableMaker
 import prog8.code.ast.*
 import prog8.code.core.*
@@ -7,6 +8,7 @@ import prog8.code.target.VMTarget
 import prog8.codegen.vm.VmAssemblyProgram
 import prog8.codegen.vm.VmCodeGen
 import prog8.intermediate.IRSubroutine
+import prog8.intermediate.Opcode
 
 class TestVmCodeGen: FunSpec({
 
@@ -39,7 +41,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         val sub = PtSub("start", emptyList(), null, Position.DUMMY)
         sub.add(PtVariable("pi", DataType.UBYTE, ZeropageWish.DONTCARE, PtNumber(DataType.UBYTE, 0.0, Position.DUMMY), null, Position.DUMMY))
         sub.add(PtVariable("particleX", DataType.ARRAY_UB, ZeropageWish.DONTCARE, null, 3u, Position.DUMMY))
@@ -89,7 +91,7 @@ class TestVmCodeGen: FunSpec({
         program.add(block)
 
         // define the "cx16.r0" virtual register
-        val cx16block = PtBlock("cx16", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val cx16block = PtBlock("cx16", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         cx16block.add(PtMemMapped("r0", DataType.UWORD, 100u, null, Position.DUMMY))
         program.add(cx16block)
 
@@ -98,7 +100,7 @@ class TestVmCodeGen: FunSpec({
         val errors = ErrorReporterForTests()
         val result = codegen.generate(program, st, options, errors) as VmAssemblyProgram
         val irChunks = (result.irProgram.blocks.first().children.single() as IRSubroutine).chunks
-        irChunks.size shouldBeGreaterThan 4
+        irChunks.size shouldBe 1
     }
 
     test("float comparison expressions against zero") {
@@ -118,7 +120,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         val sub = PtSub("start", emptyList(), null, Position.DUMMY)
         sub.add(PtVariable("f1", DataType.FLOAT, ZeropageWish.DONTCARE, null, null, Position.DUMMY))
         val if1 = PtIfElse(Position.DUMMY)
@@ -181,7 +183,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         val sub = PtSub("start", emptyList(), null, Position.DUMMY)
         sub.add(PtVariable("f1", DataType.FLOAT, ZeropageWish.DONTCARE, null, null, Position.DUMMY))
         val if1 = PtIfElse(Position.DUMMY)
@@ -240,7 +242,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         val sub = PtSub("start", emptyList(), null, Position.DUMMY)
         sub.add(PtVariable("f1", DataType.FLOAT, ZeropageWish.DONTCARE, null, null, Position.DUMMY))
         val if1 = PtIfElse(Position.DUMMY)
@@ -267,7 +269,7 @@ class TestVmCodeGen: FunSpec({
         val errors = ErrorReporterForTests()
         val result = codegen.generate(program, st, options, errors) as VmAssemblyProgram
         val irChunks = (result.irProgram.blocks.first().children.single() as IRSubroutine).chunks
-        irChunks.size shouldBeGreaterThan 4
+        irChunks.size shouldBe 1
     }
 
     test("integer comparison expressions against zero") {
@@ -287,7 +289,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         val sub = PtSub("start", emptyList(), null, Position.DUMMY)
         sub.add(PtVariable("sb1", DataType.BYTE, ZeropageWish.DONTCARE, null, null, Position.DUMMY))
         val if1 = PtIfElse(Position.DUMMY)
@@ -350,7 +352,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         val sub = PtSub("start", emptyList(), null, Position.DUMMY)
         sub.add(PtVariable("sb1", DataType.BYTE, ZeropageWish.DONTCARE, null, null, Position.DUMMY))
         val if1 = PtIfElse(Position.DUMMY)
@@ -409,7 +411,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
         val sub = PtSub("start", emptyList(), null, Position.DUMMY)
         sub.add(PtVariable("ub1", DataType.UBYTE, ZeropageWish.DONTCARE, null, null, Position.DUMMY))
         val if1 = PtIfElse(Position.DUMMY)
@@ -436,7 +438,36 @@ class TestVmCodeGen: FunSpec({
         val errors = ErrorReporterForTests()
         val result = codegen.generate(program, st, options, errors) as VmAssemblyProgram
         val irChunks = (result.irProgram.blocks.first().children.single() as IRSubroutine).chunks
-        irChunks.size shouldBeGreaterThan 4
+        irChunks.size shouldBe 1
     }
 
+    test("romsub allowed in ir-codegen") {
+//main {
+//    romsub $5000 = routine()
+//
+//    sub start()  {
+//        routine()
+//    }
+//}
+        val codegen = VmCodeGen()
+        val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
+        val block = PtBlock("main", null, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val romsub = PtAsmSub("routine", 0x5000u, setOf(CpuRegister.Y), emptyList(), emptyList(), false, Position.DUMMY)
+        block.add(romsub)
+        val sub = PtSub("start", emptyList(), null, Position.DUMMY)
+        val call = PtFunctionCall("main.routine", true, DataType.UNDEFINED, Position.DUMMY)
+        sub.add(call)
+        block.add(sub)
+        program.add(block)
+
+        val options = getTestOptions()
+        val st = SymbolTableMaker(program, options).make()
+        val errors = ErrorReporterForTests()
+        val result = codegen.generate(program, st, options, errors) as VmAssemblyProgram
+        val irChunks = (result.irProgram.blocks.first().children.single() as IRSubroutine).chunks
+        irChunks.size shouldBe 1
+        val callInstr = irChunks.single().instructions.single()
+        callInstr.opcode shouldBe Opcode.CALL
+        callInstr.address shouldBe 0x5000
+    }
 })

codeOptimizers/src/prog8/optimizer/ConstantFoldingOptimizer.kt

@@ -3,16 +3,17 @@ package prog8.optimizer
 import prog8.ast.Node
 import prog8.ast.Program
 import prog8.ast.base.FatalAstException
-import prog8.ast.base.UndefinedSymbolError
 import prog8.ast.expressions.*
 import prog8.ast.maySwapOperandOrder
 import prog8.ast.statements.ForLoop
+import prog8.ast.statements.RepeatLoop
 import prog8.ast.statements.VarDecl
 import prog8.ast.statements.VarDeclType
 import prog8.ast.walk.AstWalker
 import prog8.ast.walk.IAstModification
 import prog8.code.core.AssociativeOperators
 import prog8.code.core.DataType
+import kotlin.math.floor
 
 
 class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
@@ -307,7 +308,7 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
         val rangeTo = iterableRange.to as? NumericLiteral
         if(rangeFrom==null || rangeTo==null) return noModifications
 
-        val loopvar = forLoop.loopVar.targetVarDecl(program) ?: throw UndefinedSymbolError(forLoop.loopVar)
+        val loopvar = forLoop.loopVar.targetVarDecl(program) ?: return noModifications
 
         val stepLiteral = iterableRange.step as? NumericLiteral
         when(loopvar.datatype) {
@@ -364,6 +365,16 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
         return noModifications
     }
 
+    override fun after(repeatLoop: RepeatLoop, parent: Node): Iterable<IAstModification> {
+        val count = (repeatLoop.iterations as? NumericLiteral)?.number
+        if(count!=null && floor(count)!=count) {
+            val integer = NumericLiteral.optimalInteger(count.toInt(), repeatLoop.position)
+            repeatLoop.iterations = integer
+            integer.linkParents(repeatLoop)
+        }
+        return noModifications
+    }
+
     private class ShuffleOperands(val expr: BinaryExpression,
                                   val exprOperator: String?,
                                   val subExpr: BinaryExpression,

codeOptimizers/src/prog8/optimizer/ConstantIdentifierReplacer.kt

@@ -53,7 +53,8 @@ class VarConstantValueTypeAdjuster(private val program: Program, private val err
         }
 
         if(to==null) {
-            if(!range.to.inferType(program).isInteger)
+            val toType = range.to.inferType(program)
+            if(toType.isKnown && !range.to.inferType(program).isInteger)
                 errors.err("range expression to value must be integer", range.to.position)
         } else if(to-to.toInt()>0) {
             errors.err("range expression to value must be integer", range.to.position)
@@ -68,6 +69,124 @@ class VarConstantValueTypeAdjuster(private val program: Program, private val err
 
         return noModifications
     }
+
+    override fun after(functionCallExpr: FunctionCallExpression, parent: Node): Iterable<IAstModification> {
+        // choose specific builtin function for the given types
+        val func = functionCallExpr.target.nameInSource
+        if(func==listOf("clamp")) {
+            val t1 = functionCallExpr.args[0].inferType(program)
+            if(t1.isKnown) {
+                val replaceFunc: String
+                if(t1.isBytes) {
+                    replaceFunc = if(t1.istype(DataType.BYTE)) "clamp__byte" else "clamp__ubyte"
+                } else if(t1.isInteger) {
+                    replaceFunc = if(t1.istype(DataType.WORD)) "clamp__word" else "clamp__uword"
+                } else {
+                    errors.err("clamp builtin not supported for floats, use floats.clamp", functionCallExpr.position)
+                    return noModifications
+                }
+                return listOf(IAstModification.SetExpression({functionCallExpr.target = it as IdentifierReference},
+                    IdentifierReference(listOf(replaceFunc), functionCallExpr.target.position),
+                    functionCallExpr))
+            }
+        }
+        else if(func==listOf("min") || func==listOf("max")) {
+            val t1 = functionCallExpr.args[0].inferType(program)
+            val t2 = functionCallExpr.args[1].inferType(program)
+            if(t1.isKnown && t2.isKnown) {
+                val funcName = func[0]
+                val replaceFunc: String
+                if(t1.isBytes && t2.isBytes) {
+                    replaceFunc = if(t1.istype(DataType.BYTE) || t2.istype(DataType.BYTE))
+                        "${funcName}__byte"
+                    else
+                        "${funcName}__ubyte"
+                } else if(t1.isInteger && t2.isInteger) {
+                    replaceFunc = if(t1.istype(DataType.WORD) || t2.istype(DataType.WORD))
+                        "${funcName}__word"
+                    else
+                        "${funcName}__uword"
+                } else if(t1.isNumeric && t2.isNumeric) {
+                    errors.err("min/max not supported for floats", functionCallExpr.position)
+                    return noModifications
+                } else {
+                    errors.err("expected numeric arguments", functionCallExpr.args[0].position)
+                    return noModifications
+                }
+                return listOf(IAstModification.SetExpression({functionCallExpr.target = it as IdentifierReference},
+                    IdentifierReference(listOf(replaceFunc), functionCallExpr.target.position),
+                    functionCallExpr))
+            }
+        }
+        else if(func==listOf("abs")) {
+            val t1 = functionCallExpr.args[0].inferType(program)
+            if(t1.isKnown) {
+                val dt = t1.getOrElse { throw InternalCompilerException("invalid dt") }
+                val replaceFunc = when(dt) {
+                    DataType.BYTE -> "abs__byte"
+                    DataType.WORD -> "abs__word"
+                    DataType.FLOAT -> "abs__float"
+                    DataType.UBYTE, DataType.UWORD -> {
+                        return listOf(IAstModification.ReplaceNode(functionCallExpr, functionCallExpr.args[0], parent))
+                    }
+                    else -> {
+                        errors.err("expected numeric argument", functionCallExpr.args[0].position)
+                        return noModifications
+                    }
+                }
+                return listOf(IAstModification.SetExpression({functionCallExpr.target = it as IdentifierReference},
+                    IdentifierReference(listOf(replaceFunc), functionCallExpr.target.position),
+                    functionCallExpr))
+            }
+        }
+        else if(func==listOf("sqrt")) {
+            val t1 = functionCallExpr.args[0].inferType(program)
+            if(t1.isKnown) {
+                val dt = t1.getOrElse { throw InternalCompilerException("invalid dt") }
+                val replaceFunc = when(dt) {
+                    DataType.UBYTE -> "sqrt__ubyte"
+                    DataType.UWORD -> "sqrt__uword"
+                    DataType.FLOAT -> "sqrt__float"
+                    else -> {
+                        errors.err("expected unsigned or float numeric argument", functionCallExpr.args[0].position)
+                        return noModifications
+                    }
+                }
+                return listOf(IAstModification.SetExpression({functionCallExpr.target = it as IdentifierReference},
+                    IdentifierReference(listOf(replaceFunc), functionCallExpr.target.position),
+                    functionCallExpr))
+            }
+        }
+        return noModifications
+    }
+
+    override fun after(functionCallStatement: FunctionCallStatement, parent: Node): Iterable<IAstModification> {
+        // choose specific builtin function for the given types
+        val func = functionCallStatement.target.nameInSource
+        if(func==listOf("divmod")) {
+            val argTypes = functionCallStatement.args.map {it.inferType(program)}.toSet()
+            if(argTypes.size!=1) {
+                errors.err("expected all ubyte or all uword arguments", functionCallStatement.args[0].position)
+                return noModifications
+            }
+            val t1 = argTypes.single()
+            if(t1.isKnown) {
+                val dt = t1.getOrElse { throw InternalCompilerException("invalid dt") }
+                val replaceFunc = when(dt) {
+                    DataType.UBYTE -> "divmod__ubyte"
+                    DataType.UWORD -> "divmod__uword"
+                    else -> {
+                        errors.err("expected all ubyte or all uword arguments", functionCallStatement.args[0].position)
+                        return noModifications
+                    }
+                }
+                return listOf(IAstModification.SetExpression({functionCallStatement.target = it as IdentifierReference},
+                    IdentifierReference(listOf(replaceFunc), functionCallStatement.target.position),
+                    functionCallStatement))
+            }
+        }
+        return noModifications
+    }
 }
 
 
@@ -157,7 +276,7 @@ internal class ConstantIdentifierReplacer(private val program: Program, private
                         return listOf(IAstModification.ReplaceNode(decl.value!!, newValue, decl))
                     }
                 }
-                DataType.ARRAY_UB, DataType.ARRAY_B, DataType.ARRAY_UW, DataType.ARRAY_W -> {
+                DataType.ARRAY_UB, DataType.ARRAY_B, DataType.ARRAY_UW, DataType.ARRAY_W, DataType.ARRAY_W_SPLIT, DataType.ARRAY_UW_SPLIT -> {
                     val rangeExpr = decl.value as? RangeExpression
                     if(rangeExpr!=null) {
                         // convert the initializer range expression to an actual array

codeOptimizers/src/prog8/optimizer/ExpressionSimplifier.kt

@@ -18,7 +18,7 @@ import kotlin.math.abs
 import kotlin.math.log2
 import kotlin.math.pow
 
-// TODO add more peephole expression optimizations? Investigate what optimizations binaryen has
+// TODO add more peephole expression optimizations? Investigate what optimizations binaryen has?
 
 class ExpressionSimplifier(private val program: Program,
                            private val errors: IErrorReporter,

codeOptimizers/src/prog8/optimizer/Extensions.kt

@@ -43,9 +43,9 @@ fun Program.constantFold(errors: IErrorReporter, compTarget: ICompilationTarget)
 
 fun Program.optimizeStatements(errors: IErrorReporter,
                                functions: IBuiltinFunctions,
-                               compTarget: ICompilationTarget
+                               options: CompilationOptions
 ): Int {
-    val optimizer = StatementOptimizer(this, errors, functions, compTarget)
+    val optimizer = StatementOptimizer(this, errors, functions, options)
     optimizer.visit(this)
     val optimizationCount = optimizer.applyModifications()
 
@@ -54,8 +54,8 @@ fun Program.optimizeStatements(errors: IErrorReporter,
     return optimizationCount
 }
 
-fun Program.inlineSubroutines(): Int {
-    val inliner = Inliner(this)
+fun Program.inlineSubroutines(options: CompilationOptions): Int {
+    val inliner = Inliner(this, options)
     inliner.visit(this)
     return inliner.applyModifications()
 }

codeOptimizers/src/prog8/optimizer/Inliner.kt

@@ -8,7 +8,9 @@ import prog8.ast.statements.*
 import prog8.ast.walk.AstWalker
 import prog8.ast.walk.IAstModification
 import prog8.ast.walk.IAstVisitor
+import prog8.code.core.CompilationOptions
 import prog8.code.core.InternalCompilerException
+import prog8.code.target.VMTarget
 
 
 private  fun isEmptyReturn(stmt: Statement): Boolean = stmt is Return && stmt.value==null
@@ -16,7 +18,7 @@ private  fun isEmptyReturn(stmt: Statement): Boolean = stmt is Return && stmt.va
 
 // inliner potentially enables *ONE LINED* subroutines, wihtout to be inlined.
 
-class Inliner(val program: Program): AstWalker() {
+class Inliner(private val program: Program, private val options: CompilationOptions): AstWalker() {
 
     class DetermineInlineSubs(val program: Program): IAstVisitor {
         private val modifications = mutableListOf<IAstModification>()
@@ -211,7 +213,7 @@ class Inliner(val program: Program): AstWalker() {
 
     override fun after(functionCallStatement: FunctionCallStatement, parent: Node): Iterable<IAstModification>  {
         val sub = functionCallStatement.target.targetStatement(program) as? Subroutine
-        return if(sub==null)
+        return if(sub==null || !canInline(sub, functionCallStatement))
             noModifications
         else
             possibleInlineFcallStmt(sub, functionCallStatement, parent)
@@ -219,7 +221,7 @@ class Inliner(val program: Program): AstWalker() {
 
     override fun before(functionCallExpr: FunctionCallExpression, parent: Node): Iterable<IAstModification> {
         val sub = functionCallExpr.target.targetStatement(program) as? Subroutine
-        if(sub!=null && sub.inline && sub.parameters.isEmpty()) {
+        if(sub!=null && sub.inline && sub.parameters.isEmpty() && canInline(sub, functionCallExpr)) {
             require(sub.statements.size == 1 || (sub.statements.size == 2 && isEmptyReturn(sub.statements[1]))) {
                 "invalid inline sub at ${sub.position}"
             }
@@ -246,5 +248,17 @@ class Inliner(val program: Program): AstWalker() {
         return noModifications
     }
 
+    private fun canInline(sub: Subroutine, fcall: IFunctionCall): Boolean {
+        if(!sub.inline)
+            return false
+        if(options.compTarget.name!=VMTarget.NAME) {
+            val stmt = sub.statements.single()
+            if (stmt is IFunctionCall) {
+                val existing = (fcall as Node).definingScope.lookup(stmt.target.nameInSource.take(1))
+                return existing !is VarDecl
+            }
+        }
+        return true
+    }
 }
 

codeOptimizers/src/prog8/optimizer/StatementOptimizer.kt

@@ -13,7 +13,7 @@ import kotlin.math.floor
 class StatementOptimizer(private val program: Program,
                          private val errors: IErrorReporter,
                          private val functions: IBuiltinFunctions,
-                         private val compTarget: ICompilationTarget
+                         private val options: CompilationOptions
 ) : AstWalker() {
 
     override fun after(functionCallStatement: FunctionCallStatement, parent: Node): Iterable<IAstModification> {
@@ -39,7 +39,7 @@ class StatementOptimizer(private val program: Program,
                 if(string!=null) {
                     val pos = functionCallStatement.position
                     if (string.value.length == 1) {
-                        val firstCharEncoded = compTarget.encodeString(string.value, string.encoding)[0]
+                        val firstCharEncoded = options.compTarget.encodeString(string.value, string.encoding)[0]
                         val chrout = FunctionCallStatement(
                             IdentifierReference(listOf("txt", "chrout"), pos),
                             mutableListOf(NumericLiteral(DataType.UBYTE, firstCharEncoded.toDouble(), pos)),
@@ -51,7 +51,7 @@ class StatementOptimizer(private val program: Program,
                             IAstModification.Remove(stringDecl, stringDecl.parent as IStatementContainer)
                         )
                     } else if (string.value.length == 2) {
-                        val firstTwoCharsEncoded = compTarget.encodeString(string.value.take(2), string.encoding)
+                        val firstTwoCharsEncoded = options.compTarget.encodeString(string.value.take(2), string.encoding)
                         val chrout1 = FunctionCallStatement(
                             IdentifierReference(listOf("txt", "chrout"), pos),
                             mutableListOf(NumericLiteral(DataType.UBYTE, firstTwoCharsEncoded[0].toDouble(), pos)),
@@ -108,6 +108,16 @@ class StatementOptimizer(private val program: Program,
             }
         }
 
+        // remove obvious dangling elses (else after a return)
+        if(ifElse.elsepart.isNotEmpty() && ifElse.truepart.statements.singleOrNull() is Return) {
+            val elsePart = AnonymousScope(ifElse.elsepart.statements, ifElse.elsepart.position)
+            if(options.slowCodegenWarnings)
+                errors.warn("else can be omitted", ifElse.elsepart.position)
+            return listOf(
+                IAstModification.ReplaceNode(ifElse.elsepart, AnonymousScope(mutableListOf(), ifElse.elsepart.position), ifElse),
+                IAstModification.InsertAfter(ifElse, elsePart, parent as IStatementContainer)
+            )
+        }
         return noModifications
     }
 
@@ -141,7 +151,7 @@ class StatementOptimizer(private val program: Program,
                 val size = sv.value.length
                 if(size==1) {
                     // loop over string of length 1 -> just assign the single character
-                    val character = compTarget.encodeString(sv.value, sv.encoding)[0]
+                    val character = options.compTarget.encodeString(sv.value, sv.encoding)[0]
                     val byte = NumericLiteral(DataType.UBYTE, character.toDouble(), iterable.position)
                     val scope = AnonymousScope(mutableListOf(), forLoop.position)
                     scope.statements.add(Assignment(AssignTarget(forLoop.loopVar, null, null, forLoop.position), byte, AssignmentOrigin.OPTIMIZER, forLoop.position))
@@ -166,16 +176,6 @@ class StatementOptimizer(private val program: Program,
             }
         }
 
-        val iterationCount = forLoop.constIterationCount(program)
-        if(iterationCount!=null) {
-            val loopName = forLoop.loopVar.nameInSource
-            if(!forLoop.iterable.referencesIdentifier(loopName) && !forLoop.body.referencesIdentifier(loopName)) {
-                errors.warn("for loop can be replaced with repeat loop, possibly also remove the loop variable", forLoop.position)
-                val repeat = RepeatLoop(NumericLiteral.optimalNumeric(iterationCount, forLoop.position), forLoop.body, forLoop.position)
-                return listOf(IAstModification.ReplaceNode(forLoop, repeat, parent))
-            }
-        }
-
         return noModifications
     }
 
@@ -326,7 +326,7 @@ class StatementOptimizer(private val program: Program,
                         if (rightCv == 0.0) {
                             return listOf(IAstModification.Remove(assignment, parent as IStatementContainer))
                         } else if (targetDt in IntegerDatatypes && floor(rightCv) == rightCv) {
-                            if (vardeclDt != VarDeclType.MEMORY && rightCv in 1.0..3.0 && compTarget.name!=VMTarget.NAME) {
+                            if (vardeclDt != VarDeclType.MEMORY && rightCv in 1.0..3.0 && options.compTarget.name!=VMTarget.NAME) {
                                 // replace by several INCs if it's not a memory address (inc on a memory mapped register doesn't work very well)
                                 val incs = AnonymousScope(mutableListOf(), assignment.position)
                                 repeat(rightCv.toInt()) {
@@ -340,7 +340,7 @@ class StatementOptimizer(private val program: Program,
                         if (rightCv == 0.0) {
                             return listOf(IAstModification.Remove(assignment, parent as IStatementContainer))
                         } else if (targetDt in IntegerDatatypes && floor(rightCv) == rightCv) {
-                            if (vardeclDt != VarDeclType.MEMORY && rightCv in 1.0..3.0 && compTarget.name!=VMTarget.NAME) {
+                            if (vardeclDt != VarDeclType.MEMORY && rightCv in 1.0..3.0 && options.compTarget.name!=VMTarget.NAME) {
                                 // replace by several DECs if it's not a memory address (dec on a memory mapped register doesn't work very well)
                                 val decs = AnonymousScope(mutableListOf(), assignment.position)
                                 repeat(rightCv.toInt()) {
@@ -391,4 +391,47 @@ class StatementOptimizer(private val program: Program,
         else
             noModifications
     }
+
+    override fun after(whenStmt: When, parent: Node): Iterable<IAstModification> {
+
+        fun replaceWithIf(condition: Expression, trueBlock: AnonymousScope, elseBlock: AnonymousScope?): List<IAstModification> {
+            val ifStmt = IfElse(condition, trueBlock, elseBlock ?: AnonymousScope(mutableListOf(), whenStmt.position), whenStmt.position)
+            errors.warn("for boolean condition a normal if statement is preferred", whenStmt.position)
+            return listOf(IAstModification.ReplaceNode(whenStmt, ifStmt, parent))
+        }
+
+        if(whenStmt.condition.inferType(program).isBool) {
+            if(whenStmt.choices.all { it.values?.size==1 }) {
+                if (whenStmt.choices.all { it.values!!.single().constValue(program)!!.number in arrayOf(0.0, 1.0) }) {
+                    // it's a when statement on booleans that can just be replaced by an if or if..else.
+                    if (whenStmt.choices.size == 1) {
+                        return if(whenStmt.choices[0].values!![0].constValue(program)!!.number==1.0) {
+                            replaceWithIf(whenStmt.condition, whenStmt.choices[0].statements, null)
+                        } else {
+                            val notCondition = BinaryExpression(whenStmt.condition, "==", NumericLiteral(DataType.UBYTE, 0.0, whenStmt.condition.position), whenStmt.condition.position)
+                            replaceWithIf(notCondition, whenStmt.choices[0].statements, null)
+                        }
+                    } else if (whenStmt.choices.size == 2) {
+                        var trueBlock: AnonymousScope? = null
+                        var elseBlock: AnonymousScope? = null
+                        if(whenStmt.choices[0].values!![0].constValue(program)!!.number==1.0) {
+                            trueBlock = whenStmt.choices[0].statements
+                        } else {
+                            elseBlock = whenStmt.choices[0].statements
+                        }
+                        if(whenStmt.choices[1].values!![0].constValue(program)!!.number==1.0) {
+                            trueBlock = whenStmt.choices[1].statements
+                        } else {
+                            elseBlock = whenStmt.choices[1].statements
+                        }
+                        if(trueBlock!=null && elseBlock!=null) {
+                            return replaceWithIf(whenStmt.condition, trueBlock, elseBlock)
+                        }
+                    }
+                }
+            }
+        }
+        return noModifications
+    }
+
 }

codeOptimizers/src/prog8/optimizer/UnusedCodeRemover.kt

@@ -75,7 +75,9 @@ class UnusedCodeRemover(private val program: Program,
                     if(!block.statements.any { it is Subroutine && it.hasBeenInlined })
                         errors.warn("removing unused block '${block.name}'", block.position)
                 }
-                program.removeInternedStringsFromRemovedBlock(block)
+                if(!block.statements.any { it is Subroutine && it.hasBeenInlined }) {
+                    program.removeInternedStringsFromRemovedBlock(block)
+                }
                 return listOf(IAstModification.Remove(block, parent as IStatementContainer))
             }
         }
@@ -157,6 +159,12 @@ class UnusedCodeRemover(private val program: Program,
         return noModifications
     }
 
+    override fun after(assignment: Assignment, parent: Node): Iterable<IAstModification> {
+        if(assignment.target isSameAs assignment.value)
+            return listOf(IAstModification.Remove(assignment, parent as IStatementContainer))
+        return noModifications
+    }
+
     private fun deduplicateAssignments(statements: List<Statement>, scope: IStatementContainer): List<IAstModification> {
         // removes 'duplicate' assignments that assign the same target directly after another, unless it is a function call
         val linesToRemove = mutableListOf<Assignment>()

compiler/build.gradle

@@ -1,9 +1,10 @@
 plugins {
     id 'java'
     id 'application'
-    id "org.jetbrains.kotlin.jvm"
+    id 'org.jetbrains.kotlin.jvm'
     id 'com.github.johnrengelman.shadow' version '7.1.2'
-    id "io.kotest" version "0.3.9"
+    id 'io.kotest' version '0.3.9'
+    id 'com.peterabeles.gversion' version '1.10.2'
 }
 
 java {
@@ -24,8 +25,6 @@ compileTestKotlin {
     }
 }
 
-def prog8version = rootProject.file('compiler/res/version.txt').text.trim()
-
 dependencies {
     implementation project(':codeCore')
     implementation project(':codeOptimizers')
@@ -34,14 +33,14 @@ dependencies {
     implementation project(':codeGenIntermediate')
     implementation project(':codeGenExperimental')
     implementation project(':virtualmachine')
-    implementation "org.antlr:antlr4-runtime:4.12.0"
+    implementation "org.antlr:antlr4-runtime:4.13.0"
     implementation "org.jetbrains.kotlin:kotlin-stdlib-jdk8"
     // implementation "org.jetbrains.kotlin:kotlin-reflect"
     implementation 'org.jetbrains.kotlinx:kotlinx-cli:0.3.5'
-    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.16"
+    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18"
 
     testImplementation project(':intermediate')
-    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.5.5'
+    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.6.2'
 }
 
 configurations.all {
@@ -82,7 +81,7 @@ application {
 
 shadowJar {
     archiveBaseName = 'prog8compiler'
-    archiveVersion = prog8version
+    archiveVersion = version
     // minimize()
 }
 
@@ -100,4 +99,16 @@ test {
     }
 }
 
+gversion {
+    srcDir       = "src/"           // path is relative to the sub-project by default
+    classPackage = "prog8.buildversion"
+    className    = "BuildVersion"
+    language     = "kotlin"
+}
+
+
 build.finalizedBy installDist, installShadowDist
+
+compileKotlin.dependsOn createVersionFile    // , failDirtyNotSnapshot
+compileJava.dependsOn createVersionFile
+

compiler/res/prog8lib/atari/syslib.p8

@@ -2,44 +2,41 @@
 ; Including memory registers, I/O registers, Basic and Kernal subroutines.
 
 atari {
+    %option no_symbol_prefixing
 
         &uword  NMI_VEC         = $FFFA     ; 6502 nmi vector, determined by the kernal if banked in
         &uword  RESET_VEC       = $FFFC     ; 6502 reset vector, determined by the kernal if banked in
         &uword  IRQ_VEC         = $FFFE     ; 6502 interrupt vector, determined by the kernal if banked in
 
-; ---- kernal routines ----
-; TODO
-
-
-asmsub  init_system()  {
-    ; Initializes the machine to a sane starting state.
-    ; Called automatically by the loader program logic.
-    ; TODO
-    %asm {{
-        sei
-        cld
-        clc
-        ; TODO reset screen mode etc etc
-        clv
-        cli
-        rts
-    }}
 }
 
-asmsub  init_system_phase2()  {
-    %asm {{
-        rts     ; no phase 2 steps on the Atari
-    }}
-}
-
-}
-
-
 sys {
     ; ------- lowlevel system routines --------
 
+    %option no_symbol_prefixing
+
     const ubyte target = 8         ;  compilation target specifier.  64 = C64, 128 = C128,  16 = CommanderX16, 8 = atari800XL
 
+    asmsub  init_system()  {
+        ; Initializes the machine to a sane starting state.
+        ; Called automatically by the loader program logic.
+        ; TODO
+        %asm {{
+            sei
+            cld
+            clc
+            ; TODO reset screen mode etc etc
+            clv
+            cli
+            rts
+        }}
+    }
+
+    asmsub  init_system_phase2()  {
+        %asm {{
+            rts     ; no phase 2 steps on the Atari
+        }}
+    }
 
     asmsub  reset_system()  {
         ; Soft-reset the system back to initial power-on Basic prompt.
@@ -188,6 +185,19 @@ _longcopy
         }}
     }
 
+    inline asmsub irqsafe_set_irqd() {
+        %asm {{
+        php
+        sei
+        }}
+    }
+
+    inline asmsub irqsafe_clear_irqd() {
+        %asm {{
+        plp
+        }}
+    }
+
     inline asmsub exit(ubyte returnvalue @A) {
         ; -- immediately exit the program with a return code in the A register
         ;    TODO
@@ -208,6 +218,7 @@ _longcopy
 }
 
 cx16 {
+    %option no_symbol_prefixing
 
     ; the sixteen virtual 16-bit registers that the CX16 has defined in the zeropage
     ; they are simulated on the Atari as well but their location in memory is different
@@ -313,4 +324,31 @@ cx16 {
     &byte r13sH = $1b1b
     &byte r14sH = $1b1d
     &byte r15sH = $1b1f
+
+    asmsub save_virtual_registers() clobbers(A,Y) {
+        %asm {{
+            ldy  #31
+    -       lda  cx16.r0,y
+            sta  _cx16_vreg_storage,y
+            dey
+            bpl  -
+            rts
+
+    _cx16_vreg_storage
+            .word 0,0,0,0,0,0,0,0
+            .word 0,0,0,0,0,0,0,0
+        }}
+    }
+
+    asmsub restore_virtual_registers() clobbers(A,Y) {
+        %asm {{
+            ldy  #31
+    -       lda  save_virtual_registers._cx16_vreg_storage,y
+            sta  cx16.r0,y
+            dey
+            bpl  -
+            rts
+        }}
+    }
+
 }

compiler/res/prog8lib/atari/textio.p8

@@ -6,6 +6,8 @@
 
 txt {
 
+    %option no_symbol_prefixing
+
 const ubyte DEFAULT_WIDTH = 40
 const ubyte DEFAULT_HEIGHT = 24
 
@@ -70,7 +72,7 @@ sub uppercase() {
     ; TODO
 }
 
-asmsub  scroll_left  (ubyte alsocolors @ Pc) clobbers(A, Y)  {
+asmsub  scroll_left  (bool alsocolors @ Pc) clobbers(A, Y)  {
 	; ---- scroll the whole screen 1 character to the left
 	;      contents of the rightmost column are unchanged, you should clear/refill this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -81,7 +83,7 @@ asmsub  scroll_left  (ubyte alsocolors @ Pc) clobbers(A, Y)  {
         }}
 }
 
-asmsub  scroll_right  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_right  (bool alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character to the right
 	;      contents of the leftmost column are unchanged, you should clear/refill this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -91,7 +93,7 @@ asmsub  scroll_right  (ubyte alsocolors @ Pc) clobbers(A)  {
         }}
 }
 
-asmsub  scroll_up  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_up  (bool alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character up
 	;      contents of the bottom row are unchanged, you should refill/clear this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -101,7 +103,7 @@ asmsub  scroll_up  (ubyte alsocolors @ Pc) clobbers(A)  {
         }}
 }
 
-asmsub  scroll_down  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_down  (bool alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character down
 	;      contents of the top row are unchanged, you should refill/clear this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -118,7 +120,6 @@ romsub $F2Fd = waitkey()
 asmsub chrout(ubyte char @ A) {
 	%asm {{
 		sta  _tmp_outchar+1
-		pha
 		txa
 		pha
 		tya
@@ -130,7 +131,6 @@ _tmp_outchar
 		tay
 		pla
 		tax
-		pla
 		rts
 	}}
 }
@@ -210,7 +210,7 @@ asmsub  print_b  (byte value @ A) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_ubhex  (ubyte value @ A, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in hex form (if Carry is set, a radix prefix '$' is printed as well)
 	%asm {{
 		stx  P8ZP_SCRATCH_REG
@@ -228,7 +228,7 @@ asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_ubbin  (ubyte value @ A, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in binary form (if Carry is set, a radix prefix '%' is printed as well)
 	%asm {{
 		stx  P8ZP_SCRATCH_REG
@@ -249,7 +249,7 @@ asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_uwbin  (uword value @ AY, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in binary form (if Carry is set, a radix prefix '%' is printed as well)
 	%asm {{
 		pha
@@ -261,7 +261,7 @@ asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_uwhex  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_uwhex  (uword value @ AY, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in hexadecimal form (4 digits)
 	;      (if Carry is set, a radix prefix '$' is printed as well)
 	%asm {{

compiler/res/prog8lib/c128/floats.asm

@@ -1,664 +0,0 @@
-; --- low level floating point assembly routines for the C128
-;     these are almost all identical to the C64 except for a few details
-;     so we have to have a separate library file for the C128 unfortunately.
-
-
-FL_ONE_const	.byte  129     			; 1.0
-FL_ZERO_const	.byte  0,0,0,0,0		; 0.0
-FL_LOG2_const	.byte  $80, $31, $72, $17, $f8	; log(2)
-
-
-floats_store_reg	.byte  0		; temp storage
-
-
-ub2float	.proc
-		; -- convert ubyte in SCRATCH_ZPB1 to float at address A/Y
-		;    clobbers A, Y
-		stx  P8ZP_SCRATCH_REG
-		sta  P8ZP_SCRATCH_W2
-		sty  P8ZP_SCRATCH_W2+1
-		ldy  P8ZP_SCRATCH_B1
-		lda  #0
-		jsr  GIVAYF
-_fac_to_mem	ldx  P8ZP_SCRATCH_W2
-		ldy  P8ZP_SCRATCH_W2+1
-		jsr  MOVMF
-		ldx  P8ZP_SCRATCH_REG
-		rts
-		.pend
-
-b2float		.proc
-		; -- convert byte in SCRATCH_ZPB1 to float at address A/Y
-		;    clobbers A, Y
-		stx  P8ZP_SCRATCH_REG
-		sta  P8ZP_SCRATCH_W2
-		sty  P8ZP_SCRATCH_W2+1
-		lda  P8ZP_SCRATCH_B1
-		jsr  FREADSA
-		jmp  ub2float._fac_to_mem
-		.pend
-
-uw2float	.proc
-		; -- convert uword in SCRATCH_ZPWORD1 to float at address A/Y
-		stx  P8ZP_SCRATCH_REG
-		sta  P8ZP_SCRATCH_W2
-		sty  P8ZP_SCRATCH_W2+1
-		lda  P8ZP_SCRATCH_W1
-		ldy  P8ZP_SCRATCH_W1+1
-		jsr  GIVUAYFAY
-		jmp  ub2float._fac_to_mem
-		.pend
-
-w2float		.proc
-		; -- convert word in SCRATCH_ZPWORD1 to float at address A/Y
-		stx  P8ZP_SCRATCH_REG
-		sta  P8ZP_SCRATCH_W2
-		sty  P8ZP_SCRATCH_W2+1
-		ldy  P8ZP_SCRATCH_W1
-		lda  P8ZP_SCRATCH_W1+1
-		jsr  GIVAYF
-		jmp  ub2float._fac_to_mem
-		.pend
-
-
-cast_from_uw	.proc
-		; -- uword in A/Y into float var at (P8ZP_SCRATCH_W2)
-		stx  P8ZP_SCRATCH_REG
-		jsr  GIVUAYFAY
-		jmp  ub2float._fac_to_mem
-		.pend
-
-
-cast_from_w	.proc
-		; -- word in A/Y into float var at (P8ZP_SCRATCH_W2)
-		stx  P8ZP_SCRATCH_REG
-		jsr  GIVAYFAY
-		jmp  ub2float._fac_to_mem
-		.pend
-
-
-cast_from_ub	.proc
-		; -- ubyte in Y into float var at (P8ZP_SCRATCH_W2)
-		stx  P8ZP_SCRATCH_REG
-		jsr  FREADUY
-		jmp  ub2float._fac_to_mem
-		.pend
-
-
-cast_from_b	.proc
-		; -- byte in A into float var at (P8ZP_SCRATCH_W2)
-		stx  P8ZP_SCRATCH_REG
-		jsr  FREADSA
-		jmp  ub2float._fac_to_mem
-		.pend
-
-cast_as_uw_into_ya	.proc               ; also used for float 2 ub
-		; -- cast float at A/Y to uword into Y/A
-		jsr  MOVFM
-		jmp  cast_FAC1_as_uw_into_ya
-		.pend
-
-cast_as_w_into_ay	.proc               ; also used for float 2 b
-		; -- cast float at A/Y to word into A/Y
-		jsr  MOVFM
-		jmp  cast_FAC1_as_w_into_ay
-		.pend
-
-cast_FAC1_as_uw_into_ya	.proc               ; also used for float 2 ub
-		; -- cast fac1 to uword into Y/A
-		stx  P8ZP_SCRATCH_REG
-		jsr  GETADR     ; into Y/A
-		ldx  P8ZP_SCRATCH_REG
-		rts
-		.pend
-
-cast_FAC1_as_w_into_ay	.proc               ; also used for float 2 b
-		; -- cast fac1 to word into A/Y
-		stx  P8ZP_SCRATCH_REG
-		jsr  AYINT
-		ldy  floats.AYINT_facmo
-		lda  floats.AYINT_facmo+1
-		ldx  P8ZP_SCRATCH_REG
-		rts
-		.pend
-
-
-stack_b2float	.proc
-		; -- b2float operating on the stack
-		inx
-		lda  P8ESTACK_LO,x
-		stx  P8ZP_SCRATCH_REG
-		jsr  FREADSA
-		jmp  push_fac1._internal
-		.pend
-
-stack_w2float	.proc
-		; -- w2float operating on the stack
-		inx
-		ldy  P8ESTACK_LO,x
-		lda  P8ESTACK_HI,x
-		stx  P8ZP_SCRATCH_REG
-		jsr  GIVAYF
-		jmp  push_fac1._internal
-		.pend
-
-stack_ub2float	.proc
-		; -- ub2float operating on the stack
-		inx
-		lda  P8ESTACK_LO,x
-		stx  P8ZP_SCRATCH_REG
-		tay
-		lda  #0
-		jsr  GIVAYF
-		jmp  push_fac1._internal
-		.pend
-
-stack_uw2float	.proc
-		; -- uw2float operating on the stack
-		inx
-		lda  P8ESTACK_LO,x
-		ldy  P8ESTACK_HI,x
-		stx  P8ZP_SCRATCH_REG
-		jsr  GIVUAYFAY
-		jmp  push_fac1._internal
-		.pend
-
-stack_float2w	.proc               ; also used for float2b
-		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG
-		jsr  AYINT
-		ldx  P8ZP_SCRATCH_REG
-		lda  floats.AYINT_facmo
-		sta  P8ESTACK_HI,x
-		lda  floats.AYINT_facmo+1
-		sta  P8ESTACK_LO,x
-		dex
-		rts
-		.pend
-
-stack_float2uw	.proc               ; also used for float2ub
-		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG
-		jsr  GETADR
-		ldx  P8ZP_SCRATCH_REG
-		sta  P8ESTACK_HI,x
-		tya
-		sta  P8ESTACK_LO,x
-		dex
-		rts
-		.pend
-
-push_float	.proc
-		; ---- push mflpt5 in A/Y onto stack
-		; (taking 3 stack positions = 6 bytes of which 1 is padding)
-		sta  P8ZP_SCRATCH_W1
-		sty  P8ZP_SCRATCH_W1+1
-		ldy  #0
-		lda  (P8ZP_SCRATCH_W1),y
-		sta  P8ESTACK_LO,x
-		iny
-		lda  (P8ZP_SCRATCH_W1),y
-		sta  P8ESTACK_HI,x
-		dex
-		iny
-		lda  (P8ZP_SCRATCH_W1),y
-		sta  P8ESTACK_LO,x
-		iny
-		lda  (P8ZP_SCRATCH_W1),y
-		sta  P8ESTACK_HI,x
-		dex
-		iny
-		lda  (P8ZP_SCRATCH_W1),y
-		sta  P8ESTACK_LO,x
-		dex
-		rts
-		.pend
-
-pop_float	.proc
-		; ---- pops mflpt5 from stack to memory A/Y
-		; (frees 3 stack positions = 6 bytes of which 1 is padding)
-		sta  P8ZP_SCRATCH_W1
-		sty  P8ZP_SCRATCH_W1+1
-		ldy  #4
-		inx
-		lda  P8ESTACK_LO,x
-		sta  (P8ZP_SCRATCH_W1),y
-		dey
-		inx
-		lda  P8ESTACK_HI,x
-		sta  (P8ZP_SCRATCH_W1),y
-		dey
-		lda  P8ESTACK_LO,x
-		sta  (P8ZP_SCRATCH_W1),y
-		dey
-		inx
-		lda  P8ESTACK_HI,x
-		sta  (P8ZP_SCRATCH_W1),y
-		dey
-		lda  P8ESTACK_LO,x
-		sta  (P8ZP_SCRATCH_W1),y
-		rts
-		.pend
-
-pop_float_fac1	.proc
-		; -- pops float from stack into FAC1
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		jsr  pop_float
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		jmp  MOVFM
-		.pend
-
-copy_float	.proc
-		; -- copies the 5 bytes of the mflt value pointed to by SCRATCH_W1,
-		;    into the 5 bytes pointed to by A/Y.  Clobbers A,Y.
-		sta  _target+1
-		sty  _target+2
-		ldy  #4
-_loop		lda  (P8ZP_SCRATCH_W1),y
-_target		sta  $ffff,y			; modified
-		dey
-		bpl  _loop
-		rts
-		.pend
-
-inc_var_f	.proc
-		; -- add 1 to float pointed to by A/Y
-		sta  P8ZP_SCRATCH_W1
-		sty  P8ZP_SCRATCH_W1+1
-		stx  P8ZP_SCRATCH_REG
-		jsr  MOVFM
-		lda  #<FL_ONE_const
-		ldy  #>FL_ONE_const
-		jsr  FADD
-		ldx  P8ZP_SCRATCH_W1
-		ldy  P8ZP_SCRATCH_W1+1
-		jsr  MOVMF
-		ldx  P8ZP_SCRATCH_REG
-		rts
-		.pend
-
-dec_var_f	.proc
-		; -- subtract 1 from float pointed to by A/Y
-		sta  P8ZP_SCRATCH_W1
-		sty  P8ZP_SCRATCH_W1+1
-		stx  P8ZP_SCRATCH_REG
-		lda  #<FL_ONE_const
-		ldy  #>FL_ONE_const
-		jsr  MOVFM
-		lda  P8ZP_SCRATCH_W1
-		ldy  P8ZP_SCRATCH_W1+1
-		jsr  FSUB
-		ldx  P8ZP_SCRATCH_W1
-		ldy  P8ZP_SCRATCH_W1+1
-		jsr  MOVMF
-		ldx  P8ZP_SCRATCH_REG
-		rts
-		.pend
-
-
-pop_2_floats_f2_in_fac1	.proc
-		; -- pop 2 floats from stack, load the second one in FAC1 as well
-		lda  #<fmath_float2
-		ldy  #>fmath_float2
-		jsr  pop_float
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		jsr  pop_float
-		lda  #<fmath_float2
-		ldy  #>fmath_float2
-		jmp  MOVFM
-		.pend
-
-
-fmath_float1	.byte 0,0,0,0,0	; storage for a mflpt5 value
-fmath_float2	.byte 0,0,0,0,0	; storage for a mflpt5 value
-
-
-push_fac1	.proc
-		; -- push the float in FAC1 onto the stack
-		stx  P8ZP_SCRATCH_REG
-_internal	ldx  #<fmath_float1
-		ldy  #>fmath_float1
-		jsr  MOVMF
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		ldx  P8ZP_SCRATCH_REG
-		jmp  push_float
-		.pend
-
-div_f		.proc
-		; -- push f1/f2 on stack
-		jsr  pop_2_floats_f2_in_fac1
-		stx  P8ZP_SCRATCH_REG
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		jsr  FDIV
-		jmp  push_fac1._internal
-		.pend
-
-add_f		.proc
-		; -- push f1+f2 on stack
-		jsr  pop_2_floats_f2_in_fac1
-		stx  P8ZP_SCRATCH_REG
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		jsr  FADD
-		jmp  push_fac1._internal
-		.pend
-
-sub_f		.proc
-		; -- push f1-f2 on stack
-		jsr  pop_2_floats_f2_in_fac1
-		stx  P8ZP_SCRATCH_REG
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		jsr  FSUB
-		jmp  push_fac1._internal
-		.pend
-
-mul_f		.proc
-		; -- push f1*f2 on stack
-		jsr  pop_2_floats_f2_in_fac1
-		stx  P8ZP_SCRATCH_REG
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		jsr  FMULT
-		jmp  push_fac1._internal
-		.pend
-
-neg_f		.proc
-		; -- toggle the sign bit on the stack
-		lda  P8ESTACK_HI+3,x
-		eor  #$80
-		sta  P8ESTACK_HI+3,x
-		rts
-		.pend
-
-var_fac1_less_f	.proc
-		; -- is the float in FAC1 < the variable AY?
-		stx  P8ZP_SCRATCH_REG
-		jsr  FCOMP
-		ldx  P8ZP_SCRATCH_REG
-		cmp  #255
-		beq  +
-		lda  #0
-		rts
-+		lda  #1
-		rts
-		.pend
-
-var_fac1_lesseq_f	.proc
-		; -- is the float in FAC1 <= the variable AY?
-		stx  P8ZP_SCRATCH_REG
-		jsr  FCOMP
-		ldx  P8ZP_SCRATCH_REG
-		cmp  #0
-		beq  +
-		cmp  #255
-		beq  +
-		lda  #0
-		rts
-+		lda  #1
-		rts
-		.pend
-
-var_fac1_greater_f	.proc
-		; -- is the float in FAC1 > the variable AY?
-		stx  P8ZP_SCRATCH_REG
-		jsr  FCOMP
-		ldx  P8ZP_SCRATCH_REG
-		cmp  #1
-		beq  +
-		lda  #0
-		rts
-+		lda  #1
-		rts
-		.pend
-
-var_fac1_greatereq_f	.proc
-		; -- is the float in FAC1 >= the variable AY?
-		stx  P8ZP_SCRATCH_REG
-		jsr  FCOMP
-		ldx  P8ZP_SCRATCH_REG
-		cmp  #0
-		beq  +
-		cmp  #1
-		beq  +
-		lda  #0
-		rts
-+		lda  #1
-		rts
-		.pend
-
-var_fac1_notequal_f	.proc
-		; -- are the floats numbers in FAC1 and the variable AY *not* identical?
-		stx  P8ZP_SCRATCH_REG
-		jsr  FCOMP
-		ldx  P8ZP_SCRATCH_REG
-		and  #1
-		rts
-		.pend
-
-vars_equal_f	.proc
-		; -- are the mflpt5 numbers in P8ZP_SCRATCH_W1 and AY identical?
-		sta  P8ZP_SCRATCH_W2
-		sty  P8ZP_SCRATCH_W2+1
-		ldy  #0
-		lda  (P8ZP_SCRATCH_W1),y
-		cmp  (P8ZP_SCRATCH_W2),y
-		bne  _false
-		iny
-		lda  (P8ZP_SCRATCH_W1),y
-		cmp  (P8ZP_SCRATCH_W2),y
-		bne  _false
-		iny
-		lda  (P8ZP_SCRATCH_W1),y
-		cmp  (P8ZP_SCRATCH_W2),y
-		bne  _false
-		iny
-		lda  (P8ZP_SCRATCH_W1),y
-		cmp  (P8ZP_SCRATCH_W2),y
-		bne  _false
-		iny
-		lda  (P8ZP_SCRATCH_W1),y
-		cmp  (P8ZP_SCRATCH_W2),y
-		bne  _false
-		lda  #1
-		rts
-_false		lda  #0
-		rts
-		.pend
-
-equal_f		.proc
-		; -- are the two mflpt5 numbers on the stack identical?
-		inx
-		inx
-		inx
-		inx
-		lda  P8ESTACK_LO-3,x
-		cmp  P8ESTACK_LO,x
-		bne  _equals_false
-		lda  P8ESTACK_LO-2,x
-		cmp  P8ESTACK_LO+1,x
-		bne  _equals_false
-		lda  P8ESTACK_LO-1,x
-		cmp  P8ESTACK_LO+2,x
-		bne  _equals_false
-		lda  P8ESTACK_HI-2,x
-		cmp  P8ESTACK_HI+1,x
-		bne  _equals_false
-		lda  P8ESTACK_HI-1,x
-		cmp  P8ESTACK_HI+2,x
-		bne  _equals_false
-_equals_true	lda  #1
-_equals_store	inx
-		sta  P8ESTACK_LO+1,x
-		rts
-_equals_false	lda  #0
-		beq  _equals_store
-		.pend
-
-notequal_f	.proc
-		; -- are the two mflpt5 numbers on the stack different?
-		jsr  equal_f
-		eor  #1		; invert the result
-		sta  P8ESTACK_LO+1,x
-		rts
-		.pend
-
-vars_less_f	.proc
-		; -- is float in AY < float in P8ZP_SCRATCH_W2 ?
-		jsr  MOVFM
-		lda  P8ZP_SCRATCH_W2
-		ldy  P8ZP_SCRATCH_W2+1
-		stx  P8ZP_SCRATCH_REG
-		jsr  FCOMP
-		ldx  P8ZP_SCRATCH_REG
-		cmp  #255
-		bne  +
-		lda  #1
-		rts
-+		lda  #0
-		rts
-		.pend
-
-vars_lesseq_f	.proc
-		; -- is float in AY <= float in P8ZP_SCRATCH_W2 ?
-		jsr  MOVFM
-		lda  P8ZP_SCRATCH_W2
-		ldy  P8ZP_SCRATCH_W2+1
-		stx  P8ZP_SCRATCH_REG
-		jsr  FCOMP
-		ldx  P8ZP_SCRATCH_REG
-		cmp  #255
-		bne  +
--		lda  #1
-		rts
-+		cmp  #0
-		beq  -
-		lda  #0
-		rts
-		.pend
-
-less_f		.proc
-		; -- is f1 < f2?
-		jsr  compare_floats
-		cmp  #255
-		beq  compare_floats._return_true
-		bne  compare_floats._return_false
-		.pend
-
-
-lesseq_f	.proc
-		; -- is f1 <= f2?
-		jsr  compare_floats
-		cmp  #255
-		beq  compare_floats._return_true
-		cmp  #0
-		beq  compare_floats._return_true
-		bne  compare_floats._return_false
-		.pend
-
-greater_f	.proc
-		; -- is f1 > f2?
-		jsr  compare_floats
-		cmp  #1
-		beq  compare_floats._return_true
-		bne  compare_floats._return_false
-		.pend
-
-greatereq_f	.proc
-		; -- is f1 >= f2?
-		jsr  compare_floats
-		cmp  #1
-		beq  compare_floats._return_true
-		cmp  #0
-		beq  compare_floats._return_true
-		bne  compare_floats._return_false
-		.pend
-
-compare_floats	.proc
-		lda  #<fmath_float2
-		ldy  #>fmath_float2
-		jsr  pop_float
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		jsr  pop_float
-		lda  #<fmath_float1
-		ldy  #>fmath_float1
-		jsr  MOVFM		; fac1 = flt1
-		lda  #<fmath_float2
-		ldy  #>fmath_float2
-		stx  P8ZP_SCRATCH_REG
-		jsr  FCOMP		; A = flt1 compared with flt2 (0=equal, 1=flt1>flt2, 255=flt1<flt2)
-		ldx  P8ZP_SCRATCH_REG
-		rts
-_return_false	lda  #0
-_return_result  sta  P8ESTACK_LO,x
-		dex
-		rts
-_return_true	lda  #1
-		bne  _return_result
-		.pend
-
-set_array_float_from_fac1	.proc
-		; -- set the float in FAC1 in the array (index in A, array in P8ZP_SCRATCH_W1)
-		sta  P8ZP_SCRATCH_B1
-		asl  a
-		asl  a
-		clc
-		adc  P8ZP_SCRATCH_B1
-		ldy  P8ZP_SCRATCH_W1+1
-		clc
-		adc  P8ZP_SCRATCH_W1
-		bcc  +
-		iny
-+		stx  floats_store_reg
-		tax
-		jsr  MOVMF
-		ldx  floats_store_reg
-		rts
-		.pend
-
-
-set_0_array_float	.proc
-		; -- set a float in an array to zero (index in A, array in P8ZP_SCRATCH_W1)
-		sta  P8ZP_SCRATCH_B1
-		asl  a
-		asl  a
-		clc
-		adc  P8ZP_SCRATCH_B1
-		tay
-		lda  #0
-		sta  (P8ZP_SCRATCH_W1),y
-		iny
-		sta  (P8ZP_SCRATCH_W1),y
-		iny
-		sta  (P8ZP_SCRATCH_W1),y
-		iny
-		sta  (P8ZP_SCRATCH_W1),y
-		iny
-		sta  (P8ZP_SCRATCH_W1),y
-		rts
-		.pend
-
-
-set_array_float		.proc
-		; -- set a float in an array to a value (index in A, float in P8ZP_SCRATCH_W1, array in P8ZP_SCRATCH_W2)
-		sta  P8ZP_SCRATCH_B1
-		asl  a
-		asl  a
-		clc
-		adc  P8ZP_SCRATCH_B1
-		adc  P8ZP_SCRATCH_W2
-		ldy  P8ZP_SCRATCH_W2+1
-		bcc  +
-		iny
-+		jmp  copy_float
-			; -- copies the 5 bytes of the mflt value pointed to by SCRATCH_ZPWORD1,
-			;    into the 5 bytes pointed to by A/Y.  Clobbers A,Y.
-		.pend
-
-

compiler/res/prog8lib/c128/floats.p8

@@ -1,158 +0,0 @@
-; Prog8 definitions for floating point handling on the Commodore 128
-
-%option enable_floats
-%import floats_functions
-
-floats {
-	; ---- this block contains C-128 compatible floating point related functions ----
-
-        const float  PI     = 3.141592653589793
-        const float  TWOPI  = 6.283185307179586
-
-
-; ---- ROM float functions ----
-
-		; note: the fac1 and fac2 are working registers and take 6 bytes each,
-		; floats in memory  (and rom) are stored in 5-byte MFLPT packed format.
-
-; note: fac1/2 might get clobbered even if not mentioned in the function's name.
-; note: for subtraction and division, the left operand is in fac2, the right operand in fac1.
-
-romsub $af00 = AYINT() clobbers(A,X,Y)          ; fac1-> signed word in 102-103 ($66-$67) MSB FIRST. (might throw ILLEGAL QUANTITY)
-
-; GIVAYF: signed word in Y/A (note different lsb/msb order) -> float in fac1
-; there is also floats.GIVUAYFAY - unsigned word in A/Y (lo/hi) to fac1
-; (tip: use GIVAYFAY to use A/Y input; lo/hi switched to normal order)
-romsub $af03 = GIVAYF(ubyte lo @ Y, ubyte hi @ A) clobbers(A,X,Y)
-
-romsub $af06 = FOUT() clobbers(X) -> uword @ AY             ; fac1 -> string, address returned in AY
-; romsub $af09 = VAL_1() clobbers(A,X,Y)                      ; convert ASCII string to floating point [not yet implemented!!!]
-
-; fac1 -> unsigned word in Y/A (might throw ILLEGAL QUANTITY) (result also in $14/15)
-; (tip: use GETADRAY to get A/Y output; lo/hi switched to normal little endian order)
-romsub $af0c = GETADR() clobbers(X) -> ubyte @ Y, ubyte @ A
-romsub $af0f = FLOATC() clobbers(A,X,Y)                     ; convert address to floating point
-
-romsub $af12 = FSUB(uword mflpt @ AY) clobbers(A,X,Y)       ; fac1 = mflpt from A/Y - fac1
-romsub $af15 = FSUBT() clobbers(A,X,Y)                      ; fac1 = fac2-fac1   mind the order of the operands         NOTE: use FSUBT2() instead!
-romsub $af18 = FADD(uword mflpt @ AY) clobbers(A,X,Y)       ; fac1 += mflpt value from A/Y
-romsub $af1b = FADDT() clobbers(A,X,Y)                      ; fac1 += fac2      NOTE: use FADDT2() instead!
-romsub $af1e = FMULT(uword mflpt @ AY) clobbers(A,X,Y)      ; fac1 *= mflpt value from A/Y
-romsub $af21 = FMULTT() clobbers(A,X,Y)                     ; fac1 *= fac2      NOTE: use FMULTT2() instead!
-romsub $af24 = FDIV(uword mflpt @ AY) clobbers(A,X,Y)       ; fac1 = mflpt in A/Y / fac1
-romsub $af27 = FDIVT() clobbers(A,X,Y)                      ; fac1 = fac2/fac1  mind the order of the operands  NOTE: use FDIVT2() instead!
-romsub $af2a = LOG() clobbers(A,X,Y)                        ; fac1 = LN(fac1)  (natural log)
-romsub $af2d = INT() clobbers(A,X,Y)                        ; INT() truncates, use FADDH first to round instead of trunc
-romsub $af30 = SQR() clobbers(A,X,Y)                        ; fac1 = SQRT(fac1)
-romsub $af33 = NEGOP() clobbers(A)                          ; switch the sign of fac1 (fac1 = -fac1)
-romsub $af36 = FPWR(uword mflpt @ AY) clobbers(A,X,Y)       ; fac1 = fac2 ** float in A/Y
-romsub $af39 = FPWRT() clobbers(A,X,Y)                      ; fac1 = fac2 ** fac1           NOTE: use FPWRT2() instead!
-romsub $af3c = EXP() clobbers(A,X,Y)                        ; fac1 = EXP(fac1)  (e ** fac1)
-romsub $af3f = COS() clobbers(A,X,Y)                        ; fac1 = COS(fac1)
-romsub $af42 = SIN() clobbers(A,X,Y)                        ; fac1 = SIN(fac1)
-romsub $af45 = TAN() clobbers(A,X,Y)                        ; fac1 = TAN(fac1)
-romsub $af48 = ATN() clobbers(A,X,Y)                        ; fac1 = ATN(fac1)
-romsub $af4b = ROUND() clobbers(A,X,Y)                      ; round fac1
-romsub $af4e = ABS() clobbers(A,X,Y)                        ; fac1 = ABS(fac1)
-romsub $af51 = SIGN() clobbers(X,Y) -> ubyte @ A            ; SIGN(fac1) to A, $ff, $0, $1 for negative, zero, positive
-romsub $af54 = FCOMP(uword mflpt @ AY) clobbers(X,Y) -> ubyte @ A   ; A = compare fac1 to mflpt in A/Y, 0=equal 1=fac1 is greater, 255=fac1 is less than
-romsub $af57 = RND_0() clobbers(A,X,Y)                      ; fac1 = RND(fac1) float random number generator
-romsub $af57 = RND() clobbers(A,X,Y)                        ; alias for RND_0
-romsub $af5a = CONUPK(uword mflpt @ AY) clobbers(A,X,Y)     ; load mflpt value from memory  in A/Y into fac2
-romsub $af5d = ROMUPK(uword mflpt @ AY) clobbers(A,X,Y)     ; load mflpt value from memory in current bank in A/Y into fac2
-romsub $af60 = MOVFRM(uword mflpt @ AY) clobbers(A,X,Y)     ; load mflpt value from memory in A/Y into fac1  (use MOVFM instead)
-romsub $af63 = MOVFM(uword mflpt @ AY) clobbers(A,X,Y)      ; load mflpt value from memory  in A/Y into fac1
-romsub $af66 = MOVMF(uword mflpt @ XY) clobbers(A,X,Y)      ; store fac1 to memory  X/Y as 5-byte mflpt
-romsub $af69 = MOVFA() clobbers(A,X)                        ; copy fac2 to fac1
-romsub $af6c = MOVAF() clobbers(A,X)                        ; copy fac1 to fac2  (rounded)
-
-; X16 additions TODO so.... not on c128 !?
-romsub $af6f = FADDH() clobbers(A,X,Y)                      ; fac1 += 0.5, for rounding- call this before INT
-romsub $af72 = FADDT2() clobbers(A,X,Y)                     ; fac1 += fac2
-romsub $af75 = ZEROFC() clobbers(A,X,Y)                     ; fac1 = 0
-romsub $af78 = NORMAL() clobbers(A,X,Y)                     ; normalize fac1 (?)
-romsub $af7b = NEGFAC() clobbers(A)                         ; switch the sign of fac1 (fac1 = -fac1) (juse use NEGOP() instead!)
-romsub $af7e = FMULTT2() clobbers(A,X,Y)                    ; fac1 *= fac2
-romsub $af81 = MUL10() clobbers(A,X,Y)                      ; fac1 *= 10
-romsub $af84 = DIV10() clobbers(A,X,Y)                      ; fac1 /= 10 , CAUTION: result is always positive!
-romsub $af87 = FDIVT2() clobbers(A,X,Y)                     ; fac1 = fac2/fac1  mind the order of the operands
-romsub $af8a = MOVEF() clobbers(A,X)                        ; copy fac1 to fac2
-romsub $af8d = SGN() clobbers(A,X,Y)                        ; fac1 = SGN(fac1), result of SIGN (-1, 0 or 1)
-romsub $af90 = FLOAT() clobbers(A,X,Y)                      ; FAC = (s8).A
-romsub $af93 = FLOATS() clobbers(A,X,Y)                     ; FAC = (s16)facho+1:facho
-romsub $af9C = QINT() clobbers(A,X,Y)                       ; facho:facho+1:facho+2:facho+3 = u32(FAC)
-romsub $af9f = FINLOG(byte value @A) clobbers (A, X, Y)     ; fac1 += signed byte in A
-romsub $afa5 = FPWRT2() clobbers(A,X,Y)                     ; fac1 = fac2 ** fac1
-
-asmsub  FREADSA  (byte value @A) clobbers(A,X,Y) {
-    ; ---- 8 bit signed A -> float in fac1
-    %asm {{
-        tay
-        bpl  +
-        lda  #$ff
-        jmp  GIVAYF
-+       lda  #0
-        jmp  GIVAYF
-    }}
-}
-
-asmsub  GIVUAYFAY  (uword value @ AY) clobbers(A,X,Y)  {
-	; ---- unsigned 16 bit word in A/Y (lo/hi) to fac1
-	%asm {{
-	phx
-	sty  $64        ; facmo
-	sta  $65        ; facmo+1
-	ldx  #$90
-	sec
-	jsr  FLOATC
-	plx
-	rts
-	}}
-}
-
-asmsub  GIVAYFAY  (uword value @ AY) clobbers(A,X,Y)  {
-	; ---- signed 16 bit word in A/Y (lo/hi) to float in fac1
-	%asm {{
-		sta  P8ZP_SCRATCH_B1
-		tya
-		ldy  P8ZP_SCRATCH_B1
-		jmp  GIVAYF		; this uses the inverse order, Y/A
-	}}
-}
-
-asmsub  GETADRAY  () clobbers(X) -> uword @ AY  {
-	; ---- fac1 to unsigned word in A/Y
-	%asm {{
-		jsr  GETADR		; this uses the inverse order, Y/A
-		sta  P8ZP_SCRATCH_B1
-		tya
-		ldy  P8ZP_SCRATCH_B1
-		rts
-	}}
-}
-
-asmsub  FREADUY (ubyte value @Y) {
-    ; -- 8 bit unsigned Y -> float in fac1
-    %asm {{
-        lda  #0
-        jmp  GIVAYF
-    }}
-}
-
-&uword AYINT_facmo = $66      ; $66/$67 contain result of AYINT
-
-sub rndf() -> float {
-    %asm {{
-        stx  P8ZP_SCRATCH_REG
-        lda  #1
-        jsr  RND_0
-        ldx  P8ZP_SCRATCH_REG
-        rts
-    }}
-}
-
-
-%asminclude "library:c128/floats.asm"
-%asminclude "library:c64/floats_funcs.asm"
-
-}

compiler/res/prog8lib/c128/graphics.p8

@@ -6,6 +6,8 @@
 ; TODO c128 actually implement the graphics routines. Ideally a way to 'borrow' the code form the C64 version without just copy-pasting that here?
 
 graphics {
+    %option no_symbol_prefixing
+
     const uword WIDTH = 320
     const ubyte HEIGHT = 200
 

compiler/res/prog8lib/c128/syslib.p8

@@ -1,17 +1,18 @@
 ; Prog8 definitions for the Commodore-128
 ; Including memory registers, I/O registers, Basic and Kernal subroutines.
 
-c64 {
+cbm {
+    ; Commodore (CBM) common variables, vectors and kernal routines
+    %option no_symbol_prefixing
 
         &ubyte  TIME_HI         = $a0       ; software jiffy clock, hi byte
         &ubyte  TIME_MID        = $a1       ;  .. mid byte
         &ubyte  TIME_LO         = $a2       ;    .. lo byte. Updated by IRQ every 1/60 sec
         &ubyte  STATUS          = $90       ; kernal status variable for I/O
         &ubyte  STKEY           = $91       ; various keyboard statuses (updated by IRQ)
-        ;;&ubyte  SFDX            = $cb       ; current key pressed (matrix value) (updated by IRQ)     // TODO c128 ??
-
-        &ubyte  COLOR           = $00f1     ; cursor color
-        ;;&ubyte  HIBASE          = $0288     ; screen base address / 256 (hi-byte of screen memory address)        // TODO c128 ??
+        &ubyte  SHFLAG          = $d3       ; various modifier key status (updated by IRQ)
+        &ubyte  SFDX            = $d4       ; current key pressed (matrix value) (updated by IRQ)
+        &ubyte  COLOR           = $f1       ; cursor color
 
         &uword  IERROR          = $0300
         &uword  IMAIN           = $0302
@@ -48,6 +49,100 @@ c64 {
         const  uword  Screen    = $0400     ; to have this as an array[40*25] the compiler would have to support array size > 255
         const  uword  Colors    = $d800     ; to have this as an array[40*25] the compiler would have to support array size > 255
 
+; ---- kernal routines, these are the same as on the Commodore-64 (hence the same block name) ----
+
+; STROUT --> use txt.print
+; CLEARSCR -> use txt.clear_screen
+; HOMECRSR -> use txt.home or txt.plot
+
+romsub $FA65 = IRQDFRT() clobbers(A,X,Y)                        ; default IRQ routine
+romsub $FF33 = IRQDFEND() clobbers(A,X,Y)                       ; default IRQ end/cleanup
+
+; TODO c128 a bunch of kernal routines are missing here that are specific to the c128
+
+romsub $FF81 = CINT() clobbers(A,X,Y)                           ; (alias: SCINIT) initialize screen editor and video chip
+romsub $FF84 = IOINIT() clobbers(A, X)                          ; initialize I/O devices (CIA, SID, IRQ)
+romsub $FF87 = RAMTAS() clobbers(A,X,Y)                         ; initialize RAM, tape buffer, screen
+romsub $FF8A = RESTOR() clobbers(A,X,Y)                         ; restore default I/O vectors
+romsub $FF8D = VECTOR(uword userptr @ XY, bool dir @ Pc) clobbers(A,Y)     ; read/set I/O vector table
+romsub $FF90 = SETMSG(ubyte value @ A)                          ; set Kernal message control flag
+romsub $FF93 = SECOND(ubyte address @ A) clobbers(A)            ; (alias: LSTNSA) send secondary address after LISTEN
+romsub $FF96 = TKSA(ubyte address @ A) clobbers(A)              ; (alias: TALKSA) send secondary address after TALK
+romsub $FF99 = MEMTOP(uword address @ XY, bool dir @ Pc) -> uword @ XY     ; read/set top of memory  pointer
+romsub $FF9C = MEMBOT(uword address @ XY, bool dir @ Pc) -> uword @ XY     ; read/set bottom of memory  pointer
+romsub $FF9F = SCNKEY() clobbers(A,X,Y)                         ; scan the keyboard
+romsub $FFA2 = SETTMO(ubyte timeout @ A)                        ; set time-out flag for IEEE bus
+romsub $FFA5 = ACPTR() -> ubyte @ A                             ; (alias: IECIN) input byte from serial bus
+romsub $FFA8 = CIOUT(ubyte databyte @ A)                        ; (alias: IECOUT) output byte to serial bus
+romsub $FFAB = UNTLK() clobbers(A)                              ; command serial bus device to UNTALK
+romsub $FFAE = UNLSN() clobbers(A)                              ; command serial bus device to UNLISTEN
+romsub $FFB1 = LISTEN(ubyte device @ A) clobbers(A)             ; command serial bus device to LISTEN
+romsub $FFB4 = TALK(ubyte device @ A) clobbers(A)               ; command serial bus device to TALK
+romsub $FFB7 = READST() -> ubyte @ A                            ; read I/O status word
+romsub $FFBA = SETLFS(ubyte logical @ A, ubyte device @ X, ubyte secondary @ Y)   ; set logical file parameters
+romsub $FFBD = SETNAM(ubyte namelen @ A, str filename @ XY)     ; set filename parameters
+romsub $FFC0 = OPEN() clobbers(X,Y) -> bool @Pc, ubyte @A      ; (via 794 ($31A)) open a logical file
+romsub $FFC3 = CLOSE(ubyte logical @ A) clobbers(A,X,Y)         ; (via 796 ($31C)) close a logical file
+romsub $FFC6 = CHKIN(ubyte logical @ X) clobbers(A,X) -> bool @Pc    ; (via 798 ($31E)) define an input channel
+romsub $FFC9 = CHKOUT(ubyte logical @ X) clobbers(A,X)          ; (via 800 ($320)) define an output channel
+romsub $FFCC = CLRCHN() clobbers(A,X)                           ; (via 802 ($322)) restore default devices
+romsub $FFCF = CHRIN() clobbers(X, Y) -> ubyte @ A   ; (via 804 ($324)) input a character (for keyboard, read a whole line from the screen) A=byte read.
+romsub $FFD2 = CHROUT(ubyte char @ A)                           ; (via 806 ($326)) output a character
+romsub $FFD5 = LOAD(ubyte verify @ A, uword address @ XY) -> bool @Pc, ubyte @ A, uword @ XY     ; (via 816 ($330)) load from device
+romsub $FFD8 = SAVE(ubyte zp_startaddr @ A, uword endaddr @ XY) -> bool @ Pc, ubyte @ A          ; (via 818 ($332)) save to a device
+romsub $FFDB = SETTIM(ubyte low @ A, ubyte middle @ X, ubyte high @ Y)      ; set the software clock
+romsub $FFDE = RDTIM() -> ubyte @ A, ubyte @ X, ubyte @ Y       ; read the software clock (A=lo,X=mid,Y=high)
+romsub $FFE1 = STOP() clobbers(X) -> bool @ Pz, ubyte @ A      ; (via 808 ($328)) check the STOP key (and some others in A)
+romsub $FFE4 = GETIN() clobbers(X,Y) -> bool @Pc, ubyte @ A    ; (via 810 ($32A)) get a character
+romsub $FFE7 = CLALL() clobbers(A,X)                            ; (via 812 ($32C)) close all files
+romsub $FFEA = UDTIM() clobbers(A,X)                            ; update the software clock
+romsub $FFED = SCREEN() -> ubyte @ X, ubyte @ Y                 ; read number of screen rows and columns
+romsub $FFF0 = PLOT(ubyte col @ Y, ubyte row @ X, bool dir @ Pc) -> ubyte @ X, ubyte @ Y       ; read/set position of cursor on screen.  Use txt.plot for a 'safe' wrapper that preserves X.
+romsub $FFF3 = IOBASE() -> uword @ XY                           ; read base address of I/O devices
+
+; ---- end of C64 compatible ROM kernal routines ----
+
+; ---- utilities -----
+
+asmsub STOP2() -> ubyte @A  {
+    ; -- check if STOP key was pressed, returns true if so.  More convenient to use than STOP() because that only sets the carry status flag.
+    %asm {{
+        txa
+        pha
+        jsr  cbm.STOP
+        beq  +
+        pla
+        tax
+        lda  #0
+        rts
++       pla
+        tax
+        lda  #1
+        rts
+    }}
+}
+
+asmsub RDTIM16() -> uword @AY {
+    ; --  like RDTIM() but only returning the lower 16 bits in AY for convenience
+    %asm {{
+        stx  P8ZP_SCRATCH_REG
+        jsr  cbm.RDTIM
+        pha
+        txa
+        tay
+        pla
+        ldx  P8ZP_SCRATCH_REG
+        rts
+    }}
+}
+
+
+}
+
+c64 {
+        ; C64 I/O registers (VIC, SID, CIA)
+    %option no_symbol_prefixing
+
         ; the default locations of the 8 sprite pointers (store address of sprite / 64)
         &ubyte  SPRPTR0         = 2040
         &ubyte  SPRPTR1         = 2041
@@ -197,95 +292,93 @@ c64 {
 
     ; ---- end of SID registers ----
 
-; ---- kernal routines, these are the same as on the Commodore-64 (hence the same block name) ----
+}
 
-; STROUT --> use txt.print
-; CLEARSCR -> use txt.clear_screen
-; HOMECRSR -> use txt.home or txt.plot
+c128 {
+; ---- C128 specific registers ----
+    %option no_symbol_prefixing
 
-romsub $FA65 = IRQDFRT() clobbers(A,X,Y)                        ; default IRQ routine
-romsub $FF33 = IRQDFEND() clobbers(A,X,Y)                       ; default IRQ end/cleanup
+    &ubyte  VM1     = $0A2C         ; shadow for VUC $d018 in text mode
+    &ubyte  VM2     = $0A2D         ; shadow for VIC $d018 in bitmap screen mode
+    &ubyte  VM3     = $0A2E         ; starting page for VDC screen mem
+    &ubyte  VM4     = $0A2F         ; starting page for VDC attribute mem
 
-; TODO c128 a bunch of kernal routines are missing here that are specific to the c128
 
-romsub $FF81 = CINT() clobbers(A,X,Y)                           ; (alias: SCINIT) initialize screen editor and video chip
-romsub $FF84 = IOINIT() clobbers(A, X)                          ; initialize I/O devices (CIA, SID, IRQ)
-romsub $FF87 = RAMTAS() clobbers(A,X,Y)                         ; initialize RAM, tape buffer, screen
-romsub $FF8A = RESTOR() clobbers(A,X,Y)                         ; restore default I/O vectors
-romsub $FF8D = VECTOR(uword userptr @ XY, ubyte dir @ Pc) clobbers(A,Y)     ; read/set I/O vector table
-romsub $FF90 = SETMSG(ubyte value @ A)                          ; set Kernal message control flag
-romsub $FF93 = SECOND(ubyte address @ A) clobbers(A)            ; (alias: LSTNSA) send secondary address after LISTEN
-romsub $FF96 = TKSA(ubyte address @ A) clobbers(A)              ; (alias: TALKSA) send secondary address after TALK
-romsub $FF99 = MEMTOP(uword address @ XY, ubyte dir @ Pc) -> uword @ XY     ; read/set top of memory  pointer
-romsub $FF9C = MEMBOT(uword address @ XY, ubyte dir @ Pc) -> uword @ XY     ; read/set bottom of memory  pointer
-romsub $FF9F = SCNKEY() clobbers(A,X,Y)                         ; scan the keyboard
-romsub $FFA2 = SETTMO(ubyte timeout @ A)                        ; set time-out flag for IEEE bus
-romsub $FFA5 = ACPTR() -> ubyte @ A                             ; (alias: IECIN) input byte from serial bus
-romsub $FFA8 = CIOUT(ubyte databyte @ A)                        ; (alias: IECOUT) output byte to serial bus
-romsub $FFAB = UNTLK() clobbers(A)                              ; command serial bus device to UNTALK
-romsub $FFAE = UNLSN() clobbers(A)                              ; command serial bus device to UNLISTEN
-romsub $FFB1 = LISTEN(ubyte device @ A) clobbers(A)             ; command serial bus device to LISTEN
-romsub $FFB4 = TALK(ubyte device @ A) clobbers(A)               ; command serial bus device to TALK
-romsub $FFB7 = READST() -> ubyte @ A                            ; read I/O status word
-romsub $FFBA = SETLFS(ubyte logical @ A, ubyte device @ X, ubyte secondary @ Y)   ; set logical file parameters
-romsub $FFBD = SETNAM(ubyte namelen @ A, str filename @ XY)     ; set filename parameters
-romsub $FFC0 = OPEN() clobbers(X,Y) -> ubyte @Pc, ubyte @A      ; (via 794 ($31A)) open a logical file
-romsub $FFC3 = CLOSE(ubyte logical @ A) clobbers(A,X,Y)         ; (via 796 ($31C)) close a logical file
-romsub $FFC6 = CHKIN(ubyte logical @ X) clobbers(A,X) -> ubyte @Pc    ; (via 798 ($31E)) define an input channel
-romsub $FFC9 = CHKOUT(ubyte logical @ X) clobbers(A,X)          ; (via 800 ($320)) define an output channel
-romsub $FFCC = CLRCHN() clobbers(A,X)                           ; (via 802 ($322)) restore default devices
-romsub $FFCF = CHRIN() clobbers(X, Y) -> ubyte @ A   ; (via 804 ($324)) input a character (for keyboard, read a whole line from the screen) A=byte read.
-romsub $FFD2 = CHROUT(ubyte char @ A)                           ; (via 806 ($326)) output a character
-romsub $FFD5 = LOAD(ubyte verify @ A, uword address @ XY) -> ubyte @Pc, ubyte @ A, uword @ XY     ; (via 816 ($330)) load from device
-romsub $FFD8 = SAVE(ubyte zp_startaddr @ A, uword endaddr @ XY) -> ubyte @ Pc, ubyte @ A          ; (via 818 ($332)) save to a device
-romsub $FFDB = SETTIM(ubyte low @ A, ubyte middle @ X, ubyte high @ Y)      ; set the software clock
-romsub $FFDE = RDTIM() -> ubyte @ A, ubyte @ X, ubyte @ Y       ; read the software clock (A=lo,X=mid,Y=high)
-romsub $FFE1 = STOP() clobbers(X) -> ubyte @ Pz, ubyte @ A      ; (via 808 ($328)) check the STOP key (and some others in A)
-romsub $FFE4 = GETIN() clobbers(X,Y) -> ubyte @Pc, ubyte @ A    ; (via 810 ($32A)) get a character
-romsub $FFE7 = CLALL() clobbers(A,X)                            ; (via 812 ($32C)) close all files
-romsub $FFEA = UDTIM() clobbers(A,X)                            ; update the software clock
-romsub $FFED = SCREEN() -> ubyte @ X, ubyte @ Y                 ; read number of screen rows and columns
-romsub $FFF0 = PLOT(ubyte col @ Y, ubyte row @ X, ubyte dir @ Pc) -> ubyte @ X, ubyte @ Y       ; read/set position of cursor on screen.  Use txt.plot for a 'safe' wrapper that preserves X.
-romsub $FFF3 = IOBASE() -> uword @ XY                           ; read base address of I/O devices
+; ---- C128 specific system utility routines: ----
 
-; ---- end of C64 compatible ROM kernal routines ----
-
-; ---- utilities -----
-
-asmsub STOP2() -> ubyte @A  {
-    ; -- check if STOP key was pressed, returns true if so.  More convenient to use than STOP() because that only sets the carry status flag.
+asmsub  disable_basic() clobbers(A) {
     %asm {{
-        txa
-        pha
-        jsr  c64.STOP
-        beq  +
-        pla
-        tax
+        lda $0a04   ; disable BASIC shadow registers
+        and #$fe
+        sta $0a04
+
+        lda #$01    ; disable BASIC IRQ service routine
+        sta $12fd
+
+        lda #$ff    ; disable screen editor IRQ setup
+        sta $d8
+
+        lda #$b7    ; skip programmable function key check
+        sta $033c
+
+        lda #$0e    ; bank out BASIC ROM
+        sta $ff00
+        rts
+    }}
+}
+
+; ---- end of C128 specific system utility routines ----
+
+}
+
+sys {
+    ; ------- lowlevel system routines --------
+    %option no_symbol_prefixing
+
+    const ubyte target = 128         ;  compilation target specifier.  64 = C64, 128 = C128,  16 = CommanderX16.
+
+asmsub  init_system()  {
+    ; Initializes the machine to a sane starting state.
+    ; Called automatically by the loader program logic.
+    ; This means that the BASIC, KERNAL and CHARGEN ROMs are banked in,
+    ; the VIC, SID and CIA chips are reset, screen is cleared, and the default IRQ is set.
+    ; Also a different color scheme is chosen to identify ourselves a little.
+    ; Uppercase charset is activated, and all three registers set to 0, status flags cleared.
+    %asm {{
+        sei
+        cld
         lda  #0
-        rts
-+       pla
-        tax
-        lda  #1
+        sta  $ff00      ; select default bank 15
+        jsr  cbm.IOINIT
+        jsr  cbm.RESTOR
+        jsr  cbm.CINT
+        lda  #6
+        sta  c64.EXTCOL
+        lda  #7
+        sta  cbm.COLOR
+        lda  #0
+        sta  c64.BGCOL0
+        jsr  disable_runstop_and_charsetswitch
+        clc
+        clv
+        cli
         rts
     }}
 }
 
-asmsub RDTIM16() -> uword @AY {
-    ; --  like RDTIM() but only returning the lower 16 bits in AY for convenience
+asmsub  init_system_phase2()  {
     %asm {{
-        stx  P8ZP_SCRATCH_REG
-        jsr  c64.RDTIM
-        pha
-        txa
-        tay
-        pla
-        ldx  P8ZP_SCRATCH_REG
-        rts
+        rts     ; no phase 2 steps on the C128
     }}
 }
 
+asmsub  cleanup_at_exit() {
+    ; executed when the main subroutine does rts
+    %asm {{
+        jmp  sys.enable_runstop_and_charsetswitch
+    }}
+}
 
-; ---- system utility routines that are essentially the same as on the C64: -----
 asmsub  disable_runstop_and_charsetswitch() clobbers(A) {
     %asm {{
         lda  #$80
@@ -306,18 +399,18 @@ asmsub  enable_runstop_and_charsetswitch() clobbers(A) {
     }}
 }
 
-asmsub  set_irq(uword handler @AY, ubyte useKernal @Pc) clobbers(A)  {
+asmsub  set_irq(uword handler @AY, bool useKernal @Pc) clobbers(A)  {
 	%asm {{
 	        sta  _modified+1
 	        sty  _modified+2
 	        lda  #0
-	        adc  #0
+	        rol  a
 	        sta  _use_kernal
 		sei
 		lda  #<_irq_handler
-		sta  c64.CINV
+		sta  cbm.CINV
 		lda  #>_irq_handler
-		sta  c64.CINV+1
+		sta  cbm.CINV+1
 		cli
 		rts
 _irq_handler    jsr  _irq_handler_init
@@ -335,7 +428,7 @@ _modified	jsr  $ffff                      ; modified
 		tax
 		pla
 		rti
-+		jmp  c64.IRQDFRT		; continue with normal kernal irq routine
++		jmp  cbm.IRQDFRT		; continue with normal kernal irq routine
 
 _use_kernal     .byte  0
 
@@ -387,10 +480,10 @@ IRQ_SCRATCH_ZPWORD2	.word  0
 asmsub  restore_irq() clobbers(A) {
 	%asm {{
 		sei
-		lda  #<c64.IRQDFRT
-		sta  c64.CINV
-		lda  #>c64.IRQDFRT
-		sta  c64.CINV+1
+		lda  #<cbm.IRQDFRT
+		sta  cbm.CINV
+		lda  #>cbm.IRQDFRT
+		sta  cbm.CINV+1
 		lda  #0
 		sta  c64.IREQMASK	; disable raster irq
 		lda  #%10000001
@@ -400,21 +493,21 @@ asmsub  restore_irq() clobbers(A) {
 	}}
 }
 
-asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0, ubyte useKernal @Pc) clobbers(A) {
+asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0, bool useKernal @Pc) clobbers(A) {
 	%asm {{
 	        sta  _modified+1
 	        sty  _modified+2
 	        lda  #0
-	        adc  #0
+	        rol  a
 	        sta  set_irq._use_kernal
 		lda  cx16.r0
 		ldy  cx16.r0+1
 		sei
 		jsr  _setup_raster_irq
 		lda  #<_raster_irq_handler
-		sta  c64.CINV
+		sta  cbm.CINV
 		lda  #>_raster_irq_handler
-		sta  c64.CINV+1
+		sta  cbm.CINV+1
 		cli
 		rts
 
@@ -422,8 +515,8 @@ _raster_irq_handler
 		jsr  set_irq._irq_handler_init
 _modified	jsr  $ffff              ; modified
 		jsr  set_irq._irq_handler_end
-                lda  #$ff
-                sta  c64.VICIRQ			; acknowledge raster irq
+        lda  #$ff
+        sta  c64.VICIRQ			; acknowledge raster irq
 		lda  set_irq._use_kernal
 		bne  +
 		; end irq processing - don't use kernal's irq handling
@@ -433,7 +526,7 @@ _modified	jsr  $ffff              ; modified
 		tax
 		pla
 		rti
-+		jmp  c64.IRQDFRT                ; continue with kernal irq routine
++		jmp  cbm.IRQDFRT                ; continue with kernal irq routine
 
 _setup_raster_irq
 		pha
@@ -457,101 +550,13 @@ _setup_raster_irq
 	}}
 }
 
-}
-
-c128 {
-; ---- C128 specific registers ----
-
-    &ubyte  VM1     = $0A2C         ; shadow for VUC $d018 in text mode
-    &ubyte  VM2     = $0A2D         ; shadow for VIC $d018 in bitmap screen mode
-    &ubyte  VM3     = $0A2E         ; starting page for VDC screen mem
-    &ubyte  VM4     = $0A2F         ; starting page for VDC attribute mem
-
-
-; ---- C128 specific system utility routines: ----
-
-asmsub  init_system()  {
-    ; Initializes the machine to a sane starting state.
-    ; Called automatically by the loader program logic.
-    ; This means that the BASIC, KERNAL and CHARGEN ROMs are banked in,
-    ; the VIC, SID and CIA chips are reset, screen is cleared, and the default IRQ is set.
-    ; Also a different color scheme is chosen to identify ourselves a little.
-    ; Uppercase charset is activated, and all three registers set to 0, status flags cleared.
-    %asm {{
-        sei
-        cld
-        ;;lda  #%00101111                     ; TODO c128 ram and rom bank selection how?
-        ;;sta  $00
-        ;;lda  #%00100111
-        ;;sta  $01
-        jsr  c64.IOINIT
-        jsr  c64.RESTOR
-        jsr  c64.CINT
-        lda  #6
-        sta  c64.EXTCOL
-        lda  #7
-        sta  c64.COLOR
-        lda  #0
-        sta  c64.BGCOL0
-        jsr  c64.disable_runstop_and_charsetswitch
-        clc
-        clv
-        cli
-        rts
-    }}
-}
-
-asmsub  init_system_phase2()  {
-    %asm {{
-        rts     ; no phase 2 steps on the C128
-    }}
-}
-
-asmsub  cleanup_at_exit() {
-    ; executed when the main subroutine does rts
-    %asm {{
-        jmp  c64.enable_runstop_and_charsetswitch
-    }}
-}
-
-asmsub  disable_basic() clobbers(A) {
-    %asm {{
-        lda $0a04   ; disable BASIC shadow registers
-        and #$fe
-        sta $0a04
-
-        lda #$01    ; disable BASIC IRQ service routine
-        sta $12fd
-
-        lda #$ff    ; disable screen editor IRQ setup
-        sta $d8
-
-        lda #$b7    ; skip programmable function key check
-        sta $033c
-
-        lda #$0e    ; bank out BASIC ROM
-        sta $ff00
-        rts
-    }}
-}
-
-; ---- end of C128 specific system utility routines ----
-
-}
-
-sys {
-    ; ------- lowlevel system routines --------
-
-    const ubyte target = 128         ;  compilation target specifier.  64 = C64, 128 = C128,  16 = CommanderX16.
-
-
     asmsub  reset_system()  {
         ; Soft-reset the system back to initial power-on Basic prompt.
         %asm {{
             sei
-            ;lda  #14
-            ;sta  $01        ; bank the kernal in       TODO c128 how to do this?
-            jmp  (c64.RESET_VEC)
+            lda  #0
+            sta  $ff00      ; default bank 15
+            jmp  (cbm.RESET_VEC)
         }}
     }
 
@@ -569,9 +574,9 @@ _loop       lda  P8ZP_SCRATCH_W1
             ldx  P8ZP_SCRATCH_B1
             rts
 
-+           lda  c64.TIME_LO
++           lda  cbm.TIME_LO
             sta  P8ZP_SCRATCH_B1
--           lda  c64.TIME_LO
+-           lda  cbm.TIME_LO
             cmp  P8ZP_SCRATCH_B1
             beq  -
 
@@ -728,13 +733,26 @@ _longcopy
         }}
     }
 
+    inline asmsub irqsafe_set_irqd() {
+        %asm {{
+        php
+        sei
+        }}
+    }
+
+    inline asmsub irqsafe_clear_irqd() {
+        %asm {{
+        plp
+        }}
+    }
+
     inline asmsub exit(ubyte returnvalue @A) {
         ; -- immediately exit the program with a return code in the A register
         %asm {{
-            ;lda  #14
-            ;sta  $01        ; bank the kernal in       TODO c128 how to do this?
-            jsr  c64.CLRCHN		; reset i/o channels
-            jsr  c64.enable_runstop_and_charsetswitch
+            lda  #0
+            sta  $ff00          ; default bank 15
+            jsr  cbm.CLRCHN		; reset i/o channels
+            jsr  sys.enable_runstop_and_charsetswitch
             ldx  prog8_lib.orig_stackpointer
             txs
             rts		; return to original caller
@@ -751,6 +769,7 @@ _longcopy
 }
 
 cx16 {
+    %option no_symbol_prefixing
 
     ; the sixteen virtual 16-bit registers that the CX16 has defined in the zeropage
     ; they are simulated on the C128 as well but their location in memory is different
@@ -860,4 +879,30 @@ cx16 {
     &byte r13sH = $1b1b
     &byte r14sH = $1b1d
     &byte r15sH = $1b1f
+
+    asmsub save_virtual_registers() clobbers(A,Y) {
+        %asm {{
+            ldy  #31
+    -       lda  cx16.r0,y
+            sta  _cx16_vreg_storage,y
+            dey
+            bpl  -
+            rts
+
+    _cx16_vreg_storage
+            .word 0,0,0,0,0,0,0,0
+            .word 0,0,0,0,0,0,0,0
+        }}
+    }
+
+    asmsub restore_virtual_registers() clobbers(A,Y) {
+        %asm {{
+            ldy  #31
+    -       lda  save_virtual_registers._cx16_vreg_storage,y
+            sta  cx16.r0,y
+            dey
+            bpl  -
+            rts
+        }}
+    }
 }

compiler/res/prog8lib/c128/textio.p8

@@ -5,6 +5,7 @@
 
 
 txt {
+    %option no_symbol_prefixing
 
 const ubyte DEFAULT_WIDTH = 40
 const ubyte DEFAULT_HEIGHT = 25
@@ -30,10 +31,10 @@ asmsub column(ubyte col @A) clobbers(A, X, Y) {
     ; ---- set the cursor on the given column (starting with 0) on the current line
     %asm {{
         sec
-        jsr  c64.PLOT
+        jsr  cbm.PLOT
         tay
         clc
-        jmp  c64.PLOT
+        jmp  cbm.PLOT
     }}
 }
 
@@ -57,10 +58,10 @@ asmsub  clear_screenchars (ubyte char @ A) clobbers(Y)  {
 	;      (assumes screen matrix is at the default address)
 	%asm {{
 		ldy  #250
--		sta  c64.Screen+250*0-1,y
-		sta  c64.Screen+250*1-1,y
-		sta  c64.Screen+250*2-1,y
-		sta  c64.Screen+250*3-1,y
+-		sta  cbm.Screen+250*0-1,y
+		sta  cbm.Screen+250*1-1,y
+		sta  cbm.Screen+250*2-1,y
+		sta  cbm.Screen+250*3-1,y
 		dey
 		bne  -
 		rts
@@ -72,10 +73,10 @@ asmsub  clear_screencolors (ubyte color @ A) clobbers(Y)  {
 	;      (assumes color matrix is at the default address)
 	%asm {{
 		ldy  #250
--		sta  c64.Colors+250*0-1,y
-		sta  c64.Colors+250*1-1,y
-		sta  c64.Colors+250*2-1,y
-		sta  c64.Colors+250*3-1,y
+-		sta  cbm.Colors+250*0-1,y
+		sta  cbm.Colors+250*1-1,y
+		sta  cbm.Colors+250*2-1,y
+		sta  cbm.Colors+250*3-1,y
 		dey
 		bne  -
 		rts
@@ -83,7 +84,7 @@ asmsub  clear_screencolors (ubyte color @ A) clobbers(Y)  {
 }
 
 sub color (ubyte txtcol) {
-    c64.COLOR = txtcol
+    cbm.COLOR = txtcol
 }
 
 sub lowercase() {
@@ -96,7 +97,7 @@ sub uppercase() {
     c128.VM1 &= ~2
 }
 
-asmsub  scroll_left  (ubyte alsocolors @ Pc) clobbers(A, Y)  {
+asmsub  scroll_left  (bool alsocolors @ Pc) clobbers(A, Y)  {
 	; ---- scroll the whole screen 1 character to the left
 	;      contents of the rightmost column are unchanged, you should clear/refill this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -110,10 +111,10 @@ asmsub  scroll_left  (ubyte alsocolors @ Pc) clobbers(A, Y)  {
 		ldy  #38
 -
         .for row=0, row<=24, row+=1
-            lda  c64.Screen + 40*row + 1,x
-            sta  c64.Screen + 40*row + 0,x
-            lda  c64.Colors + 40*row + 1,x
-            sta  c64.Colors + 40*row + 0,x
+            lda  cbm.Screen + 40*row + 1,x
+            sta  cbm.Screen + 40*row + 0,x
+            lda  cbm.Colors + 40*row + 1,x
+            sta  cbm.Colors + 40*row + 0,x
         .next
 		inx
 		dey
@@ -125,8 +126,8 @@ _scroll_screen  ; scroll only the screen memory
 		ldy  #38
 -
         .for row=0, row<=24, row+=1
-            lda  c64.Screen + 40*row + 1,x
-            sta  c64.Screen + 40*row + 0,x
+            lda  cbm.Screen + 40*row + 1,x
+            sta  cbm.Screen + 40*row + 0,x
         .next
 		inx
 		dey
@@ -137,7 +138,7 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-asmsub  scroll_right  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_right  (bool alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character to the right
 	;      contents of the leftmost column are unchanged, you should clear/refill this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -149,10 +150,10 @@ asmsub  scroll_right  (ubyte alsocolors @ Pc) clobbers(A)  {
 		ldx  #38
 -
         .for row=0, row<=24, row+=1
-            lda  c64.Screen + 40*row + 0,x
-            sta  c64.Screen + 40*row + 1,x
-            lda  c64.Colors + 40*row + 0,x
-            sta  c64.Colors + 40*row + 1,x
+            lda  cbm.Screen + 40*row + 0,x
+            sta  cbm.Screen + 40*row + 1,x
+            lda  cbm.Colors + 40*row + 0,x
+            sta  cbm.Colors + 40*row + 1,x
         .next
 		dex
 		bpl  -
@@ -162,8 +163,8 @@ _scroll_screen  ; scroll only the screen memory
 		ldx  #38
 -
         .for row=0, row<=24, row+=1
-            lda  c64.Screen + 40*row + 0,x
-            sta  c64.Screen + 40*row + 1,x
+            lda  cbm.Screen + 40*row + 0,x
+            sta  cbm.Screen + 40*row + 1,x
         .next
 		dex
 		bpl  -
@@ -173,7 +174,7 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-asmsub  scroll_up  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_up  (bool alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character up
 	;      contents of the bottom row are unchanged, you should refill/clear this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -185,10 +186,10 @@ asmsub  scroll_up  (ubyte alsocolors @ Pc) clobbers(A)  {
 		ldx #39
 -
         .for row=1, row<=24, row+=1
-            lda  c64.Screen + 40*row,x
-            sta  c64.Screen + 40*(row-1),x
-            lda  c64.Colors + 40*row,x
-            sta  c64.Colors + 40*(row-1),x
+            lda  cbm.Screen + 40*row,x
+            sta  cbm.Screen + 40*(row-1),x
+            lda  cbm.Colors + 40*row,x
+            sta  cbm.Colors + 40*(row-1),x
         .next
 		dex
 		bpl  -
@@ -198,8 +199,8 @@ _scroll_screen  ; scroll only the screen memory
 		ldx #39
 -
         .for row=1, row<=24, row+=1
-            lda  c64.Screen + 40*row,x
-            sta  c64.Screen + 40*(row-1),x
+            lda  cbm.Screen + 40*row,x
+            sta  cbm.Screen + 40*(row-1),x
         .next
 		dex
 		bpl  -
@@ -209,7 +210,7 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-asmsub  scroll_down  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_down  (bool alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character down
 	;      contents of the top row are unchanged, you should refill/clear this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -221,10 +222,10 @@ asmsub  scroll_down  (ubyte alsocolors @ Pc) clobbers(A)  {
 		ldx #39
 -
         .for row=23, row>=0, row-=1
-            lda  c64.Colors + 40*row,x
-            sta  c64.Colors + 40*(row+1),x
-            lda  c64.Screen + 40*row,x
-            sta  c64.Screen + 40*(row+1),x
+            lda  cbm.Colors + 40*row,x
+            sta  cbm.Colors + 40*(row+1),x
+            lda  cbm.Screen + 40*row,x
+            sta  cbm.Screen + 40*(row+1),x
         .next
 		dex
 		bpl  -
@@ -234,8 +235,8 @@ _scroll_screen  ; scroll only the screen memory
 		ldx #39
 -
         .for row=23, row>=0, row-=1
-            lda  c64.Screen + 40*row,x
-            sta  c64.Screen + 40*(row+1),x
+            lda  cbm.Screen + 40*row,x
+            sta  cbm.Screen + 40*(row+1),x
         .next
 		dex
 		bpl  -
@@ -245,20 +246,20 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use c64.CHROUT directly ofcourse.
+romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use cbm.CHROUT directly ofcourse.
 
 asmsub  print (str text @ AY) clobbers(A,Y)  {
 	; ---- print null terminated string from A/Y
 	; note: the compiler contains an optimization that will replace
 	;       a call to this subroutine with a string argument of just one char,
-	;       by just one call to c64.CHROUT of that single char.
+	;       by just one call to cbm.CHROUT of that single char.
 	%asm {{
 		sta  P8ZP_SCRATCH_B1
 		sty  P8ZP_SCRATCH_REG
 		ldy  #0
 -		lda  (P8ZP_SCRATCH_B1),y
 		beq  +
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		bne  -
 +		rts
@@ -272,11 +273,11 @@ asmsub  print_ub0  (ubyte value @ A) clobbers(A,Y)  {
 		jsr  conv.ubyte2decimal
 		pha
 		tya
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		pla
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		txa
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
@@ -292,16 +293,16 @@ _print_byte_digits
 		cpy  #'0'
 		beq  +
 		tya
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		pla
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		jmp  _ones
 +       pla
         cmp  #'0'
         beq  _ones
-        jsr  c64.CHROUT
+        jsr  cbm.CHROUT
 _ones   txa
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
@@ -315,45 +316,45 @@ asmsub  print_b  (byte value @ A) clobbers(A,Y)  {
 		cmp  #0
 		bpl  +
 		lda  #'-'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 +		pla
 		jsr  conv.byte2decimal
 		jmp  print_ub._print_byte_digits
 	}}
 }
 
-asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_ubhex  (ubyte value @ A, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in hex form (if Carry is set, a radix prefix '$' is printed as well)
 	%asm {{
 		stx  P8ZP_SCRATCH_REG
 		bcc  +
 		pha
 		lda  #'$'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		pla
 +		jsr  conv.ubyte2hex
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		tya
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
 
-asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_ubbin  (ubyte value @ A, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in binary form (if Carry is set, a radix prefix '%' is printed as well)
 	%asm {{
 		stx  P8ZP_SCRATCH_REG
 		sta  P8ZP_SCRATCH_B1
 		bcc  +
 		lda  #'%'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 +		ldy  #8
 -		lda  #'0'
 		asl  P8ZP_SCRATCH_B1
 		bcc  +
 		lda  #'1'
-+		jsr  c64.CHROUT
++		jsr  cbm.CHROUT
 		dey
 		bne  -
 		ldx  P8ZP_SCRATCH_REG
@@ -361,7 +362,7 @@ asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_uwbin  (uword value @ AY, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in binary form (if Carry is set, a radix prefix '%' is printed as well)
 	%asm {{
 		pha
@@ -373,7 +374,7 @@ asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_uwhex  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_uwhex  (uword value @ AY, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in hexadecimal form (4 digits)
 	;      (if Carry is set, a radix prefix '$' is printed as well)
 	%asm {{
@@ -394,7 +395,7 @@ asmsub  print_uw0  (uword value @ AY) clobbers(A,Y)  {
 		ldy  #0
 -		lda  conv.uword2decimal.decTenThousands,y
         beq  +
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		bne  -
 +		ldx  P8ZP_SCRATCH_REG
@@ -417,14 +418,14 @@ asmsub  print_uw  (uword value @ AY) clobbers(A,Y)  {
 		bne  -
 
 _gotdigit
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		lda  conv.uword2decimal.decTenThousands,y
 		bne  _gotdigit
 		rts
 _allzero
         lda  #'0'
-        jmp  c64.CHROUT
+        jmp  cbm.CHROUT
 	}}
 }
 
@@ -435,7 +436,7 @@ asmsub  print_w  (word value @ AY) clobbers(A,Y)  {
 		bpl  +
 		pha
 		lda  #'-'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		tya
 		eor  #255
 		tay
@@ -457,7 +458,7 @@ asmsub  input_chars  (uword buffer @ AY) clobbers(A) -> ubyte @ Y  {
 		sta  P8ZP_SCRATCH_W1
 		sty  P8ZP_SCRATCH_W1+1
 		ldy  #0				; char counter = 0
--		jsr  c64.CHRIN
+-		jsr  cbm.CHRIN
 		cmp  #$0d			; return (ascii 13) pressed?
 		beq  +				; yes, end.
 		sta  (P8ZP_SCRATCH_W1),y	; else store char in buffer
@@ -588,7 +589,7 @@ asmsub  plot  (ubyte col @ Y, ubyte row @ A) clobbers(A) {
 		stx  P8ZP_SCRATCH_REG
 		tax
 		clc
-		jsr  c64.PLOT
+		jsr  cbm.PLOT
 		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
@@ -597,7 +598,7 @@ asmsub  plot  (ubyte col @ Y, ubyte row @ A) clobbers(A) {
 asmsub width() clobbers(X,Y) -> ubyte @A {
     ; -- returns the text screen width (number of columns)
     %asm {{
-        jsr  c64.SCREEN
+        jsr  cbm.SCREEN
         txa
         rts
     }}
@@ -606,7 +607,7 @@ asmsub width() clobbers(X,Y) -> ubyte @A {
 asmsub height() clobbers(X, Y) -> ubyte @A {
     ; -- returns the text screen height (number of rows)
     %asm {{
-        jsr  c64.SCREEN
+        jsr  cbm.SCREEN
         tya
         rts
     }}

compiler/res/prog8lib/c64/floats.p8

@@ -5,6 +5,7 @@
 
 floats {
 	; ---- this block contains C-64 floating point related functions ----
+    %option no_symbol_prefixing
 
         const float  PI     = 3.141592653589793
         const float  TWOPI  = 6.283185307179586

compiler/res/prog8lib/c64/floats_funcs.asm

@@ -144,3 +144,19 @@ func_all_f_stack	.proc
 		jsr  a_times_5
 		jmp  prog8_lib.func_all_b_stack
 		.pend
+
+func_abs_f_into_FAC1    .proc
+        stx  P8ZP_SCRATCH_REG
+        jsr  MOVFM
+        jsr  ABS
+        ldx  P8ZP_SCRATCH_REG
+        rts
+	.pend
+
+func_sqrt_into_FAC1     .proc
+        stx  P8ZP_SCRATCH_REG
+        jsr  MOVFM
+        jsr  SQR
+        ldx  P8ZP_SCRATCH_REG
+        rts
+	.pend

compiler/res/prog8lib/c64/graphics.p8

@@ -5,6 +5,8 @@
 ; assumes bitmap screen memory is $2000-$3fff
 
 graphics {
+    %option no_symbol_prefixing
+
     const uword BITMAP_ADDRESS = $2000
     const uword WIDTH = 320
     const ubyte HEIGHT = 200

compiler/res/prog8lib/c64/syslib.p8

@@ -1,13 +1,20 @@
 ; Prog8 definitions for the Commodore-64
 ; Including memory registers, I/O registers, Basic and Kernal subroutines.
 
-c64 {
+
+cbm {
+
+    ; Commodore (CBM) common variables, vectors and kernal routines
+
+    %option no_symbol_prefixing
+
         &ubyte  TIME_HI         = $a0       ; software jiffy clock, hi byte
         &ubyte  TIME_MID        = $a1       ;  .. mid byte
         &ubyte  TIME_LO         = $a2       ;    .. lo byte. Updated by IRQ every 1/60 sec
         &ubyte  STATUS          = $90       ; kernal status variable for I/O
         &ubyte  STKEY           = $91       ; various keyboard statuses (updated by IRQ)
         &ubyte  SFDX            = $cb       ; current key pressed (matrix value) (updated by IRQ)
+        &ubyte  SHFLAG          = $028d     ; various modifier key status (updated by IRQ)
 
         &ubyte  COLOR           = $0286     ; cursor color
         &ubyte  HIBASE          = $0288     ; screen base address / 256 (hi-byte of screen memory address)
@@ -49,6 +56,93 @@ c64 {
         const  uword  Screen    = $0400     ; to have this as an array[40*25] the compiler would have to support array size > 255
         const  uword  Colors    = $d800     ; to have this as an array[40*25] the compiler would have to support array size > 255
 
+
+; ---- CBM ROM kernal routines (C64 addresses) ----
+
+romsub $AB1E = STROUT(uword strptr @ AY) clobbers(A, X, Y)      ; print null-terminated string (use txt.print instead)
+romsub $E544 = CLEARSCR() clobbers(A,X,Y)                       ; clear the screen
+romsub $E566 = HOMECRSR() clobbers(A,X,Y)                       ; cursor to top left of screen
+romsub $EA31 = IRQDFRT() clobbers(A,X,Y)                        ; default IRQ routine
+romsub $EA81 = IRQDFEND() clobbers(A,X,Y)                       ; default IRQ end/cleanup
+romsub $FF81 = CINT() clobbers(A,X,Y)                           ; (alias: SCINIT) initialize screen editor and video chip
+romsub $FF84 = IOINIT() clobbers(A, X)                          ; initialize I/O devices (CIA, SID, IRQ)
+romsub $FF87 = RAMTAS() clobbers(A,X,Y)                         ; initialize RAM, tape buffer, screen
+romsub $FF8A = RESTOR() clobbers(A,X,Y)                         ; restore default I/O vectors
+romsub $FF8D = VECTOR(uword userptr @ XY, bool dir @ Pc) clobbers(A,Y)     ; read/set I/O vector table
+romsub $FF90 = SETMSG(ubyte value @ A)                          ; set Kernal message control flag
+romsub $FF93 = SECOND(ubyte address @ A) clobbers(A)            ; (alias: LSTNSA) send secondary address after LISTEN
+romsub $FF96 = TKSA(ubyte address @ A) clobbers(A)              ; (alias: TALKSA) send secondary address after TALK
+romsub $FF99 = MEMTOP(uword address @ XY, bool dir @ Pc) -> uword @ XY     ; read/set top of memory  pointer
+romsub $FF9C = MEMBOT(uword address @ XY, bool dir @ Pc) -> uword @ XY     ; read/set bottom of memory  pointer
+romsub $FF9F = SCNKEY() clobbers(A,X,Y)                         ; scan the keyboard
+romsub $FFA2 = SETTMO(ubyte timeout @ A)                        ; set time-out flag for IEEE bus
+romsub $FFA5 = ACPTR() -> ubyte @ A                             ; (alias: IECIN) input byte from serial bus
+romsub $FFA8 = CIOUT(ubyte databyte @ A)                        ; (alias: IECOUT) output byte to serial bus
+romsub $FFAB = UNTLK() clobbers(A)                              ; command serial bus device to UNTALK
+romsub $FFAE = UNLSN() clobbers(A)                              ; command serial bus device to UNLISTEN
+romsub $FFB1 = LISTEN(ubyte device @ A) clobbers(A)             ; command serial bus device to LISTEN
+romsub $FFB4 = TALK(ubyte device @ A) clobbers(A)               ; command serial bus device to TALK
+romsub $FFB7 = READST() -> ubyte @ A                            ; read I/O status word
+romsub $FFBA = SETLFS(ubyte logical @ A, ubyte device @ X, ubyte secondary @ Y)   ; set logical file parameters
+romsub $FFBD = SETNAM(ubyte namelen @ A, str filename @ XY)     ; set filename parameters
+romsub $FFC0 = OPEN() clobbers(X,Y) -> bool @Pc, ubyte @A      ; (via 794 ($31A)) open a logical file
+romsub $FFC3 = CLOSE(ubyte logical @ A) clobbers(A,X,Y)         ; (via 796 ($31C)) close a logical file
+romsub $FFC6 = CHKIN(ubyte logical @ X) clobbers(A,X) -> bool @Pc    ; (via 798 ($31E)) define an input channel
+romsub $FFC9 = CHKOUT(ubyte logical @ X) clobbers(A,X)          ; (via 800 ($320)) define an output channel
+romsub $FFCC = CLRCHN() clobbers(A,X)                           ; (via 802 ($322)) restore default devices
+romsub $FFCF = CHRIN() clobbers(X, Y) -> ubyte @ A   ; (via 804 ($324)) input a character (for keyboard, read a whole line from the screen) A=byte read.
+romsub $FFD2 = CHROUT(ubyte char @ A)                           ; (via 806 ($326)) output a character
+romsub $FFD5 = LOAD(ubyte verify @ A, uword address @ XY) -> bool @Pc, ubyte @ A, uword @ XY     ; (via 816 ($330)) load from device
+romsub $FFD8 = SAVE(ubyte zp_startaddr @ A, uword endaddr @ XY) -> bool @ Pc, ubyte @ A          ; (via 818 ($332)) save to a device
+romsub $FFDB = SETTIM(ubyte low @ A, ubyte middle @ X, ubyte high @ Y)      ; set the software clock
+romsub $FFDE = RDTIM() -> ubyte @ A, ubyte @ X, ubyte @ Y       ; read the software clock (A=lo,X=mid,Y=high)
+romsub $FFE1 = STOP() clobbers(X) -> bool @ Pz, ubyte @ A      ; (via 808 ($328)) check the STOP key (and some others in A)
+romsub $FFE4 = GETIN() clobbers(X,Y) -> bool @Pc, ubyte @ A    ; (via 810 ($32A)) get a character
+romsub $FFE7 = CLALL() clobbers(A,X)                            ; (via 812 ($32C)) close all files
+romsub $FFEA = UDTIM() clobbers(A,X)                            ; update the software clock
+romsub $FFED = SCREEN() -> ubyte @ X, ubyte @ Y                 ; read number of screen rows and columns
+romsub $FFF0 = PLOT(ubyte col @ Y, ubyte row @ X, bool dir @ Pc) -> ubyte @ X, ubyte @ Y       ; read/set position of cursor on screen.  Use txt.plot for a 'safe' wrapper that preserves X.
+romsub $FFF3 = IOBASE() -> uword @ XY                           ; read base address of I/O devices
+
+asmsub STOP2() -> ubyte @A  {
+    ; -- check if STOP key was pressed, returns true if so.  More convenient to use than STOP() because that only sets the carry status flag.
+    %asm {{
+        txa
+        pha
+        jsr  cbm.STOP
+        beq  +
+        pla
+        tax
+        lda  #0
+        rts
++       pla
+        tax
+        lda  #1
+        rts
+    }}
+}
+
+asmsub RDTIM16() -> uword @AY {
+    ; --  like RDTIM() but only returning the lower 16 bits in AY for convenience
+    %asm {{
+        stx  P8ZP_SCRATCH_REG
+        jsr  cbm.RDTIM
+        pha
+        txa
+        tay
+        pla
+        ldx  P8ZP_SCRATCH_REG
+        rts
+    }}
+}
+
+}
+
+c64 {
+        ; C64 I/O registers (VIC, SID, CIA)
+
+    %option no_symbol_prefixing
+
         ; the default locations of the 8 sprite pointers (store address of sprite / 64)
         &ubyte  SPRPTR0         = 2040
         &ubyte  SPRPTR1         = 2041
@@ -198,93 +292,15 @@ c64 {
 
 ; ---- end of SID registers ----
 
-
-; ---- C64 ROM kernal routines ----
-
-romsub $AB1E = STROUT(uword strptr @ AY) clobbers(A, X, Y)      ; print null-terminated string (use txt.print instead)
-romsub $E544 = CLEARSCR() clobbers(A,X,Y)                       ; clear the screen
-romsub $E566 = HOMECRSR() clobbers(A,X,Y)                       ; cursor to top left of screen
-romsub $EA31 = IRQDFRT() clobbers(A,X,Y)                        ; default IRQ routine
-romsub $EA81 = IRQDFEND() clobbers(A,X,Y)                       ; default IRQ end/cleanup
-romsub $FF81 = CINT() clobbers(A,X,Y)                           ; (alias: SCINIT) initialize screen editor and video chip
-romsub $FF84 = IOINIT() clobbers(A, X)                          ; initialize I/O devices (CIA, SID, IRQ)
-romsub $FF87 = RAMTAS() clobbers(A,X,Y)                         ; initialize RAM, tape buffer, screen
-romsub $FF8A = RESTOR() clobbers(A,X,Y)                         ; restore default I/O vectors
-romsub $FF8D = VECTOR(uword userptr @ XY, ubyte dir @ Pc) clobbers(A,Y)     ; read/set I/O vector table
-romsub $FF90 = SETMSG(ubyte value @ A)                          ; set Kernal message control flag
-romsub $FF93 = SECOND(ubyte address @ A) clobbers(A)            ; (alias: LSTNSA) send secondary address after LISTEN
-romsub $FF96 = TKSA(ubyte address @ A) clobbers(A)              ; (alias: TALKSA) send secondary address after TALK
-romsub $FF99 = MEMTOP(uword address @ XY, ubyte dir @ Pc) -> uword @ XY     ; read/set top of memory  pointer
-romsub $FF9C = MEMBOT(uword address @ XY, ubyte dir @ Pc) -> uword @ XY     ; read/set bottom of memory  pointer
-romsub $FF9F = SCNKEY() clobbers(A,X,Y)                         ; scan the keyboard
-romsub $FFA2 = SETTMO(ubyte timeout @ A)                        ; set time-out flag for IEEE bus
-romsub $FFA5 = ACPTR() -> ubyte @ A                             ; (alias: IECIN) input byte from serial bus
-romsub $FFA8 = CIOUT(ubyte databyte @ A)                        ; (alias: IECOUT) output byte to serial bus
-romsub $FFAB = UNTLK() clobbers(A)                              ; command serial bus device to UNTALK
-romsub $FFAE = UNLSN() clobbers(A)                              ; command serial bus device to UNLISTEN
-romsub $FFB1 = LISTEN(ubyte device @ A) clobbers(A)             ; command serial bus device to LISTEN
-romsub $FFB4 = TALK(ubyte device @ A) clobbers(A)               ; command serial bus device to TALK
-romsub $FFB7 = READST() -> ubyte @ A                            ; read I/O status word
-romsub $FFBA = SETLFS(ubyte logical @ A, ubyte device @ X, ubyte secondary @ Y)   ; set logical file parameters
-romsub $FFBD = SETNAM(ubyte namelen @ A, str filename @ XY)     ; set filename parameters
-romsub $FFC0 = OPEN() clobbers(X,Y) -> ubyte @Pc, ubyte @A      ; (via 794 ($31A)) open a logical file
-romsub $FFC3 = CLOSE(ubyte logical @ A) clobbers(A,X,Y)         ; (via 796 ($31C)) close a logical file
-romsub $FFC6 = CHKIN(ubyte logical @ X) clobbers(A,X) -> ubyte @Pc    ; (via 798 ($31E)) define an input channel
-romsub $FFC9 = CHKOUT(ubyte logical @ X) clobbers(A,X)          ; (via 800 ($320)) define an output channel
-romsub $FFCC = CLRCHN() clobbers(A,X)                           ; (via 802 ($322)) restore default devices
-romsub $FFCF = CHRIN() clobbers(X, Y) -> ubyte @ A   ; (via 804 ($324)) input a character (for keyboard, read a whole line from the screen) A=byte read.
-romsub $FFD2 = CHROUT(ubyte char @ A)                           ; (via 806 ($326)) output a character
-romsub $FFD5 = LOAD(ubyte verify @ A, uword address @ XY) -> ubyte @Pc, ubyte @ A, uword @ XY     ; (via 816 ($330)) load from device
-romsub $FFD8 = SAVE(ubyte zp_startaddr @ A, uword endaddr @ XY) -> ubyte @ Pc, ubyte @ A          ; (via 818 ($332)) save to a device
-romsub $FFDB = SETTIM(ubyte low @ A, ubyte middle @ X, ubyte high @ Y)      ; set the software clock
-romsub $FFDE = RDTIM() -> ubyte @ A, ubyte @ X, ubyte @ Y       ; read the software clock (A=lo,X=mid,Y=high)
-romsub $FFE1 = STOP() clobbers(X) -> ubyte @ Pz, ubyte @ A      ; (via 808 ($328)) check the STOP key (and some others in A)
-romsub $FFE4 = GETIN() clobbers(X,Y) -> ubyte @Pc, ubyte @ A    ; (via 810 ($32A)) get a character
-romsub $FFE7 = CLALL() clobbers(A,X)                            ; (via 812 ($32C)) close all files
-romsub $FFEA = UDTIM() clobbers(A,X)                            ; update the software clock
-romsub $FFED = SCREEN() -> ubyte @ X, ubyte @ Y                 ; read number of screen rows and columns
-romsub $FFF0 = PLOT(ubyte col @ Y, ubyte row @ X, ubyte dir @ Pc) -> ubyte @ X, ubyte @ Y       ; read/set position of cursor on screen.  Use txt.plot for a 'safe' wrapper that preserves X.
-romsub $FFF3 = IOBASE() -> uword @ XY                           ; read base address of I/O devices
-
-; ---- end of C64 ROM kernal routines ----
-
-; ---- utilities -----
-
-asmsub STOP2() -> ubyte @A  {
-    ; -- check if STOP key was pressed, returns true if so.  More convenient to use than STOP() because that only sets the carry status flag.
-    %asm {{
-        txa
-        pha
-        jsr  c64.STOP
-        beq  +
-        pla
-        tax
-        lda  #0
-        rts
-+       pla
-        tax
-        lda  #1
-        rts
-    }}
 }
 
-asmsub RDTIM16() -> uword @AY {
-    ; --  like RDTIM() but only returning the lower 16 bits in AY for convenience
-    %asm {{
-        stx  P8ZP_SCRATCH_REG
-        jsr  c64.RDTIM
-        pha
-        txa
-        tay
-        pla
-        ldx  P8ZP_SCRATCH_REG
-        rts
-    }}
-}
+sys {
+    ; ------- lowlevel system routines --------
 
+    %option no_symbol_prefixing
 
+    const ubyte target = 64         ;  compilation target specifier.  64 = C64, 128 = C128, 16 = CommanderX16.
 
-; ---- C64 specific system utility routines: ----
 
 asmsub  init_system()  {
     ; Initializes the machine to a sane starting state.
@@ -300,13 +316,13 @@ asmsub  init_system()  {
         sta  $00
         lda  #%00100111
         sta  $01
-        jsr  c64.IOINIT
-        jsr  c64.RESTOR
-        jsr  c64.CINT
+        jsr  cbm.IOINIT
+        jsr  cbm.RESTOR
+        jsr  cbm.CINT
         lda  #6
         sta  c64.EXTCOL
         lda  #7
-        sta  c64.COLOR
+        sta  cbm.COLOR
         lda  #0
         sta  c64.BGCOL0
         jsr  disable_runstop_and_charsetswitch
@@ -326,7 +342,7 @@ asmsub  init_system_phase2()  {
 asmsub  cleanup_at_exit() {
     ; executed when the main subroutine does rts
     %asm {{
-        jmp  c64.enable_runstop_and_charsetswitch
+        jmp  sys.enable_runstop_and_charsetswitch
     }}
 }
 
@@ -350,18 +366,18 @@ asmsub  enable_runstop_and_charsetswitch() clobbers(A) {
     }}
 }
 
-asmsub  set_irq(uword handler @AY, ubyte useKernal @Pc) clobbers(A)  {
+asmsub  set_irq(uword handler @AY, bool useKernal @Pc) clobbers(A)  {
 	%asm {{
 	        sta  _modified+1
 	        sty  _modified+2
 	        lda  #0
-	        adc  #0
+	        rol  a
 	        sta  _use_kernal
 		sei
 		lda  #<_irq_handler
-		sta  c64.CINV
+		sta  cbm.CINV
 		lda  #>_irq_handler
-		sta  c64.CINV+1
+		sta  cbm.CINV+1
 		cli
 		rts
 _irq_handler    jsr  _irq_handler_init
@@ -379,7 +395,7 @@ _modified	jsr  $ffff                      ; modified
 		tax
 		pla
 		rti
-+		jmp  c64.IRQDFRT		; continue with normal kernal irq routine
++		jmp  cbm.IRQDFRT		; continue with normal kernal irq routine
 
 _use_kernal     .byte  0
 
@@ -431,10 +447,10 @@ IRQ_SCRATCH_ZPWORD2	.word  0
 asmsub  restore_irq() clobbers(A) {
 	%asm {{
 		sei
-		lda  #<c64.IRQDFRT
-		sta  c64.CINV
-		lda  #>c64.IRQDFRT
-		sta  c64.CINV+1
+		lda  #<cbm.IRQDFRT
+		sta  cbm.CINV
+		lda  #>cbm.IRQDFRT
+		sta  cbm.CINV+1
 		lda  #0
 		sta  c64.IREQMASK	; disable raster irq
 		lda  #%10000001
@@ -444,21 +460,21 @@ asmsub  restore_irq() clobbers(A) {
 	}}
 }
 
-asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0, ubyte useKernal @Pc) clobbers(A) {
+asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0, bool useKernal @Pc) clobbers(A) {
 	%asm {{
 	        sta  _modified+1
 	        sty  _modified+2
 	        lda  #0
-	        adc  #0
+	        rol  a
 	        sta  set_irq._use_kernal
 		lda  cx16.r0
 		ldy  cx16.r0+1
 		sei
 		jsr  _setup_raster_irq
 		lda  #<_raster_irq_handler
-		sta  c64.CINV
+		sta  cbm.CINV
 		lda  #>_raster_irq_handler
-		sta  c64.CINV+1
+		sta  cbm.CINV+1
 		cli
 		rts
 
@@ -466,8 +482,8 @@ _raster_irq_handler
 		jsr  set_irq._irq_handler_init
 _modified	jsr  $ffff              ; modified
 		jsr  set_irq._irq_handler_end
-                lda  #$ff
-                sta  c64.VICIRQ			; acknowledge raster irq
+        lda  #$ff
+        sta  c64.VICIRQ			; acknowledge raster irq
 		lda  set_irq._use_kernal
 		bne  +
 		; end irq processing - don't use kernal's irq handling
@@ -477,7 +493,7 @@ _modified	jsr  $ffff              ; modified
 		tax
 		pla
 		rti
-+		jmp  c64.IRQDFRT                ; continue with kernal irq routine
++		jmp  cbm.IRQDFRT                ; continue with kernal irq routine
 
 _setup_raster_irq
 		pha
@@ -501,23 +517,14 @@ _setup_raster_irq
 	}}
 }
 
-; ---- end of C64 specific system utility routines ----
 
-}
-
-sys {
-    ; ------- lowlevel system routines --------
-
-    const ubyte target = 64         ;  compilation target specifier.  64 = C64, 128 = C128, 16 = CommanderX16.
-
-
-    asmsub  reset_system()  {
+    asmsub reset_system()  {
         ; Soft-reset the system back to initial power-on Basic prompt.
         %asm {{
             sei
             lda  #14
             sta  $01        ; bank the kernal in
-            jmp  (c64.RESET_VEC)
+            jmp  (cbm.RESET_VEC)
         }}
     }
 
@@ -535,9 +542,9 @@ _loop       lda  P8ZP_SCRATCH_W1
             ldx  P8ZP_SCRATCH_B1
             rts
 
-+           lda  c64.TIME_LO
++           lda  cbm.TIME_LO
             sta  P8ZP_SCRATCH_B1
--           lda  c64.TIME_LO
+-           lda  cbm.TIME_LO
             cmp  P8ZP_SCRATCH_B1
             beq  -
 
@@ -694,13 +701,26 @@ _longcopy
         }}
     }
 
+    inline asmsub irqsafe_set_irqd() {
+        %asm {{
+        php
+        sei
+        }}
+    }
+
+    inline asmsub irqsafe_clear_irqd() {
+        %asm {{
+        plp
+        }}
+    }
+
     inline asmsub exit(ubyte returnvalue @A) {
         ; -- immediately exit the program with a return code in the A register
         %asm {{
             lda  #31
             sta  $01        ; bank the kernal in
-            jsr  c64.CLRCHN		; reset i/o channels
-            jsr  c64.enable_runstop_and_charsetswitch
+            jsr  cbm.CLRCHN		; reset i/o channels
+            jsr  sys.enable_runstop_and_charsetswitch
             ldx  prog8_lib.orig_stackpointer
             txs
             rts		; return to original caller
@@ -718,6 +738,8 @@ _longcopy
 
 cx16 {
 
+    %option no_symbol_prefixing
+
     ; the sixteen virtual 16-bit registers that the CX16 has defined in the zeropage
     ; they are simulated on the C64 as well but their location in memory is different
     ; (because there's no room for them in the zeropage)
@@ -825,4 +847,31 @@ cx16 {
     &byte r13sH = $cf1b
     &byte r14sH = $cf1d
     &byte r15sH = $cf1f
+
+    asmsub save_virtual_registers() clobbers(A,Y) {
+        %asm {{
+            ldy  #31
+    -       lda  cx16.r0,y
+            sta  _cx16_vreg_storage,y
+            dey
+            bpl  -
+            rts
+
+    _cx16_vreg_storage
+            .word 0,0,0,0,0,0,0,0
+            .word 0,0,0,0,0,0,0,0
+        }}
+    }
+
+    asmsub restore_virtual_registers() clobbers(A,Y) {
+        %asm {{
+            ldy  #31
+    -       lda  save_virtual_registers._cx16_vreg_storage,y
+            sta  cx16.r0,y
+            dey
+            bpl  -
+            rts
+        }}
+    }
+
 }

compiler/res/prog8lib/c64/textio.p8

@@ -6,6 +6,8 @@
 
 txt {
 
+    %option no_symbol_prefixing
+
 const ubyte DEFAULT_WIDTH = 40
 const ubyte DEFAULT_HEIGHT = 25
 
@@ -30,10 +32,10 @@ asmsub column(ubyte col @A) clobbers(A, X, Y) {
     ; ---- set the cursor on the given column (starting with 0) on the current line
     %asm {{
         sec
-        jsr  c64.PLOT
+        jsr  cbm.PLOT
         tay
         clc
-        jmp  c64.PLOT
+        jmp  cbm.PLOT
     }}
 }
 
@@ -57,10 +59,10 @@ asmsub  clear_screenchars (ubyte char @ A) clobbers(Y)  {
 	;      (assumes screen matrix is at the default address)
 	%asm {{
 		ldy  #250
--		sta  c64.Screen+250*0-1,y
-		sta  c64.Screen+250*1-1,y
-		sta  c64.Screen+250*2-1,y
-		sta  c64.Screen+250*3-1,y
+-		sta  cbm.Screen+250*0-1,y
+		sta  cbm.Screen+250*1-1,y
+		sta  cbm.Screen+250*2-1,y
+		sta  cbm.Screen+250*3-1,y
 		dey
 		bne  -
 		rts
@@ -72,10 +74,10 @@ asmsub  clear_screencolors (ubyte color @ A) clobbers(Y)  {
 	;      (assumes color matrix is at the default address)
 	%asm {{
 		ldy  #250
--		sta  c64.Colors+250*0-1,y
-		sta  c64.Colors+250*1-1,y
-		sta  c64.Colors+250*2-1,y
-		sta  c64.Colors+250*3-1,y
+-		sta  cbm.Colors+250*0-1,y
+		sta  cbm.Colors+250*1-1,y
+		sta  cbm.Colors+250*2-1,y
+		sta  cbm.Colors+250*3-1,y
 		dey
 		bne  -
 		rts
@@ -83,7 +85,7 @@ asmsub  clear_screencolors (ubyte color @ A) clobbers(Y)  {
 }
 
 sub color (ubyte txtcol) {
-    c64.COLOR = txtcol
+    cbm.COLOR = txtcol
 }
 
 sub lowercase() {
@@ -94,7 +96,7 @@ sub uppercase() {
     c64.VMCSB &= ~2
 }
 
-asmsub  scroll_left  (ubyte alsocolors @ Pc) clobbers(A, Y)  {
+asmsub  scroll_left  (bool alsocolors @ Pc) clobbers(A, Y)  {
 	; ---- scroll the whole screen 1 character to the left
 	;      contents of the rightmost column are unchanged, you should clear/refill this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -108,10 +110,10 @@ asmsub  scroll_left  (ubyte alsocolors @ Pc) clobbers(A, Y)  {
 		ldy  #38
 -
         .for row=0, row<=24, row+=1
-            lda  c64.Screen + 40*row + 1,x
-            sta  c64.Screen + 40*row + 0,x
-            lda  c64.Colors + 40*row + 1,x
-            sta  c64.Colors + 40*row + 0,x
+            lda  cbm.Screen + 40*row + 1,x
+            sta  cbm.Screen + 40*row + 0,x
+            lda  cbm.Colors + 40*row + 1,x
+            sta  cbm.Colors + 40*row + 0,x
         .next
 		inx
 		dey
@@ -123,8 +125,8 @@ _scroll_screen  ; scroll only the screen memory
 		ldy  #38
 -
         .for row=0, row<=24, row+=1
-            lda  c64.Screen + 40*row + 1,x
-            sta  c64.Screen + 40*row + 0,x
+            lda  cbm.Screen + 40*row + 1,x
+            sta  cbm.Screen + 40*row + 0,x
         .next
 		inx
 		dey
@@ -135,7 +137,7 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-asmsub  scroll_right  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_right  (bool alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character to the right
 	;      contents of the leftmost column are unchanged, you should clear/refill this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -147,10 +149,10 @@ asmsub  scroll_right  (ubyte alsocolors @ Pc) clobbers(A)  {
 		ldx  #38
 -
         .for row=0, row<=24, row+=1
-            lda  c64.Screen + 40*row + 0,x
-            sta  c64.Screen + 40*row + 1,x
-            lda  c64.Colors + 40*row + 0,x
-            sta  c64.Colors + 40*row + 1,x
+            lda  cbm.Screen + 40*row + 0,x
+            sta  cbm.Screen + 40*row + 1,x
+            lda  cbm.Colors + 40*row + 0,x
+            sta  cbm.Colors + 40*row + 1,x
         .next
 		dex
 		bpl  -
@@ -160,8 +162,8 @@ _scroll_screen  ; scroll only the screen memory
 		ldx  #38
 -
         .for row=0, row<=24, row+=1
-            lda  c64.Screen + 40*row + 0,x
-            sta  c64.Screen + 40*row + 1,x
+            lda  cbm.Screen + 40*row + 0,x
+            sta  cbm.Screen + 40*row + 1,x
         .next
 		dex
 		bpl  -
@@ -171,7 +173,7 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-asmsub  scroll_up  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_up  (bool alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character up
 	;      contents of the bottom row are unchanged, you should refill/clear this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -183,10 +185,10 @@ asmsub  scroll_up  (ubyte alsocolors @ Pc) clobbers(A)  {
 		ldx #39
 -
         .for row=1, row<=24, row+=1
-            lda  c64.Screen + 40*row,x
-            sta  c64.Screen + 40*(row-1),x
-            lda  c64.Colors + 40*row,x
-            sta  c64.Colors + 40*(row-1),x
+            lda  cbm.Screen + 40*row,x
+            sta  cbm.Screen + 40*(row-1),x
+            lda  cbm.Colors + 40*row,x
+            sta  cbm.Colors + 40*(row-1),x
         .next
 		dex
 		bpl  -
@@ -196,8 +198,8 @@ _scroll_screen  ; scroll only the screen memory
 		ldx #39
 -
         .for row=1, row<=24, row+=1
-            lda  c64.Screen + 40*row,x
-            sta  c64.Screen + 40*(row-1),x
+            lda  cbm.Screen + 40*row,x
+            sta  cbm.Screen + 40*(row-1),x
         .next
 		dex
 		bpl  -
@@ -207,7 +209,7 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-asmsub  scroll_down  (ubyte alsocolors @ Pc) clobbers(A)  {
+asmsub  scroll_down  (bool alsocolors @ Pc) clobbers(A)  {
 	; ---- scroll the whole screen 1 character down
 	;      contents of the top row are unchanged, you should refill/clear this yourself
 	;      Carry flag determines if screen color data must be scrolled too
@@ -219,10 +221,10 @@ asmsub  scroll_down  (ubyte alsocolors @ Pc) clobbers(A)  {
 		ldx #39
 -
         .for row=23, row>=0, row-=1
-            lda  c64.Colors + 40*row,x
-            sta  c64.Colors + 40*(row+1),x
-            lda  c64.Screen + 40*row,x
-            sta  c64.Screen + 40*(row+1),x
+            lda  cbm.Colors + 40*row,x
+            sta  cbm.Colors + 40*(row+1),x
+            lda  cbm.Screen + 40*row,x
+            sta  cbm.Screen + 40*(row+1),x
         .next
 		dex
 		bpl  -
@@ -232,8 +234,8 @@ _scroll_screen  ; scroll only the screen memory
 		ldx #39
 -
         .for row=23, row>=0, row-=1
-            lda  c64.Screen + 40*row,x
-            sta  c64.Screen + 40*(row+1),x
+            lda  cbm.Screen + 40*row,x
+            sta  cbm.Screen + 40*(row+1),x
         .next
 		dex
 		bpl  -
@@ -243,20 +245,20 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use c64.CHROUT directly ofcourse.
+romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use cbm.CHROUT directly ofcourse.
 
 asmsub  print (str text @ AY) clobbers(A,Y)  {
 	; ---- print null terminated string from A/Y
 	; note: the compiler contains an optimization that will replace
 	;       a call to this subroutine with a string argument of just one char,
-	;       by just one call to c64.CHROUT of that single char.
+	;       by just one call to cbm.CHROUT of that single char.
 	%asm {{
 		sta  P8ZP_SCRATCH_B1
 		sty  P8ZP_SCRATCH_REG
 		ldy  #0
 -		lda  (P8ZP_SCRATCH_B1),y
 		beq  +
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		bne  -
 +		rts
@@ -270,11 +272,11 @@ asmsub  print_ub0  (ubyte value @ A) clobbers(A,Y)  {
 		jsr  conv.ubyte2decimal
 		pha
 		tya
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		pla
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		txa
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
@@ -290,16 +292,16 @@ _print_byte_digits
 		cpy  #'0'
 		beq  +
 		tya
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		pla
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		jmp  _ones
 +       pla
         cmp  #'0'
         beq  _ones
-        jsr  c64.CHROUT
+        jsr  cbm.CHROUT
 _ones   txa
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
@@ -313,45 +315,45 @@ asmsub  print_b  (byte value @ A) clobbers(A,Y)  {
 		cmp  #0
 		bpl  +
 		lda  #'-'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 +		pla
 		jsr  conv.byte2decimal
 		jmp  print_ub._print_byte_digits
 	}}
 }
 
-asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_ubhex  (ubyte value @ A, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in hex form (if Carry is set, a radix prefix '$' is printed as well)
 	%asm {{
 		stx  P8ZP_SCRATCH_REG
 		bcc  +
 		pha
 		lda  #'$'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		pla
 +		jsr  conv.ubyte2hex
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		tya
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
 
-asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_ubbin  (ubyte value @ A, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in binary form (if Carry is set, a radix prefix '%' is printed as well)
 	%asm {{
 		stx  P8ZP_SCRATCH_REG
 		sta  P8ZP_SCRATCH_B1
 		bcc  +
 		lda  #'%'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 +		ldy  #8
 -		lda  #'0'
 		asl  P8ZP_SCRATCH_B1
 		bcc  +
 		lda  #'1'
-+		jsr  c64.CHROUT
++		jsr  cbm.CHROUT
 		dey
 		bne  -
 		ldx  P8ZP_SCRATCH_REG
@@ -359,7 +361,7 @@ asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_uwbin  (uword value @ AY, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in binary form (if Carry is set, a radix prefix '%' is printed as well)
 	%asm {{
 		pha
@@ -371,7 +373,7 @@ asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_uwhex  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_uwhex  (uword value @ AY, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in hexadecimal form (4 digits)
 	;      (if Carry is set, a radix prefix '$' is printed as well)
 	%asm {{
@@ -392,7 +394,7 @@ asmsub  print_uw0  (uword value @ AY) clobbers(A,Y)  {
 		ldy  #0
 -		lda  conv.uword2decimal.decTenThousands,y
         beq  +
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		bne  -
 +		ldx  P8ZP_SCRATCH_REG
@@ -415,14 +417,14 @@ asmsub  print_uw  (uword value @ AY) clobbers(A,Y)  {
 		bne  -
 
 _gotdigit
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		lda  conv.uword2decimal.decTenThousands,y
 		bne  _gotdigit
 		rts
 _allzero
         lda  #'0'
-        jmp  c64.CHROUT
+        jmp  cbm.CHROUT
 	}}
 }
 
@@ -433,7 +435,7 @@ asmsub  print_w  (word value @ AY) clobbers(A,Y)  {
 		bpl  +
 		pha
 		lda  #'-'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		tya
 		eor  #255
 		tay
@@ -455,7 +457,7 @@ asmsub  input_chars  (uword buffer @ AY) clobbers(A) -> ubyte @ Y  {
 		sta  P8ZP_SCRATCH_W1
 		sty  P8ZP_SCRATCH_W1+1
 		ldy  #0				; char counter = 0
--		jsr  c64.CHRIN
+-		jsr  cbm.CHRIN
 		cmp  #$0d			; return (ascii 13) pressed?
 		beq  +				; yes, end.
 		sta  (P8ZP_SCRATCH_W1),y	; else store char in buffer
@@ -586,7 +588,7 @@ asmsub  plot  (ubyte col @ Y, ubyte row @ A) clobbers(A) {
 		stx  P8ZP_SCRATCH_REG
 		tax
 		clc
-		jsr  c64.PLOT
+		jsr  cbm.PLOT
 		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
@@ -595,7 +597,7 @@ asmsub  plot  (ubyte col @ Y, ubyte row @ A) clobbers(A) {
 asmsub width() clobbers(X,Y) -> ubyte @A {
     ; -- returns the text screen width (number of columns)
     %asm {{
-        jsr  c64.SCREEN
+        jsr  cbm.SCREEN
         txa
         rts
     }}
@@ -604,7 +606,7 @@ asmsub width() clobbers(X,Y) -> ubyte @A {
 asmsub height() clobbers(X, Y) -> ubyte @A {
     ; -- returns the text screen height (number of rows)
     %asm {{
-        jsr  c64.SCREEN
+        jsr  cbm.SCREEN
         tya
         rts
     }}

compiler/res/prog8lib/conv.p8

@@ -2,6 +2,8 @@
 
 conv {
 
+    %option no_symbol_prefixing
+
 ; ----- number conversions to decimal strings ----
 
     str  @shared string_out = "????????????????"       ; result buffer for the string conversion routines
@@ -506,8 +508,7 @@ asmsub  ubyte2decimal  (ubyte value @A) -> ubyte @Y, ubyte @A, ubyte @X  {
 	; ---- A to decimal string in Y/A/X  (100s in Y, 10s in A, 1s in X)
 	%asm {{
 		ldy  #uword2decimal.ASCII_0_OFFSET
-		bne  uword2decimal.hex_try200
-		rts
+		jmp  uword2decimal.hex_try200
 	}}
 }
 

compiler/res/prog8lib/cx16/cx16diskio.p8

@@ -1,234 +0,0 @@
-; Cx16 specific disk drive I/O routines.
-
-%import diskio
-%import string
-
-cx16diskio {
-
-    ; Same as diskio.load() but with additional bank parameter to select the Ram bank to load into.
-    ; Use kernal LOAD routine to load the given program file in memory.
-    ; This is similar to Basic's  LOAD "filename",drive  /  LOAD "filename",drive,1
-    ; If you don't give an address_override, the location in memory is taken from the 2-byte file header.
-    ; If you specify a custom address_override, the first 2 bytes in the file are ignored
-    ; and the rest is loaded at the given location in memory.
-    ; Returns the end load address+1 if successful or 0 if a load error occurred.
-    ; You can use the load_size() function to calcuate the size of the file that was loaded.
-    sub load(ubyte drivenumber, uword filenameptr, ubyte bank, uword address_override) -> uword {
-        cx16.rambank(bank)
-        return diskio.internal_load_routine(drivenumber, filenameptr, address_override, false)
-    }
-
-    ; Same as diskio.load_raw() but with additional bank parameter to select the Ram bank to load into.
-    ; Use kernal LOAD routine to load the given file in memory.
-    ; INCLUDING the first 2 bytes in the file: no program header is assumed in the file.
-    ; The load address is mandatory. Returns the number of bytes loaded.
-    ; If you load into regular system ram, use cx16.getrambank() for the bank argument,
-    ; or alternatively make sure to reset the correct ram bank yourself after the load!
-    ; Returns the end load address+1 if successful or 0 if a load error occurred.
-    ; You can use the load_size() function to calcuate the size of the file that was loaded.
-    sub load_raw(ubyte drivenumber, uword filenameptr, ubyte bank, uword address_override) -> uword {
-        cx16.rambank(bank)
-        return diskio.internal_load_routine(drivenumber, filenameptr, address_override, true)
-    }
-
-    ; For use directly after a load or load_raw call (don't mess with the ram bank yet):
-    ; Calculates the number of bytes loaded (files > 64Kb ar truncated to 16 bits)
-    sub load_size(ubyte startbank, uword startaddress, uword endaddress) -> uword {
-        return $2000 * (cx16.getrambank() - startbank) + endaddress - startaddress
-    }
-
-    asmsub vload(str name @R0, ubyte drivenumber @Y, ubyte bank @A, uword address @R1) -> ubyte @A {
-        ; -- like the basic command VLOAD "filename",drivenumber,bank,address
-        ;    loads a file into Vera's video memory in the given bank:address, returns success in A
-        ;    the file has to have the usual 2 byte header (which will be skipped)
-        %asm {{
-            clc
-internal_vload:
-            phx
-            pha
-            tya
-            tax
-            bcc +
-            ldy  #%00000010     ; headerless load mode
-            bne  ++
-+           ldy  #0             ; normal load mode
-+           lda  #1
-            jsr  c64.SETLFS
-            lda  cx16.r0
-            ldy  cx16.r0+1
-            jsr  prog8_lib.strlen
-            tya
-            ldx  cx16.r0
-            ldy  cx16.r0+1
-            jsr  c64.SETNAM
-            pla
-            clc
-            adc  #2
-            ldx  cx16.r1
-            ldy  cx16.r1+1
-            stz  P8ZP_SCRATCH_B1
-            jsr  c64.LOAD
-            bcs  +
-            inc  P8ZP_SCRATCH_B1
-    +       jsr  c64.CLRCHN
-            lda  #1
-            jsr  c64.CLOSE
-            plx
-            lda  P8ZP_SCRATCH_B1
-            rts
-        }}
-    }
-
-    asmsub vload_raw(str name @R0, ubyte drivenumber @Y, ubyte bank @A, uword address @R1) -> ubyte @A {
-        ; -- like the basic command BVLOAD "filename",drivenumber,bank,address
-        ;    loads a file into Vera's video memory in the given bank:address, returns success in A
-        ;    the file is read fully including the first two bytes.
-        %asm {{
-            sec
-            jmp  vload.internal_vload
-        }}
-    }
-
-    ; Replacement function that makes use of fast block read capability of the X16,
-    ; and can wrap over multiple ram banks while reading.
-    ; Use this in place of regular diskio.f_read() on X16.
-    sub f_read(uword bufferpointer, uword num_bytes) -> uword {
-        ; -- read from the currently open file, up to the given number of bytes.
-        ;    returns the actual number of bytes read.  (checks for End-of-file and error conditions)
-        if not diskio.iteration_in_progress or not num_bytes
-            return 0
-
-        diskio.list_blocks = 0     ; we reuse this variable for the total number of bytes read
-
-        ; commander X16 supports fast block-read via macptr() kernal call
-        uword size
-        while num_bytes {
-            size = 255
-            if num_bytes<size
-                size = num_bytes
-            size = cx16.macptr(lsb(size), bufferpointer, false)
-            if_cs
-                goto byte_read_loop     ; macptr block read not supported, do fallback loop
-            diskio.list_blocks += size
-            bufferpointer += size
-            if msb(bufferpointer) == $c0
-                bufferpointer = mkword($a0, lsb(bufferpointer))  ; wrap over bank boundary
-            num_bytes -= size
-            if c64.READST() & $40 {
-                diskio.f_close()       ; end of file, close it
-                break
-            }
-        }
-        return diskio.list_blocks  ; number of bytes read
-
-byte_read_loop:         ; fallback if macptr() isn't supported on the device
-        %asm {{
-            lda  bufferpointer
-            sta  m_in_buffer+1
-            lda  bufferpointer+1
-            sta  m_in_buffer+2
-        }}
-        while num_bytes {
-            if c64.READST() {
-                diskio.f_close()
-                if c64.READST() & $40    ; eof?
-                    return diskio.list_blocks   ; number of bytes read
-                return 0  ; error.
-            }
-            %asm {{
-                jsr  c64.CHRIN
-m_in_buffer     sta  $ffff
-                inc  m_in_buffer+1
-                bne  +
-                inc  m_in_buffer+2
-+
-            }}
-            diskio.list_blocks++
-            num_bytes--
-        }
-        return diskio.list_blocks  ; number of bytes read
-    }
-
-    ; replacement function that makes use of fast block read capability of the X16
-    ; use this in place of regular diskio.f_read_all() on X16
-    sub f_read_all(uword bufferpointer) -> uword {
-        ; -- read the full contents of the file, returns number of bytes read.
-        if not diskio.iteration_in_progress
-            return 0
-
-        uword total_read = 0
-        while not c64.READST() {
-            cx16.r0 = cx16diskio.f_read(bufferpointer, 256)
-            total_read += cx16.r0
-            bufferpointer += cx16.r0
-        }
-        return total_read
-    }
-
-
-    sub chdir(ubyte drivenumber, str path) {
-        ; -- change current directory.
-        diskio.list_filename[0] = 'c'
-        diskio.list_filename[1] = 'd'
-        diskio.list_filename[2] = ':'
-        void string.copy(path, &diskio.list_filename+3)
-        diskio.send_command(drivenumber, diskio.list_filename)
-    }
-
-    sub mkdir(ubyte drivenumber, str name) {
-        ; -- make a new subdirectory.
-        diskio.list_filename[0] = 'm'
-        diskio.list_filename[1] = 'd'
-        diskio.list_filename[2] = ':'
-        void string.copy(name, &diskio.list_filename+3)
-        diskio.send_command(drivenumber, diskio.list_filename)
-    }
-
-    sub rmdir(ubyte drivenumber, str name) {
-        ; -- remove a subdirectory.
-        void string.find(name, '*')
-        if_cs
-            return    ; refuse to act on a wildcard *
-        diskio.list_filename[0] = 'r'
-        diskio.list_filename[1] = 'd'
-        diskio.list_filename[2] = ':'
-        void string.copy(name, &diskio.list_filename+3)
-        diskio.send_command(drivenumber, diskio.list_filename)
-    }
-
-    sub relabel(ubyte drivenumber, str name) {
-        ; -- change the disk label.
-        diskio.list_filename[0] = 'r'
-        diskio.list_filename[1] = '-'
-        diskio.list_filename[2] = 'h'
-        diskio.list_filename[3] = ':'
-        void string.copy(name, &diskio.list_filename+4)
-        diskio.send_command(drivenumber, diskio.list_filename)
-    }
-
-    sub f_seek(uword pos_hiword, uword pos_loword) {
-        ; -- seek in the reading file opened with f_open, to the given 32-bits position
-        ubyte[6] command = ['p',0,0,0,0,0]
-        command[1] = 12       ; f_open uses channel 12
-        command[2] = lsb(pos_loword)
-        command[3] = msb(pos_loword)
-        command[4] = lsb(pos_hiword)
-        command[5] = msb(pos_hiword)
-    send_command:
-        c64.SETNAM(sizeof(command), &command)
-        c64.SETLFS(15, diskio.last_drivenumber, 15)
-        void c64.OPEN()
-        c64.CLOSE(15)
-    }
-
-    ; TODO see if we can get this to work as well:
-;    sub f_seek_w(uword pos_hiword, uword pos_loword) {
-;        ; -- seek in the output file opened with f_open_w, to the given 32-bits position
-;        cx16diskio.f_seek.command[1] = 13       ; f_open_w uses channel 13
-;        cx16diskio.f_seek.command[2] = lsb(pos_loword)
-;        cx16diskio.f_seek.command[3] = msb(pos_loword)
-;        cx16diskio.f_seek.command[4] = lsb(pos_hiword)
-;        cx16diskio.f_seek.command[5] = msb(pos_hiword)
-;        goto cx16diskio.f_seek.send_command
-;    }
-}

compiler/res/prog8lib/cx16/diskio.p8

@@ -0,0 +1,797 @@
+; Commander X16 disk drive I/O routines.
+; Largely compatible with the default C64 ones, but adds more stuff specific to the X16 as well.
+
+%import textio
+%import string
+%import syslib
+
+diskio {
+    %option no_symbol_prefixing
+
+    ubyte drivenumber = 8
+
+    sub set_drive(ubyte number) {
+        drivenumber = number
+    }
+
+    sub directory() -> bool {
+        ; -- Prints the directory contents to the screen. Returns success.
+
+        cbm.SETNAM(1, "$")
+        cbm.SETLFS(12, drivenumber, 0)
+        ubyte status = 1
+        void cbm.OPEN()          ; open 12,8,0,"$"
+        if_cs
+            goto io_error
+        void cbm.CHKIN(12)        ; use #12 as input channel
+        if_cs
+            goto io_error
+
+        repeat 4 {
+            void cbm.CHRIN()     ; skip the 4 prologue bytes
+        }
+
+        ; while not stop key pressed / EOF encountered, read data.
+        status = cbm.READST()
+        if status!=0 {
+            status = 1
+            goto io_error
+        }
+
+        while status==0 {
+            ubyte low = cbm.CHRIN()
+            ubyte high = cbm.CHRIN()
+            txt.print_uw(mkword(high, low))
+            txt.spc()
+            ubyte @zp char
+            repeat {
+                char = cbm.CHRIN()
+                if char==0
+                    break
+                txt.chrout(char)
+            }
+            txt.nl()
+            void cbm.CHRIN()     ; skip 2 bytes
+            void cbm.CHRIN()
+            status = cbm.READST()
+            if cbm.STOP2()
+                break
+        }
+        status = cbm.READST()
+
+io_error:
+        cbm.CLRCHN()        ; restore default i/o devices
+        cbm.CLOSE(12)
+
+        if status and status & $40 == 0 {            ; bit 6=end of file
+            txt.print("\ni/o error, status: ")
+            txt.print_ub(status)
+            txt.nl()
+            return false
+        }
+
+        return true
+    }
+
+    sub diskname() -> uword {
+        ; returns disk label name or 0 if error
+        cbm.SETNAM(3, "$")
+        cbm.SETLFS(12, diskio.drivenumber, 0)
+        ubyte status = 1
+        void cbm.OPEN()          ; open 12,8,0,"$=c"
+        if_cs
+            goto io_error
+        void cbm.CHKIN(12)        ; use #12 as input channel
+        if_cs
+            goto io_error
+
+        while cbm.CHRIN()!='"' {
+            ; skip up to entry name
+        }
+
+        cx16.r0 = &diskio.list_filename
+        repeat {
+            @(cx16.r0) = cbm.CHRIN()
+            if @(cx16.r0)=='"' {
+                @(cx16.r0) = ' '
+                while @(cx16.r0)==' ' and cx16.r0>=&diskio.list_filename {
+                    @(cx16.r0) = 0
+                    cx16.r0--
+                }
+                break
+            }
+            cx16.r0++
+        }
+        status = cbm.READST()
+
+io_error:
+        cbm.CLRCHN()
+        cbm.CLOSE(12)
+        if status and status & $40 == 0
+            return 0
+        return diskio.list_filename
+    }
+
+    ; internal variables for the iterative file lister / loader
+    bool list_skip_disk_name
+    uword list_pattern
+    uword list_blocks
+    bool iteration_in_progress = false
+    str list_filetype = "???"       ; prg, seq, dir
+    str list_filename = "?" * 50
+
+    ; ----- get a list of files (uses iteration functions internally) -----
+
+    sub list_filenames(uword pattern_ptr, uword filenames_buffer, uword filenames_buf_size) -> ubyte {
+        ; -- fill the provided buffer with the names of the files on the disk (until buffer is full).
+        ;    Files in the buffer are separeted by a 0 byte. You can provide an optional pattern to match against.
+        ;    After the last filename one additional 0 byte is placed to indicate the end of the list.
+        ;    Returns number of files (it skips 'dir' entries i.e. subdirectories).
+        ;    Also sets carry on exit: Carry clear = all files returned, Carry set = directory has more files that didn't fit in the buffer.
+        uword buffer_start = filenames_buffer
+        ubyte files_found = 0
+        if lf_start_list(pattern_ptr) {
+            while lf_next_entry() {
+                if list_filetype!="dir" {
+                    filenames_buffer += string.copy(list_filename, filenames_buffer) + 1
+                    files_found++
+                    if filenames_buffer - buffer_start > filenames_buf_size-20 {
+                        @(filenames_buffer)=0
+                        lf_end_list()
+                        sys.set_carry()
+                        return files_found
+                    }
+                }
+            }
+            lf_end_list()
+        }
+        @(filenames_buffer)=0
+        sys.clear_carry()
+        return files_found
+    }
+
+    ; ----- iterative file lister functions (uses io channel 12) -----
+
+    sub lf_start_list(uword pattern_ptr) -> bool {
+        ; -- start an iterative file listing with optional pattern matching.
+        ;    note: only a single iteration loop can be active at a time!
+        lf_end_list()
+        list_pattern = pattern_ptr
+        list_skip_disk_name = true
+        iteration_in_progress = true
+
+        cbm.SETNAM(1, "$")
+        cbm.SETLFS(12, drivenumber, 0)
+        void cbm.OPEN()          ; open 12,8,0,"$"
+        if_cs
+            goto io_error
+        void cbm.CHKIN(12)        ; use #12 as input channel
+        if_cs
+            goto io_error
+
+        repeat 4 {
+            void cbm.CHRIN()     ; skip the 4 prologue bytes
+        }
+
+        if cbm.READST()==0
+            return true
+
+io_error:
+        lf_end_list()
+        return false
+    }
+
+    sub lf_next_entry() -> bool {
+        ; -- retrieve the next entry from an iterative file listing session.
+        ;    results will be found in list_blocks, list_filename, and list_filetype.
+        ;    if it returns false though, there are no more entries (or an error occurred).
+
+        if not iteration_in_progress
+            return false
+
+        repeat {
+            void cbm.CHKIN(12)        ; use #12 as input channel again
+
+            uword nameptr = &list_filename
+            ubyte blocks_lsb = cbm.CHRIN()
+            ubyte blocks_msb = cbm.CHRIN()
+
+            if cbm.READST()
+                goto close_end
+
+            list_blocks = mkword(blocks_msb, blocks_lsb)
+
+            ; read until the filename starts after the first "
+            while cbm.CHRIN()!='\"'  {
+                if cbm.READST()
+                    goto close_end
+            }
+
+            ; read the filename
+            repeat {
+                ubyte char = cbm.CHRIN()
+                if char==0
+                    break
+                if char=='\"'
+                    break
+                @(nameptr) = char
+                nameptr++
+            }
+
+            @(nameptr) = 0
+
+            do {
+                cx16.r15L = cbm.CHRIN()
+            } until cx16.r15L!=' '      ; skip blanks up to 3 chars entry type
+            list_filetype[0] = cx16.r15L
+            list_filetype[1] = cbm.CHRIN()
+            list_filetype[2] = cbm.CHRIN()
+            while cbm.CHRIN() {
+                ; read the rest of the entry until the end
+            }
+
+            void cbm.CHRIN()     ; skip 2 bytes
+            void cbm.CHRIN()
+
+            if not list_skip_disk_name {
+                if not list_pattern
+                    return true
+                if string.pattern_match(list_filename, list_pattern)
+                    return true
+            }
+            list_skip_disk_name = false
+        }
+
+close_end:
+        lf_end_list()
+        return false
+    }
+
+    sub lf_end_list() {
+        ; -- end an iterative file listing session (close channels).
+        if iteration_in_progress {
+            cbm.CLRCHN()
+            cbm.CLOSE(12)
+            iteration_in_progress = false
+        }
+    }
+
+
+    ; ----- iterative file loader functions (uses io channel 12) -----
+
+    sub f_open(uword filenameptr) -> bool {
+        ; -- open a file for iterative reading with f_read
+        ;    note: only a single iteration loop can be active at a time!
+        f_close()
+
+        cbm.SETNAM(string.length(filenameptr), filenameptr)
+        cbm.SETLFS(12, drivenumber, 12)     ; note: has to be 12,x,12 because otherwise f_seek doesn't work
+        void cbm.OPEN()          ; open 12,8,12,"filename"
+        if_cc {
+            if cbm.READST()==0 {
+                iteration_in_progress = true
+                void cbm.CHKIN(12)          ; use #12 as input channel
+                if_cc {
+                    void cbm.CHRIN()        ; read first byte to test for file not found
+                    if not cbm.READST() {
+                        cbm.CLOSE(12)           ; close file because we already consumed first byte
+                        void cbm.OPEN()         ; re-open the file
+                        void cbm.CHKIN(12)
+                        return true
+                    }
+                }
+            }
+        }
+        f_close()
+        return false
+    }
+
+    ; optimized for Commander X16 to use MACPTR block read kernal call
+    sub f_read(uword bufferpointer, uword num_bytes) -> uword {
+        ; -- read from the currently open file, up to the given number of bytes.
+        ;    returns the actual number of bytes read.  (checks for End-of-file and error conditions)
+        if not iteration_in_progress or not num_bytes
+            return 0
+
+        list_blocks = 0     ; we reuse this variable for the total number of bytes read
+
+        ; commander X16 supports fast block-read via macptr() kernal call
+        uword size
+        while num_bytes {
+            size = 255
+            if num_bytes<size
+                size = num_bytes
+            size = cx16.macptr(lsb(size), bufferpointer, false)
+            if_cs
+                goto byte_read_loop     ; macptr block read not supported, do fallback loop
+            list_blocks += size
+            bufferpointer += size
+            if msb(bufferpointer) == $c0
+                bufferpointer = mkword($a0, lsb(bufferpointer))  ; wrap over bank boundary
+            num_bytes -= size
+            if cbm.READST() & $40 {
+                f_close()       ; end of file, close it
+                break
+            }
+        }
+        return list_blocks  ; number of bytes read
+
+byte_read_loop:         ; fallback if macptr() isn't supported on the device
+        %asm {{
+            lda  bufferpointer
+            sta  m_in_buffer+1
+            lda  bufferpointer+1
+            sta  m_in_buffer+2
+        }}
+        while num_bytes {
+            if cbm.READST() {
+                f_close()
+                if cbm.READST() & $40    ; eof?
+                    return list_blocks   ; number of bytes read
+                return 0  ; error.
+            }
+            %asm {{
+                jsr  cbm.CHRIN
+m_in_buffer     sta  $ffff
+                inc  m_in_buffer+1
+                bne  +
+                inc  m_in_buffer+2
++
+            }}
+            list_blocks++
+            num_bytes--
+        }
+        return list_blocks  ; number of bytes read
+    }
+
+    ; optimized for Commander X16 to use MACPTR block read kernal call
+    sub f_read_all(uword bufferpointer) -> uword {
+        ; -- read the full contents of the file, returns number of bytes read.
+        if not iteration_in_progress
+            return 0
+
+        uword total_read = 0
+        while not cbm.READST() {
+            cx16.r0 = f_read(bufferpointer, 256)
+            total_read += cx16.r0
+            bufferpointer += cx16.r0
+        }
+        return total_read
+    }
+
+    asmsub f_readline(uword bufptr @AY) clobbers(X) -> ubyte @Y {
+        ; Routine to read text lines from a text file. Lines must be less than 255 characters.
+        ; Reads characters from the input file UNTIL a newline or return character (or EOF).
+        ; The line read will be 0-terminated in the buffer (and not contain the end of line character).
+        ; The length of the line is returned in Y. Note that an empty line is okay and is length 0!
+        ; I/O error status should be checked by the caller itself via READST() routine.
+        %asm {{
+            sta  P8ZP_SCRATCH_W1
+            sty  P8ZP_SCRATCH_W1+1
+            ldx  #12
+            jsr  cbm.CHKIN              ; use channel 12 again for input
+            ldy  #0
+_loop       jsr  cbm.CHRIN
+            sta  (P8ZP_SCRATCH_W1),y
+            beq  _end
+            iny
+            cmp  #$0a
+            beq  _line_end
+            cmp  #$0d
+            bne  _loop
+_line_end   dey     ; get rid of the trailing end-of-line char
+            lda  #0
+            sta  (P8ZP_SCRATCH_W1),y
+_end        rts
+        }}
+    }
+
+    sub f_close() {
+        ; -- end an iterative file loading session (close channels).
+        if iteration_in_progress {
+            cbm.CLRCHN()
+            cbm.CLOSE(12)
+            iteration_in_progress = false
+        }
+    }
+
+
+    ; ----- iterative file writing functions (uses io channel 13) -----
+
+    sub f_open_w(uword filenameptr) -> bool {
+        ; -- open a file for iterative writing with f_write
+        f_close_w()
+
+        cbm.SETNAM(string.length(filenameptr), filenameptr)
+        cbm.SETLFS(13, drivenumber, 1)
+        void cbm.OPEN()             ; open 13,8,1,"filename"
+        if_cc {
+            cbm.CHKOUT(13)          ; use #13 as output channel
+            return not cbm.READST()
+        }
+        f_close_w()
+        return false
+    }
+
+    sub f_write(uword bufferpointer, uword num_bytes) -> bool {
+        ; -- write the given number of bytes to the currently open file
+        if num_bytes!=0 {
+            cbm.CHKOUT(13)        ; use #13 as output channel again
+            repeat num_bytes {
+                cbm.CHROUT(@(bufferpointer))
+                bufferpointer++
+            }
+            return not cbm.READST()
+        }
+        return true
+    }
+
+    sub f_close_w() {
+        ; -- end an iterative file writing session (close channels).
+        cbm.CLRCHN()
+        cbm.CLOSE(13)
+    }
+
+
+    ; ---- other functions ----
+
+    sub status() -> uword {
+        ; -- retrieve the disk drive's current status message
+        uword messageptr = &list_filename
+        cbm.SETNAM(0, list_filename)
+        cbm.SETLFS(15, drivenumber, 15)
+        void cbm.OPEN()          ; open 15,8,15
+        if_cs
+            goto io_error
+        void cbm.CHKIN(15)        ; use #15 as input channel
+        if_cs
+            goto io_error
+
+        while not cbm.READST() {
+            cx16.r5L = cbm.CHRIN()
+            if cx16.r5L=='\r' or cx16.r5L=='\n'
+                break
+            @(messageptr) = cx16.r5L
+            messageptr++
+        }
+        @(messageptr) = 0
+
+done:
+        cbm.CLRCHN()        ; restore default i/o devices
+        cbm.CLOSE(15)
+        return list_filename
+
+io_error:
+        list_filename = "?disk error"
+        goto done
+    }
+
+
+    ; saves a block of memory to disk, including the default 2 byte prg header.
+    sub save(uword filenameptr, uword address, uword size) -> bool {
+        return internal_save_routine(filenameptr, address, size, false)
+    }
+
+    ; like save() but omits the 2 byte prg header.
+    sub save_raw(uword filenameptr, uword address, uword size) -> bool {
+        return internal_save_routine(filenameptr, address, size, true)
+    }
+
+    sub internal_save_routine(uword filenameptr, uword address, uword size, bool headerless) -> bool {
+        cbm.SETNAM(string.length(filenameptr), filenameptr)
+        cbm.SETLFS(1, drivenumber, 0)
+        uword @shared end_address = address + size
+        cx16.r0L = 0
+
+        %asm {{
+            lda  address
+            sta  P8ZP_SCRATCH_W1
+            lda  address+1
+            sta  P8ZP_SCRATCH_W1+1
+            stx  P8ZP_SCRATCH_REG
+            ldx  end_address
+            ldy  end_address+1
+            lda  headerless
+            beq  +
+            lda  #<P8ZP_SCRATCH_W1
+            jsr  cx16.BSAVE
+            bra  ++
++           lda  #<P8ZP_SCRATCH_W1
+            jsr  cbm.SAVE
++           php
+            ldx  P8ZP_SCRATCH_REG
+            plp
+        }}
+
+        if_cc
+            cx16.r0L = cbm.READST()==0
+
+        cbm.CLRCHN()
+        cbm.CLOSE(1)
+
+        return cx16.r0L
+    }
+
+    ; Use kernal LOAD routine to load the given program file in memory.
+    ; This is similar to Basic's  LOAD "filename",drive  /  LOAD "filename",drive,1
+    ; If you don't give an address_override, the location in memory is taken from the 2-byte file header.
+    ; If you specify a custom address_override, the first 2 bytes in the file are ignored
+    ; and the rest is loaded at the given location in memory.
+    ; Returns the end load address+1 if successful or 0 if a load error occurred.
+    ; NOTE: when the load is larger than 64Kb and/or spans multiple RAM banks
+    ;       (which is possible on the Commander X16), the returned size is not correct,
+    ;       because it doesn't take the number of ram banks into account.
+    ;       You can use the load_size() function to calcuate the size in this case.
+    ; NOTE: data is read into the current Ram bank if you're reading into banked ram.
+    ;       if you require loading into another ram bank, you have to set that
+    ;       yourself using cx16.rambank(bank) before calling load().
+    sub load(uword filenameptr, uword address_override) -> uword {
+        return internal_load_routine(filenameptr, address_override, false)
+    }
+
+    ; Identical to load(), but DOES INCLUDE the first 2 bytes in the file.
+    ; No program header is assumed in the file. Everything is loaded.
+    ; See comments on load() for more details.
+    sub load_raw(uword filenameptr, uword address) -> uword {
+        return internal_load_routine(filenameptr, address, true)
+    }
+
+
+    sub internal_load_routine(uword filenameptr, uword address_override, bool headerless) -> uword {
+        cbm.SETNAM(string.length(filenameptr), filenameptr)
+        ubyte secondary = 1
+        cx16.r1 = 0
+        if address_override
+            secondary = 0
+        if headerless
+            secondary |= %00000010  ; activate cx16 kernal headerless load support
+        cbm.SETLFS(1, drivenumber, secondary)
+        %asm {{
+            stx  P8ZP_SCRATCH_REG
+            lda  #0
+            ldx  address_override
+            ldy  address_override+1
+            jsr  cbm.LOAD
+            bcs  +
+            stx  cx16.r1
+            sty  cx16.r1+1
++           ldx  P8ZP_SCRATCH_REG
+        }}
+
+        cbm.CLRCHN()
+        cbm.CLOSE(1)
+        return cx16.r1
+    }
+
+    sub delete(uword filenameptr) {
+        ; -- delete a file on the drive
+        list_filename[0] = 's'
+        list_filename[1] = ':'
+        ubyte flen = string.copy(filenameptr, &list_filename+2)
+        cbm.SETNAM(flen+2, list_filename)
+        cbm.SETLFS(1, drivenumber, 15)
+        void cbm.OPEN()
+        cbm.CLRCHN()
+        cbm.CLOSE(1)
+    }
+
+    sub rename(uword oldfileptr, uword newfileptr) {
+        ; -- rename a file on the drive
+        list_filename[0] = 'r'
+        list_filename[1] = ':'
+        ubyte flen_new = string.copy(newfileptr, &list_filename+2)
+        list_filename[flen_new+2] = '='
+        ubyte flen_old = string.copy(oldfileptr, &list_filename+3+flen_new)
+        cbm.SETNAM(3+flen_new+flen_old, list_filename)
+        cbm.SETLFS(1, drivenumber, 15)
+        void cbm.OPEN()
+        cbm.CLRCHN()
+        cbm.CLOSE(1)
+    }
+
+    sub send_command(uword commandptr) {
+        ; -- send a dos command to the drive
+        cbm.SETNAM(string.length(commandptr), commandptr)
+        cbm.SETLFS(15, drivenumber, 15)
+        void cbm.OPEN()
+        cbm.CLRCHN()
+        cbm.CLOSE(15)
+    }
+
+
+    ; CommanderX16 extensions over the basic C64/C128 diskio routines:
+
+    ; For use directly after a load or load_raw call (don't mess with the ram bank yet):
+    ; Calculates the number of bytes loaded (files > 64Kb ar truncated to 16 bits)
+    sub load_size(ubyte startbank, uword startaddress, uword endaddress) -> uword {
+        return $2000 * (cx16.getrambank() - startbank) + endaddress - startaddress
+    }
+
+    asmsub vload(str name @R0, ubyte bank @A, uword address @R1) -> ubyte @A {
+        ; -- like the basic command VLOAD "filename",drivenumber,bank,address
+        ;    loads a file into Vera's video memory in the given bank:address, returns success in A
+        ;    the file has to have the usual 2 byte header (which will be skipped)
+        %asm {{
+            clc
+internal_vload:
+            phx
+            pha
+            ldx  drivenumber
+            bcc +
+            ldy  #%00000010     ; headerless load mode
+            bne  ++
++           ldy  #0             ; normal load mode
++           lda  #1
+            jsr  cbm.SETLFS
+            lda  cx16.r0
+            ldy  cx16.r0+1
+            jsr  prog8_lib.strlen
+            tya
+            ldx  cx16.r0
+            ldy  cx16.r0+1
+            jsr  cbm.SETNAM
+            pla
+            clc
+            adc  #2
+            ldx  cx16.r1
+            ldy  cx16.r1+1
+            stz  P8ZP_SCRATCH_B1
+            jsr  cbm.LOAD
+            bcs  +
+            inc  P8ZP_SCRATCH_B1
+    +       jsr  cbm.CLRCHN
+            lda  #1
+            jsr  cbm.CLOSE
+            plx
+            lda  P8ZP_SCRATCH_B1
+            rts
+        }}
+    }
+
+    asmsub vload_raw(str name @R0, ubyte bank @A, uword address @R1) -> ubyte @A {
+        ; -- like the basic command BVLOAD "filename",drivenumber,bank,address
+        ;    loads a file into Vera's video memory in the given bank:address, returns success in A
+        ;    the file is read fully including the first two bytes.
+        %asm {{
+            sec
+            jmp  vload.internal_vload
+        }}
+    }
+
+    sub chdir(str path) {
+        ; -- change current directory.
+        list_filename[0] = 'c'
+        list_filename[1] = 'd'
+        list_filename[2] = ':'
+        void string.copy(path, &list_filename+3)
+        send_command(list_filename)
+    }
+
+    sub mkdir(str name) {
+        ; -- make a new subdirectory.
+        list_filename[0] = 'm'
+        list_filename[1] = 'd'
+        list_filename[2] = ':'
+        void string.copy(name, &list_filename+3)
+        send_command(list_filename)
+    }
+
+    sub rmdir(str name) {
+        ; -- remove a subdirectory.
+        void string.find(name, '*')
+        if_cs
+            return    ; refuse to act on a wildcard *
+        list_filename[0] = 'r'
+        list_filename[1] = 'd'
+        list_filename[2] = ':'
+        void string.copy(name, &list_filename+3)
+        send_command(list_filename)
+    }
+
+    sub curdir() -> uword {
+        ; return current directory name or 0 if error
+        ; special X16 dos command to only return the current path in the entry list (R42+)
+        const ubyte MAX_PATH_LEN=80
+        uword reversebuffer = memory("curdir_buffer", MAX_PATH_LEN, 0)
+        cx16.r12 = reversebuffer + MAX_PATH_LEN-1
+        @(cx16.r12)=0
+        cbm.SETNAM(3, "$=c")
+        cbm.SETLFS(12, diskio.drivenumber, 0)
+        void cbm.OPEN()          ; open 12,8,0,"$=c"
+        if_cs
+            goto io_error
+        void cbm.CHKIN(12)        ; use #12 as input channel
+        if_cs
+            goto io_error
+
+        repeat 6 {
+            void cbm.CHRIN()
+        }
+        while cbm.CHRIN() {
+            ; skip first line (drive label)
+        }
+        while cbm.CHRIN()!='"' {
+            ; skip to first name
+        }
+        ubyte status = cbm.READST()
+        cx16.r10 = &list_filename
+        while status==0 {
+            repeat {
+                @(cx16.r10) = cbm.CHRIN()
+                if @(cx16.r10)==0
+                    break
+                cx16.r10++
+            }
+            while @(cx16.r10)!='"' and cx16.r10>=&list_filename {
+                @(cx16.r10)=0
+                cx16.r10--
+            }
+            @(cx16.r10)=0
+            prepend(list_filename)
+            cx16.r10 = &list_filename
+            while cbm.CHRIN()!='"' and status==0 {
+                status = cbm.READST()
+                ; skipping up to next entry name
+            }
+        }
+
+io_error:
+        cbm.CLRCHN()
+        cbm.CLOSE(12)
+        if status and status & $40 == 0
+            return 0
+        if @(cx16.r12)==0 {
+            cx16.r12--
+            @(cx16.r12)='/'
+        }
+        return cx16.r12
+
+        sub prepend(str dir) {
+            if dir[0]=='/' and dir[1]==0
+                return
+            cx16.r9L = string.length(dir)
+            cx16.r12 -= cx16.r9L
+            sys.memcopy(dir, cx16.r12, cx16.r9L)
+            cx16.r12--
+            @(cx16.r12)='/'
+        }
+    }
+
+    sub relabel(str name) {
+        ; -- change the disk label.
+        list_filename[0] = 'r'
+        list_filename[1] = '-'
+        list_filename[2] = 'h'
+        list_filename[3] = ':'
+        void string.copy(name, &list_filename+4)
+        send_command(list_filename)
+    }
+
+    sub f_seek(uword pos_hiword, uword pos_loword) {
+        ; -- seek in the reading file opened with f_open, to the given 32-bits position
+        ubyte[6] command = ['p',0,0,0,0,0]
+        command[1] = 12       ; f_open uses channel 12
+        command[2] = lsb(pos_loword)
+        command[3] = msb(pos_loword)
+        command[4] = lsb(pos_hiword)
+        command[5] = msb(pos_hiword)
+    send_command:
+        cbm.SETNAM(sizeof(command), &command)
+        cbm.SETLFS(15, drivenumber, 15)
+        void cbm.OPEN()
+        cbm.CLOSE(15)
+    }
+
+
+    ; NOTE: f_seek_w() doesn't work reliably right now. I only manage to corrupt the fat32 filesystem on the sdcard with it...
+;    sub f_seek_w(uword pos_hiword, uword pos_loword) {
+;        ; -- seek in the output file opened with f_open_w, to the given 32-bits position
+;        diskio.f_seek.command[1] = 13       ; f_open_w uses channel 13
+;        diskio.f_seek.command[2] = lsb(pos_loword)
+;        diskio.f_seek.command[3] = msb(pos_loword)
+;        diskio.f_seek.command[4] = lsb(pos_hiword)
+;        diskio.f_seek.command[5] = msb(pos_hiword)
+;        goto diskio.f_seek.send_command
+;    }
+
+}

compiler/res/prog8lib/cx16/floats.p8

@@ -5,7 +5,8 @@
 
 floats {
 	; ---- this block contains C-64 compatible floating point related functions ----
-	;      the addresses are from cx16 V39 emulator and roms! they won't work on older versions.
+
+    %option no_symbol_prefixing
 
         const float  PI     = 3.141592653589793
         const float  TWOPI  = 6.283185307179586

compiler/res/prog8lib/cx16/gfx2.p8

@@ -13,16 +13,19 @@
 ; SCREEN MODE LIST:
 ;   mode 0 = reset back to default text mode
 ;   mode 1 = bitmap 320 x 240 monochrome
-;   mode 2 = bitmap 320 x 240 x 4c (TODO not yet implemented)
-;   mode 3 = bitmap 320 x 240 x 16c (TODO not yet implemented)
+;   mode 2 = bitmap 320 x 240 x 4c (not yet implemented: just use 256c, there's enough vram for that)
+;   mode 3 = bitmap 320 x 240 x 16c (not yet implemented: just use 256c, there's enough vram for that)
 ;   mode 4 = bitmap 320 x 240 x 256c  (like SCREEN $80 but using this api instead of kernal)
 ;   mode 5 = bitmap 640 x 480 monochrome
 ;   mode 6 = bitmap 640 x 480 x 4c
 ;   higher color dephts in highres are not supported due to lack of VRAM
 
+; TODO remove the phx/plx pairs in non-stack compiler version
 
 gfx2 {
 
+    %option no_symbol_prefixing
+
     ; read-only control variables:
     ubyte active_mode = 0
     uword width = 0
@@ -44,7 +47,7 @@ gfx2 {
                 height = 240
                 bpp = 1
             }
-            ; TODO modes 2, 3 not yet implemented
+            ; TODO modes 2, 3
             4 -> {
                 ; lores 256c
                 cx16.VERA_DC_VIDEO = (cx16.VERA_DC_VIDEO & %11001111) | %00100000      ; enable only layer 1
@@ -84,7 +87,7 @@ gfx2 {
             else -> {
                 ; back to default text mode
                 cx16.r15L = cx16.VERA_DC_VIDEO & %00000111 ; retain chroma + output mode
-                c64.CINT()
+                cbm.CINT()
                 cx16.VERA_DC_VIDEO = (cx16.VERA_DC_VIDEO & %11111000) | cx16.r15L
                 width = 0
                 height = 0
@@ -107,7 +110,7 @@ gfx2 {
                 repeat 240/2/8
                     cs_innerloop640()
             }
-            ; TODO mode 2, 3
+            ; TODO modes 2, 3
             4 -> {
                 ; lores 256c
                 repeat 240/2
@@ -155,6 +158,9 @@ gfx2 {
     }
 
     sub horizontal_line(uword x, uword y, uword length, ubyte color) {
+        ubyte[9] masked_ends   = [ 0, %10000000, %11000000, %11100000, %11110000, %11111000, %11111100, %11111110, %11111111]
+        ubyte[9] masked_starts = [ 0, %00000001, %00000011, %00000111, %00001111, %00011111, %00111111, %01111111, %11111111]
+
         if length==0
             return
         when active_mode {
@@ -163,12 +169,13 @@ gfx2 {
                 ubyte separate_pixels = (8-lsb(x)) & 7
                 if separate_pixels as uword > length
                     separate_pixels = lsb(length)
-                repeat separate_pixels {
-                    ; TODO optimize this by writing a masked byte in 1 go
-                    plot(x, y, color)
-                    x++
+                if separate_pixels {
+                    position(x,y)
+                    cx16.VERA_ADDR_H &= %00000111   ; vera auto-increment off
+                    cx16.VERA_DATA0 = cx16.VERA_DATA0 | masked_starts[separate_pixels]
+                    length -= separate_pixels
+                    x += separate_pixels
                 }
-                length -= separate_pixels
                 if length {
                     position(x, y)
                     separate_pixels = lsb(length) & 7
@@ -205,11 +212,9 @@ _loop                   lda  length
                         bra  _loop
 _done
                     }}
-                    repeat separate_pixels {
-                        ; TODO optimize this by writing a masked byte in 1 go
-                        plot(x, y, color)
-                        x++
-                    }
+
+                    cx16.VERA_ADDR_H &= %00000111   ; vera auto-increment off
+                    cx16.VERA_DATA0 = cx16.VERA_DATA0 | masked_ends[separate_pixels]
                 }
                 cx16.VERA_ADDR_H &= %00000111   ; vera auto-increment off again
             }
@@ -236,8 +241,7 @@ _done
                 }}
             }
             6 -> {
-                ; highres 4c
-                ; TODO also mostly usable for lores 4c?
+                ; highres 4c ....also mostly usable for mode 2, lores 4c?
                 color &= 3
                 ubyte[4] colorbits
                 ubyte ii
@@ -366,8 +370,14 @@ _done
                     return
                 position2(x,y,true)
                 set_both_strides(13)    ; 160 increment = 1 line in 640 px 4c mode
-                color &= 3
-                color <<= gfx2.plot.shift4c[lsb(x) & 3]         ; TODO with lookup table
+                ;; color &= 3
+                ;; color <<= gfx2.plot.shift4c[lsb(x) & 3]
+                cx16.r2L = lsb(x) & 3
+                when color & 3 {
+                    1 -> color = gfx2.plot.shiftedleft_4c_1[cx16.r2L]
+                    2 -> color = gfx2.plot.shiftedleft_4c_2[cx16.r2L]
+                    3 -> color = gfx2.plot.shiftedleft_4c_3[cx16.r2L]
+                }
                 ubyte @shared mask = gfx2.plot.mask4c[lsb(x) & 3]
                 repeat lheight {
                     %asm {{
@@ -547,10 +557,13 @@ _done
         }
     }
 
-    sub plot(uword @zp x, uword y, ubyte color) {
+    sub plot(uword @zp x, uword @zp y, ubyte @zp color) {
         ubyte[8] @shared bits = [128, 64, 32, 16, 8, 4, 2, 1]
         ubyte[4] @shared mask4c = [%00111111, %11001111, %11110011, %11111100]
         ubyte[4] @shared shift4c = [6,4,2,0]
+        ubyte[4] shiftedleft_4c_1 = [1<<6, 1<<4, 1<<2, 1<<0]
+        ubyte[4] shiftedleft_4c_2 = [2<<6, 2<<4, 2<<2, 2<<0]
+        ubyte[4] shiftedleft_4c_3 = [3<<6, 3<<4, 3<<2, 3<<0]
 
         when active_mode {
             1 -> {
@@ -587,7 +600,7 @@ _done
                     }}
                 }
             }
-            ; TODO mode 2,3
+            ; TODO modes 2, 3
             4 -> {
                 ; lores 256c
                 void addr_mul_24_for_lores_256c(y, x)      ; 24 bits result is in r0 and r1L (highest byte)
@@ -639,12 +652,16 @@ _done
                 }
             }
             6 -> {
-                ; highres 4c
-                ; TODO also mostly usable for lores 4c?
+                ; highres 4c   ....also mostly usable for mode 2, lores 4c?
                 void addr_mul_24_for_highres_4c(y, x)      ; 24 bits result is in r0 and r1L (highest byte)
                 cx16.r2L = lsb(x) & 3       ; xbits
-                color &= 3
-                color <<= shift4c[cx16.r2L]     ; TODO with lookup table
+                ; color &= 3
+                ; color <<= shift4c[cx16.r2L]
+                when color & 3 {
+                    1 -> color = shiftedleft_4c_1[cx16.r2L]
+                    2 -> color = shiftedleft_4c_2[cx16.r2L]
+                    3 -> color = shiftedleft_4c_3[cx16.r2L]
+                }
                 %asm {{
                     stz  cx16.VERA_CTRL
                     lda  cx16.r1L
@@ -689,7 +706,7 @@ _done
 +
                 }}
             }
-            ; TODO mode 2 and 3
+            ; TODO modes 2, 3
             4 -> {
                 ; lores 256c
                 void addr_mul_24_for_lores_256c(y, x)      ; 24 bits result is in r0 and r1L (highest byte)
@@ -743,27 +760,138 @@ _done
                     sta  cx16.r0L
                 }}
                 cx16.r1L = lsb(x) & 3
-                cx16.r0L >>= gfx2.plot.shift4c[cx16.r1L]        ; TODO with lookup table
+                cx16.r0L >>= gfx2.plot.shift4c[cx16.r1L]
                 return cx16.r0L & 3
             }
             else -> return 0
         }
     }
 
+    sub fill(word @zp x, word @zp y, ubyte new_color) {
+        ; Non-recursive scanline flood fill.
+        ; based loosely on code found here https://www.codeproject.com/Articles/6017/QuickFill-An-efficient-flood-fill-algorithm
+        ; with the fixes applied to the seedfill_4 routine as mentioned in the comments.
+        const ubyte MAXDEPTH = 48
+        word[MAXDEPTH] @split @shared stack_xl
+        word[MAXDEPTH] @split @shared stack_xr
+        word[MAXDEPTH] @split @shared stack_y
+        byte[MAXDEPTH] @shared stack_dy
+        cx16.r12L = 0       ; stack pointer
+        word x1
+        word x2
+        byte dy
+        cx16.r10L = new_color
+        sub push_stack(word sxl, word sxr, word sy, byte sdy) {
+            if cx16.r12L==MAXDEPTH
+                return
+            cx16.r0s = sy+sdy
+            if cx16.r0s>=0 and cx16.r0s<=height-1 {
+;;                stack_xl[cx16.r12L] = sxl
+;;                stack_xr[cx16.r12L] = sxr
+;;                stack_y[cx16.r12L] = sy
+;;                stack_dy[cx16.r12L] = sdy
+;;                cx16.r12L++
+                %asm {{
+                    ldy  cx16.r12L
+                    lda  sxl
+                    sta  stack_xl_lsb,y
+                    lda  sxl+1
+                    sta  stack_xl_msb,y
+                    lda  sxr
+                    sta  stack_xr_lsb,y
+                    lda  sxr+1
+                    sta  stack_xr_msb,y
+                    lda  sy
+                    sta  stack_y_lsb,y
+                    lda  sy+1
+                    sta  stack_y_msb,y
+                    ldy  cx16.r12L
+                    lda  sdy
+                    sta  stack_dy,y
+                    inc  cx16.r12L
+                }}
+            }
+        }
+        sub pop_stack() {
+;;            cx16.r12L--
+;;            x1 = stack_xl[cx16.r12L]
+;;            x2 = stack_xr[cx16.r12L]
+;;            y = stack_y[cx16.r12L]
+;;            dy = stack_dy[cx16.r12L]
+            %asm {{
+                dec  cx16.r12L
+                ldy  cx16.r12L
+                lda  stack_xl_lsb,y
+                sta  x1
+                lda  stack_xl_msb,y
+                sta  x1+1
+                lda  stack_xr_lsb,y
+                sta  x2
+                lda  stack_xr_msb,y
+                sta  x2+1
+                lda  stack_y_lsb,y
+                sta  y
+                lda  stack_y_msb,y
+                sta  y+1
+                ldy  cx16.r12L
+                lda  stack_dy,y
+                sta  dy
+            }}
+            y+=dy
+        }
+        cx16.r11L = pget(x as uword, y as uword)        ; old_color
+        if cx16.r11L == cx16.r10L
+            return
+        if x<0 or x > width-1 or y<0 or y > height-1
+            return
+        push_stack(x, x, y, 1)
+        push_stack(x, x, y + 1, -1)
+        word left = 0
+        while cx16.r12L {
+            pop_stack()
+            x = x1
+            while x >= 0 and pget(x as uword, y as uword) == cx16.r11L {
+                plot(x as uword, y as uword, cx16.r10L)
+                x--
+            }
+            if x>= x1
+                goto skip
+
+            left = x + 1
+            if left < x1
+                push_stack(left, x1 - 1, y, -dy)
+            x = x1 + 1
+
+            do {
+                while x <= width-1 and pget(x as uword, y as uword) == cx16.r11L {
+                    plot(x as uword, y as uword, cx16.r10L)
+                    x++
+                }
+                push_stack(left, x - 1, y, dy)
+                if x > x2 + 1
+                    push_stack(x2 + 1, x - 1, y, -dy)
+skip:
+                x++
+                while x <= x2 and pget(x as uword, y as uword) != cx16.r11L
+                    x++
+                left = x
+            } until x>x2
+        }
+    }
+
     sub position(uword @zp x, uword y) {
-        ubyte bank
         when active_mode {
             1 -> {
                 ; lores monochrome
                 cx16.r0 = y*(320/8) + x/8
                 cx16.vaddr(0, cx16.r0, 0, 1)
             }
-            ; TODO modes 2,3
+            ; TODO modes 2, 3
             4 -> {
                 ; lores 256c
                 void addr_mul_24_for_lores_256c(y, x)      ; 24 bits result is in r0 and r1L (highest byte)
-                bank = lsb(cx16.r1)
-                cx16.vaddr(bank, cx16.r0, 0, 1)
+                cx16.r2L = cx16.r1L
+                cx16.vaddr(cx16.r2L, cx16.r0, 0, 1)
             }
             5 -> {
                 ; highres monochrome
@@ -773,24 +901,16 @@ _done
             6 -> {
                 ; highres 4c
                 void addr_mul_24_for_highres_4c(y, x)      ; 24 bits result is in r0 and r1L (highest byte)
-                bank = lsb(cx16.r1)
-                cx16.vaddr(bank, cx16.r0, 0, 1)
+                cx16.r2L = cx16.r1L
+                cx16.vaddr(cx16.r2L, cx16.r0, 0, 1)
             }
         }
     }
 
     sub position2(uword @zp x, uword y, bool also_port_1) {
         position(x, y)
-        if also_port_1 {
-            when active_mode {
-                1, 5 -> cx16.vaddr(0, cx16.r0, 1, 1)
-                ; TODO modes 2, 3
-                4, 6 -> {
-                    ubyte bank = lsb(cx16.r1)
-                    cx16.vaddr(bank, cx16.r0, 1, 1)
-                }
-            }
-        }
+        if also_port_1
+            cx16.vaddr_clone(0)
     }
 
     inline asmsub next_pixel(ubyte color @A) {
@@ -865,48 +985,96 @@ _done
     sub text(uword @zp x, uword y, ubyte color, uword sctextptr) {
         ; -- Write some text at the given pixel position. The text string must be in screencode encoding (not petscii!).
         ;    You must also have called text_charset() first to select and prepare the character set to use.
-        ;    NOTE: in monochrome (1bpp) screen modes, x position is currently constrained to multiples of 8 !  TODO allow per-pixel horizontal positioning
-        ; TODO draw whole horizontal spans using vera auto increment if possible, instead of per-character columns
         uword chardataptr
+        ubyte[8] @shared char_bitmap_bytes_left
+        ubyte[8] @shared char_bitmap_bytes_right
+
         when active_mode {
             1, 5 -> {
                 ; monochrome mode, either resolution
-                cx16.r2 = 40
-                if active_mode==5
-                    cx16.r2 = 80
-                while @(sctextptr) {
-                    chardataptr = charset_addr + (@(sctextptr) as uword)*8
-                    cx16.vaddr(charset_bank, chardataptr, 1, 1)
-                    position(x,y)
+                cx16.r3 = sctextptr
+                while @(cx16.r3) {
+                    chardataptr = charset_addr + @(cx16.r3) * $0008
+                    ; copy the character bitmap into RAM
+                    cx16.vaddr_autoincr(charset_bank, chardataptr, 0, 1)
                     %asm {{
-                        lda  cx16.VERA_ADDR_H
-                        and  #%111              ; don't auto-increment, we have to do that manually because of the ora
-                        sta  cx16.VERA_ADDR_H
-                        lda  color
+                        ; pre-shift the bits
+                        phx ; TODO remove in non-stack version
+                        lda  text.x
+                        and  #7
                         sta  P8ZP_SCRATCH_B1
-                        ldy  #8
--                       lda  P8ZP_SCRATCH_B1
-                        bne  +                  ; white color, plot normally
-                        lda  cx16.VERA_DATA1
-                        eor  #255               ; black color, keep only the other pixels
-                        and  cx16.VERA_DATA0
-                        bra  ++
-+                       lda  cx16.VERA_DATA0
-                        ora  cx16.VERA_DATA1
-+                       sta  cx16.VERA_DATA0
-                        lda  cx16.VERA_ADDR_L
-                        clc
-                        adc  cx16.r2
-                        sta  cx16.VERA_ADDR_L
-                        bcc  +
-                        inc  cx16.VERA_ADDR_M
-+                       inc  x
-                        bne  +
-                        inc  x+1
-+                       dey
+                        ldy  #0
+-                       lda  cx16.VERA_DATA0
+                        stz  P8ZP_SCRATCH_REG
+                        ldx  P8ZP_SCRATCH_B1
+                        cpx  #0
+                        beq  +
+-                       lsr  a
+                        ror  P8ZP_SCRATCH_REG
+                        dex
                         bne  -
++                       sta  char_bitmap_bytes_left,y
+                        lda  P8ZP_SCRATCH_REG
+                        sta  char_bitmap_bytes_right,y
+                        iny
+                        cpy  #8
+                        bne  --
+                        plx     ; TODO remove in non-stack version
                     }}
-                    sctextptr++
+                    ; left part of shifted char
+                    position2(x, y, true)
+                    set_autoincrs_mode1_or_5()
+                    if color {
+                        %asm {{
+                            ldy  #0
+-                           lda  char_bitmap_bytes_left,y
+                            ora  cx16.VERA_DATA1
+                            sta  cx16.VERA_DATA0
+                            iny
+                            cpy  #8
+                            bne  -
+                        }}
+                    } else {
+                        %asm {{
+                            ldy  #0
+-                           lda  char_bitmap_bytes_left,y
+                            eor  #255
+                            and  cx16.VERA_DATA1
+                            sta  cx16.VERA_DATA0
+                            iny
+                            cpy  #8
+                            bne  -
+                        }}
+                    }
+                    ; right part of shifted char
+                    if lsb(x) & 7 {
+                        position2(x+8, y, true)
+                        set_autoincrs_mode1_or_5()
+                        if color {
+                            %asm {{
+                                ldy  #0
+    -                           lda  char_bitmap_bytes_right,y
+                                ora  cx16.VERA_DATA1
+                                sta  cx16.VERA_DATA0
+                                iny
+                                cpy  #8
+                                bne  -
+                            }}
+                        } else {
+                            %asm {{
+                                ldy  #0
+    -                           lda  char_bitmap_bytes_right,y
+                                eor  #255
+                                and  cx16.VERA_DATA1
+                                sta  cx16.VERA_DATA0
+                                iny
+                                cpy  #8
+                                bne  -
+                            }}
+                        }
+                    }
+                    cx16.r3++
+                    x += 8
                 }
             }
             4 -> {
@@ -918,7 +1086,7 @@ _done
                         position(x,y)
                         y++
                         %asm {{
-                            phx
+                            phx ; TODO remove in non-stack version
                             ldx  color
                             lda  cx16.VERA_DATA1
                             sta  P8ZP_SCRATCH_B1
@@ -930,7 +1098,7 @@ _done
 +                           lda  cx16.VERA_DATA0    ; don't write a pixel, but do advance to the next address
 +                           dey
                             bne  -
-                            plx
+                            plx ; TODO remove in non-stack version
                         }}
                     }
                     x+=8
@@ -942,30 +1110,80 @@ _done
                 ; hires 4c
                 ; we're going to use a few cx16 registers to make sure every variable is in zeropage in the inner loop.
                 cx16.r11L = color
-                cx16.r8 = y
                 while @(sctextptr) {
                     chardataptr = charset_addr + (@(sctextptr) as uword)*8
+                    cx16.vaddr(charset_bank, chardataptr, 1, true)  ; for reading the chardata from Vera data channel 1
+                    position(x, y)              ; only calculated once, we update vera address in the loop instead
+                    cx16.VERA_ADDR_H &= $0f     ; no auto increment
                     repeat 8 {
-                        ; TODO rewrite this inner loop partly in assembly
-                        ;      requires expanding the charbits to 2-bits per pixel (based on color)
-                        ;      also it's way more efficient to draw whole horizontal spans instead of per-character
-                        cx16.r9L = cx16.vpeek(charset_bank, chardataptr)  ; get the 8 horizontal character bits
+                        cx16.r10L = cx16.VERA_DATA1  ; get the next 8 horizontal character bits
                         cx16.r7 = x
                         repeat 8 {
-                            cx16.r9L <<= 1
-                            if_cs
-                                plot(cx16.r7, cx16.r8, cx16.r11L)
+                            cx16.r10L <<= 1
+                            if_cs {
+                                cx16.r2L = cx16.r7L & 3       ; xbits
+                                when cx16.r11L & 3 {
+                                    1 -> cx16.r12L = gfx2.plot.shiftedleft_4c_1[cx16.r2L]
+                                    2 -> cx16.r12L = gfx2.plot.shiftedleft_4c_2[cx16.r2L]
+                                    3 -> cx16.r12L = gfx2.plot.shiftedleft_4c_3[cx16.r2L]
+                                    else -> cx16.r12L = 0
+                                }
+                                cx16.VERA_DATA0 = cx16.VERA_DATA0 & gfx2.plot.mask4c[cx16.r2L] | cx16.r12L
+                            }
                             cx16.r7++
+                            if (cx16.r7 & 3) == 0 {
+                                ; increment the pixel address by one
+                                %asm {{
+                                    stz  cx16.VERA_CTRL
+                                    clc
+                                    lda  cx16.VERA_ADDR_L
+                                    adc  #1
+                                    sta  cx16.VERA_ADDR_L
+                                    lda  cx16.VERA_ADDR_M
+                                    adc  #0
+                                    sta  cx16.VERA_ADDR_M
+                                    lda  cx16.VERA_ADDR_H
+                                    adc  #0
+                                    sta  cx16.VERA_ADDR_H
+                                }}
+                            }
                         }
-                        chardataptr++
-                        cx16.r8++
+
+                        ; increment pixel address to the next line
+                        %asm {{
+                            stz  cx16.VERA_CTRL
+                            clc
+                            lda  cx16.VERA_ADDR_L
+                            adc  #(640-8)/4
+                            sta  cx16.VERA_ADDR_L
+                            lda  cx16.VERA_ADDR_M
+                            adc  #0
+                            sta  cx16.VERA_ADDR_M
+                            lda  cx16.VERA_ADDR_H
+                            adc  #0
+                            sta  cx16.VERA_ADDR_H
+                        }}
                     }
                     x+=8
-                    cx16.r8-=8
                     sctextptr++
                 }
             }
         }
+
+        sub set_autoincrs_mode1_or_5() {
+            ; set autoincrements to go to next pixel row (40 or 80 increment)
+            if active_mode==1 {
+                cx16.VERA_CTRL = 0
+                cx16.VERA_ADDR_H = cx16.VERA_ADDR_H & $0f | (11<<4)
+                cx16.VERA_CTRL = 1
+                cx16.VERA_ADDR_H = cx16.VERA_ADDR_H & $0f | (11<<4)
+            } else {
+                cx16.VERA_CTRL = 0
+                cx16.VERA_ADDR_H = cx16.VERA_ADDR_H & $0f | (12<<4)
+                cx16.VERA_CTRL = 1
+                cx16.VERA_ADDR_H = cx16.VERA_ADDR_H & $0f | (12<<4)
+            }
+        }
     }
 
     asmsub cs_innerloop640() clobbers(Y) {

compiler/res/prog8lib/cx16/graphics.p8

@@ -10,6 +10,8 @@
 
 
 graphics {
+    %option no_symbol_prefixing
+
     const uword WIDTH = 320
     const ubyte HEIGHT = 240
 

compiler/res/prog8lib/cx16/palette.p8

@@ -2,9 +2,10 @@
 ; Should you want to restore the default palette, you have to reinitialize the Vera yourself.
 
 palette {
+    %option no_symbol_prefixing
 
     uword vera_palette_ptr
-    ubyte c
+    ubyte cc
 
     sub set_color(ubyte index, uword color) {
         vera_palette_ptr = $fa00+(index as uword * 2)
@@ -79,13 +80,13 @@ palette {
     sub set_grayscale() {
         vera_palette_ptr = $fa00
         repeat 16 {
-            c=0
+            cc=0
             repeat 16 {
-                cx16.vpoke(1, vera_palette_ptr, c)
+                cx16.vpoke(1, vera_palette_ptr, cc)
                 vera_palette_ptr++
-                cx16.vpoke(1, vera_palette_ptr, c)
+                cx16.vpoke(1, vera_palette_ptr, cc)
                 vera_palette_ptr++
-                c += $11
+                cc += $11
             }
         }
     }
@@ -150,11 +151,11 @@ palette {
     sub set_c64pepto() {
         vera_palette_ptr = $fa00
         repeat 16 {
-            for c in 0 to 15 {
-                uword cc = C64_colorpalette_pepto[c]
-                cx16.vpoke(1, vera_palette_ptr, lsb(cc))     ; G, B
+            for cc in 0 to 15 {
+                uword ccp = C64_colorpalette_pepto[cc]
+                cx16.vpoke(1, vera_palette_ptr, lsb(ccp))     ; G, B
                 vera_palette_ptr++
-                cx16.vpoke(1, vera_palette_ptr, msb(cc))     ; R
+                cx16.vpoke(1, vera_palette_ptr, msb(ccp))     ; R
                 vera_palette_ptr++
             }
         }
@@ -163,11 +164,11 @@ palette {
     sub set_c64light() {
         vera_palette_ptr = $fa00
         repeat 16 {
-            for c in 0 to 15 {
-                uword cc = C64_colorpalette_light[c]
-                cx16.vpoke(1, vera_palette_ptr, lsb(cc))     ; G, B
+            for cc in 0 to 15 {
+                uword ccp = C64_colorpalette_light[cc]
+                cx16.vpoke(1, vera_palette_ptr, lsb(ccp))     ; G, B
                 vera_palette_ptr++
-                cx16.vpoke(1, vera_palette_ptr, msb(cc))     ; R
+                cx16.vpoke(1, vera_palette_ptr, msb(ccp))     ; R
                 vera_palette_ptr++
             }
         }
@@ -176,11 +177,11 @@ palette {
     sub set_c64dark() {
         vera_palette_ptr = $fa00
         repeat 16 {
-            for c in 0 to 15 {
-                uword cc = C64_colorpalette_dark[c]
-                cx16.vpoke(1, vera_palette_ptr, lsb(cc))     ; G, B
+            for cc in 0 to 15 {
+                uword ccp = C64_colorpalette_dark[cc]
+                cx16.vpoke(1, vera_palette_ptr, lsb(ccp))     ; G, B
                 vera_palette_ptr++
-                cx16.vpoke(1, vera_palette_ptr, msb(cc))     ; R
+                cx16.vpoke(1, vera_palette_ptr, msb(ccp))     ; R
                 vera_palette_ptr++
             }
         }

compiler/res/prog8lib/cx16/psg.p8

@@ -1,6 +1,8 @@
 %import syslib
 
 psg {
+    %option no_symbol_prefixing
+
     ; $1F9C0 - $1F9FF 	16 blocks of 4 PSG registers (16 voices)
     ; 00  frequency word LSB
     ; 01  frequency word MSB.    freqword = HERZ / 0.3725290298461914
@@ -22,6 +24,7 @@ psg {
         ;    waveform = one of PULSE,SAWTOOTH,TRIANGLE,NOISE.
         ;    pulsewidth = 0-63.  Specifies the pulse width for waveform=PULSE.
         envelope_states[voice_num] = 255
+        sys.irqsafe_set_irqd()
         cx16.r0 = $f9c2 + voice_num * 4
         cx16.VERA_CTRL = 0
         cx16.VERA_ADDR_L = lsb(cx16.r0)
@@ -32,6 +35,7 @@ psg {
         cx16.VERA_DATA0 = waveform | pulsewidth
         envelope_volumes[voice_num] = mkword(volume, 0)
         envelope_maxvolumes[voice_num] = volume
+        sys.irqsafe_clear_irqd()
     }
 
 ;    sub freq_hz(ubyte voice_num, float hertz) {
@@ -44,48 +48,50 @@ psg {
     sub freq(ubyte voice_num, uword vera_freq) {
         ; -- Changes the frequency of the voice's sound.
         ;    voice_num = 0-15,  vera_freq = 0-65535  calculate this via the formula given in the Vera's PSG documentation.
-        ;    (https://github.com/commanderx16/x16-docs/blob/master/VERA%20Programmer's%20Reference.md)
-        cx16.r0 = $f9c0 + voice_num * 4
+        ;    (https://github.com/x16community/x16-docs/blob/master/VERA%20Programmer's%20Reference.md)
+        ;    Write freq MSB first and then LSB to reduce the chance on clicks
+        sys.irqsafe_set_irqd()
+        cx16.r0 = $f9c1 + voice_num * 4
         cx16.VERA_CTRL = 0
         cx16.VERA_ADDR_L = lsb(cx16.r0)
         cx16.VERA_ADDR_M = msb(cx16.r0)
         cx16.VERA_ADDR_H = 1
-        cx16.VERA_DATA0 = lsb(vera_freq)
-        cx16.VERA_ADDR_L++
         cx16.VERA_DATA0 = msb(vera_freq)
+        cx16.VERA_ADDR_L--
+        cx16.VERA_DATA0 = lsb(vera_freq)
+        sys.irqsafe_clear_irqd()
     }
 
     sub volume(ubyte voice_num, ubyte vol) {
         ; -- Modifies the volume of this voice.
         ;    voice_num = 0-15, vol = 0-63 where 0=silent, 63=loudest.
-        cx16.r0 = $f9c2 + voice_num * 4
-        cx16.vpoke(1, cx16.r0, cx16.vpeek(1, cx16.r0) & %11000000 | vol)
         envelope_volumes[voice_num] = mkword(vol, 0)
+        cx16.vpoke_mask(1, $f9c2 + voice_num * 4, %11000000, vol)
         envelope_maxvolumes[voice_num] = vol
     }
 
     sub pulse_width(ubyte voice_num, ubyte pw) {
         ; -- Modifies the pulse width of this voice (when waveform=PULSE)
         ;    voice_num = 0-15, pw = 0-63  where 0=narrow, 63=50%cycle so square wave.
-        cx16.r0 = $f9c3 + voice_num * 4
-        cx16.vpoke(1, cx16.r0, cx16.vpeek(1, cx16.r0) & %11000000 | pw)
+        cx16.vpoke_mask(1, $f9c3 + voice_num * 4, %11000000, pw)
     }
 
     sub envelope(ubyte voice_num, ubyte maxvolume, ubyte attack, ubyte sustain, ubyte release) {
         ; -- Enables AttackSustainRelease volume envelope for a voice.
         ;    Note: this requires setting up envelopes_irq() as well, read its description.
         ;    voice_num = 0-15   maxvolume = 0-63
-        ;    attack, sustain, release = 0-255 that determine the speed of the A/D/R.
-        ;    TODO describe how the speeds are calculated. For now, experiment. Higher values means *slower* enveloping.
+        ;    attack, sustain, release = 0-255 that determine the speed of the A/D/R:
+        ;    attack time:    MAXVOL/15/attack  seconds.    higher value = faster attack.
+        ;    sustain time:   sustain/60 seconds    higher sustain value = longer sustain (!).
+        ;    release time:   MAXVOL/15/release seconds.   higher vaule = faster release.
+
         envelope_states[voice_num] = 255
         envelope_attacks[voice_num] = attack
         envelope_sustains[voice_num] = sustain
         envelope_releases[voice_num] = release
-        if attack
-            attack = 0
-        else
-            attack = maxvolume     ; max volume when no attack is set
-        envelope_volumes[voice_num] = mkword(attack, 0)
+        cx16.r0 = mkword(maxvolume, 0)
+        if cx16.r0<envelope_volumes[voice_num]
+            envelope_volumes[voice_num] = cx16.r0
         envelope_maxvolumes[voice_num] = maxvolume
         envelope_states[voice_num] = 0
     }
@@ -94,7 +100,6 @@ psg {
         ; -- Shut down all PSG voices.
         for cx16.r1L in 0 to 15 {
             envelope_states[cx16.r1L] = 255
-            envelope_volumes[cx16.r1L] = 0
             volume(cx16.r1L, 0)
         }
     }
@@ -104,11 +109,11 @@ psg {
         ; you have to call this routine every 1/60th second, for example from your vsync irq handler,
         ; or just install this routine as the only irq handler if you don't have to do other things there.
         ; Example: cx16.set_irq(&psg.envelopes_irq, true)
+        ; NOTE: this routine calls save/restore_vera_context() for you, don't nest this or call it yourself!
 
         ; cx16.r0 = the volume word (volume scaled by 256)
         ; cx16.r1L = the voice number
         ; cx16.r2L = attack value
-
         pushw(cx16.r0)
         push(cx16.r1L)
         push(cx16.r2L)
@@ -148,7 +153,7 @@ psg {
         }
 
         ; set new volumes of all 16 voices, using vera stride of 4
-        cx16.push_vera_context()
+        cx16.save_vera_context()
         cx16.VERA_CTRL = 0
         cx16.VERA_ADDR_L = $c2
         cx16.VERA_ADDR_M = $f9
@@ -160,7 +165,7 @@ psg {
         for cx16.r1L in 0 to 15 {
             cx16.VERA_DATA0 = cx16.VERA_DATA1 & %11000000 | msb(envelope_volumes[cx16.r1L])
         }
-        cx16.pop_vera_context()
+        cx16.restore_vera_context()
         popw(cx16.r9)
         pop(cx16.r2L)
         pop(cx16.r1L)
@@ -168,7 +173,7 @@ psg {
     }
 
     ubyte[16] envelope_states
-    uword[16] envelope_volumes      ; scaled by 256
+    uword[16] @split envelope_volumes      ; scaled by 256
     ubyte[16] envelope_attacks
     ubyte[16] envelope_sustains
     ubyte[16] envelope_releases

compiler/res/prog8lib/cx16/syslib.p8

@@ -1,9 +1,11 @@
 ; Prog8 definitions for the CommanderX16
 ; Including memory registers, I/O registers, Basic and Kernal subroutines.
 
-c64 {
+cbm {
 
-; ---- kernal routines, these are the same as on the Commodore-64 (hence the same block name) ----
+    ; Commodore (CBM) common variables, vectors and kernal routines
+
+    %option no_symbol_prefixing
 
 ; STROUT --> use txt.print
 ; CLEARSCR -> use txt.clear_screen
@@ -13,12 +15,12 @@ romsub $FF81 = CINT() clobbers(A,X,Y)                           ; (alias: SCINIT
 romsub $FF84 = IOINIT() clobbers(A, X)                          ; initialize I/O devices (CIA, SID, IRQ)
 romsub $FF87 = RAMTAS() clobbers(A,X,Y)                         ; initialize RAM, tape buffer, screen
 romsub $FF8A = RESTOR() clobbers(A,X,Y)                         ; restore default I/O vectors
-romsub $FF8D = VECTOR(uword userptr @ XY, ubyte dir @ Pc) clobbers(A,Y)     ; read/set I/O vector table
+romsub $FF8D = VECTOR(uword userptr @ XY, bool dir @ Pc) clobbers(A,Y)     ; read/set I/O vector table
 romsub $FF90 = SETMSG(ubyte value @ A)                          ; set Kernal message control flag
 romsub $FF93 = SECOND(ubyte address @ A) clobbers(A)            ; (alias: LSTNSA) send secondary address after LISTEN
 romsub $FF96 = TKSA(ubyte address @ A) clobbers(A)              ; (alias: TALKSA) send secondary address after TALK
-romsub $FF99 = MEMTOP(uword address @ XY, ubyte dir @ Pc) -> uword @ XY     ; read/set top of memory  pointer.   NOTE: as a Cx16 extension, also returns the number of RAM memory banks in register A !  See cx16.numbanks()
-romsub $FF9C = MEMBOT(uword address @ XY, ubyte dir @ Pc) -> uword @ XY     ; read/set bottom of memory  pointer
+romsub $FF99 = MEMTOP(uword address @ XY, bool dir @ Pc) -> uword @ XY     ; read/set top of memory  pointer.   NOTE: as a Cx16 extension, also returns the number of RAM memory banks in register A !  See cx16.numbanks()
+romsub $FF9C = MEMBOT(uword address @ XY, bool dir @ Pc) -> uword @ XY     ; read/set bottom of memory  pointer
 romsub $FF9F = SCNKEY() clobbers(A,X,Y)                         ; scan the keyboard
 romsub $FFA2 = SETTMO(ubyte timeout @ A)                        ; set time-out flag for IEEE bus
 romsub $FFA5 = ACPTR() -> ubyte @ A                             ; (alias: IECIN) input byte from serial bus
@@ -30,23 +32,23 @@ romsub $FFB4 = TALK(ubyte device @ A) clobbers(A)               ; command serial
 romsub $FFB7 = READST() -> ubyte @ A                            ; read I/O status word
 romsub $FFBA = SETLFS(ubyte logical @ A, ubyte device @ X, ubyte secondary @ Y)   ; set logical file parameters
 romsub $FFBD = SETNAM(ubyte namelen @ A, str filename @ XY)     ; set filename parameters
-romsub $FFC0 = OPEN() clobbers(X,Y) -> ubyte @Pc, ubyte @A      ; (via 794 ($31A)) open a logical file
+romsub $FFC0 = OPEN() clobbers(X,Y) -> bool @Pc, ubyte @A      ; (via 794 ($31A)) open a logical file
 romsub $FFC3 = CLOSE(ubyte logical @ A) clobbers(A,X,Y)         ; (via 796 ($31C)) close a logical file
-romsub $FFC6 = CHKIN(ubyte logical @ X) clobbers(A,X) -> ubyte @Pc    ; (via 798 ($31E)) define an input channel
+romsub $FFC6 = CHKIN(ubyte logical @ X) clobbers(A,X) -> bool @Pc    ; (via 798 ($31E)) define an input channel
 romsub $FFC9 = CHKOUT(ubyte logical @ X) clobbers(A,X)          ; (via 800 ($320)) define an output channel
 romsub $FFCC = CLRCHN() clobbers(A,X)                           ; (via 802 ($322)) restore default devices
 romsub $FFCF = CHRIN() clobbers(X, Y) -> ubyte @ A   ; (via 804 ($324)) input a character (for keyboard, read a whole line from the screen) A=byte read.
 romsub $FFD2 = CHROUT(ubyte char @ A)                           ; (via 806 ($326)) output a character
-romsub $FFD5 = LOAD(ubyte verify @ A, uword address @ XY) -> ubyte @Pc, ubyte @ A, uword @ XY     ; (via 816 ($330)) load from device
-romsub $FFD8 = SAVE(ubyte zp_startaddr @ A, uword endaddr @ XY) -> ubyte @ Pc, ubyte @ A          ; (via 818 ($332)) save to a device
+romsub $FFD5 = LOAD(ubyte verify @ A, uword address @ XY) -> bool @Pc, ubyte @ A, uword @ XY     ; (via 816 ($330)) load from device
+romsub $FFD8 = SAVE(ubyte zp_startaddr @ A, uword endaddr @ XY) clobbers (X, Y) -> bool @ Pc, ubyte @ A       ; (via 818 ($332)) save to a device
 romsub $FFDB = SETTIM(ubyte low @ A, ubyte middle @ X, ubyte high @ Y)      ; set the software clock
 romsub $FFDE = RDTIM() -> ubyte @ A, ubyte @ X, ubyte @ Y       ; read the software clock (A=lo,X=mid,Y=high)
-romsub $FFE1 = STOP() clobbers(X) -> ubyte @ Pz, ubyte @ A      ; (via 808 ($328)) check the STOP key (and some others in A)
-romsub $FFE4 = GETIN() clobbers(X,Y) -> ubyte @Pc, ubyte @ A    ; (via 810 ($32A)) get a character
+romsub $FFE1 = STOP() clobbers(X) -> bool @ Pz, ubyte @ A      ; (via 808 ($328)) check the STOP key (and some others in A)
+romsub $FFE4 = GETIN() clobbers(X,Y) -> bool @Pc, ubyte @ A    ; (via 810 ($32A)) get a character
 romsub $FFE7 = CLALL() clobbers(A,X)                            ; (via 812 ($32C)) close all files
 romsub $FFEA = UDTIM() clobbers(A,X)                            ; update the software clock
 romsub $FFED = SCREEN() -> ubyte @ X, ubyte @ Y                 ; read number of screen rows and columns
-romsub $FFF0 = PLOT(ubyte col @ Y, ubyte row @ X, ubyte dir @ Pc) -> ubyte @ X, ubyte @ Y       ; read/set position of cursor on screen.  Use txt.plot for a 'safe' wrapper that preserves X.
+romsub $FFF0 = PLOT(ubyte col @ Y, ubyte row @ X, bool dir @ Pc) -> ubyte @ X, ubyte @ Y       ; read/set position of cursor on screen.  Use txt.plot for a 'safe' wrapper that preserves X.
 romsub $FFF3 = IOBASE() -> uword @ XY                           ; read base address of I/O devices
 
 ; ---- utility
@@ -55,7 +57,7 @@ asmsub STOP2() -> ubyte @A  {
     ; -- check if STOP key was pressed, returns true if so.  More convenient to use than STOP() because that only sets the carry status flag.
     %asm {{
         phx
-        jsr  c64.STOP
+        jsr  cbm.STOP
         beq  +
         plx
         lda  #0
@@ -67,10 +69,14 @@ asmsub STOP2() -> ubyte @A  {
 }
 
 asmsub RDTIM16() -> uword @AY {
-    ; --  like RDTIM() but only returning the lower 16 bits in AY for convenience
+    ; --  like RDTIM() but only returning the lower 16 bits in AY for convenience. Also avoids ram bank issue for irqs.
     %asm {{
         phx
-        jsr  c64.RDTIM
+        php
+        sei
+        jsr  cbm.RDTIM
+        plp
+        cli
         pha
         txa
         tay
@@ -84,38 +90,40 @@ asmsub RDTIM16() -> uword @AY {
 
 cx16 {
 
+    %option no_symbol_prefixing
+
 ; irq, system and hardware vectors:
-    &uword  IERROR          = $0300
-    &uword  IMAIN           = $0302
-    &uword  ICRNCH          = $0304
-    &uword  IQPLOP          = $0306
-    &uword  IGONE           = $0308
-    &uword  IEVAL           = $030a
-    &ubyte  SAREG           = $030c     ; register storage for A for SYS calls
-    &ubyte  SXREG           = $030d     ; register storage for X for SYS calls
-    &ubyte  SYREG           = $030e     ; register storage for Y for SYS calls
-    &ubyte  SPREG           = $030f     ; register storage for P (status register) for SYS calls
-    &uword  USRADD          = $0311     ; vector for the USR() basic command
+    &uword  IERROR      = $0300
+    &uword  IMAIN       = $0302
+    &uword  ICRNCH      = $0304
+    &uword  IQPLOP      = $0306
+    &uword  IGONE       = $0308
+    &uword  IEVAL       = $030a
+    &ubyte  SAREG       = $030c     ; register storage for A for SYS calls
+    &ubyte  SXREG       = $030d     ; register storage for X for SYS calls
+    &ubyte  SYREG       = $030e     ; register storage for Y for SYS calls
+    &ubyte  SPREG       = $030f     ; register storage for P (status register) for SYS calls
+    &uword  USRADD      = $0311     ; vector for the USR() basic command
     ; $0313 is unused.
-    &uword  CINV            = $0314     ; IRQ vector (in ram)
-    &uword  CBINV           = $0316     ; BRK vector (in ram)
-    &uword  NMINV           = $0318     ; NMI vector (in ram)
-    &uword  IOPEN           = $031a
-    &uword  ICLOSE          = $031c
-    &uword  ICHKIN          = $031e
-    &uword  ICKOUT          = $0320
-    &uword  ICLRCH          = $0322
-    &uword  IBASIN          = $0324
-    &uword  IBSOUT          = $0326
-    &uword  ISTOP           = $0328
-    &uword  IGETIN          = $032a
-    &uword  ICLALL          = $032c
-    &uword  KEYHDL          = $032e     ; keyboard scan code handler see examples/cx16/keyboardhandler.p8
-    &uword  ILOAD           = $0330
-    &uword  ISAVE           = $0332
-    &uword  NMI_VEC         = $FFFA     ; 65c02 nmi vector, determined by the kernal if banked in
-    &uword  RESET_VEC       = $FFFC     ; 65c02 reset vector, determined by the kernal if banked in
-    &uword  IRQ_VEC         = $FFFE     ; 65c02 interrupt vector, determined by the kernal if banked in
+    &uword  CINV        = $0314     ; IRQ vector (in ram)
+    &uword  CBINV       = $0316     ; BRK vector (in ram)
+    &uword  NMINV       = $0318     ; NMI vector (in ram)
+    &uword  IOPEN       = $031a
+    &uword  ICLOSE      = $031c
+    &uword  ICHKIN      = $031e
+    &uword  ICKOUT      = $0320
+    &uword  ICLRCH      = $0322
+    &uword  IBASIN      = $0324
+    &uword  IBSOUT      = $0326
+    &uword  ISTOP       = $0328
+    &uword  IGETIN      = $032a
+    &uword  ICLALL      = $032c
+    &uword  KEYHDL      = $032e     ; keyboard scan code handler see examples/cx16/keyboardhandler.p8
+    &uword  ILOAD       = $0330
+    &uword  ISAVE       = $0332
+    &uword  NMI_VEC     = $FFFA     ; 65c02 nmi vector, determined by the kernal if banked in
+    &uword  RESET_VEC   = $FFFC     ; 65c02 reset vector, determined by the kernal if banked in
+    &uword  IRQ_VEC     = $FFFE     ; 65c02 interrupt vector, determined by the kernal if banked in
 
 
 ; the sixteen virtual 16-bit registers in both normal unsigned mode and signed mode (s)
@@ -271,44 +279,44 @@ cx16 {
 ; I/O
 
     const uword  VIA1_BASE   = $9f00                  ;VIA 6522 #1
-    &ubyte  via1prb	= VIA1_BASE + 0
-    &ubyte  via1pra	= VIA1_BASE + 1
-    &ubyte  via1ddrb	= VIA1_BASE + 2
-    &ubyte  via1ddra	= VIA1_BASE + 3
-    &ubyte  via1t1l	= VIA1_BASE + 4
-    &ubyte  via1t1h	= VIA1_BASE + 5
-    &ubyte  via1t1ll	= VIA1_BASE + 6
-    &ubyte  via1t1lh	= VIA1_BASE + 7
-    &ubyte  via1t2l	= VIA1_BASE + 8
-    &ubyte  via1t2h	= VIA1_BASE + 9
-    &ubyte  via1sr	= VIA1_BASE + 10
-    &ubyte  via1acr	= VIA1_BASE + 11
-    &ubyte  via1pcr	= VIA1_BASE + 12
-    &ubyte  via1ifr	= VIA1_BASE + 13
-    &ubyte  via1ier	= VIA1_BASE + 14
-    &ubyte  via1ora	= VIA1_BASE + 15
+    &ubyte  via1prb    = VIA1_BASE + 0
+    &ubyte  via1pra    = VIA1_BASE + 1
+    &ubyte  via1ddrb   = VIA1_BASE + 2
+    &ubyte  via1ddra   = VIA1_BASE + 3
+    &ubyte  via1t1l    = VIA1_BASE + 4
+    &ubyte  via1t1h    = VIA1_BASE + 5
+    &ubyte  via1t1ll   = VIA1_BASE + 6
+    &ubyte  via1t1lh   = VIA1_BASE + 7
+    &ubyte  via1t2l    = VIA1_BASE + 8
+    &ubyte  via1t2h    = VIA1_BASE + 9
+    &ubyte  via1sr     = VIA1_BASE + 10
+    &ubyte  via1acr    = VIA1_BASE + 11
+    &ubyte  via1pcr    = VIA1_BASE + 12
+    &ubyte  via1ifr    = VIA1_BASE + 13
+    &ubyte  via1ier    = VIA1_BASE + 14
+    &ubyte  via1ora    = VIA1_BASE + 15
 
     const uword  VIA2_BASE   = $9f10                  ;VIA 6522 #2
-    &ubyte  via2prb	= VIA2_BASE + 0
-    &ubyte  via2pra	= VIA2_BASE + 1
-    &ubyte  via2ddrb	= VIA2_BASE + 2
-    &ubyte  via2ddra	= VIA2_BASE + 3
-    &ubyte  via2t1l	= VIA2_BASE + 4
-    &ubyte  via2t1h	= VIA2_BASE + 5
-    &ubyte  via2t1ll	= VIA2_BASE + 6
-    &ubyte  via2t1lh	= VIA2_BASE + 7
-    &ubyte  via2t2l	= VIA2_BASE + 8
-    &ubyte  via2t2h	= VIA2_BASE + 9
-    &ubyte  via2sr	= VIA2_BASE + 10
-    &ubyte  via2acr	= VIA2_BASE + 11
-    &ubyte  via2pcr	= VIA2_BASE + 12
-    &ubyte  via2ifr	= VIA2_BASE + 13
-    &ubyte  via2ier	= VIA2_BASE + 14
-    &ubyte  via2ora	= VIA2_BASE + 15
+    &ubyte  via2prb    = VIA2_BASE + 0
+    &ubyte  via2pra    = VIA2_BASE + 1
+    &ubyte  via2ddrb   = VIA2_BASE + 2
+    &ubyte  via2ddra   = VIA2_BASE + 3
+    &ubyte  via2t1l    = VIA2_BASE + 4
+    &ubyte  via2t1h    = VIA2_BASE + 5
+    &ubyte  via2t1ll   = VIA2_BASE + 6
+    &ubyte  via2t1lh   = VIA2_BASE + 7
+    &ubyte  via2t2l    = VIA2_BASE + 8
+    &ubyte  via2t2h    = VIA2_BASE + 9
+    &ubyte  via2sr     = VIA2_BASE + 10
+    &ubyte  via2acr    = VIA2_BASE + 11
+    &ubyte  via2pcr    = VIA2_BASE + 12
+    &ubyte  via2ifr    = VIA2_BASE + 13
+    &ubyte  via2ier    = VIA2_BASE + 14
+    &ubyte  via2ora    = VIA2_BASE + 15
 
 ; YM-2151 sound chip
     &ubyte  YM_ADDRESS	= $9f40
-    &ubyte  YM_DATA	= $9f41
+    &ubyte  YM_DATA	    = $9f41
 
     const uword  extdev	= $9f60
 
@@ -320,7 +328,7 @@ cx16 {
 romsub $ff4a = close_all(ubyte device @A)  clobbers(A,X,Y)
 romsub $ff59 = lkupla(ubyte la @A)  clobbers(A,X,Y)
 romsub $ff5c = lkupsa(ubyte sa @Y)  clobbers(A,X,Y)
-romsub $ff5f = screen_mode(ubyte mode @A, ubyte getCurrent @Pc)  clobbers(A, X, Y) -> ubyte @Pc
+romsub $ff5f = screen_mode(ubyte mode @A, bool getCurrent @Pc)  clobbers(X, Y) -> ubyte @A, bool @Pc        ; note: X,Y size result is not supported, use SCREEN or get_screen_mode routine for that
 romsub $ff62 = screen_set_charset(ubyte charset @A, uword charsetptr @XY)  clobbers(A,X,Y)      ; incompatible with C128  dlchr()
 ; not yet supported: romsub $ff65 = pfkey()  clobbers(A,X,Y)
 romsub $ff6e = jsrfar()  ; following word = address to call, byte after that=rom/ram bank it is in
@@ -337,12 +345,12 @@ romsub $ff23 = GRAPH_clear()  clobbers(A,X,Y)
 romsub $ff26 = GRAPH_set_window(uword x @R0, uword y @R1, uword width @R2, uword height @R3)  clobbers(A,X,Y)
 romsub $ff29 = GRAPH_set_colors(ubyte stroke @A, ubyte fill @X, ubyte background @Y)  clobbers (A,X,Y)
 romsub $ff2c = GRAPH_draw_line(uword x1 @R0, uword y1 @R1, uword x2 @R2, uword y2 @R3)  clobbers(A,X,Y)
-romsub $ff2f = GRAPH_draw_rect(uword x @R0, uword y @R1, uword width @R2, uword height @R3, uword cornerradius @R4, ubyte fill @Pc)  clobbers(A,X,Y)
+romsub $ff2f = GRAPH_draw_rect(uword x @R0, uword y @R1, uword width @R2, uword height @R3, uword cornerradius @R4, bool fill @Pc)  clobbers(A,X,Y)
 romsub $ff32 = GRAPH_move_rect(uword sx @R0, uword sy @R1, uword tx @R2, uword ty @R3, uword width @R4, uword height @R5)  clobbers(A,X,Y)
-romsub $ff35 = GRAPH_draw_oval(uword x @R0, uword y @R1, uword width @R2, uword height @R3, ubyte fill @Pc)  clobbers(A,X,Y)
+romsub $ff35 = GRAPH_draw_oval(uword x @R0, uword y @R1, uword width @R2, uword height @R3, bool fill @Pc)  clobbers(A,X,Y)
 romsub $ff38 = GRAPH_draw_image(uword x @R0, uword y @R1, uword ptr @R2, uword width @R3, uword height @R4)  clobbers(A,X,Y)
 romsub $ff3b = GRAPH_set_font(uword fontptr @R0)  clobbers(A,X,Y)
-romsub $ff3e = GRAPH_get_char_size(ubyte baseline @A, ubyte width @X, ubyte height_or_style @Y, ubyte is_control @Pc)  clobbers(A,X,Y)
+romsub $ff3e = GRAPH_get_char_size(ubyte baseline @A, ubyte width @X, ubyte height_or_style @Y, bool is_control @Pc)  clobbers(A,X,Y)
 romsub $ff41 = GRAPH_put_char(uword x @R0, uword y @R1, ubyte char @A)  clobbers(A,X,Y)
 romsub $ff41 = GRAPH_put_next_char(ubyte char @A)  clobbers(A,X,Y)     ; alias for the routine above that doesn't reset the position of the initial character
 
@@ -364,16 +372,17 @@ romsub $ff1a = FB_filter_pixels(uword pointer @ R0, uword count @R1)  clobbers(A
 romsub $ff1d = FB_move_pixels(uword sx @R0, uword sy @R1, uword tx @R2, uword ty @R3, uword count @R4)  clobbers(A,X,Y)
 
 ; misc
-romsub $fec6 = i2c_read_byte(ubyte device @X, ubyte offset @Y) clobbers (X,Y) -> ubyte @A, ubyte @Pc
-romsub $fec9 = i2c_write_byte(ubyte device @X, ubyte offset @Y, ubyte data @A) clobbers (A,X,Y) -> ubyte @Pc
-romsub $fef0 = sprite_set_image(uword pixels @R0, uword mask @R1, ubyte bpp @R2, ubyte number @A, ubyte width @X, ubyte height @Y, ubyte apply_mask @Pc)  clobbers(A,X,Y) -> ubyte @Pc
+romsub $FEBA = BSAVE(ubyte zp_startaddr @ A, uword endaddr @ XY) clobbers (X, Y) -> bool @ Pc, ubyte @ A      ; like cbm.SAVE, but omits the 2-byte prg header
+romsub $fec6 = i2c_read_byte(ubyte device @X, ubyte offset @Y) clobbers (X,Y) -> ubyte @A, bool @Pc
+romsub $fec9 = i2c_write_byte(ubyte device @X, ubyte offset @Y, ubyte data @A) clobbers (A,X,Y) -> bool @Pc
+romsub $fef0 = sprite_set_image(uword pixels @R0, uword mask @R1, ubyte bpp @R2, ubyte number @A, ubyte width @X, ubyte height @Y, bool apply_mask @Pc)  clobbers(A,X,Y) -> bool @Pc
 romsub $fef3 = sprite_set_position(uword x @R0, uword y @R1, ubyte number @A)  clobbers(A,X,Y)
 romsub $fee4 = memory_fill(uword address @R0, uword num_bytes @R1, ubyte value @A)  clobbers(A,X,Y)
 romsub $fee7 = memory_copy(uword source @R0, uword target @R1, uword num_bytes @R2)  clobbers(A,X,Y)
 romsub $feea = memory_crc(uword address @R0, uword num_bytes @R1)  clobbers(A,X,Y) -> uword @R2
 romsub $feed = memory_decompress(uword input @R0, uword output @R1)  clobbers(A,X,Y) -> uword @R1       ; last address +1 is result in R1
 romsub $fedb = console_init(uword x @R0, uword y @R1, uword width @R2, uword height @R3)  clobbers(A,X,Y)
-romsub $fede = console_put_char(ubyte char @A, ubyte wrapping @Pc)  clobbers(A,X,Y)
+romsub $fede = console_put_char(ubyte char @A, bool wrapping @Pc)  clobbers(A,X,Y)
 romsub $fee1 = console_get_char()  clobbers(X,Y) -> ubyte @A
 romsub $fed8 = console_put_image(uword pointer @R0, uword width @R1, uword height @R2)  clobbers(A,X,Y)
 romsub $fed5 = console_set_paging_message(uword msgptr @R0)  clobbers(A,X,Y)
@@ -381,7 +390,7 @@ romsub $fecf = entropy_get() -> ubyte @A, ubyte @X, ubyte @Y
 romsub $fecc = monitor()  clobbers(A,X,Y)
 
 romsub $ff44 = macptr(ubyte length @A, uword buffer @XY, bool dontAdvance @Pc)  clobbers(A) -> bool @Pc, uword @XY
-romsub $ff47 = enter_basic(ubyte cold_or_warm @Pc)  clobbers(A,X,Y)
+romsub $ff47 = enter_basic(bool cold_or_warm @Pc)  clobbers(A,X,Y)
 romsub $ff4d = clock_set_date_time(uword yearmonth @R0, uword dayhours @R1, uword minsecs @R2, ubyte jiffies @R3)  clobbers(A, X, Y)
 romsub $ff50 = clock_get_date_time()  clobbers(A, X, Y)  -> uword @R0, uword @R1, uword @R2, ubyte @R3   ; result registers see clock_set_date_time()
 
@@ -398,18 +407,37 @@ romsub $ff53 = joystick_scan()  clobbers(A, X, Y)
 romsub $ff56 = joystick_get(ubyte joynr @A) -> ubyte @A, ubyte @X, ubyte @Y
 romsub $ff56 = joystick_get2(ubyte joynr @A) clobbers(Y) -> uword @AX   ; alternative to above to not have the hassle to deal with multiple return values
 
-; Audio (bank 10)
+; Audio (rom bank 10)
 romsub $C04B = psg_init() clobbers(A,X,Y)
-romsub $C063 = ym_init() clobbers(A,X,Y) -> ubyte @Pc                   ; (re)init YM chip
-romsub $C066 = ym_loaddefpatches() clobbers(A,X,Y) -> ubyte @Pc         ; load default YM patches
-romsub $C09F = audio_init() clobbers(A,X,Y) -> ubyte @Pc                ; (re)initialize PSG and YM audio chips
+romsub $C063 = ym_init() clobbers(A,X,Y) -> bool @Pc                   ; (re)init YM chip
+romsub $C066 = ym_loaddefpatches() clobbers(A,X,Y) -> bool @Pc         ; load default YM patches
+romsub $C09F = audio_init() clobbers(A,X,Y) -> bool @Pc                ; (re)initialize PSG and YM audio chips
 ; TODO: add more of the audio routines?
 
 
+asmsub set_screen_mode(ubyte mode @A) clobbers(A,X,Y) -> bool @Pc {
+    ; -- convenience wrapper for screen_mode() to just set a new mode (and return success)
+    %asm {{
+        clc
+        jmp  screen_mode
+    }}
+}
+
+asmsub get_screen_mode() -> byte @A, byte @X, byte @Y {
+    ; -- convenience wrapper for screen_mode() to just get the current mode in A, and size in characters in X+Y
+    ;    this does need a piece of inlined asm to call it ans store the result values if you call this from prog8 code
+    ;    Note: you can also just do the SEC yourself and simply call screen_mode() directly,
+    ;          or use the existing SCREEN kernal routine for just getting the size in characters.
+    %asm {{
+        sec
+        jmp  screen_mode
+    }}
+}
+
 asmsub kbdbuf_clear() {
     ; -- convenience helper routine to clear the keyboard buffer
     %asm {{
--       jsr  c64.GETIN
+-       jsr  cbm.GETIN
         bne  -
         rts
     }}
@@ -482,7 +510,7 @@ asmsub numbanks() -> uword @AY {
     %asm {{
         phx
         sec
-        jsr  c64.MEMTOP
+        jsr  cbm.MEMTOP
         ldy  #0
         cmp  #0
         bne  +
@@ -511,6 +539,8 @@ asmsub vpeek(ubyte bank @A, uword address @XY) -> ubyte @A {
 
 asmsub vaddr(ubyte bank @A, uword address @R0, ubyte addrsel @R1, byte autoIncrOrDecrByOne @Y) clobbers(A) {
         ; -- setup the VERA's data address register 0 or 1
+        ;    with optional auto increment or decrement of 1.
+        ;    Note that the vaddr_autoincr() and vaddr_autodecr() routines allow to set all possible strides, not just 1.
         %asm {{
             and  #1
             pha
@@ -534,6 +564,105 @@ asmsub vaddr(ubyte bank @A, uword address @R0, ubyte addrsel @R1, byte autoIncrO
         }}
 }
 
+asmsub vaddr_clone(ubyte port @A) clobbers (A,X,Y) {
+    ; -- clones Vera addresses from the given source port to the other one.
+    ;    leaves CTRL on the destination port.
+    %asm {{
+        sta  VERA_CTRL
+        ldx  VERA_ADDR_L
+        ldy  VERA_ADDR_H
+        phy
+        ldy  VERA_ADDR_M
+        eor  #1
+        sta  VERA_CTRL
+        stx  VERA_ADDR_L
+        sty  VERA_ADDR_M
+        ply
+        sty  VERA_ADDR_H
+        rts
+    }}
+}
+
+asmsub vaddr_autoincr(ubyte bank @A, uword address @R0, ubyte addrsel @R1, uword autoIncrAmount @R2) clobbers(A,Y) {
+        ; -- setup the VERA's data address register 0 or 1
+        ;    including setting up optional auto increment amount.
+        ;    Specifiying an unsupported amount results in amount of zero. See the Vera docs about what amounts are possible.
+        %asm {{
+            jsr  _setup
+            lda  cx16.r2H
+            ora  cx16.r2L
+            beq  +
+            jsr  _determine_incr_bits
++           ora  P8ZP_SCRATCH_REG
+            sta  cx16.VERA_ADDR_H
+            rts
+
+_setup      and  #1
+            sta  P8ZP_SCRATCH_REG
+            lda  cx16.r1
+            and  #1
+            sta  cx16.VERA_CTRL
+            lda  cx16.r0
+            sta  cx16.VERA_ADDR_L
+            lda  cx16.r0+1
+            sta  cx16.VERA_ADDR_M
+            rts
+
+_determine_incr_bits
+            lda  cx16.r2H
+            bne  _large
+            lda  cx16.r2L
+            ldy  #13
+-           cmp  _strides_lsb,y
+            beq  +
+            dey
+            bpl  -
++           tya
+            asl  a
+            asl  a
+            asl  a
+            asl  a
+            rts
+_large      ora  cx16.r2L
+            cmp  #1         ; 256
+            bne  +
+            lda  #9<<4
+            rts
++           cmp  #2         ; 512
+            bne  +
+            lda  #10<<4
+            rts
++           cmp  #65        ; 320
+            bne  +
+            lda  #14<<4
+            rts
++           cmp  #130       ; 640
+            bne  +
+            lda  #15<<4
+            rts
++           lda  #0
+            rts
+_strides_lsb    .byte   0,1,2,4,8,16,32,64,128,255,255,40,80,160,255,255
+        }}
+}
+
+asmsub vaddr_autodecr(ubyte bank @A, uword address @R0, ubyte addrsel @R1, uword autoDecrAmount @R2) clobbers(A,Y) {
+        ; -- setup the VERA's data address register 0 or 1
+        ;    including setting up optional auto decrement amount.
+        ;    Specifiying an unsupported amount results in amount of zero. See the Vera docs about what amounts are possible.
+        %asm {{
+            jsr  vaddr_autoincr._setup
+            lda  cx16.r2H
+            ora  cx16.r2L
+            beq  +
+            jsr  vaddr_autoincr._determine_incr_bits
+            ora  #%00001000         ; autodecrement
++           ora  P8ZP_SCRATCH_REG
+            sta  cx16.VERA_ADDR_H
+            rts
+        }}
+}
+
 asmsub vpoke(ubyte bank @A, uword address @R0, ubyte value @Y) clobbers(A) {
     ; -- write a single byte to VERA's video memory
     ;    note: inefficient when writing multiple sequential bytes!
@@ -604,42 +733,111 @@ asmsub vpoke_xor(ubyte bank @A, uword address @R0, ubyte value @Y) clobbers (A)
     }}
 }
 
-
-sub FB_set_pixels_from_buf(uword buffer, uword count) {
+asmsub vpoke_mask(ubyte bank @A, uword address @R0, ubyte mask @X, ubyte value @Y) clobbers (A) {
+    ; -- bitwise or a single byte to the value already in the VERA's video memory at that location
+    ;    after applying the and-mask. Note: inefficient when writing multiple sequential bytes!
     %asm {{
-            ; -- This is replacement code for the normal FB_set_pixels subroutine in ROM
-            ;    However that routine contains a bug in the current v38 ROM that makes it crash when count > 255.
-            ;    So the code below replaces that. Once the ROM is patched this routine is no longer necessary.
-            ;    See https://github.com/commanderx16/x16-rom/issues/179
-            phx
-            lda  buffer
-            ldy  buffer+1
-            sta  P8ZP_SCRATCH_W1
-            sty  P8ZP_SCRATCH_W1+1
-            jsr  _pixels
-            plx
-            rts
-
-_pixels     lda  count+1
-            beq  +
-            ldx  #0
--           jsr  _loop
-            inc  P8ZP_SCRATCH_W1+1
-            dec  count+1
-            bne  -
-
-+           ldx  count
-_loop       ldy  #0
--           lda  (P8ZP_SCRATCH_W1),y
-            sta  cx16.VERA_DATA0
-            iny
-            dex
-            bne  -
-            rts
-        }}
+        sty  P8ZP_SCRATCH_B1
+        stz  cx16.VERA_CTRL
+        and  #1
+        sta  cx16.VERA_ADDR_H
+        lda  cx16.r0
+        sta  cx16.VERA_ADDR_L
+        lda  cx16.r0+1
+        sta  cx16.VERA_ADDR_M
+        txa
+        and  cx16.VERA_DATA0
+        ora  P8ZP_SCRATCH_B1
+        sta  cx16.VERA_DATA0
+        rts
+    }}
 }
 
-; ---- system stuff -----
+asmsub save_virtual_registers() clobbers(A,Y) {
+    %asm {{
+        ldy  #31
+-       lda  cx16.r0,y
+        sta  _cx16_vreg_storage,y
+        dey
+        bpl  -
+        rts
+
+_cx16_vreg_storage
+        .word 0,0,0,0,0,0,0,0
+        .word 0,0,0,0,0,0,0,0
+    }}
+}
+
+asmsub restore_virtual_registers() clobbers(A,Y) {
+    %asm {{
+        ldy  #31
+-       lda  save_virtual_registers._cx16_vreg_storage,y
+        sta  cx16.r0,y
+        dey
+        bpl  -
+        rts
+    }}
+}
+
+
+asmsub save_vera_context() clobbers(A) {
+    ; -- use this at the start of your IRQ handler if it uses Vera registers, to save the state
+    %asm {{
+        ; note cannot store this on cpu hardware stack because this gets called as a subroutine
+        lda  cx16.VERA_ADDR_L
+        sta  _vera_storage
+        lda  cx16.VERA_ADDR_M
+        sta  _vera_storage+1
+        lda  cx16.VERA_ADDR_H
+        sta  _vera_storage+2
+        lda  cx16.VERA_CTRL
+        sta  _vera_storage+3
+        eor  #1
+        sta  _vera_storage+7
+        sta  cx16.VERA_CTRL
+        lda  cx16.VERA_ADDR_L
+        sta  _vera_storage+4
+        lda  cx16.VERA_ADDR_M
+        sta  _vera_storage+5
+        lda  cx16.VERA_ADDR_H
+        sta  _vera_storage+6
+        rts
+_vera_storage:  .byte 0,0,0,0,0,0,0,0
+    }}
+}
+
+asmsub restore_vera_context() clobbers(A) {
+    ; -- use this at the end of your IRQ handler if it uses Vera registers, to restore the state
+    %asm {{
+        lda  cx16.save_vera_context._vera_storage+7
+        sta  cx16.VERA_CTRL
+        lda  cx16.save_vera_context._vera_storage+6
+        sta  cx16.VERA_ADDR_H
+        lda  cx16.save_vera_context._vera_storage+5
+        sta  cx16.VERA_ADDR_M
+        lda  cx16.save_vera_context._vera_storage+4
+        sta  cx16.VERA_ADDR_L
+        lda  cx16.save_vera_context._vera_storage+3
+        sta  cx16.VERA_CTRL
+        lda  cx16.save_vera_context._vera_storage+2
+        sta  cx16.VERA_ADDR_H
+        lda  cx16.save_vera_context._vera_storage+1
+        sta  cx16.VERA_ADDR_M
+        lda  cx16.save_vera_context._vera_storage+0
+        sta  cx16.VERA_ADDR_L
+        rts
+    }}
+}
+
+}
+
+sys {
+    ; ------- lowlevel system routines --------
+
+    %option no_symbol_prefixing
+
+    const ubyte target = 16         ;  compilation target specifier.  64 = C64,  128 = C128,  16 = CommanderX16.
+
 asmsub  init_system()  {
     ; Initializes the machine to a sane starting state.
     ; Called automatically by the loader program logic.
@@ -650,29 +848,30 @@ asmsub  init_system()  {
         tay
         jsr  cx16.mouse_config  ; disable mouse
         cld
-        lda  VERA_DC_VIDEO
+        lda  cx16.VERA_DC_VIDEO
         and  #%00000111 ; retain chroma + output mode
         sta  P8ZP_SCRATCH_REG
         lda  #$0a
         sta  $01        ; rom bank 10 (audio)
-        jsr  audio_init ; silence
+        jsr  cx16.audio_init ; silence
         stz  $01        ; rom bank 0 (kernal)
-        jsr  c64.IOINIT
-        jsr  c64.RESTOR
-        jsr  c64.CINT
-        lda  VERA_DC_VIDEO
+        jsr  cbm.IOINIT
+        jsr  cbm.RESTOR
+        jsr  cbm.CINT
+        lda  cx16.VERA_DC_VIDEO
         and  #%11111000
         ora  P8ZP_SCRATCH_REG
-        sta  VERA_DC_VIDEO  ; restore old output mode
+        sta  cx16.VERA_DC_VIDEO  ; restore old output mode
         lda  #$90       ; black
-        jsr  c64.CHROUT
+        jsr  cbm.CHROUT
         lda  #1
-        sta  $00        ; select ram bank 1
-        jsr  c64.CHROUT ; swap fg/bg
+        jsr  cbm.CHROUT ; swap fg/bg
         lda  #$9e       ; yellow
-        jsr  c64.CHROUT
+        jsr  cbm.CHROUT
         lda  #147       ; clear screen
-        jsr  c64.CHROUT
+        jsr  cbm.CHROUT
+        lda  #PROG8_VARSHIGH_RAMBANK
+        sta  $00    ; select ram bank
         lda  #0
         tax
         tay
@@ -690,6 +889,8 @@ asmsub  init_system_phase2()  {
         sta  restore_irq._orig_irqvec
         lda  cx16.CINV+1
         sta  restore_irq._orig_irqvec+1
+        lda  #PROG8_VARSHIGH_RAMBANK
+        sta  $00    ; select ram bank
         cli
         rts
     }}
@@ -707,12 +908,12 @@ asmsub  cleanup_at_exit() {
     }}
 }
 
-asmsub  set_irq(uword handler @AY, ubyte useKernal @Pc) clobbers(A)  {
+asmsub  set_irq(uword handler @AY, bool useKernal @Pc) clobbers(A)  {
 	%asm {{
 	        sta  _modified+1
 	        sty  _modified+2
 	        lda  #0
-	        adc  #0
+	        rol  a
 	        sta  _use_kernal
 		sei
 		lda  #<_irq_handler
@@ -786,57 +987,6 @@ IRQ_SCRATCH_ZPWORD2	.word  0
 		}}
 }
 
-asmsub push_vera_context() clobbers(A) {
-    ; -- use this at the start of your IRQ handler if it uses Vera registers, to save the state
-    %asm {{
-        ; note cannot store this on cpu hardware stack because this gets called as a subroutine
-        lda  cx16.VERA_ADDR_L
-        sta  _vera_storage
-        lda  cx16.VERA_ADDR_M
-        sta  _vera_storage+1
-        lda  cx16.VERA_ADDR_H
-        sta  _vera_storage+2
-        lda  cx16.VERA_CTRL
-        sta  _vera_storage+3
-        eor  #1
-        sta  cx16.VERA_CTRL
-        lda  cx16.VERA_ADDR_L
-        sta  _vera_storage+4
-        lda  cx16.VERA_ADDR_M
-        sta  _vera_storage+5
-        lda  cx16.VERA_ADDR_H
-        sta  _vera_storage+6
-        lda  cx16.VERA_CTRL
-        sta  _vera_storage+7
-        rts
-_vera_storage:  .byte 0,0,0,0,0,0,0,0
-    }}
-}
-
-asmsub pop_vera_context() clobbers(A) {
-    ; -- use this at the end of your IRQ handler if it uses Vera registers, to restore the state
-    %asm {{
-        lda  cx16.push_vera_context._vera_storage+7
-        sta  cx16.VERA_CTRL
-        lda  cx16.push_vera_context._vera_storage+6
-        sta  cx16.VERA_ADDR_H
-        lda  cx16.push_vera_context._vera_storage+5
-        sta  cx16.VERA_ADDR_M
-        lda  cx16.push_vera_context._vera_storage+4
-        sta  cx16.VERA_ADDR_L
-        lda  cx16.push_vera_context._vera_storage+3
-        sta  cx16.VERA_CTRL
-        lda  cx16.push_vera_context._vera_storage+2
-        sta  cx16.VERA_ADDR_H
-        lda  cx16.push_vera_context._vera_storage+1
-        sta  cx16.VERA_ADDR_M
-        lda  cx16.push_vera_context._vera_storage+0
-        sta  cx16.VERA_ADDR_L
-        rts
-    }}
-}
-
-
 asmsub  restore_irq() clobbers(A) {
 	%asm {{
 	    sei
@@ -906,27 +1056,33 @@ asmsub  set_rasterline(uword line @AY) {
     }}
 }
 
-}
-
-
-sys {
-    ; ------- lowlevel system routines --------
-
-    const ubyte target = 16         ;  compilation target specifier.  64 = C64,  128 = C128,  16 = CommanderX16.
-
-
     asmsub reset_system() {
         ; Soft-reset the system back to initial power-on Basic prompt.
+        ; We do this via the SMC so that a true reset is performed that also resets the Vera fully.
         %asm {{
             sei
-            stz  $01                        ; bank the kernal in
-            jmp  (cx16.RESET_VEC)
+            ldx #$42
+            ldy #2
+            lda #0
+            jsr  cx16.i2c_write_byte
+            bra  *
         }}
     }
 
+    sub poweroff_system() {
+        ; use the SMC to shutdown the computer
+        void cx16.i2c_write_byte($42, $01, $00)
+    }
+
+    sub set_leds_brightness(ubyte activity, ubyte power) {
+        void cx16.i2c_write_byte($42, $04, power)
+        void cx16.i2c_write_byte($42, $05, activity)
+    }
+
     asmsub wait(uword jiffies @AY) {
         ; --- wait approximately the given number of jiffies (1/60th seconds) (N or N+1)
         ;     note: the system irq handler has to be active for this to work as it depends on the system jiffy clock
+        ;     note: this routine cannot be used from inside a irq handler
         %asm {{
             phx
             sta  P8ZP_SCRATCH_W1
@@ -938,9 +1094,13 @@ _loop       lda  P8ZP_SCRATCH_W1
             plx
             rts
 
-+           jsr  c64.RDTIM
++           sei
+            jsr  cbm.RDTIM
+            cli
             sta  P8ZP_SCRATCH_B1
--           jsr  c64.RDTIM
+-           sei
+            jsr  cbm.RDTIM
+            cli
             cmp  P8ZP_SCRATCH_B1
             beq  -
 
@@ -1080,10 +1240,23 @@ _longcopy
         }}
     }
 
+    inline asmsub irqsafe_set_irqd() {
+        %asm {{
+        php
+        sei
+        }}
+    }
+
+    inline asmsub irqsafe_clear_irqd() {
+        %asm {{
+        plp
+        }}
+    }
+
     inline asmsub exit(ubyte returnvalue @A) {
         ; -- immediately exit the program with a return code in the A register
         %asm {{
-            jsr  c64.CLRCHN		; reset i/o channels
+            jsr  cbm.CLRCHN		; reset i/o channels
             ldx  prog8_lib.orig_stackpointer
             txs
             rts		; return to original caller

compiler/res/prog8lib/cx16/textio.p8

@@ -6,6 +6,8 @@
 
 txt {
 
+    %option no_symbol_prefixing
+
 const ubyte DEFAULT_WIDTH = 80
 const ubyte DEFAULT_HEIGHT = 60
 
@@ -33,10 +35,10 @@ asmsub column(ubyte col @A) clobbers(A, X, Y) {
     ; ---- set the cursor on the given column (starting with 0) on the current line
     %asm {{
         sec
-        jsr  c64.PLOT
+        jsr  cbm.PLOT
         tay
         clc
-        jmp  c64.PLOT
+        jmp  cbm.PLOT
     }}
 }
 
@@ -46,7 +48,7 @@ asmsub  fill_screen (ubyte char @ A, ubyte color @ Y) clobbers(A)  {
         sty  _ly+1
         phx
         pha
-        jsr  c64.SCREEN             ; get dimensions in X/Y
+        jsr  cbm.SCREEN             ; get dimensions in X/Y
         txa
         lsr  a
         lsr  a
@@ -94,7 +96,7 @@ asmsub  clear_screenchars (ubyte char @ A) clobbers(Y)  {
 	%asm {{
         phx
         pha
-        jsr  c64.SCREEN             ; get dimensions in X/Y
+        jsr  cbm.SCREEN             ; get dimensions in X/Y
         txa
         lsr  a
         lsr  a
@@ -125,7 +127,7 @@ asmsub  clear_screencolors (ubyte color @ A) clobbers(Y)  {
 	%asm {{
         phx
         sta  _la+1
-        jsr  c64.SCREEN             ; get dimensions in X/Y
+        jsr  cbm.SCREEN             ; get dimensions in X/Y
         txa
         lsr  a
         lsr  a
@@ -158,35 +160,35 @@ ubyte[16] color_to_charcode = [$90,$05,$1c,$9f,$9c,$1e,$1f,$9e,$81,$95,$96,$97,$
 
 sub color (ubyte txtcol) {
     txtcol &= 15
-    c64.CHROUT(color_to_charcode[txtcol])
+    cbm.CHROUT(color_to_charcode[txtcol])
 }
 
 sub color2 (ubyte txtcol, ubyte bgcol) {
     txtcol &= 15
     bgcol &= 15
-    c64.CHROUT(color_to_charcode[bgcol])
-    c64.CHROUT(1)       ; switch fg and bg colors
-    c64.CHROUT(color_to_charcode[txtcol])
+    cbm.CHROUT(color_to_charcode[bgcol])
+    cbm.CHROUT(1)       ; switch fg and bg colors
+    cbm.CHROUT(color_to_charcode[txtcol])
 }
 
 sub lowercase() {
-    c64.CHROUT($0e)
+    cbm.CHROUT($0e)
     ; this is not 100% compatible: cx16.screen_set_charset(3, 0)  ; lowercase petscii charset
 }
 
 sub uppercase() {
-    c64.CHROUT($8e)
+    cbm.CHROUT($8e)
     ; this is not 100% compatible: cx16.screen_set_charset(2, 0)  ; uppercase petscii charset
 }
 
 sub iso() {
-    c64.CHROUT($0f)
+    cbm.CHROUT($0f)
     ; This doesn't enable it completely: cx16.screen_set_charset(1, 0)  ; iso charset
 }
 
 sub iso_off() {
     ; -- you have to call this first when switching back from iso charset to regular charset.
-    c64.CHROUT($8f)
+    cbm.CHROUT($8f)
 }
 
 
@@ -195,7 +197,7 @@ asmsub  scroll_left() clobbers(A, Y)  {
 	;      contents of the rightmost column are unchanged, you should clear/refill this yourself
 	%asm {{
 	    phx
-	    jsr  c64.SCREEN
+	    jsr  cbm.SCREEN
 	    dex
 	    stx  _lx+1
         dey
@@ -236,12 +238,12 @@ _lx     ldx  #0                ; modified
 	}}
 }
 
-asmsub  scroll_right() clobbers(A)  {
+asmsub  scroll_right() clobbers(A,Y)  {
 	; ---- scroll the whole screen 1 character to the right
 	;      contents of the leftmost column are unchanged, you should clear/refill this yourself
 	%asm {{
 	    phx
-	    jsr  c64.SCREEN
+	    jsr  cbm.SCREEN
 	    dex
 	    stx  _lx+1
 	    txa
@@ -295,7 +297,7 @@ asmsub  scroll_up() clobbers(A, Y)  {
 	;      contents of the bottom row are unchanged, you should refill/clear this yourself
 	%asm {{
 	    phx
-	    jsr  c64.SCREEN
+	    jsr  cbm.SCREEN
 	    stx  _nextline+1
 	    dey
         sty  P8ZP_SCRATCH_B1
@@ -345,7 +347,7 @@ asmsub  scroll_down() clobbers(A, Y)  {
 	;      contents of the top row are unchanged, you should refill/clear this yourself
 	%asm {{
 	    phx
-	    jsr  c64.SCREEN
+	    jsr  cbm.SCREEN
 	    stx  _nextline+1
 	    dey
         sty  P8ZP_SCRATCH_B1
@@ -396,20 +398,20 @@ _nextline
 	}}
 }
 
-romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use c64.CHROUT directly ofcourse.
+romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use cbm.CHROUT directly ofcourse.
 
 asmsub  print (str text @ AY) clobbers(A,Y)  {
 	; ---- print null terminated string from A/Y
 	; note: the compiler contains an optimization that will replace
 	;       a call to this subroutine with a string argument of just one char,
-	;       by just one call to c64.CHROUT of that single char.
+	;       by just one call to cbm.CHROUT of that single char.
 	%asm {{
 		sta  P8ZP_SCRATCH_B1
 		sty  P8ZP_SCRATCH_REG
 		ldy  #0
 -		lda  (P8ZP_SCRATCH_B1),y
 		beq  +
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		bne  -
 +		rts
@@ -423,11 +425,11 @@ asmsub  print_ub0  (ubyte value @ A) clobbers(A,Y)  {
 		jsr  conv.ubyte2decimal
 		pha
 		tya
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		pla
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		txa
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		plx
 		rts
 	}}
@@ -443,16 +445,16 @@ _print_byte_digits
 		cpy  #'0'
 		beq  +
 		tya
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		pla
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		bra  _ones
 +       pla
         cmp  #'0'
         beq  _ones
-        jsr  c64.CHROUT
+        jsr  cbm.CHROUT
 _ones   txa
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		plx
 		rts
 	}}
@@ -466,45 +468,45 @@ asmsub  print_b  (byte value @ A) clobbers(A,Y)  {
 		cmp  #0
 		bpl  +
 		lda  #'-'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 +		pla
 		jsr  conv.byte2decimal
 		bra  print_ub._print_byte_digits
 	}}
 }
 
-asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_ubhex  (ubyte value @ A, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in hex form (if Carry is set, a radix prefix '$' is printed as well)
 	%asm {{
 		phx
 		bcc  +
 		pha
 		lda  #'$'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		pla
 +		jsr  conv.ubyte2hex
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		tya
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		plx
 		rts
 	}}
 }
 
-asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_ubbin  (ubyte value @ A, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in binary form (if Carry is set, a radix prefix '%' is printed as well)
 	%asm {{
 		phx
 		sta  P8ZP_SCRATCH_B1
 		bcc  +
 		lda  #'%'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 +		ldy  #8
 -		lda  #'0'
 		asl  P8ZP_SCRATCH_B1
 		bcc  +
 		lda  #'1'
-+		jsr  c64.CHROUT
++		jsr  cbm.CHROUT
 		dey
 		bne  -
 		plx
@@ -512,7 +514,7 @@ asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_uwbin  (uword value @ AY, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in binary form (if Carry is set, a radix prefix '%' is printed as well)
 	%asm {{
 		pha
@@ -524,7 +526,7 @@ asmsub  print_uwbin  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
 	}}
 }
 
-asmsub  print_uwhex  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
+asmsub  print_uwhex  (uword value @ AY, bool prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in hexadecimal form (4 digits)
 	;      (if Carry is set, a radix prefix '$' is printed as well)
 	%asm {{
@@ -545,7 +547,7 @@ asmsub  print_uw0  (uword value @ AY) clobbers(A,Y)  {
 		ldy  #0
 -		lda  conv.uword2decimal.decTenThousands,y
         beq  +
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		bne  -
 +		plx
@@ -568,14 +570,14 @@ asmsub  print_uw  (uword value @ AY) clobbers(A,Y)  {
 		bne  -
 
 _gotdigit
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		lda  conv.uword2decimal.decTenThousands,y
 		bne  _gotdigit
 		rts
 _allzero
         lda  #'0'
-        jmp  c64.CHROUT
+        jmp  cbm.CHROUT
 	}}
 }
 
@@ -586,7 +588,7 @@ asmsub  print_w  (word value @ AY) clobbers(A,Y)  {
 		bpl  +
 		pha
 		lda  #'-'
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		tya
 		eor  #255
 		tay
@@ -607,7 +609,7 @@ asmsub  input_chars  (uword buffer @ AY) clobbers(A) -> ubyte @ Y  {
 		sta  P8ZP_SCRATCH_W1
 		sty  P8ZP_SCRATCH_W1+1
 		ldy  #0				; char counter = 0
--		jsr  c64.CHRIN
+-		jsr  cbm.CHRIN
 		cmp  #$0d			; return (ascii 13) pressed?
 		beq  +				; yes, end.
 		sta  (P8ZP_SCRATCH_W1),y	; else store char in buffer
@@ -769,7 +771,7 @@ asmsub  plot  (ubyte col @ Y, ubyte row @ A) clobbers(A) {
 		phx
 		tax
 		clc
-		jsr  c64.PLOT
+		jsr  cbm.PLOT
 		plx
 		rts
 	}}
@@ -778,7 +780,7 @@ asmsub  plot  (ubyte col @ Y, ubyte row @ A) clobbers(A) {
 asmsub width() clobbers(X,Y) -> ubyte @A {
     ; -- returns the text screen width (number of columns)
     %asm {{
-        jsr  c64.SCREEN
+        jsr  cbm.SCREEN
         txa
         rts
     }}
@@ -787,7 +789,7 @@ asmsub width() clobbers(X,Y) -> ubyte @A {
 asmsub height() clobbers(X, Y) -> ubyte @A {
     ; -- returns the text screen height (number of rows)
     %asm {{
-        jsr  c64.SCREEN
+        jsr  cbm.SCREEN
         tya
         rts
     }}

compiler/res/prog8lib/cx16logo.p8

@@ -3,6 +3,8 @@
 %import textio
 
 cx16logo {
+    %option no_symbol_prefixing
+
     sub logo_at(ubyte column, ubyte row) {
         uword strptr
         for strptr in logo_lines {

compiler/res/prog8lib/diskio.p8

@@ -1,59 +1,66 @@
-; C64 and Cx16 disk drive I/O routines.
+; C64/C128 disk drive I/O routines.
 
 %import textio
 %import string
 %import syslib
 
 diskio {
+    %option no_symbol_prefixing
 
-    sub directory(ubyte drivenumber) -> bool {
-        ; -- Prints the directory contents of disk drive 8-11 to the screen. Returns success.
+    ubyte drivenumber = 8
 
-        c64.SETNAM(1, "$")
-        c64.SETLFS(12, drivenumber, 0)
+    sub set_drive(ubyte number) {
+        drivenumber = number
+    }
+
+    sub directory() -> bool {
+        ; -- Prints the directory contents to the screen. Returns success.
+
+        cbm.SETNAM(1, "$")
+        cbm.SETLFS(12, drivenumber, 0)
         ubyte status = 1
-        void c64.OPEN()          ; open 12,8,0,"$"
+        void cbm.OPEN()          ; open 12,8,0,"$"
         if_cs
             goto io_error
-        void c64.CHKIN(12)        ; use #12 as input channel
+        void cbm.CHKIN(12)        ; use #12 as input channel
         if_cs
             goto io_error
 
         repeat 4 {
-            void c64.CHRIN()     ; skip the 4 prologue bytes
+            void cbm.CHRIN()     ; skip the 4 prologue bytes
         }
 
         ; while not stop key pressed / EOF encountered, read data.
-        status = c64.READST()
+        status = cbm.READST()
         if status!=0 {
             status = 1
             goto io_error
         }
 
         while status==0 {
-            ubyte low = c64.CHRIN()
-            ubyte high = c64.CHRIN()
+            ubyte low = cbm.CHRIN()
+            ubyte high = cbm.CHRIN()
             txt.print_uw(mkword(high, low))
             txt.spc()
             ubyte @zp char
             repeat {
-                char = c64.CHRIN()
+                char = cbm.CHRIN()
                 if char==0
                     break
                 txt.chrout(char)
             }
             txt.nl()
-            void c64.CHRIN()     ; skip 2 bytes
-            void c64.CHRIN()
-            status = c64.READST()
-            if c64.STOP2()
+            void cbm.CHRIN()     ; skip 2 bytes
+            void cbm.CHRIN()
+            status = cbm.READST()
+            if cbm.STOP2()
                 break
         }
-        status = c64.READST()
+        status = cbm.READST()
 
 io_error:
-        c64.CLRCHN()        ; restore default i/o devices
-        c64.CLOSE(12)
+        cbm.CLRCHN()        ; restore default i/o devices
+        cbm.CLOSE(12)
 
         if status and status & $40 == 0 {            ; bit 6=end of file
             txt.print("\ni/o error, status: ")
@@ -65,37 +72,43 @@ io_error:
         return true
     }
 
-    sub diskname(ubyte drivenumber) -> uword {
+    sub diskname() -> uword {
         ; -- Returns pointer to disk name string or 0 if failure.
 
-        c64.SETNAM(1, "$")
-        c64.SETLFS(12, drivenumber, 0)
+        cbm.SETNAM(1, "$")
+        cbm.SETLFS(12, drivenumber, 0)
         ubyte okay = false
-        void c64.OPEN()          ; open 12,8,0,"$"
+        void cbm.OPEN()          ; open 12,8,0,"$"
         if_cs
             goto io_error
-        void c64.CHKIN(12)        ; use #12 as input channel
+        void cbm.CHKIN(12)        ; use #12 as input channel
         if_cs
             goto io_error
 
-        repeat 6 {
-            void c64.CHRIN()     ; skip the 6 prologue bytes
+        while cbm.CHRIN()!='"' {
+            ; skip up to entry name
         }
-        if c64.READST()!=0
+        if cbm.READST()!=0
             goto io_error
 
         cx16.r0 = &list_filename
         repeat {
-            @(cx16.r0) = c64.CHRIN()
-            if @(cx16.r0)==0
+            @(cx16.r0) = cbm.CHRIN()
+            if @(cx16.r0)=='"' {
+                @(cx16.r0) = ' '
+                while @(cx16.r0)==' ' and cx16.r0>=&diskio.list_filename {
+                    @(cx16.r0) = 0
+                    cx16.r0--
+                }
                 break
+            }
             cx16.r0++
         }
         okay = true
 
 io_error:
-        c64.CLRCHN()        ; restore default i/o devices
-        c64.CLOSE(12)
+        cbm.CLRCHN()        ; restore default i/o devices
+        cbm.CLOSE(12)
         if okay
             return &list_filename
         return 0
@@ -106,13 +119,12 @@ io_error:
     uword list_pattern
     uword list_blocks
     bool iteration_in_progress = false
-    ubyte last_drivenumber = 8       ; which drive was last used for a f_open operation?
     str list_filetype = "???"       ; prg, seq, dir
     str list_filename = "?" * 50
 
     ; ----- get a list of files (uses iteration functions internally) -----
 
-    sub list_filenames(ubyte drivenumber, uword pattern_ptr, uword filenames_buffer, uword filenames_buf_size) -> ubyte {
+    sub list_filenames(uword pattern_ptr, uword filenames_buffer, uword filenames_buf_size) -> ubyte {
         ; -- fill the provided buffer with the names of the files on the disk (until buffer is full).
         ;    Files in the buffer are separeted by a 0 byte. You can provide an optional pattern to match against.
         ;    After the last filename one additional 0 byte is placed to indicate the end of the list.
@@ -120,10 +132,10 @@ io_error:
         ;    Also sets carry on exit: Carry clear = all files returned, Carry set = directory has more files that didn't fit in the buffer.
         uword buffer_start = filenames_buffer
         ubyte files_found = 0
-        if lf_start_list(drivenumber, pattern_ptr) {
+        if lf_start_list(pattern_ptr) {
             while lf_next_entry() {
                 if list_filetype!="dir" {
-                    filenames_buffer += string.copy(diskio.list_filename, filenames_buffer) + 1
+                    filenames_buffer += string.copy(list_filename, filenames_buffer) + 1
                     files_found++
                     if filenames_buffer - buffer_start > filenames_buf_size-20 {
                         @(filenames_buffer)=0
@@ -142,7 +154,7 @@ io_error:
 
     ; ----- iterative file lister functions (uses io channel 12) -----
 
-    sub lf_start_list(ubyte drivenumber, uword pattern_ptr) -> bool {
+    sub lf_start_list(uword pattern_ptr) -> bool {
         ; -- start an iterative file listing with optional pattern matching.
         ;    note: only a single iteration loop can be active at a time!
         lf_end_list()
@@ -150,20 +162,20 @@ io_error:
         list_skip_disk_name = true
         iteration_in_progress = true
 
-        c64.SETNAM(1, "$")
-        c64.SETLFS(12, drivenumber, 0)
-        void c64.OPEN()          ; open 12,8,0,"$"
+        cbm.SETNAM(1, "$")
+        cbm.SETLFS(12, drivenumber, 0)
+        void cbm.OPEN()          ; open 12,8,0,"$"
         if_cs
             goto io_error
-        void c64.CHKIN(12)        ; use #12 as input channel
+        void cbm.CHKIN(12)        ; use #12 as input channel
         if_cs
             goto io_error
 
         repeat 4 {
-            void c64.CHRIN()     ; skip the 4 prologue bytes
+            void cbm.CHRIN()     ; skip the 4 prologue bytes
         }
 
-        if c64.READST()==0
+        if cbm.READST()==0
             return true
 
 io_error:
@@ -180,26 +192,26 @@ io_error:
             return false
 
         repeat {
-            void c64.CHKIN(12)        ; use #12 as input channel again
+            void cbm.CHKIN(12)        ; use #12 as input channel again
 
             uword nameptr = &list_filename
-            ubyte blocks_lsb = c64.CHRIN()
-            ubyte blocks_msb = c64.CHRIN()
+            ubyte blocks_lsb = cbm.CHRIN()
+            ubyte blocks_msb = cbm.CHRIN()
 
-            if c64.READST()
+            if cbm.READST()
                 goto close_end
 
             list_blocks = mkword(blocks_msb, blocks_lsb)
 
             ; read until the filename starts after the first "
-            while c64.CHRIN()!='\"'  {
-                if c64.READST()
+            while cbm.CHRIN()!='\"'  {
+                if cbm.READST()
                     goto close_end
             }
 
             ; read the filename
             repeat {
-                ubyte char = c64.CHRIN()
+                ubyte char = cbm.CHRIN()
                 if char==0
                     break
                 if char=='\"'
@@ -211,17 +223,17 @@ io_error:
             @(nameptr) = 0
 
             do {
-                cx16.r15L = c64.CHRIN()
+                cx16.r15L = cbm.CHRIN()
             } until cx16.r15L!=' '      ; skip blanks up to 3 chars entry type
             list_filetype[0] = cx16.r15L
-            list_filetype[1] = c64.CHRIN()
-            list_filetype[2] = c64.CHRIN()
-            while c64.CHRIN() {
+            list_filetype[1] = cbm.CHRIN()
+            list_filetype[2] = cbm.CHRIN()
+            while cbm.CHRIN() {
                 ; read the rest of the entry until the end
             }
 
-            void c64.CHRIN()     ; skip 2 bytes
-            void c64.CHRIN()
+            void cbm.CHRIN()     ; skip 2 bytes
+            void cbm.CHRIN()
 
             if not list_skip_disk_name {
                 if not list_pattern
@@ -240,8 +252,8 @@ close_end:
     sub lf_end_list() {
         ; -- end an iterative file listing session (close channels).
         if iteration_in_progress {
-            c64.CLRCHN()
-            c64.CLOSE(12)
+            cbm.CLRCHN()
+            cbm.CLOSE(12)
             iteration_in_progress = false
         }
     }
@@ -249,25 +261,24 @@ close_end:
 
     ; ----- iterative file loader functions (uses io channel 12) -----
 
-    sub f_open(ubyte drivenumber, uword filenameptr) -> bool {
+    sub f_open(uword filenameptr) -> bool {
         ; -- open a file for iterative reading with f_read
         ;    note: only a single iteration loop can be active at a time!
         f_close()
 
-        c64.SETNAM(string.length(filenameptr), filenameptr)
-        c64.SETLFS(12, drivenumber, 12)     ; note: has to be 12,x,12 because otherwise f_seek doesn't work
-        last_drivenumber = drivenumber
-        void c64.OPEN()          ; open 12,8,12,"filename"
+        cbm.SETNAM(string.length(filenameptr), filenameptr)
+        cbm.SETLFS(12, drivenumber, 12)     ; note: has to be 12,x,12 because otherwise f_seek doesn't work
+        void cbm.OPEN()          ; open 12,8,12,"filename"
         if_cc {
-            if c64.READST()==0 {
+            if cbm.READST()==0 {
                 iteration_in_progress = true
-                void c64.CHKIN(12)          ; use #12 as input channel
+                void cbm.CHKIN(12)          ; use #12 as input channel
                 if_cc {
-                    void c64.CHRIN()        ; read first byte to test for file not found
-                    if not c64.READST() {
-                        c64.CLOSE(12)           ; close file because we already consumed first byte
-                        void c64.OPEN()         ; re-open the file
-                        void c64.CHKIN(12)
+                    void cbm.CHRIN()        ; read first byte to test for file not found
+                    if not cbm.READST() {
+                        cbm.CLOSE(12)           ; close file because we already consumed first byte
+                        void cbm.OPEN()         ; re-open the file
+                        void cbm.CHKIN(12)
                         return true
                     }
                 }
@@ -280,9 +291,6 @@ close_end:
     sub f_read(uword bufferpointer, uword num_bytes) -> uword {
         ; -- read from the currently open file, up to the given number of bytes.
         ;    returns the actual number of bytes read.  (checks for End-of-file and error conditions)
-        ;    NOTE: on systems with banked ram (such as Commander X16) this routine DOES NOT
-        ;          automatically load into subsequent banks if it reaches a bank boundary!
-        ;          Consider using cx16diskio.f_read() on X16.
         if not iteration_in_progress or not num_bytes
             return 0
 
@@ -295,14 +303,14 @@ close_end:
             sta  m_in_buffer+2
         }}
         while num_bytes {
-            if c64.READST() {
+            if cbm.READST() {
                 f_close()
-                if c64.READST() & $40    ; eof?
+                if cbm.READST() & $40    ; eof?
                     return list_blocks   ; number of bytes read
                 return 0  ; error.
             }
             %asm {{
-                jsr  c64.CHRIN
+                jsr  cbm.CHRIN
 m_in_buffer     sta  $ffff
                 inc  m_in_buffer+1
                 bne  +
@@ -317,12 +325,11 @@ m_in_buffer     sta  $ffff
 
     sub f_read_all(uword bufferpointer) -> uword {
         ; -- read the full contents of the file, returns number of bytes read.
-        ;    Note: Consider using cx16diskio.f_read_all() on X16!
         if not iteration_in_progress
             return 0
 
         uword total_read = 0
-        while not c64.READST() {
+        while not cbm.READST() {
             cx16.r0 = f_read(bufferpointer, 256)
             total_read += cx16.r0
             bufferpointer += cx16.r0
@@ -340,9 +347,9 @@ m_in_buffer     sta  $ffff
             sta  P8ZP_SCRATCH_W1
             sty  P8ZP_SCRATCH_W1+1
             ldx  #12
-            jsr  c64.CHKIN              ; use channel 12 again for input
+            jsr  cbm.CHKIN              ; use channel 12 again for input
             ldy  #0
-_loop       jsr  c64.CHRIN
+_loop       jsr  cbm.CHRIN
             sta  (P8ZP_SCRATCH_W1),y
             beq  _end
             iny
@@ -357,12 +364,11 @@ _end        rts
         }}
     }
 
-
     sub f_close() {
         ; -- end an iterative file loading session (close channels).
         if iteration_in_progress {
-            c64.CLRCHN()
-            c64.CLOSE(12)
+            cbm.CLRCHN()
+            cbm.CLOSE(12)
             iteration_in_progress = false
         }
     }
@@ -370,16 +376,16 @@ _end        rts
 
     ; ----- iterative file writing functions (uses io channel 13) -----
 
-    sub f_open_w(ubyte drivenumber, uword filenameptr) -> bool {
+    sub f_open_w(uword filenameptr) -> bool {
         ; -- open a file for iterative writing with f_write
         f_close_w()
 
-        c64.SETNAM(string.length(filenameptr), filenameptr)
-        c64.SETLFS(13, drivenumber, 1)
-        void c64.OPEN()             ; open 13,8,1,"filename"
+        cbm.SETNAM(string.length(filenameptr), filenameptr)
+        cbm.SETLFS(13, drivenumber, 1)
+        void cbm.OPEN()             ; open 13,8,1,"filename"
         if_cc {
-            c64.CHKOUT(13)          ; use #13 as output channel
-            return not c64.READST()
+            cbm.CHKOUT(13)          ; use #13 as output channel
+            return not cbm.READST()
         }
         f_close_w()
         return false
@@ -388,39 +394,39 @@ _end        rts
     sub f_write(uword bufferpointer, uword num_bytes) -> bool {
         ; -- write the given number of bytes to the currently open file
         if num_bytes!=0 {
-            c64.CHKOUT(13)        ; use #13 as output channel again
+            cbm.CHKOUT(13)        ; use #13 as output channel again
             repeat num_bytes {
-                c64.CHROUT(@(bufferpointer))
+                cbm.CHROUT(@(bufferpointer))
                 bufferpointer++
             }
-            return not c64.READST()
+            return not cbm.READST()
         }
         return true
     }
 
     sub f_close_w() {
         ; -- end an iterative file writing session (close channels).
-        c64.CLRCHN()
-        c64.CLOSE(13)
+        cbm.CLRCHN()
+        cbm.CLOSE(13)
     }
 
 
     ; ---- other functions ----
 
-    sub status(ubyte drivenumber) -> uword {
+    sub status() -> uword {
         ; -- retrieve the disk drive's current status message
         uword messageptr = &list_filename
-        c64.SETNAM(0, list_filename)
-        c64.SETLFS(15, drivenumber, 15)
-        void c64.OPEN()          ; open 15,8,15
+        cbm.SETNAM(0, list_filename)
+        cbm.SETLFS(15, drivenumber, 15)
+        void cbm.OPEN()          ; open 15,8,15
         if_cs
             goto io_error
-        void c64.CHKIN(15)        ; use #15 as input channel
+        void cbm.CHKIN(15)        ; use #15 as input channel
         if_cs
             goto io_error
 
-        while not c64.READST() {
-            cx16.r5L = c64.CHRIN()
+        while not cbm.READST() {
+            cx16.r5L = cbm.CHRIN()
             if cx16.r5L=='\r' or cx16.r5L=='\n'
                 break
             @(messageptr) = cx16.r5L
@@ -429,8 +435,8 @@ _end        rts
         @(messageptr) = 0
 
 done:
-        c64.CLRCHN()        ; restore default i/o devices
-        c64.CLOSE(15)
+        cbm.CLRCHN()        ; restore default i/o devices
+        cbm.CLOSE(15)
         return list_filename
 
 io_error:
@@ -438,9 +444,9 @@ io_error:
         goto done
     }
 
-    sub save(ubyte drivenumber, uword filenameptr, uword address, uword size) -> bool {
-        c64.SETNAM(string.length(filenameptr), filenameptr)
-        c64.SETLFS(1, drivenumber, 0)
+    sub save(uword filenameptr, uword address, uword size) -> bool {
+        cbm.SETNAM(string.length(filenameptr), filenameptr)
+        cbm.SETLFS(1, drivenumber, 0)
         uword @shared end_address = address + size
         cx16.r0L = 0
 
@@ -453,17 +459,17 @@ io_error:
             lda  #<P8ZP_SCRATCH_W1
             ldx  end_address
             ldy  end_address+1
-            jsr  c64.SAVE
+            jsr  cbm.SAVE
             php
             ldx  P8ZP_SCRATCH_REG
             plp
         }}
 
         if_cc
-            cx16.r0L = c64.READST()==0
+            cx16.r0L = cbm.READST()==0
 
-        c64.CLRCHN()
-        c64.CLOSE(1)
+        cbm.CLRCHN()
+        cbm.CLOSE(1)
 
         return cx16.r0L
     }
@@ -474,99 +480,77 @@ io_error:
     ; If you specify a custom address_override, the first 2 bytes in the file are ignored
     ; and the rest is loaded at the given location in memory.
     ; Returns the end load address+1 if successful or 0 if a load error occurred.
-    ; NOTE: when the load is larger than 64Kb and/or spans multiple RAM banks
-    ;       (which is possible on the Commander X16), the returned size is not correct,
-    ;       because it doesn't take the number of ram banks into account.
-    ;       Also consider using cx16diskio.load() instead  on the Commander X16.
-    sub load(ubyte drivenumber, uword filenameptr, uword address_override) -> uword {
-        return internal_load_routine(drivenumber, filenameptr, address_override, false)
-    }
-
-    ; Use kernal LOAD routine to load the given file in memory.
-    ; INCLUDING the first 2 bytes in the file: no program header is assumed in the file.
-    ; This is different from Basic's LOAD instruction which always skips the first two bytes.
-    ; The load address is mandatory.
-    ; Returns the end load address+1 if successful or 0 if a load error occurred.
-    ; NOTE: when the load is larger than 64Kb and/or spans multiple RAM banks
-    ;       (which is possible on the Commander X16), the returned size is not correct,
-    ;       because it doesn't take the number of ram banks into account.
-    ;       Also consider using cx16diskio.load_raw() instead on the Commander X16.
-    sub load_raw(ubyte drivenumber, uword filenameptr, uword address) -> uword {
-        if sys.target==16   ; are we on commander X16?
-            return internal_load_routine(drivenumber, filenameptr, address, true)
-        ; fallback to reading the 2 header bytes separately
-        if not f_open(drivenumber, filenameptr)
-            return 0
-        cx16.r1 = f_read(address, 2)
-        f_close()
-        if cx16.r1!=2
-            return 0
-        address += 2
-        return load(drivenumber, filenameptr, address)
-    }
-
-
-    ; Internal routine, only to be used on Commander X16 platform if headerless=true,
-    ; because this routine uses kernal support for that to load headerless files.
-    ; On C64 it will always be called with headerless=false.
-    sub internal_load_routine(ubyte drivenumber, uword filenameptr, uword address_override, bool headerless) -> uword {
-        c64.SETNAM(string.length(filenameptr), filenameptr)
+    sub load(uword filenameptr, uword address_override) -> uword {
+        cbm.SETNAM(string.length(filenameptr), filenameptr)
         ubyte secondary = 1
         cx16.r1 = 0
         if address_override
             secondary = 0
-        if headerless
-            secondary |= %00000010  ; activate cx16 kernal headerless load support
-        c64.SETLFS(1, drivenumber, secondary)
+        cbm.SETLFS(1, drivenumber, secondary)
         %asm {{
             stx  P8ZP_SCRATCH_REG
             lda  #0
             ldx  address_override
             ldy  address_override+1
-            jsr  c64.LOAD
+            jsr  cbm.LOAD
             bcs  +
             stx  cx16.r1
             sty  cx16.r1+1
 +           ldx  P8ZP_SCRATCH_REG
         }}
 
-        c64.CLRCHN()
-        c64.CLOSE(1)
+        cbm.CLRCHN()
+        cbm.CLOSE(1)
         return cx16.r1
     }
 
-    sub delete(ubyte drivenumber, uword filenameptr) {
+    ; Identical to load(), but DOES INCLUDE the first 2 bytes in the file.
+    ; No program header is assumed in the file. Everything is loaded.
+    ; See comments on load() for more details.
+    sub load_raw(uword filenameptr, uword address) -> uword {
+        ; read the 2 header bytes separately to skip them
+        if not f_open(filenameptr)
+            return 0
+        cx16.r1 = f_read(address, 2)
+        f_close()
+        if cx16.r1!=2
+            return 0
+        address += 2
+        return load(filenameptr, address)
+    }
+
+    sub delete(uword filenameptr) {
         ; -- delete a file on the drive
         list_filename[0] = 's'
         list_filename[1] = ':'
         ubyte flen = string.copy(filenameptr, &list_filename+2)
-        c64.SETNAM(flen+2, list_filename)
-        c64.SETLFS(1, drivenumber, 15)
-        void c64.OPEN()
-        c64.CLRCHN()
-        c64.CLOSE(1)
+        cbm.SETNAM(flen+2, list_filename)
+        cbm.SETLFS(1, drivenumber, 15)
+        void cbm.OPEN()
+        cbm.CLRCHN()
+        cbm.CLOSE(1)
     }
 
-    sub rename(ubyte drivenumber, uword oldfileptr, uword newfileptr) {
+    sub rename(uword oldfileptr, uword newfileptr) {
         ; -- rename a file on the drive
         list_filename[0] = 'r'
         list_filename[1] = ':'
         ubyte flen_new = string.copy(newfileptr, &list_filename+2)
         list_filename[flen_new+2] = '='
         ubyte flen_old = string.copy(oldfileptr, &list_filename+3+flen_new)
-        c64.SETNAM(3+flen_new+flen_old, list_filename)
-        c64.SETLFS(1, drivenumber, 15)
-        void c64.OPEN()
-        c64.CLRCHN()
-        c64.CLOSE(1)
+        cbm.SETNAM(3+flen_new+flen_old, list_filename)
+        cbm.SETLFS(1, drivenumber, 15)
+        void cbm.OPEN()
+        cbm.CLRCHN()
+        cbm.CLOSE(1)
     }
 
-    sub send_command(ubyte drivenumber, uword commandptr) {
+    sub send_command(uword commandptr) {
         ; -- send a dos command to the drive
-        c64.SETNAM(string.length(commandptr), commandptr)
-        c64.SETLFS(15, drivenumber, 15)
-        void c64.OPEN()
-        c64.CLRCHN()
-        c64.CLOSE(15)
+        cbm.SETNAM(string.length(commandptr), commandptr)
+        cbm.SETLFS(15, drivenumber, 15)
+        void cbm.OPEN()
+        cbm.CLRCHN()
+        cbm.CLOSE(15)
     }
 }

compiler/res/prog8lib/floats_functions.p8

@@ -1,6 +1,6 @@
 floats {
     ; the floating point functions shared across compiler targets
-    %option merge
+    %option merge, no_symbol_prefixing
 
 sub print_f(float value) {
 	; ---- prints the floating point value (without a newline).
@@ -15,7 +15,7 @@ sub print_f(float value) {
 		ldy  #0
 -		lda  (P8ZP_SCRATCH_W1),y
 		beq  +
-		jsr  c64.CHROUT
+		jsr  cbm.CHROUT
 		iny
 		bne  -
 +		ldx  floats_store_reg
@@ -39,18 +39,6 @@ sub pow(float value, float power) -> float {
     }}
 }
 
-sub fabs(float value) -> float {
-    %asm {{
-        stx  P8ZP_SCRATCH_REG
-        lda  #<value
-        ldy  #>value
-        jsr  MOVFM
-        jsr  ABS
-        ldx  P8ZP_SCRATCH_REG
-        rts
-    }}
-}
-
 sub sin(float angle) -> float {
     %asm {{
         lda  #<angle
@@ -128,18 +116,6 @@ sub log2(float value) -> float {
     }}
 }
 
-sub sqrt(float value) -> float {
-    %asm {{
-        lda  #<value
-        ldy  #>value
-        jsr  MOVFM
-        stx  P8ZP_SCRATCH_REG
-        jsr  SQR
-        ldx  P8ZP_SCRATCH_REG
-        rts
-    }}
-}
-
 sub rad(float angle) -> float {
     ; -- convert degrees to radians (d * pi / 180)
     %asm {{
@@ -237,4 +213,27 @@ sub rndseedf(float seed) {
     }}
 }
 
+
+sub minf(float f1, float f2) -> float {
+    if f1<f2
+        return f1
+    return f2
+}
+
+
+sub maxf(float f1, float f2) -> float {
+    if f1>f2
+        return f1
+    return f2
+}
+
+
+sub clampf(float value, float minimum, float maximum) -> float {
+    if value>maximum
+      value=maximum
+    if value>minimum
+      return value
+    return minimum
+}
+
 }

compiler/res/prog8lib/math.asm

@@ -51,8 +51,7 @@ multiply_bytes_into_word	.proc
 multiply_words	.proc
 	; -- multiply two 16-bit words into a 32-bit result  (signed and unsigned)
 	;      input: A/Y = first 16-bit number, P8ZP_SCRATCH_W1 in ZP = second 16-bit number
-	;      output: multiply_words.result  4-bytes/32-bits product, LSB order (low-to-high)
-	;      clobbers: A
+	;      output: multiply_words.result  4-bytes/32-bits product, LSB order (low-to-high)  low 16 bits also in AY.
 
 		sta  P8ZP_SCRATCH_W2
 		sty  P8ZP_SCRATCH_W2+1
@@ -79,6 +78,8 @@ mult16		lda  #0
 		dex
 		bne  -
 		ldx  P8ZP_SCRATCH_REG
+		lda  result
+		ldy  result+1
 		rts
 
 result		.byte  0,0,0,0

compiler/res/prog8lib/math.p8

@@ -1,6 +1,8 @@
 ; Internal Math library routines - always included by the compiler
 
 math {
+    %option no_symbol_prefixing
+
 	%asminclude "library:math.asm"
 
     asmsub sin8u(ubyte angle @A) clobbers(Y) -> ubyte @A {
@@ -95,4 +97,296 @@ _sinecosR8	.char  trunc(127.0 * sin(range(180+45) * rad(360.0/180.0)))
             rts
         }}
     }
+
+
+sub direction_sc(byte x1, byte y1, byte x2, byte y2) -> ubyte {
+    ; From a pair of signed coordinates around the origin, calculate discrete direction between 0 and 23 into A.
+    cx16.r0L = 3        ; quadrant
+    cx16.r1sL = x2-x1   ; xdelta
+    if_neg {
+        cx16.r0L--
+        cx16.r1sL = -cx16.r1sL
+    }
+    cx16.r2sL = y2-y1   ; ydelta
+    if_neg {
+        cx16.r0L-=2
+        cx16.r2sL = -cx16.r2sL
+    }
+    return direction_qd(cx16.r0L, cx16.r1L, cx16.r2L)
+}
+
+sub direction(ubyte x1, ubyte y1, ubyte x2, ubyte y2) -> ubyte {
+    ; From a pair of positive coordinates, calculate discrete direction between 0 and 23 into A.
+    cx16.r0L = 3        ; quadrant
+    if x2>=x1 {
+        cx16.r1L = x2-x1
+    } else {
+        cx16.r1L = x1-x2
+        cx16.r0L--
+    }
+    if y2>=y1 {
+        cx16.r2L = y2-y1
+    } else {
+        cx16.r2L = y1-y2
+        cx16.r0L -= 2
+    }
+    return direction_qd(cx16.r0L, cx16.r1L, cx16.r2L)
+}
+
+asmsub direction_qd(ubyte quadrant @A, ubyte xdelta @X, ubyte ydelta @Y) -> ubyte @A {
+    ;Arctan  https://github.com/dustmop/arctan24
+    ; From a pair of X/Y deltas (both >=0), and quadrant 0-3, calculate discrete direction between 0 and 23 into A.
+    ;  .reg:a @in  quadrant Number 0 to 3.
+    ;  .reg:x @in  x_delta Delta for x direction.
+    ;  .reg:y @in  y_delta Delta for y direction.
+    ; Returns A as the direction (0-23).
+
+    %asm {{
+x_delta = cx16.r0L
+y_delta = cx16.r1L
+quadrant = cx16.r2L
+half_value = cx16.r3L
+region_number = cx16.r4L
+small = cx16.r5L
+large = cx16.r5H
+
+  sta quadrant
+  sty y_delta
+  stx x_delta
+  cpx y_delta
+  bcs _XGreaterOrEqualY
+
+_XLessY:
+  lda #16
+  sta region_number
+  stx small
+  sty large
+  bne _DetermineRegion
+
+_XGreaterOrEqualY:
+  lda #0
+  sta region_number
+  stx large
+  sty small
+
+_DetermineRegion:
+  ; set A = small * 2.5
+  lda small
+  lsr a
+  sta half_value
+  lda small
+  asl a
+  bcs _SmallerQuotient
+  clc
+  adc half_value
+  bcs _SmallerQuotient
+  cmp large
+  bcc _LargerQuotient
+
+; S * 2.5 > L
+_SmallerQuotient:
+  ; set A = S * 1.25
+  lsr half_value
+  lda small
+  clc
+  adc half_value
+  cmp large
+  bcc _Region1 ; if S * 1.25 < L then goto Region1 (L / S > 1.25)
+  bcs _Region0 ;                                   (L / S < 1.25)
+
+; S * 2.5 < L
+_LargerQuotient:
+  ; set A = S * 7.5
+  lda small
+  asl a
+  asl a
+  asl a
+  bcs _Region2
+  sec
+  sbc half_value
+  cmp large
+  bcc _Region3 ; if S * 7.5 < L then goto Region3 (L / S > 7.5)
+  jmp _Region2 ;                                  (L / S < 7.5)
+
+_Region0:
+  ; L / S < 1.25. d=3,9,15,21
+  jmp _LookupResult
+
+_Region1:
+  ; 1.25 < L / S < 2.5. d=2,4,8,10,14,16,20,22
+  lda region_number
+  clc
+  adc #4
+  sta region_number
+  bpl _LookupResult
+
+_Region2:
+  ; 2.5 < L / S < 7.5. d=1,5,7,11,13,17,19,23
+  lda region_number
+  clc
+  adc #8
+  sta region_number
+  bpl _LookupResult
+
+_Region3:
+  ; 7.5 < L / S. d=0,6,12,18
+  lda region_number
+  clc
+  adc #12
+  sta region_number
+
+_LookupResult:
+  lda quadrant
+  clc
+  adc region_number
+  tax
+  lda _quadrant_region_to_direction,x
+  rts
+
+_quadrant_region_to_direction:
+  .byte  9, 3,15,21
+  .byte 10, 2,14,22
+  .byte 11, 1,13,23
+  .byte 12, 0,12, 0
+  .byte  9, 3,15,21
+  .byte  8, 4,16,20
+  .byte  7, 5,17,19
+  .byte  6, 6,18,18
+
+    }}
+}
+
+asmsub atan2(ubyte x1 @R0, ubyte y1 @R1, ubyte x2 @R2, ubyte y2 @R3) -> ubyte @A {
+    ;; Calculate the angle, in a 256-degree circle, between two points into A.
+    ;; The points (x1, y1) and (x2, y2) have to use *unsigned coordinates only* from the positive quadrant in the carthesian plane!
+    ;; https://www.codebase64.org/doku.php?id=base:8bit_atan2_8-bit_angle
+    ;; This uses 2 large lookup tables so uses a lot of memory but is super fast.
+
+    %asm {{
+
+x1 = cx16.r0L
+y1 = cx16.r1L
+x2 = cx16.r2L
+y2 = cx16.r3L
+octant = cx16.r4L			;; temporary zeropage variable
+
+		lda x1
+		sec
+		sbc x2
+		bcs *+4
+		eor #$ff
+		tax
+		rol octant
+
+		lda y1
+		sec
+		sbc y2
+		bcs *+4
+		eor #$ff
+		tay
+		rol octant
+
+		lda log2_tab,x
+		sec
+		sbc log2_tab,y
+		bcc *+4
+		eor #$ff
+		tax
+
+		lda octant
+		rol a
+		and #%111
+		tay
+
+		lda atan_tab,x
+		eor octant_adjust,y
+		rts
+
+octant_adjust
+		.byte %00111111		;; x+,y+,|x|>|y|
+		.byte %00000000		;; x+,y+,|x|<|y|
+		.byte %11000000		;; x+,y-,|x|>|y|
+		.byte %11111111		;; x+,y-,|x|<|y|
+		.byte %01000000		;; x-,y+,|x|>|y|
+		.byte %01111111		;; x-,y+,|x|<|y|
+		.byte %10111111		;; x-,y-,|x|>|y|
+		.byte %10000000		;; x-,y-,|x|<|y|
+
+
+		;;;;;;;; atan(2^(x/32))*128/pi ;;;;;;;;
+
+atan_tab
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$00,$00,$00
+		.byte $00,$00,$00,$00,$00,$01,$01,$01
+		.byte $01,$01,$01,$01,$01,$01,$01,$01
+		.byte $01,$01,$01,$01,$01,$01,$01,$01
+		.byte $01,$01,$01,$01,$01,$01,$01,$01
+		.byte $01,$01,$01,$01,$01,$02,$02,$02
+		.byte $02,$02,$02,$02,$02,$02,$02,$02
+		.byte $02,$02,$02,$02,$02,$02,$02,$02
+		.byte $03,$03,$03,$03,$03,$03,$03,$03
+		.byte $03,$03,$03,$03,$03,$04,$04,$04
+		.byte $04,$04,$04,$04,$04,$04,$04,$04
+		.byte $05,$05,$05,$05,$05,$05,$05,$05
+		.byte $06,$06,$06,$06,$06,$06,$06,$06
+		.byte $07,$07,$07,$07,$07,$07,$08,$08
+		.byte $08,$08,$08,$08,$09,$09,$09,$09
+		.byte $09,$0a,$0a,$0a,$0a,$0b,$0b,$0b
+		.byte $0b,$0c,$0c,$0c,$0c,$0d,$0d,$0d
+		.byte $0d,$0e,$0e,$0e,$0e,$0f,$0f,$0f
+		.byte $10,$10,$10,$11,$11,$11,$12,$12
+		.byte $12,$13,$13,$13,$14,$14,$15,$15
+		.byte $15,$16,$16,$17,$17,$17,$18,$18
+		.byte $19,$19,$19,$1a,$1a,$1b,$1b,$1c
+		.byte $1c,$1c,$1d,$1d,$1e,$1e,$1f,$1f
+
+
+		;;;;;;;; log2(x)*32 ;;;;;;;;
+
+log2_tab
+		.byte $00,$00,$20,$32,$40,$4a,$52,$59
+		.byte $60,$65,$6a,$6e,$72,$76,$79,$7d
+		.byte $80,$82,$85,$87,$8a,$8c,$8e,$90
+		.byte $92,$94,$96,$98,$99,$9b,$9d,$9e
+		.byte $a0,$a1,$a2,$a4,$a5,$a6,$a7,$a9
+		.byte $aa,$ab,$ac,$ad,$ae,$af,$b0,$b1
+		.byte $b2,$b3,$b4,$b5,$b6,$b7,$b8,$b9
+		.byte $b9,$ba,$bb,$bc,$bd,$bd,$be,$bf
+		.byte $c0,$c0,$c1,$c2,$c2,$c3,$c4,$c4
+		.byte $c5,$c6,$c6,$c7,$c7,$c8,$c9,$c9
+		.byte $ca,$ca,$cb,$cc,$cc,$cd,$cd,$ce
+		.byte $ce,$cf,$cf,$d0,$d0,$d1,$d1,$d2
+		.byte $d2,$d3,$d3,$d4,$d4,$d5,$d5,$d5
+		.byte $d6,$d6,$d7,$d7,$d8,$d8,$d9,$d9
+		.byte $d9,$da,$da,$db,$db,$db,$dc,$dc
+		.byte $dd,$dd,$dd,$de,$de,$de,$df,$df
+		.byte $df,$e0,$e0,$e1,$e1,$e1,$e2,$e2
+		.byte $e2,$e3,$e3,$e3,$e4,$e4,$e4,$e5
+		.byte $e5,$e5,$e6,$e6,$e6,$e7,$e7,$e7
+		.byte $e7,$e8,$e8,$e8,$e9,$e9,$e9,$ea
+		.byte $ea,$ea,$ea,$eb,$eb,$eb,$ec,$ec
+		.byte $ec,$ec,$ed,$ed,$ed,$ed,$ee,$ee
+		.byte $ee,$ee,$ef,$ef,$ef,$ef,$f0,$f0
+		.byte $f0,$f1,$f1,$f1,$f1,$f1,$f2,$f2
+		.byte $f2,$f2,$f3,$f3,$f3,$f3,$f4,$f4
+		.byte $f4,$f4,$f5,$f5,$f5,$f5,$f5,$f6
+		.byte $f6,$f6,$f6,$f7,$f7,$f7,$f7,$f7
+		.byte $f8,$f8,$f8,$f8,$f9,$f9,$f9,$f9
+		.byte $f9,$fa,$fa,$fa,$fa,$fa,$fb,$fb
+		.byte $fb,$fb,$fb,$fc,$fc,$fc,$fc,$fc
+		.byte $fd,$fd,$fd,$fd,$fd,$fd,$fe,$fe
+		.byte $fe,$fe,$fe,$ff,$ff,$ff,$ff,$ff
+
+    }}
+}
+
 }

compiler/res/prog8lib/prog8_funcs.asm

@@ -86,16 +86,15 @@ func_all_w_stack	.proc
 
 abs_b_stack	.proc
 	; -- push abs(A) on stack (as unsigned word)
-		jsr  abs_b_into_AY
+		jsr  abs_b_into_A
 		sta  P8ESTACK_LO,x
 		stz  P8ESTACK_HI,x
 		dex
 		rts
 		.pend
 
-abs_b_into_AY	.proc
-	; -- AY = abs(A)  (abs always returns unsigned word)
-		ldy  #0
+abs_b_into_A	.proc
+	; -- A = abs(A)
 		cmp  #0
 		bmi  +
 		rts
@@ -556,3 +555,111 @@ func_pokew   .proc
 	sta  (P8ZP_SCRATCH_W1),y
 	rts
 	.pend
+
+
+func_clamp_byte .proc
+	; signed value in A, result in A
+	; minimum in P8ZP_SCRATCH_W1
+	; maximum in P8ZP_SCRATCH_W1+1
+	tay
+	sec
+	sbc  P8ZP_SCRATCH_W1+1
+	bvc  +
+	eor  #$80
++       bmi  +
+	lda  P8ZP_SCRATCH_W1+1
+	tay
+	jmp  ++
++       tya
++	sec
+	sbc  P8ZP_SCRATCH_W1
+	bvc  +
+	eor  #$80
++       bmi  +
+	tya
+	rts
++       lda  P8ZP_SCRATCH_W1
+	rts
+	.pend
+
+
+func_clamp_ubyte .proc
+	; value in A, result in A
+	; minimum in P8ZP_SCRATCH_W1
+	; maximum in P8ZP_SCRATCH_W1+1
+	cmp  P8ZP_SCRATCH_W1+1
+	bcc  +
+	lda  P8ZP_SCRATCH_W1+1
++       cmp  P8ZP_SCRATCH_W1
+	bcc  +
+	rts
++       lda  P8ZP_SCRATCH_W1
+	rts
+	.pend
+
+func_clamp_word .proc
+	; signed value in AY, result in AY
+	; minimum in P8ZP_SCRATCH_W1
+	; maximum in P8ZP_SCRATCH_W2
+	sta  P8ZP_SCRATCH_B1
+	sty  P8ZP_SCRATCH_REG
+	ldy  P8ZP_SCRATCH_W2+1
+	lda  P8ZP_SCRATCH_W2
+	cmp  P8ZP_SCRATCH_B1
+	tya
+	sbc  P8ZP_SCRATCH_REG
+	bvc  +
+	eor  #$80
++       bpl  +
+	lda  P8ZP_SCRATCH_W2
+	ldy  P8ZP_SCRATCH_W2+1
+	sta  P8ZP_SCRATCH_B1
+	sty  P8ZP_SCRATCH_REG
++	ldy  P8ZP_SCRATCH_W1+1
+	lda  P8ZP_SCRATCH_W1
+	cmp  P8ZP_SCRATCH_B1
+	tya
+	sbc  P8ZP_SCRATCH_REG
+	bvc  +
+	eor  #$80
++       bpl  +
+	ldy  P8ZP_SCRATCH_REG
+	lda  P8ZP_SCRATCH_B1
+	rts
++	ldy  P8ZP_SCRATCH_W1+1
+	lda  P8ZP_SCRATCH_W1
+	rts
+	.pend
+
+func_clamp_uword .proc
+	; value in AY, result in AY
+	; minimum in P8ZP_SCRATCH_W1
+	; maximum in P8ZP_SCRATCH_W2
+	sta  P8ZP_SCRATCH_B1
+	sty  P8ZP_SCRATCH_REG
+	cpy  P8ZP_SCRATCH_W2+1
+	bcc  ++
+	bne  +
+	cmp  P8ZP_SCRATCH_W2
+	bcc  ++
++       beq  +
+	lda  P8ZP_SCRATCH_W2
+	ldy  P8ZP_SCRATCH_W2+1
+	sta  P8ZP_SCRATCH_B1
+	sty  P8ZP_SCRATCH_REG
++	ldy  P8ZP_SCRATCH_REG
+	lda  P8ZP_SCRATCH_B1
+	cpy  P8ZP_SCRATCH_W1+1
+	bcc  ++
+	bne  +
+	cmp  P8ZP_SCRATCH_W1
+	bcc  ++
++       beq  +
+	ldy  P8ZP_SCRATCH_REG
+	lda  P8ZP_SCRATCH_B1
+	rts
++	ldy  P8ZP_SCRATCH_W1+1
+	lda  P8ZP_SCRATCH_W1
+	rts
+
+	.pend

compiler/res/prog8lib/prog8_lib.asm

@@ -167,9 +167,8 @@ mul_word	.proc
 		lda  P8ESTACK_LO+1,x
 		ldy  P8ESTACK_HI+1,x
 		jsr  math.multiply_words
-		lda  math.multiply_words.result
 		sta  P8ESTACK_LO+1,x
-		lda  math.multiply_words.result+1
+		tya
 		sta  P8ESTACK_HI+1,x
 		rts
 		.pend

compiler/res/prog8lib/prog8_lib.p8

@@ -1,6 +1,8 @@
 ; Internal library routines - always included by the compiler
 
 prog8_lib {
+    %option no_symbol_prefixing
+
 	%asminclude "library:prog8_lib.asm"
 	%asminclude "library:prog8_funcs.asm"
 }

compiler/res/prog8lib/string.p8

@@ -1,6 +1,7 @@
 ; 0-terminated string manipulation routines.
 
 string {
+    %option no_symbol_prefixing
 
     asmsub length(uword string @AY) clobbers(A) -> ubyte @Y {
         ; Returns the number of bytes in the string.
@@ -127,7 +128,7 @@ _startloop	dey
         }}
     }
 
-    asmsub find(uword string @R0, ubyte character @A) -> ubyte @A, ubyte @Pc {
+    asmsub find(uword string @R0, ubyte character @A) -> ubyte @A, bool @Pc {
         ; Locates the first position of the given character in the string,
         ; returns Carry set if found + index in A, or A=0 + Carry clear if not found.
         %asm {{
@@ -292,7 +293,7 @@ str = P8ZP_SCRATCH_W1
 	sta  modify_pattern2+2
 	jsr  _match
 	lda  #0
-	adc  #0
+	rol  a
 	ldx  P8ZP_SCRATCH_REG
 	rts
 

compiler/res/prog8lib/test_stack.p8

@@ -3,6 +3,7 @@
 %import textio
 
 test_stack {
+    %option no_symbol_prefixing
 
     asmsub test() {
         %asm {{

compiler/res/prog8lib/virtual/conv.p8

@@ -106,19 +106,21 @@ sub  str_uwhex  (uword value) {
 
 sub  str_uw0  (uword value) {
     ; ---- convert the uword in A/Y in decimal string form, with left padding 0s (5 positions total)
-    ubyte tenthousands = (value / 10000) as ubyte
-    value -= 10000*tenthousands
-    ubyte thousands = (value / 1000) as ubyte
-    value -= 1000*thousands
-    ubyte hundreds = (value / 100) as ubyte
-    value -= 100 as uword * hundreds
-    ubyte tens = (value / 10) as ubyte
-    value -= 10*tens
+    uword value2 = value/10
+    ubyte digits = value-value2*10 as ubyte
+    uword value3 = value2/10
+    ubyte tens = value2-value3*10 as ubyte
+    uword value4 = value3/10
+    ubyte hundreds = value3-value4*10 as ubyte
+    uword value5 = value4/10
+    ubyte thousands = value4-value5*10 as ubyte
+    uword value6 = value5/10
+    ubyte tenthousands = value5-value6*10 as ubyte
     string_out[0] = tenthousands+'0'
     string_out[1] = thousands+'0'
     string_out[2] = hundreds+'0'
     string_out[3] = tens+'0'
-    string_out[4] = value as ubyte + '0'
+    string_out[4] = digits+'0'
     string_out[5] = 0
 }
 
@@ -139,14 +141,16 @@ sub  str_w  (word value)  {
 }
 
 sub internal_str_uw(uword value, uword out_ptr) {
-    ubyte tenthousands = (value / 10000) as ubyte
-    value -= 10000*tenthousands
-    ubyte thousands = (value / 1000) as ubyte
-    value -= 1000*thousands
-    ubyte hundreds = (value / 100) as ubyte
-    value -= 100 as uword * hundreds
-    ubyte tens = (value / 10) as ubyte
-    value -= 10*tens
+    uword value2 = value/10
+    ubyte digits = value-value2*10 as ubyte
+    uword value3 = value2/10
+    ubyte tens = value2-value3*10 as ubyte
+    uword value4 = value3/10
+    ubyte hundreds = value3-value4*10 as ubyte
+    uword value5 = value4/10
+    ubyte thousands = value4-value5*10 as ubyte
+    uword value6 = value5/10
+    ubyte tenthousands = value5-value6*10 as ubyte
     if tenthousands
         goto output_tenthousands
     if thousands
@@ -169,7 +173,7 @@ output_tens:
     @(out_ptr) = tens+'0'
     out_ptr++
 output_ones:
-    @(out_ptr) = value as ubyte + '0'
+    @(out_ptr) = digits+'0'
     out_ptr++
     @(out_ptr) = 0
 }
@@ -196,9 +200,9 @@ sub  str2uword(str string) -> uword {
     ;    the number may NOT be preceded by a + sign and may NOT contain spaces
     ;    (any non-digit character will terminate the number string that is parsed)
     %ir {{
-        loadm.w r65500,conv.str2uword.string
-        syscall 11
-        returnreg.w r0
+        loadm.w r65535,conv.str2uword.string
+        syscall 11 (r65535.w) : r0.w
+        returnr.w r0
     }}
 }
 
@@ -207,9 +211,9 @@ sub  str2word(str string) -> word {
     ;    the number may be preceded by a + or - sign but may NOT contain spaces
     ;    (any non-digit character will terminate the number string that is parsed)
     %ir {{
-        loadm.w r65500,conv.str2word.string
-        syscall 12
-        returnreg.w r0
+        loadm.w r65535,conv.str2word.string
+        syscall 12 (r65535.w) : r0.w
+        returnr.w r0
     }}
 }
 

compiler/res/prog8lib/virtual/diskio.p8

@@ -0,0 +1 @@
+; there is no diskio yet for the VM target.

compiler/res/prog8lib/virtual/floats.p8

@@ -10,8 +10,8 @@ floats {
 sub print_f(float value) {
     ; ---- prints the floating point value (without a newline).
     %ir {{
-        loadm.f fr65500,floats.print_f.value
-        syscall 25
+        loadm.f fr65535,floats.print_f.value
+        syscall 25 (fr65535.f)
         return
     }}
 }
@@ -21,15 +21,7 @@ sub pow(float value, float power) -> float {
         loadm.f fr0,floats.pow.value
         loadm.f fr1,floats.pow.power
         fpow.f fr0,fr1
-        returnreg.f fr0
-    }}
-}
-
-sub fabs(float value) -> float {
-    %ir {{
-        loadm.f fr0,floats.fabs.value
-        fabs.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
@@ -37,7 +29,7 @@ sub sin(float angle) -> float {
     %ir {{
         loadm.f fr0,floats.sin.angle
         fsin.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
@@ -45,7 +37,7 @@ sub cos(float angle) -> float {
     %ir {{
         loadm.f fr0,floats.cos.angle
         fcos.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
@@ -53,7 +45,7 @@ sub tan(float value) -> float {
     %ir {{
         loadm.f fr0,floats.tan.value
         ftan.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
@@ -61,7 +53,7 @@ sub atan(float value) -> float {
     %ir {{
         loadm.f fr0,floats.atan.value
         fatan.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
@@ -69,7 +61,7 @@ sub ln(float value) -> float {
     %ir {{
         loadm.f fr0,floats.ln.value
         fln.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
@@ -77,15 +69,7 @@ sub log2(float value) -> float {
     %ir {{
         loadm.f fr0,floats.log2.value
         flog.f fr0,fr0
-        returnreg.f fr0
-    }}
-}
-
-sub sqrt(float value) -> float {
-    %ir {{
-        loadm.f fr0,floats.sqrt.value
-        sqrt.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
@@ -103,7 +87,7 @@ sub round(float value) -> float {
     %ir {{
         loadm.f fr0,floats.round.value
         fround.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
@@ -111,7 +95,7 @@ sub floor(float value) -> float {
     %ir {{
         loadm.f fr0,floats.floor.value
         ffloor.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
@@ -120,22 +104,46 @@ sub ceil(float value) -> float {
     %ir {{
         loadm.f fr0,floats.ceil.value
         fceil.f fr0,fr0
-        returnreg.f fr0
+        returnr.f fr0
     }}
 }
 
 sub rndf() -> float {
     %ir {{
-        syscall 35
-        returnreg.f fr0
+        syscall 35 () : fr0.f
+        returnr.f fr0
     }}
 }
 
 sub rndseedf(float seed) {
     %ir {{
-        loadm.f  fr65500,floats.rndseedf.seed
-        syscall 32
+        loadm.f  fr65535,floats.rndseedf.seed
+        syscall 32 (fr65535.f)
+        return
     }}
 }
 
+
+sub minf(float f1, float f2) -> float {
+    if f1<f2
+        return f1
+    return f2
+}
+
+
+sub maxf(float f1, float f2) -> float {
+    if f1>f2
+        return f1
+    return f2
+}
+
+
+sub clampf(float value, float minimum, float maximum) -> float {
+    if value<minimum
+        value=minimum
+    if value<maximum
+        return value
+    return maximum
+}
+
 }

compiler/res/prog8lib/virtual/math.p8

@@ -161,25 +161,82 @@ math {
 
     sub rnd() -> ubyte {
         %ir {{
-            syscall 33
-            returnreg.b r0
+            syscall 33 (): r0.b
+            returnr.b r0
         }}
     }
 
     sub rndw() -> uword {
         %ir {{
-            syscall 34
-            returnreg.w r0
+            syscall 34 (): r0.w
+            returnr.w r0
         }}
     }
 
     sub rndseed(uword seed1, uword seed2) {
         ; -- reset the pseudo RNG's seed values. Defaults are: $a55a, $7653.
         %ir {{
-            loadm.w r65500,math.rndseed.seed1
-            loadm.w r65501,math.rndseed.seed2
-            syscall 31
+            loadm.w r65534,math.rndseed.seed1
+            loadm.w r65535,math.rndseed.seed2
+            syscall 31 (r65534.w, r65535.w)
             return
         }}
     }
+
+
+sub direction(ubyte x1, ubyte y1, ubyte x2, ubyte y2) -> ubyte {
+    ; From a pair of positive coordinates, calculate discrete direction between 0 and 23 into A.
+    ; This adjusts the atan() result  so that the direction N is centered on the angle=N instead of having it as a boundary
+    ubyte angle = atan2(x1, y1, x2, y2) - 256/48
+    return 23-lsb(mkword(angle,0) / 2730)
+}
+
+sub direction_sc(byte x1, byte y1, byte x2, byte y2) -> ubyte {
+    ; From a pair of signed coordinates around the origin, calculate discrete direction between 0 and 23 into A.
+    ; shift the points into the positive quadrant
+    ubyte px1
+    ubyte py1
+    ubyte px2
+    ubyte py2
+    if x1<0 or x2<0 {
+        px1 = x1 as ubyte + 128
+        px2 = x2 as ubyte + 128
+    } else {
+        px1 = x1 as ubyte
+        px2 = x2 as ubyte
+    }
+    if y1<0 or y2<0 {
+        py1 = y1 as ubyte + 128
+        py2 = y2 as ubyte + 128
+    } else {
+        py1 = y1 as ubyte
+        py2 = y2 as ubyte
+    }
+
+    return direction(px1, py1, px2, py2)
+}
+
+sub direction_qd(ubyte quadrant, ubyte xdelta, ubyte ydelta) -> ubyte {
+    ; From a pair of X/Y deltas (both >=0), and quadrant 0-3, calculate discrete direction between 0 and 23.
+    when quadrant {
+        3 -> return direction(0, 0, xdelta, ydelta)
+        2 -> return direction(xdelta, 0, 0, ydelta)
+        1 -> return direction(0, ydelta, xdelta, 0)
+        else -> return direction(xdelta, ydelta, 0, 0)
+    }
+}
+
+sub atan2(ubyte x1, ubyte y1, ubyte x2, ubyte y2) -> ubyte {
+    ;; Calculate the angle, in a 256-degree circle, between two points into A.
+    ;; The points (x1, y1) and (x2, y2) have to use *unsigned coordinates only* from the positive quadrant in the carthesian plane!
+    %ir {{
+        loadm.b r65532,math.atan2.x1
+        loadm.b r65533,math.atan2.y1
+        loadm.b r65534,math.atan2.x2
+        loadm.b r65535,math.atan2.y2
+        syscall 44 (r65532.b, r65533.b, r65534.b, r65535.b): r0.b
+        returnr.b r0
+    }}
+}
+
 }

compiler/res/prog8lib/virtual/string.p8

@@ -84,10 +84,10 @@ string {
         ; Note that you can also directly compare strings and string values with eachother using
         ; comparison operators ==, < etcetera (it will use strcmp for you under water automatically).
         %ir {{
-            loadm.w r65500,string.compare.st1
-            loadm.w r65501,string.compare.st2
-            syscall 29
-            returnreg.b r0
+            loadm.w r65534,string.compare.st1
+            loadm.w r65535,string.compare.st2
+            syscall 29 (r65534.w, r65535.w) : r0.b
+            returnr.b r0
         }}
     }
 

compiler/res/prog8lib/virtual/syslib.p8

@@ -8,22 +8,22 @@ sys {
     sub  reset_system()  {
         ; Soft-reset the system back to initial power-on Basic prompt.
         %ir {{
-            syscall 0
+            syscall 0 ()
         }}
     }
 
     sub wait(uword jiffies) {
         ; --- wait approximately the given number of jiffies (1/60th seconds)
         %ir {{
-            loadm.w r65500,sys.wait.jiffies
-            syscall 13
+            loadm.w r65535,sys.wait.jiffies
+            syscall 13 (r65535.w)
         }}
     }
 
     sub waitvsync() {
         ; --- busy wait till the next vsync has occurred (approximately), without depending on custom irq handling.
         %ir {{
-            syscall 14
+            syscall 14()
         }}
     }
 
@@ -62,8 +62,8 @@ sys {
     sub exit(ubyte returnvalue) {
         ; -- immediately exit the program with a return code in the A register
         %ir {{
-            loadm.b r65500,sys.exit.returnvalue
-            syscall 1
+            loadm.b r65535,sys.exit.returnvalue
+            syscall 1 (r65535.b)
         }}
     }
 
@@ -82,33 +82,33 @@ sys {
 
     sub gfx_enable(ubyte mode) {
         %ir {{
-            loadm.b r65500,sys.gfx_enable.mode
-            syscall 8
+            loadm.b r65535,sys.gfx_enable.mode
+            syscall 8 (r65535.b)
         }}
     }
 
     sub gfx_clear(ubyte color) {
         %ir {{
-            loadm.b r65500,sys.gfx_clear.color
-            syscall 9
+            loadm.b r65535,sys.gfx_clear.color
+            syscall 9 (r65535.b)
         }}
     }
 
     sub gfx_plot(uword xx, uword yy, ubyte color) {
         %ir {{
-            loadm.w r65500,sys.gfx_plot.xx
-            loadm.w r65501,sys.gfx_plot.yy
-            loadm.b r65502,sys.gfx_plot.color
-            syscall 10
+            loadm.w r65533,sys.gfx_plot.xx
+            loadm.w r65534,sys.gfx_plot.yy
+            loadm.b r65535,sys.gfx_plot.color
+            syscall 10 (r65533.w, r65534.w, r65535.b)
         }}
     }
 
     sub gfx_getpixel(uword xx, uword yy) -> ubyte {
         %ir {{
-            loadm.w r65500,sys.gfx_getpixel.xx
-            loadm.w r65501,sys.gfx_getpixel.yy
-            syscall 30
-            returnreg.b r0
+            loadm.w r65534,sys.gfx_getpixel.xx
+            loadm.w r65535,sys.gfx_getpixel.yy
+            syscall 30 (r65534.w, r65535.w): r0.b
+            returnr.b r0
         }}
     }
 }
@@ -218,4 +218,43 @@ cx16 {
     &byte r13sH = $ff1d
     &byte r14sH = $ff1f
     &byte r15sH = $ff21
+
+    sub save_virtual_registers() {
+        uword[32] storage
+        storage[0] = r0
+        storage[1] = r1
+        storage[2] = r2
+        storage[3] = r3
+        storage[4] = r4
+        storage[5] = r5
+        storage[6] = r6
+        storage[7] = r7
+        storage[8] = r8
+        storage[9] = r9
+        storage[10] = r10
+        storage[11] = r11
+        storage[12] = r12
+        storage[13] = r13
+        storage[14] = r14
+        storage[15] = r15
+    }
+
+    sub restore_virtual_registers() {
+        r0 = cx16.save_virtual_registers.storage[0]
+        r1 = cx16.save_virtual_registers.storage[1]
+        r2 = cx16.save_virtual_registers.storage[2]
+        r3 = cx16.save_virtual_registers.storage[3]
+        r4 = cx16.save_virtual_registers.storage[4]
+        r5 = cx16.save_virtual_registers.storage[5]
+        r6 = cx16.save_virtual_registers.storage[6]
+        r7 = cx16.save_virtual_registers.storage[7]
+        r8 = cx16.save_virtual_registers.storage[8]
+        r9 = cx16.save_virtual_registers.storage[9]
+        r10 = cx16.save_virtual_registers.storage[10]
+        r11 = cx16.save_virtual_registers.storage[11]
+        r12 = cx16.save_virtual_registers.storage[12]
+        r13 = cx16.save_virtual_registers.storage[13]
+        r14 = cx16.save_virtual_registers.storage[14]
+        r15 = cx16.save_virtual_registers.storage[15]
+    }
 }