sponsorship link.

error sorting.
version.

.github/FUNDING.yml

@@ -0,0 +1,13 @@
+# These are supported funding model platforms
+
+#github: # Replace with up to 4 GitHub Sponsors-enabled usernames e.g., [user1, user2]
+#patreon: # Replace with a single Patreon username
+open_collective: # Replace with a single Open Collective username
+ko_fi: irmen
+#tidelift: # Replace with a single Tidelift platform-name/package-name e.g., npm/babel
+#community_bridge: # Replace with a single Community Bridge project-name e.g., cloud-foundry
+#liberapay: # Replace with a single Liberapay username
+#issuehunt: # Replace with a single IssueHunt username
+#otechie: # Replace with a single Otechie username
+#lfx_crowdfunding: # Replace with a single LFX Crowdfunding project-name e.g., cloud-foundry
+#custom: # Replace with up to 4 custom sponsorship URLs e.g., ['link1', 'link2']

.idea/kotlinc.xml

@@ -4,6 +4,6 @@
     <option name="jvmTarget" value="11" />
   </component>
   <component name="KotlinJpsPluginSettings">
-    <option name="version" value="1.9.0-release-358" />
+    <option name="version" value="1.9.20" />
   </component>
 </project>

.idea/libraries/KotlinJavaRuntime.xml

@@ -1,19 +1,23 @@
 <component name="libraryTable">
-  <library name="KotlinJavaRuntime">
+  <library name="KotlinJavaRuntime" type="repository">
+    <properties maven-id="org.jetbrains.kotlin:kotlin-stdlib-jdk8:1.9.22" />
     <CLASSES>
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-stdlib.jar!/" />
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-reflect.jar!/" />
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-test.jar!/" />
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-stdlib-jdk7.jar!/" />
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-stdlib-jdk8.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.9.22/kotlin-stdlib-jdk8-1.9.22.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.9.22/kotlin-stdlib-1.9.22.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-jdk7/1.9.22/kotlin-stdlib-jdk7-1.9.22.jar!/" />
     </CLASSES>
-    <JAVADOC />
+    <JAVADOC>
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.9.22/kotlin-stdlib-jdk8-1.9.22-javadoc.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.9.22/kotlin-stdlib-1.9.22-javadoc.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/annotations/13.0/annotations-13.0-javadoc.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.9.22/kotlin-stdlib-jdk7-1.9.22-javadoc.jar!/" />
+    </JAVADOC>
     <SOURCES>
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-stdlib-sources.jar!/" />
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-reflect-sources.jar!/" />
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-test-sources.jar!/" />
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-stdlib-jdk7-sources.jar!/" />
-      <root url="jar://$KOTLIN_BUNDLED$/lib/kotlin-stdlib-jdk8-sources.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk8/1.9.22/kotlin-stdlib-jdk8-1.9.22-sources.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib/1.9.22/kotlin-stdlib-1.9.22-sources.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/annotations/13.0/annotations-13.0-sources.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/jetbrains/kotlin/kotlin-stdlib-jdk7/1.9.22/kotlin-stdlib-jdk7-1.9.22-sources.jar!/" />
     </SOURCES>
   </library>
 </component>

.idea/libraries/io_kotest_assertions_core_jvm.xml

@@ -1,19 +1,19 @@
 <component name="libraryTable">
   <library name="io.kotest.assertions.core.jvm" type="repository">
-    <properties maven-id="io.kotest:kotest-assertions-core-jvm:5.6.2" />
+    <properties maven-id="io.kotest:kotest-assertions-core-jvm:5.8.0" />
     <CLASSES>
-      <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://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-core-jvm/5.8.0/kotest-assertions-core-jvm-5.8.0.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/kotest/kotest-assertions-shared-jvm/5.8.0/kotest-assertions-shared-jvm-5.8.0.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/opentest4j/opentest4j/1.3.0/opentest4j-1.3.0.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://$MAVEN_REPOSITORY$/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://$MAVEN_REPOSITORY$/io/kotest/kotest-common-jvm/5.8.0/kotest-common-jvm-5.8.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-api-jvm/5.8.0/kotest-assertions-api-jvm-5.8.0.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>

.idea/libraries/io_kotest_runner_junit5_jvm.xml

@@ -1,29 +1,29 @@
 <component name="libraryTable">
   <library name="io.kotest.runner.junit5.jvm" type="repository">
-    <properties maven-id="io.kotest:kotest-runner-junit5-jvm:5.6.2" />
+    <properties maven-id="io.kotest:kotest-runner-junit5-jvm:5.8.0" />
     <CLASSES>
-      <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/kotest/kotest-runner-junit5-jvm/5.8.0/kotest-runner-junit5-jvm-5.8.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-api-jvm/5.8.0/kotest-framework-api-jvm-5.8.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-shared-jvm/5.8.0/kotest-assertions-shared-jvm-5.8.0.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/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://$MAVEN_REPOSITORY$/io/kotest/kotest-common-jvm/5.8.0/kotest-common-jvm-5.8.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-engine-jvm/5.8.0/kotest-framework-engine-jvm-5.8.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/github/classgraph/classgraph/4.8.162/classgraph-4.8.162.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/opentest4j/opentest4j/1.3.0/opentest4j-1.3.0.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/5.6.2/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://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-discovery-jvm/5.8.0/kotest-framework-discovery-jvm-5.8.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-core-jvm/5.8.0/kotest-assertions-core-jvm-5.8.0.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://$MAVEN_REPOSITORY$/io/kotest/kotest-assertions-api-jvm/5.8.0/kotest-assertions-api-jvm-5.8.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-extensions-jvm/5.8.0/kotest-extensions-jvm-5.8.0.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/io/kotest/kotest-framework-concurrency-jvm/5.8.0/kotest-framework-concurrency-jvm-5.8.0.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!/" />

.idea/libraries/slf4j_simple.xml

@@ -1,9 +1,9 @@
 <component name="libraryTable">
   <library name="slf4j.simple" type="repository">
-    <properties maven-id="org.slf4j:slf4j-simple:2.0.9" />
+    <properties maven-id="org.slf4j:slf4j-simple:2.0.11" />
     <CLASSES>
-      <root url="jar://$MAVEN_REPOSITORY$/org/slf4j/slf4j-simple/2.0.9/slf4j-simple-2.0.9.jar!/" />
-      <root url="jar://$MAVEN_REPOSITORY$/org/slf4j/slf4j-api/2.0.9/slf4j-api-2.0.9.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/slf4j/slf4j-simple/2.0.11/slf4j-simple-2.0.11.jar!/" />
+      <root url="jar://$MAVEN_REPOSITORY$/org/slf4j/slf4j-api/2.0.11/slf4j-api-2.0.11.jar!/" />
     </CLASSES>
     <JAVADOC />
     <SOURCES />

README.md

@@ -1,3 +1,4 @@
+[![ko-fi](https://ko-fi.com/img/githubbutton_sm.svg)](https://ko-fi.com/H2H6S0FFF)
 [![Documentation](https://readthedocs.org/projects/prog8/badge/?version=latest)](https://prog8.readthedocs.io/)
 
 Prog8 - Structured Programming Language for 8-bit 6502/65c02 microprocessors
@@ -9,6 +10,13 @@ This is a structured programming language for the 8-bit 6502/6510/65c02 micropro
 as used in many home computers from that era. It is a medium to low level programming language,
 which aims to provide many conveniences over raw assembly code (even when using a macro assembler).
 
+**Want to buy me a coffee or a pizza perhaps?**
+
+This project was created over the last couple of years by dedicating thousands of hours of my free time to it, to make it the best I possibly can.
+If you like Prog8, and think it's worth a nice cup of hot coffee or a delicious pizza, 
+you can help me out a little bit over at [ko-fi.com/irmen](https://ko-fi.com/irmen). 
+
+
 Documentation
 -------------
 Full documentation (syntax reference, how to use the language and the compiler, etc.) can be found at:

codeCore/src/prog8/code/SymbolTableMaker.kt

@@ -174,7 +174,7 @@ class SymbolTableMaker(private val program: PtProgram, private val options: Comp
 //                }
 //                VarDeclType.MEMORY -> {
 //                    val numElements =
-//                        if(decl.datatype in ArrayDatatypes)
+//                        if(decl.isArray)
 //                            decl.arraysize!!.constIndex()
 //                        else null
 //                    val astNode = PtVariable(decl.name, decl.datatype, null, null, decl.position)

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

@@ -31,7 +31,10 @@ sealed class PtNode(val position: Position) {
 }
 
 
-class PtNodeGroup : PtNode(Position.DUMMY)
+sealed interface IPtStatementContainer
+
+
+class PtNodeGroup : PtNode(Position.DUMMY), IPtStatementContainer
 
 
 sealed class PtNamedNode(var name: String, position: Position): PtNode(position) {
@@ -64,27 +67,30 @@ class PtProgram(
         children.asSequence().filterIsInstance<PtBlock>()
 
     fun entrypoint(): PtSub? =
-        allBlocks().firstOrNull { it.name == "main" || it.name=="p8_main" }
+        allBlocks().firstOrNull { it.name == "main" || it.name=="p8b_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?
+            ?.firstOrNull { it is PtSub && (it.name == "start" || it.name=="main.start" || it.name=="p8s_start" || it.name=="p8b_main.p8s_start") } as PtSub?
 }
 
 
 class PtBlock(name: String,
-              val address: UInt?,
               val library: Boolean,
-              val forceOutput: Boolean,
-              val noSymbolPrefixing: Boolean,
-              val veraFxMuls: Boolean,
-              val alignment: BlockAlignment,
               val source: SourceCode,       // taken from the module the block is defined in.
+              val options: Options,
               position: Position
-) : PtNamedNode(name, position) {
+) : PtNamedNode(name, position), IPtStatementContainer {
     enum class BlockAlignment {
         NONE,
         WORD,
         PAGE
     }
+
+    class Options(val address: UInt? = null,
+                  val forceOutput: Boolean = false,
+                  val noSymbolPrefixing: Boolean = false,
+                  val veraFxMuls: Boolean = false,
+                  val ignoreUnused: Boolean = false,
+                  val alignment: BlockAlignment = BlockAlignment.NONE)
 }
 
 

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

@@ -3,7 +3,7 @@ package prog8.code.ast
 import prog8.code.core.*
 import java.util.*
 import kotlin.math.abs
-import kotlin.math.round
+import kotlin.math.truncate
 
 
 sealed class PtExpression(val type: DataType, position: Position) : PtNode(position) {
@@ -36,7 +36,14 @@ sealed class PtExpression(val type: DataType, position: Position) : PtNode(posit
                 return arrayIndexExpr!! isSameAs other.arrayIndexExpr!!
             }
             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 PtBinaryExpression -> {
+                if(other !is PtBinaryExpression || other.operator!=operator)
+                    false
+                else if(operator in AssociativeOperators)
+                    (other.left isSameAs left && other.right isSameAs right) || (other.left isSameAs right && other.right isSameAs left)
+                else
+                    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
             is PtMachineRegister -> other is PtMachineRegister && other.type==type && other.register==register
@@ -72,7 +79,12 @@ sealed class PtExpression(val type: DataType, position: Position) : PtNode(posit
             is PtArray -> true
             is PtArrayIndexer -> index is PtNumber || index is PtIdentifier
             is PtBinaryExpression -> false
-            is PtBuiltinFunctionCall -> name in arrayOf("msb", "lsb", "peek", "peekw", "mkword", "set_carry", "set_irqd", "clear_carry", "clear_irqd")
+            is PtBuiltinFunctionCall -> {
+                when (name) {
+                    in arrayOf("msb", "lsb", "mkword", "set_carry", "set_irqd", "clear_carry", "clear_irqd") -> this.args.all { it.isSimple() }
+                    else -> false
+                }
+            }
             is PtContainmentCheck -> false
             is PtFunctionCall -> false
             is PtIdentifier -> true
@@ -227,9 +239,9 @@ class PtNumber(type: DataType, val number: Double, position: Position) : PtExpre
         if(type==DataType.BOOL)
             throw IllegalArgumentException("bool should have become ubyte @$position")
         if(type!=DataType.FLOAT) {
-            val rounded = round(number)
-            if (rounded != number)
-                throw IllegalArgumentException("refused rounding of float to avoid loss of precision @$position")
+            val trunc = truncate(number)
+            if (trunc != number)
+                throw IllegalArgumentException("refused truncating of float to avoid loss of precision @$position")
         }
     }
 

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

@@ -53,15 +53,24 @@ fun printAst(root: PtNode, skipLibraries: Boolean, output: (text: String) -> Uni
                     "goto ${node.identifier.name}"
                 else if(node.address!=null)
                     "goto ${node.address.toHex()}"
-                else if(node.generatedLabel!=null)
-                    "goto ${node.generatedLabel}"
                 else
                     "???"
             }
             is PtAsmSub -> {
-                val params = if (node.parameters.isEmpty()) "" else "...TODO ${node.parameters.size} PARAMS..."
+                val params = node.parameters.map {
+                                val register = it.first.registerOrPair
+                                val statusflag = it.first.statusflag
+                                "${it.second.type} ${it.second.name} @${register ?: statusflag}"
+                            }.joinToString(", ")
                 val clobbers = if (node.clobbers.isEmpty()) "" else "clobbers ${node.clobbers}"
-                val returns = if (node.returns.isEmpty()) "" else (if (node.returns.size == 1) "-> ${node.returns[0].second.name.lowercase()}" else "-> ${node.returns.map { it.second.name.lowercase() }}")
+                val returns = if (node.returns.isEmpty()) "" else {
+                    "-> ${node.returns.map {
+                        val register = it.first.registerOrPair
+                        val statusflag = it.first.statusflag
+                        "${it.second} @${register ?: statusflag}"}
+                        .joinToString(", ")
+                    }"
+                }
                 val str = if (node.inline) "inline " else ""
                 if(node.address==null) {
                     str + "asmsub ${node.name}($params) $clobbers $returns"
@@ -70,8 +79,8 @@ fun printAst(root: PtNode, skipLibraries: Boolean, output: (text: String) -> Uni
                 }
             }
             is PtBlock -> {
-                val addr = if(node.address==null) "" else "@${node.address.toHex()}"
-                val align = if(node.alignment==PtBlock.BlockAlignment.NONE) "" else "align=${node.alignment}"
+                val addr = if(node.options.address==null) "" else "@${node.options.address.toHex()}"
+                val align = if(node.options.alignment==PtBlock.BlockAlignment.NONE) "" else "align=${node.options.alignment}"
                 "\nblock '${node.name}' $addr $align"
             }
             is PtConstant -> {
@@ -89,7 +98,7 @@ fun printAst(root: PtNode, skipLibraries: Boolean, output: (text: String) -> Uni
                 }
             }
             is PtSub -> {
-                val params = if (node.parameters.isEmpty()) "" else "...TODO ${node.parameters.size} PARAMS..."
+                val params = node.parameters.map { "${it.type} ${it.name}" }.joinToString(", ")
                 var str = "sub ${node.name}($params) "
                 if(node.returntype!=null)
                     str += "-> ${node.returntype.name.lowercase()}"

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

@@ -23,7 +23,7 @@ class PtSub(
     val parameters: List<PtSubroutineParameter>,
     val returntype: DataType?,
     position: Position
-) : PtNamedNode(name, position), IPtSubroutine {
+) : PtNamedNode(name, position), IPtSubroutine, IPtStatementContainer {
     init {
         // params and return value should not be str
         if(parameters.any{ it.type !in NumericDatatypes })
@@ -101,9 +101,8 @@ class PtIfElse(position: Position) : PtNode(position) {
 }
 
 
-class PtJump(val identifier: PtIdentifier?,
+class PtJump(val identifier: PtIdentifier?,     // note: even ad-hoc labels are wrapped as an Identifier to simplify code. Just use dummy type and position.
              val address: UInt?,
-             val generatedLabel: String?,
              position: Position) : PtNode(position) {
     init {
         identifier?.let {it.parent = this }

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

@@ -120,17 +120,16 @@ val BuiltinFunctions: Map<String, FSignature> = mapOf(
     "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),
+    "peekf"     to FSignature(true, listOf(FParam("address", arrayOf(DataType.UWORD))), DataType.FLOAT),
     "poke"      to FSignature(false, listOf(FParam("address", arrayOf(DataType.UWORD)), FParam("value", arrayOf(DataType.UBYTE))), null),
-    "pokemon"   to FSignature(false, listOf(FParam("address", arrayOf(DataType.UWORD)), FParam("value", arrayOf(DataType.UBYTE))), null),
     "pokew"     to FSignature(false, listOf(FParam("address", arrayOf(DataType.UWORD)), FParam("value", arrayOf(DataType.UWORD))), null),
-    "pop"       to FSignature(false, listOf(FParam("target", ByteDatatypes)), null),
-    "popw"      to FSignature(false, listOf(FParam("target", WordDatatypes)), null),
-    "push"      to FSignature(false, listOf(FParam("value", ByteDatatypes)), null),
-    "pushw"     to FSignature(false, listOf(FParam("value", WordDatatypes)), null),
+    "pokef"     to FSignature(false, listOf(FParam("address", arrayOf(DataType.UWORD)), FParam("value", arrayOf(DataType.FLOAT))), null),
+    "pokemon"   to FSignature(false, listOf(FParam("address", arrayOf(DataType.UWORD)), FParam("value", arrayOf(DataType.UBYTE))), DataType.UBYTE),
     "rsave"     to FSignature(false, emptyList(), null),
     "rrestore"  to FSignature(false, emptyList(), null),
     "memory"    to FSignature(true, listOf(FParam("name", arrayOf(DataType.STR)), FParam("size", arrayOf(DataType.UWORD)), FParam("alignment", arrayOf(DataType.UWORD))), DataType.UWORD),
     "callfar"   to FSignature(false, listOf(FParam("bank", arrayOf(DataType.UBYTE)), FParam("address", arrayOf(DataType.UWORD)), FParam("arg", arrayOf(DataType.UWORD))), DataType.UWORD),
+    "call"      to FSignature(false, listOf(FParam("address", arrayOf(DataType.UWORD))), DataType.UWORD),
 )
 
 val InplaceModifyingBuiltinFunctions = setOf("setlsb", "setmsb", "rol", "ror", "rol2", "ror2", "sort", "reverse")

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

@@ -8,6 +8,7 @@ class CompilationOptions(val output: OutputType,
                          val launcher: CbmPrgLauncherType,
                          val zeropage: ZeropageType,
                          val zpReserved: List<UIntRange>,
+                         val zpAllowed: List<UIntRange>,
                          val floats: Boolean,
                          val noSysInit: Boolean,
                          val compTarget: ICompilationTarget,
@@ -20,11 +21,19 @@ class CompilationOptions(val output: OutputType,
                          var includeSourcelines: Boolean = false,
                          var experimentalCodegen: Boolean = false,
                          var varsHighBank: Int? = null,
+                         var varsGolden: Boolean = false,
+                         var slabsHighBank: Int? = null,
+                         var slabsGolden: Boolean = false,
                          var splitWordArrays: Boolean = false,
+                         var breakpointCpuInstruction: Boolean = false,
                          var outputDir: Path = Path(""),
                          var symbolDefs: Map<String, String> = emptyMap()
 ) {
     init {
         compTarget.machine.initializeMemoryAreas(this)
     }
+
+    companion object {
+        val AllZeropageAllowed: List<UIntRange> = listOf(0u..255u)
+    }
 }

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

@@ -1,12 +1,13 @@
 package prog8.code.core
 
 enum class DataType {
-    UBYTE,              // pass by value
-    BYTE,               // pass by value
-    UWORD,              // pass by value
-    WORD,               // pass by value
-    FLOAT,              // pass by value
-    BOOL,               // pass by value
+    UBYTE,              // pass by value            8 bits unsigned
+    BYTE,               // pass by value            8 bits signed
+    UWORD,              // pass by value            16 bits unsigned
+    WORD,               // pass by value            16 bits signed
+    LONG,               // pass by value            32 bits signed
+    FLOAT,              // pass by value            machine dependent
+    BOOL,               // pass by value            bit 0 of a 8 bit byte
     STR,                // pass by reference
     ARRAY_UB,           // pass by reference
     ARRAY_B,            // pass by reference
@@ -23,11 +24,12 @@ enum class DataType {
      */
     infix fun isAssignableTo(targetType: DataType) =
         when(this) {
-            BOOL -> targetType.oneOf(BOOL, BYTE, UBYTE, WORD, UWORD, FLOAT)
-            UBYTE -> targetType.oneOf(UBYTE, WORD, UWORD, FLOAT, BOOL)
-            BYTE -> targetType.oneOf(BYTE, WORD, FLOAT)
-            UWORD -> targetType.oneOf(UWORD, FLOAT)
-            WORD -> targetType.oneOf(WORD, FLOAT)
+            BOOL -> targetType.oneOf(BOOL, BYTE, UBYTE, WORD, UWORD, LONG, FLOAT)
+            UBYTE -> targetType.oneOf(UBYTE, WORD, UWORD, LONG, FLOAT, BOOL)
+            BYTE -> targetType.oneOf(BYTE, WORD, LONG, FLOAT)
+            UWORD -> targetType.oneOf(UWORD, LONG, FLOAT)
+            WORD -> targetType.oneOf(WORD, LONG, FLOAT)
+            LONG -> targetType.oneOf(LONG, FLOAT)
             FLOAT -> targetType.oneOf(FLOAT)
             STR -> targetType.oneOf(STR, UWORD)
             in ArrayDatatypes -> targetType == this
@@ -41,7 +43,8 @@ enum class DataType {
             this == other -> false
             this in ByteDatatypes -> false
             this in WordDatatypes -> other in ByteDatatypes
-            this== STR && other== UWORD || this== UWORD && other== STR -> false
+            this == LONG -> other in ByteDatatypes+WordDatatypes
+            this == STR && other == UWORD || this == UWORD && other == STR -> false
             else -> true
         }
 
@@ -76,6 +79,13 @@ enum class RegisterOrPair {
 
     companion object {
         val names by lazy { values().map { it.toString()} }
+        fun fromCpuRegister(cpu: CpuRegister): RegisterOrPair {
+            return when(cpu) {
+                CpuRegister.A -> A
+                CpuRegister.X -> X
+                CpuRegister.Y -> Y
+            }
+        }
     }
 
     fun asCpuRegister(): CpuRegister = when(this) {
@@ -110,17 +120,17 @@ enum class BranchCondition {
     PL,     // PL == POS
     POS,
     VS,
-    VC,
+    VC
 }
 
 
 val ByteDatatypes = arrayOf(DataType.UBYTE, DataType.BYTE, DataType.BOOL)
 val WordDatatypes = arrayOf(DataType.UWORD, DataType.WORD)
-val IntegerDatatypes = arrayOf(DataType.UBYTE, DataType.BYTE, DataType.UWORD, DataType.WORD, DataType.BOOL)
-val IntegerDatatypesNoBool = arrayOf(DataType.UBYTE, DataType.BYTE, DataType.UWORD, DataType.WORD)
-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 IntegerDatatypesNoBool = arrayOf(DataType.UBYTE, DataType.BYTE, DataType.UWORD, DataType.WORD, DataType.LONG)
+val IntegerDatatypes = IntegerDatatypesNoBool + DataType.BOOL
+val NumericDatatypesNoBool = arrayOf(DataType.UBYTE, DataType.BYTE, DataType.UWORD, DataType.WORD, DataType.LONG, DataType.FLOAT)
+val NumericDatatypes = NumericDatatypesNoBool + DataType.BOOL
+val SignedDatatypes =  arrayOf(DataType.BYTE, DataType.WORD, DataType.LONG, DataType.FLOAT)
 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)
@@ -150,8 +160,10 @@ val ElementToArrayTypes = mapOf(
     DataType.WORD to DataType.ARRAY_W,
     DataType.UWORD to DataType.ARRAY_UW,
     DataType.FLOAT to DataType.ARRAY_F,
-    DataType.BOOL to DataType.ARRAY_BOOL
+    DataType.BOOL to DataType.ARRAY_BOOL,
+    DataType.STR to DataType.ARRAY_UW          // array of str is just an array of pointers
 )
+
 val Cx16VirtualRegisters = arrayOf(
     RegisterOrPair.R0, RegisterOrPair.R1, RegisterOrPair.R2, RegisterOrPair.R3,
     RegisterOrPair.R4, RegisterOrPair.R5, RegisterOrPair.R6, RegisterOrPair.R7,
@@ -159,6 +171,10 @@ val Cx16VirtualRegisters = arrayOf(
     RegisterOrPair.R12, RegisterOrPair.R13, RegisterOrPair.R14, RegisterOrPair.R15
 )
 
+val CpuRegisters = setOf(
+    RegisterOrPair.A, RegisterOrPair.X, RegisterOrPair.Y,
+    RegisterOrPair.AX, RegisterOrPair.AY, RegisterOrPair.XY
+)
 
 
 enum class OutputType {

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

@@ -3,8 +3,6 @@ package prog8.code.core
 interface ICompilationTarget: IStringEncoding, IMemSizer {
     val name: String
     val machine: IMachineDefinition
-    val supportedEncodings: Set<Encoding>
-    val defaultEncoding: Encoding
 
     override fun encodeString(str: String, encoding: Encoding): List<UByte>
     override fun decodeString(bytes: Iterable<UByte>, encoding: Encoding): String

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

@@ -3,11 +3,12 @@ package prog8.code.core
 interface IErrorReporter {
     fun err(msg: String, position: Position)
     fun warn(msg: String, position: Position)
+    fun info(msg: String, position: Position)
     fun undefined(symbol: List<String>, position: Position)
     fun noErrors(): Boolean
     fun report()
-    fun finalizeNumErrors(numErrors: Int, numWarnings: Int) {
+    fun finalizeNumErrors(numErrors: Int, numWarnings: Int, numInfos: Int) {
         if(numErrors>0)
-            throw ErrorsReportedException("There are $numErrors errors and $numWarnings warnings.")
+            throw ErrorsReportedException("There are $numErrors errors, $numWarnings warnings, and $numInfos infos.")
     }
 }

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

@@ -16,6 +16,8 @@ interface IMachineDefinition {
     val PROGRAM_LOAD_ADDRESS : UInt
     val BSSHIGHRAM_START: UInt
     val BSSHIGHRAM_END: UInt
+    val BSSGOLDENRAM_START: UInt
+    val BSSGOLDENRAM_END: UInt
 
     val cpu: CpuType
     var zeropage: Zeropage

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

@@ -9,6 +9,8 @@ enum class Encoding(val prefix: String) {
 }
 
 interface IStringEncoding {
+    val defaultEncoding: Encoding
+
     fun encodeString(str: String, encoding: Encoding): List<UByte>
     fun decodeString(bytes: Iterable<UByte>, encoding: Encoding): String
 }

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

@@ -42,6 +42,13 @@ abstract class Zeropage(options: CompilationOptions): MemoryAllocator(options) {
         }
     }
 
+    fun retainAllowed() {
+        synchronized(this) {
+            for(allowed in options.zpAllowed)
+                free.retainAll { it in allowed }
+        }
+    }
+
     fun availableBytes() = if(options.zeropage== ZeropageType.DONTUSE) 0 else free.size
     fun hasByteAvailable() = if(options.zeropage== ZeropageType.DONTUSE) false else free.isNotEmpty()
     fun hasWordAvailable(): Boolean {

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

@@ -1,12 +1,10 @@
 package prog8.code.core
 
-val AssociativeOperators = setOf("+", "*", "&", "|", "^", "==", "!=")
+val AssociativeOperators = setOf("+", "*", "&", "|", "^", "==", "!=", "and", "or", "xor")
 val ComparisonOperators = setOf("==", "!=", "<", ">", "<=", ">=")
 val LogicalOperators = setOf("and", "or", "xor", "not")
-val AugmentAssignmentOperators = setOf("+", "-", "/", "*", "&", "|", "^", "<<", ">>", "%", "and", "or", "xor")
 val BitwiseOperators = setOf("&", "|", "^", "~")
 val PrefixOperators = setOf("+", "-", "~", "not")
-// val InvalidOperatorsForBoolean = setOf("+", "-", "*", "/", "%", "<<", ">>") + BitwiseOperators
 
 fun invertedComparisonOperator(operator: String) =
     when (operator) {

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

@@ -3,6 +3,7 @@ package prog8.code.core
 import java.io.File
 import java.io.IOException
 import java.nio.file.Path
+import java.text.Normalizer
 import kotlin.io.path.Path
 import kotlin.io.path.readText
 
@@ -94,7 +95,7 @@ sealed class SourceCode {
             val normalized = path.normalize()
             origin = relative(normalized).toString()
             try {
-                text = normalized.readText()
+                text = Normalizer.normalize(normalized.readText(), Normalizer.Form.NFC)
                 name = normalized.toFile().nameWithoutExtension
             } catch (nfx: java.nio.file.NoSuchFileException) {
                 throw NoSuchFileException(normalized.toFile()).also { it.initCause(nfx) }
@@ -126,7 +127,7 @@ sealed class SourceCode {
                 )
             }
             val stream = object {}.javaClass.getResourceAsStream(normalized)
-            text = stream!!.reader().use { it.readText() }
+            text = stream!!.reader().use { Normalizer.normalize(it.readText(), Normalizer.Form.NFC) }
             name = Path(pathString).toFile().nameWithoutExtension
         }
     }

codeCore/src/prog8/code/optimize/Optimizer.kt

@@ -0,0 +1,127 @@
+package prog8.code.optimize
+
+import prog8.code.ast.*
+import prog8.code.core.*
+
+
+fun optimizeIntermediateAst(program: PtProgram, options: CompilationOptions, errors: IErrorReporter) {
+    if (!options.optimize)
+        return
+    while(errors.noErrors() && optimizeCommonSubExpressions(program, errors)>0) {
+        // keep rolling
+    }
+}
+
+
+private fun walkAst(root: PtNode, act: (node: PtNode, depth: Int) -> Boolean) {
+    fun recurse(node: PtNode, depth: Int) {
+        if(act(node, depth))
+            node.children.forEach { recurse(it, depth+1) }
+    }
+    recurse(root, 0)
+}
+
+
+private fun optimizeCommonSubExpressions(program: PtProgram, errors: IErrorReporter): Int {
+
+    fun extractableSubExpr(expr: PtExpression): Boolean {
+        val result = if(expr is PtBinaryExpression)
+            expr.type !in ByteDatatypes ||
+                    !expr.left.isSimple() ||
+                    !expr.right.isSimple() ||
+                    (expr.operator !in LogicalOperators && expr.operator !in BitwiseOperators)
+        else if (expr is PtArrayIndexer && expr.type !in ByteDatatypes)
+            true
+        else
+            !expr.isSimple()
+        return result
+    }
+
+    // for each Binaryexpression, recurse to find a common subexpression pair therein.
+    val commons = mutableMapOf<PtBinaryExpression, Pair<PtExpression, PtExpression>>()
+    walkAst(program) { node: PtNode, depth: Int ->
+        if(node is PtBinaryExpression) {
+            val subExpressions = mutableListOf<PtExpression>()
+            walkAst(node.left) { subNode: PtNode, subDepth: Int ->
+                if (subNode is PtExpression) {
+                    if(extractableSubExpr(subNode)) subExpressions.add(subNode)
+                    true
+                } else false
+            }
+            walkAst(node.right) { subNode: PtNode, subDepth: Int ->
+                if (subNode is PtExpression) {
+                    if(extractableSubExpr(subNode)) subExpressions.add(subNode)
+                    true
+                } else false
+            }
+
+            outer@for (first in subExpressions) {
+                for (second in subExpressions) {
+                    if (first!==second && first isSameAs second) {
+                        commons[node] = first to second
+                        break@outer     // do only 1 replacement at a time per binaryexpression
+                    }
+                }
+            }
+            false
+        } else true
+    }
+
+    // replace common subexpressions by a temp variable that is assigned only once.
+    // TODO: check for commonalities across multiple separate expressions in the current scope, not only inside a single line
+    commons.forEach { binexpr, (occurrence1, occurrence2) ->
+        val (stmtContainer, stmt) = findContainingStatements(binexpr)
+        val occurrence1idx = occurrence1.parent.children.indexOf(occurrence1)
+        val occurrence2idx = occurrence2.parent.children.indexOf(occurrence2)
+        val containerScopedName = findScopeName(stmtContainer)
+        val tempvarName = "subexprvar_line${binexpr.position.line}_${binexpr.hashCode().toUInt()}"
+        // TODO: some tempvars could be reused, if they are from different lines
+
+        val datatype = occurrence1.type
+        val singleReplacement1 = PtIdentifier("$containerScopedName.$tempvarName", datatype, occurrence1.position)
+        val singleReplacement2 = PtIdentifier("$containerScopedName.$tempvarName", datatype, occurrence2.position)
+        occurrence1.parent.children[occurrence1idx] = singleReplacement1
+        singleReplacement1.parent = occurrence1.parent
+        occurrence2.parent.children[occurrence2idx] = singleReplacement2
+        singleReplacement2.parent = occurrence2.parent
+
+        val tempassign = PtAssignment(binexpr.position).also { assign ->
+            assign.add(PtAssignTarget(binexpr.position).also { tgt->
+                tgt.add(PtIdentifier("$containerScopedName.$tempvarName", datatype, binexpr.position))
+            })
+            assign.add(occurrence1)
+            occurrence1.parent = assign
+        }
+        stmtContainer.children.add(stmtContainer.children.indexOf(stmt), tempassign)
+        tempassign.parent = stmtContainer
+
+        val tempvar = PtVariable(tempvarName, datatype, ZeropageWish.DONTCARE, null, null, binexpr.position)
+        stmtContainer.add(0, tempvar)
+        tempvar.parent = stmtContainer
+
+        // errors.info("common subexpressions replaced by a tempvar, maybe simplify the expression manually", binexpr.position)
+    }
+
+    return commons.size
+}
+
+
+internal fun findScopeName(node: PtNode): String {
+    var parent=node
+    while(parent !is PtNamedNode)
+        parent = parent.parent
+    return parent.scopedName
+}
+
+
+internal fun findContainingStatements(node: PtNode): Pair<PtNode, PtNode> {      // returns (parentstatementcontainer, childstatement)
+    var parent = node.parent
+    var child = node
+    while(true) {
+        if(parent is IPtStatementContainer) {
+            return parent to child
+        }
+        child=parent
+        parent=parent.parent
+    }
+}

codeCore/src/prog8/code/target/AtariTarget.kt

@@ -7,7 +7,6 @@ import prog8.code.target.atari.AtariMachineDefinition
 class AtariTarget: ICompilationTarget, IStringEncoding by Encoder, IMemSizer {
     override val name = NAME
     override val machine = AtariMachineDefinition()
-    override val supportedEncodings = setOf(Encoding.ATASCII)
     override val defaultEncoding = Encoding.ATASCII
 
     companion object {
@@ -19,7 +18,7 @@ class AtariTarget: ICompilationTarget, IStringEncoding by Encoder, IMemSizer {
             in ByteDatatypes -> 1
             in WordDatatypes, in PassByReferenceDatatypes -> 2
             DataType.FLOAT -> machine.FLOAT_MEM_SIZE
-            else -> Int.MIN_VALUE
+            else -> throw IllegalArgumentException("invalid datatype")
         }
     }
 

codeCore/src/prog8/code/target/C128Target.kt

@@ -11,7 +11,6 @@ import prog8.code.target.cbm.CbmMemorySizer
 class C128Target: ICompilationTarget, IStringEncoding by Encoder, IMemSizer by CbmMemorySizer {
     override val name = NAME
     override val machine = C128MachineDefinition()
-    override val supportedEncodings = setOf(Encoding.PETSCII, Encoding.SCREENCODES)
     override val defaultEncoding = Encoding.PETSCII
 
     companion object {

codeCore/src/prog8/code/target/C64Target.kt

@@ -11,7 +11,6 @@ import prog8.code.target.cbm.CbmMemorySizer
 class C64Target: ICompilationTarget, IStringEncoding by Encoder, IMemSizer by CbmMemorySizer {
     override val name = NAME
     override val machine = C64MachineDefinition()
-    override val supportedEncodings = setOf(Encoding.PETSCII, Encoding.SCREENCODES)
     override val defaultEncoding = Encoding.PETSCII
 
     companion object {

codeCore/src/prog8/code/target/Cx16Target.kt

@@ -11,7 +11,6 @@ import prog8.code.target.cx16.CX16MachineDefinition
 class Cx16Target: ICompilationTarget, IStringEncoding by Encoder, IMemSizer by CbmMemorySizer {
     override val name = NAME
     override val machine = CX16MachineDefinition()
-    override val supportedEncodings = setOf(Encoding.PETSCII, Encoding.SCREENCODES, Encoding.ISO)
     override val defaultEncoding = Encoding.PETSCII
 
     companion object {

codeCore/src/prog8/code/target/Encoder.kt

@@ -10,6 +10,8 @@ import prog8.code.target.cbm.PetsciiEncoding
 
 
 object Encoder: IStringEncoding {
+    override val defaultEncoding: Encoding = Encoding.ISO
+
     override fun encodeString(str: String, encoding: Encoding): List<UByte> {
         val coded = when(encoding) {
             Encoding.PETSCII -> PetsciiEncoding.encodePetscii(str, true)

codeCore/src/prog8/code/target/PETTarget.kt

@@ -11,7 +11,6 @@ import prog8.code.target.pet.PETMachineDefinition
 class PETTarget: ICompilationTarget, IStringEncoding by Encoder, IMemSizer by CbmMemorySizer {
     override val name = NAME
     override val machine = PETMachineDefinition()
-    override val supportedEncodings = setOf(Encoding.PETSCII, Encoding.SCREENCODES)
     override val defaultEncoding = Encoding.PETSCII
 
     companion object {

codeCore/src/prog8/code/target/VMTarget.kt

@@ -6,7 +6,6 @@ import prog8.code.target.virtual.VirtualMachineDefinition
 class VMTarget: ICompilationTarget, IStringEncoding by Encoder, IMemSizer {
     override val name = NAME
     override val machine = VirtualMachineDefinition()
-    override val supportedEncodings = setOf(Encoding.ISO)
     override val defaultEncoding = Encoding.ISO
 
     companion object {
@@ -18,7 +17,7 @@ class VMTarget: ICompilationTarget, IStringEncoding by Encoder, IMemSizer {
             in ByteDatatypes -> 1
             in WordDatatypes, in PassByReferenceDatatypes -> 2
             DataType.FLOAT -> machine.FLOAT_MEM_SIZE
-            else -> Int.MIN_VALUE
+            else -> throw IllegalArgumentException("invalid datatype")
         }
     }
 

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

@@ -15,6 +15,8 @@ class AtariMachineDefinition: IMachineDefinition {
 
     override val BSSHIGHRAM_START = 0u    // TODO
     override val BSSHIGHRAM_END = 0u      // TODO
+    override val BSSGOLDENRAM_START = 0u  // TODO
+    override val BSSGOLDENRAM_END = 0u    // TODO
 
     override lateinit var zeropage: Zeropage
     override lateinit var golden: GoldenRam

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

@@ -48,6 +48,7 @@ class AtariZeropage(options: CompilationOptions) : Zeropage(options) {
         free.addAll(distinctFree)
 
         removeReservedFromFreePool()
+        retainAllowed()
     }
 
     override fun allocateCx16VirtualRegisters() {

codeCore/src/prog8/code/target/c128/C128MachineDefinition.kt

@@ -17,6 +17,8 @@ class C128MachineDefinition: IMachineDefinition {
 
     override val BSSHIGHRAM_START = 0u    // TODO
     override val BSSHIGHRAM_END = 0u      // TODO
+    override val BSSGOLDENRAM_START = 0u  // TODO
+    override val BSSGOLDENRAM_END = 0u    // TODO
 
     override lateinit var zeropage: Zeropage
     override lateinit var golden: GoldenRam

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

@@ -69,6 +69,7 @@ class C128Zeropage(options: CompilationOptions) : Zeropage(options) {
         free.addAll(distinctFree)
 
         removeReservedFromFreePool()
+        retainAllowed()
     }
 
     override fun allocateCx16VirtualRegisters() {

codeCore/src/prog8/code/target/c64/C64MachineDefinition.kt

@@ -17,7 +17,9 @@ class C64MachineDefinition: IMachineDefinition {
     override val PROGRAM_LOAD_ADDRESS = 0x0801u
 
     override val BSSHIGHRAM_START = 0xc000u
-    override val BSSHIGHRAM_END = 0xd000u
+    override val BSSHIGHRAM_END = 0xcfffu
+    override val BSSGOLDENRAM_START = 0u
+    override val BSSGOLDENRAM_END = 0u
 
     override lateinit var zeropage: Zeropage
     override lateinit var golden: GoldenRam

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

@@ -75,6 +75,8 @@ class C64Zeropage(options: CompilationOptions) : Zeropage(options) {
             // in these cases there is enough space on the zero page to stick the cx16 virtual registers in there as well.
             allocateCx16VirtualRegisters()
         }
+
+        retainAllowed()
     }
 
     override fun allocateCx16VirtualRegisters() {

codeCore/src/prog8/code/target/cbm/CbmMemorySizer.kt

@@ -9,7 +9,7 @@ internal object CbmMemorySizer: IMemSizer {
             in ByteDatatypes -> 1
             in WordDatatypes, in PassByReferenceDatatypes -> 2
             DataType.FLOAT -> Mflpt5.FLOAT_MEM_SIZE
-            else -> Int.MIN_VALUE
+            else -> throw IllegalArgumentException("invalid datatype")
         }
     }
 

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

@@ -16,7 +16,9 @@ class CX16MachineDefinition: IMachineDefinition {
     override val PROGRAM_LOAD_ADDRESS = 0x0801u
 
     override val BSSHIGHRAM_START = 0xa000u     // hiram bank 1, 8Kb, assumed to be active
-    override val BSSHIGHRAM_END = 0xc000u       // rom starts here.
+    override val BSSHIGHRAM_END = 0xbfffu       // Rom starts at $c000
+    override val BSSGOLDENRAM_START = 0x0400u
+    override val BSSGOLDENRAM_END = 0x07ffu
 
     override lateinit var zeropage: Zeropage
     override lateinit var golden: GoldenRam
@@ -36,7 +38,7 @@ class CX16MachineDefinition: IMachineDefinition {
         when(selectedEmulator) {
             1 -> {
                 emulator = "x16emu"
-                extraArgs = emptyList()
+                extraArgs = listOf("-debug")
             }
             2 -> {
                 emulator = "box16"
@@ -49,7 +51,7 @@ class CX16MachineDefinition: IMachineDefinition {
         }
 
         println("\nStarting Commander X16 emulator $emulator...")
-        val cmdline = listOf(emulator, "-scale", "2", "-run", "-prg", "${programNameWithPath}.prg") + extraArgs
+        val cmdline = listOf(emulator, "-scale", "2", "-rtc", "-run", "-prg", "${programNameWithPath}.prg") + extraArgs
         val processb = ProcessBuilder(cmdline).inheritIO()
         processb.environment()["PULSE_LATENCY_MSEC"] = "10"
         val process: Process = processb.start()

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

@@ -47,8 +47,8 @@ class CX16Zeropage(options: CompilationOptions) : Zeropage(options) {
             free.addAll(distinctFree)
 
             removeReservedFromFreePool()
-
             allocateCx16VirtualRegisters()
+            retainAllowed()
         }
     }
 

codeCore/src/prog8/code/target/pet/PETMachineDefinition.kt

@@ -15,8 +15,10 @@ class PETMachineDefinition: IMachineDefinition {
     override val FLOAT_MEM_SIZE = Mflpt5.FLOAT_MEM_SIZE
     override val PROGRAM_LOAD_ADDRESS = 0x0401u
 
-    override val BSSHIGHRAM_START = 0xffffu
-    override val BSSHIGHRAM_END = 0xffffu
+    override val BSSHIGHRAM_START = 0u
+    override val BSSHIGHRAM_END = 0u
+    override val BSSGOLDENRAM_START = 0u
+    override val BSSGOLDENRAM_END = 0u
 
     override lateinit var zeropage: Zeropage
     override lateinit var golden: GoldenRam

codeCore/src/prog8/code/target/pet/PETZeropage.kt

@@ -50,6 +50,7 @@ class PETZeropage(options: CompilationOptions) : Zeropage(options) {
         free.addAll(distinctFree)
 
         removeReservedFromFreePool()
+        retainAllowed()
     }
 
     override fun allocateCx16VirtualRegisters() {

codeCore/src/prog8/code/target/virtual/VirtualMachineDefinition.kt

@@ -12,18 +12,20 @@ class VirtualMachineDefinition: IMachineDefinition {
 
     override val FLOAT_MAX_POSITIVE = Float.MAX_VALUE.toDouble()
     override val FLOAT_MAX_NEGATIVE = -Float.MAX_VALUE.toDouble()
-    override val FLOAT_MEM_SIZE = 4             // 32-bits floating point
+    override val FLOAT_MEM_SIZE = 8             // 64-bits double
     override val PROGRAM_LOAD_ADDRESS = 0u      // not actually used
 
     override val BSSHIGHRAM_START = 0u          // not actually used
     override val BSSHIGHRAM_END = 0u            // not actually used
+    override val BSSGOLDENRAM_START = 0u        // not actually used
+    override val BSSGOLDENRAM_END = 0u          // not actually used
     override lateinit var zeropage: Zeropage    // not actually used
     override lateinit var golden: GoldenRam     // not actually used
 
     override fun getFloatAsmBytes(num: Number): String {
         // little endian binary representation
-        val bits = num.toFloat().toBits().toUInt()
-        val hexStr = bits.toString(16).padStart(8, '0')
+        val bits = num.toDouble().toBits().toULong()
+        val hexStr = bits.toString(16).padStart(16, '0')
         val parts = hexStr.chunked(2).map { "\$" + it }
         return parts.joinToString(", ")
     }

codeGenCpu6502/build.gradle

@@ -30,7 +30,7 @@ dependencies {
     // implementation "org.jetbrains.kotlin:kotlin-reflect"
     implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18"
 
-    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.6.2'
+    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.8.0'
 }
 
 sourceSets {

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

@@ -1,6 +1,7 @@
 package prog8.codegen.cpu6502
 
 import com.github.michaelbull.result.fold
+import prog8.code.StNode
 import prog8.code.StNodeType
 import prog8.code.SymbolTable
 import prog8.code.SymbolTableMaker
@@ -32,7 +33,14 @@ class AsmGen6502(val prefixSymbols: Boolean): ICodeGeneratorBackend {
         val functionCallsToPrefix = mutableListOf<Pair<PtNode, Int>>()      // parent + index
 
         fun prefixNamedNode(node: PtNamedNode) {
-            node.name = "p8_${node.name}"
+            when(node) {
+                is PtAsmSub, is PtSub -> node.name = "p8s_${node.name}"
+                is PtBlock -> node.name = "p8b_${node.name}"
+                is PtLabel -> node.name = "p8l_${node.name}"
+                is PtConstant -> node.name = "p8c_${node.name}"
+                is PtVariable, is PtMemMapped, is PtSubroutineParameter -> node.name = "p8v_${node.name}"
+                else -> node.name = "p8_${node.name}"
+            }
         }
 
         fun prefixSymbols(node: PtNode) {
@@ -47,7 +55,7 @@ class AsmGen6502(val prefixSymbols: Boolean): ICodeGeneratorBackend {
                 }
                 is PtFunctionCall -> {
                     val stNode = st.lookup(node.name)!!
-                    if(stNode.astNode.definingBlock()?.noSymbolPrefixing!=true) {
+                    if(stNode.astNode.definingBlock()?.options?.noSymbolPrefixing!=true) {
                         val index = node.parent.children.indexOf(node)
                         functionCallsToPrefix += node.parent to index
                     }
@@ -58,25 +66,16 @@ class AsmGen6502(val prefixSymbols: Boolean): ICodeGeneratorBackend {
                         lookupName = lookupName.dropLast(4)
                     }
                     val stNode = st.lookup(lookupName)!!
-                    if(stNode.astNode.definingBlock()?.noSymbolPrefixing!=true) {
+                    if(stNode.astNode.definingBlock()?.options?.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
-                        }
+                    val stNode = st.lookup(node.identifier!!.name) ?: throw AssemblyError("name not found ${node.identifier}")
+                    if(stNode.astNode.definingBlock()?.options?.noSymbolPrefixing!=true) {
+                        val index = node.parent.children.indexOf(node)
+                        nodesToPrefix += node.parent to index
                     }
                 }
                 is PtBlock -> prefixNamedNode(node)
@@ -94,7 +93,7 @@ class AsmGen6502(val prefixSymbols: Boolean): ICodeGeneratorBackend {
         }
 
         program.allBlocks().forEach { block ->
-            if (!block.noSymbolPrefixing) {
+            if (!block.options.noSymbolPrefixing) {
                 prefixSymbols(block)
             }
         }
@@ -124,8 +123,20 @@ class AsmGen6502(val prefixSymbols: Boolean): ICodeGeneratorBackend {
     }
 }
 
+private fun prefixScopedName(name: String, type: Char): String {
+    if('.' !in name)
+        return "p8${type}_$name"
+    val parts = name.split('.')
+    val firstPrefixed = "p8b_${parts[0]}"
+    val lastPrefixed = "p8${type}_${parts.last()}"
+    // the parts in between are assumed to be subroutine scopes.
+    val inbetweenPrefixed = parts.drop(1).dropLast(1).map{ "p8s_$it" }
+    val prefixed = listOf(firstPrefixed) + inbetweenPrefixed + listOf(lastPrefixed)
+    return prefixed.joinToString(".")
+}
+
 private fun PtVariable.prefix(st: SymbolTable): PtVariable {
-    name = name.split('.').map {"p8_$it" }.joinToString(".")
+    name = prefixScopedName(name, 'v')
     if(value==null)
         return this
 
@@ -137,7 +148,7 @@ private fun PtVariable.prefix(st: SymbolTable): PtVariable {
                 is PtIdentifier -> newValue.add(elt.prefix(arrayValue, st))
                 is PtNumber -> newValue.add(elt)
                 is PtAddressOf -> {
-                    if(elt.definingBlock()?.noSymbolPrefixing==true)
+                    if(elt.definingBlock()?.options?.noSymbolPrefixing==true)
                         newValue.add(elt)
                     else {
                         val newAddr = PtAddressOf(elt.position)
@@ -155,19 +166,14 @@ private fun PtVariable.prefix(st: SymbolTable): PtVariable {
 }
 
 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)
-    }
+    val prefixedIdent = identifier!!.prefix(this, st)
+    val jump = PtJump(prefixedIdent, address, position)
     jump.parent = parent
     return jump
 }
 
 private fun PtFunctionCall.prefix(parent: PtNode): PtFunctionCall {
-    val newName = name.split('.').map {"p8_$it" }.joinToString(".")
+    val newName = prefixScopedName(name, 's')
     val call = PtFunctionCall(newName, void, type, position)
     call.children.addAll(children)
     call.children.forEach { it.parent = call }
@@ -176,11 +182,29 @@ private fun PtFunctionCall.prefix(parent: PtNode): PtFunctionCall {
 }
 
 private fun PtIdentifier.prefix(parent: PtNode, st: SymbolTable): PtIdentifier {
-    val target = st.lookup(name)
-    if(target?.astNode?.definingBlock()?.noSymbolPrefixing==true)
+    var target = st.lookup(name)
+    if(target?.astNode?.definingBlock()?.options?.noSymbolPrefixing==true)
         return this
 
-    val newName = name.split('.').map { "p8_$it" }.joinToString(".")
+    if(target==null) {
+        if(name.endsWith("_lsb") || name.endsWith("_msb")) {
+            target = st.lookup(name.dropLast(4))
+            if(target?.astNode?.definingBlock()?.options?.noSymbolPrefixing==true)
+                return this
+        }
+    }
+
+    val prefixType = when(target!!.type) {
+        StNodeType.BLOCK -> 'b'
+        StNodeType.SUBROUTINE, StNodeType.ROMSUB -> 's'
+        StNodeType.LABEL -> 'l'
+        StNodeType.STATICVAR, StNodeType.MEMVAR -> 'v'
+        StNodeType.CONSTANT -> 'c'
+        StNodeType.BUILTINFUNC -> 's'
+        StNodeType.MEMORYSLAB -> 'v'
+        else -> '?'
+    }
+    val newName = prefixScopedName(name, prefixType)
     val node = PtIdentifier(newName, type, position)
     node.parent = parent
     return node
@@ -296,6 +320,18 @@ class AsmGen6502Internal (
             name
     }
 
+    fun asmVariableName(st: StNode, scope: PtSub?): String {
+        val name = asmVariableName(st.scopedName)
+        if(scope==null)
+            return name
+        // remove the whole prefix and just make the variable name locally scoped (64tass scopes it to the proper .proc block)
+        val subName = scope.scopedName
+        return if (name.length>subName.length && name.startsWith(subName) && name[subName.length] == '.')
+            name.drop(subName.length+1)
+        else
+            name
+    }
+
     internal fun getTempVarName(dt: DataType): String {
         return when(dt) {
             DataType.UBYTE -> "cx16.r9L"
@@ -303,7 +339,7 @@ class AsmGen6502Internal (
             DataType.UWORD -> "cx16.r9"
             DataType.WORD -> "cx16.r9s"
             DataType.FLOAT -> "floats.floats_temp_var"
-            else -> throw AssemblyError("invalid dt $dt")
+            else -> throw AssemblyError("invalid dt")
         }
     }
 
@@ -487,40 +523,30 @@ class AsmGen6502Internal (
         }
     }
 
-    internal fun loadScaledArrayIndexIntoRegister(
-        expr: PtArrayIndexer,
-        elementDt: DataType,
-        register: CpuRegister
-    ) {
+    internal fun loadScaledArrayIndexIntoRegister(expr: PtArrayIndexer, register: CpuRegister) {
         val reg = register.toString().lowercase()
         val indexnum = expr.index.asConstInteger()
         if (indexnum != null) {
             val indexValue = if(expr.splitWords)
                 indexnum
             else
-                indexnum * options.compTarget.memorySize(elementDt)
+                indexnum * options.compTarget.memorySize(expr.type)
             out("  ld$reg  #$indexValue")
             return
         }
 
-        val indexVar = expr.index as? PtIdentifier
-            ?: throw AssemblyError("array indexer should have been replaced with a temp var @ ${expr.index.position}")
-
-        val indexName = asmVariableName(indexVar)
-
         if(expr.splitWords) {
-            when (register) {
-                CpuRegister.A -> out("  lda  $indexName")
-                CpuRegister.X -> out("  ldx  $indexName")
-                CpuRegister.Y -> out("  ldy  $indexName")
-            }
+            assignExpressionToRegister(expr.index, RegisterOrPair.fromCpuRegister(register), false)
             return
         }
 
-        when (elementDt) {
-            in ByteDatatypes -> out("  ld$reg  $indexName")
+        when (expr.type) {
+            in ByteDatatypes -> {
+                assignExpressionToRegister(expr.index, RegisterOrPair.fromCpuRegister(register), false)
+            }
             in WordDatatypes -> {
-                out("  lda  $indexName |  asl  a")
+                assignExpressionToRegister(expr.index, RegisterOrPair.A, false)
+                out("  asl  a")
                 when (register) {
                     CpuRegister.A -> {}
                     CpuRegister.X -> out(" tax")
@@ -529,12 +555,13 @@ class AsmGen6502Internal (
             }
             DataType.FLOAT -> {
                 require(options.compTarget.memorySize(DataType.FLOAT) == 5) {"invalid float size ${expr.position}"}
+                assignExpressionToRegister(expr.index, RegisterOrPair.A, false)
                 out("""
-                    lda  $indexName
+                    sta  P8ZP_SCRATCH_REG
                     asl  a
                     asl  a
                     clc
-                    adc  $indexName"""
+                    adc  P8ZP_SCRATCH_REG"""
                 )
                 when (register) {
                     CpuRegister.A -> {}
@@ -568,7 +595,7 @@ class AsmGen6502Internal (
         when(reg) {
             RegisterOrPair.A,
             RegisterOrPair.X,
-            RegisterOrPair.Y -> assignmentAsmGen.assignRegisterByte(target, reg.asCpuRegister(), target.datatype in SignedDatatypes)
+            RegisterOrPair.Y -> assignmentAsmGen.assignRegisterByte(target, reg.asCpuRegister(), target.datatype in SignedDatatypes, true)
             RegisterOrPair.AX,
             RegisterOrPair.AY,
             RegisterOrPair.XY,
@@ -582,6 +609,34 @@ class AsmGen6502Internal (
     internal fun assignExpressionTo(value: PtExpression, target: AsmAssignTarget) {
         when (target.datatype) {
             in ByteDatatypes -> {
+                if (value.asConstInteger()==0) {
+                    when(target.kind) {
+                        TargetStorageKind.VARIABLE -> {
+                            if (isTargetCpu(CpuType.CPU6502))
+                                out("lda  #0 |  sta  ${target.asmVarname}")
+                            else
+                                out("stz  ${target.asmVarname}")
+                        }
+                        TargetStorageKind.MEMORY -> {
+                            val address = target.memory!!.address.asConstInteger()
+                            if(address!=null) {
+                                if (isTargetCpu(CpuType.CPU6502))
+                                    out("lda  #0 |  sta  ${address.toHex()}")
+                                else
+                                    out("  stz  ${address.toHex()}")
+                                return
+                            }
+                        }
+                        TargetStorageKind.REGISTER -> {
+                            val zero = PtNumber(DataType.UBYTE, 0.0, value.position)
+                            zero.parent = value
+                            assignExpressionToRegister(zero, target.register!!, false)
+                            return
+                        }
+                        else -> { }
+                    }
+                }
+
                 assignExpressionToRegister(value, RegisterOrPair.A)
                 assignRegister(RegisterOrPair.A, target)
             }
@@ -661,7 +716,6 @@ class AsmGen6502Internal (
                 if (jump is PtJump) {
                     // jump somewhere if X!=0
                     val label = when {
-                        jump.generatedLabel!=null -> jump.generatedLabel!!
                         jump.identifier!=null -> asmSymbolName(jump.identifier!!)
                         jump.address!=null -> jump.address!!.toHex()
                         else -> throw AssemblyError("weird jump")
@@ -961,7 +1015,6 @@ $repeatLabel""")
 
     private fun getJumpTarget(jump: PtJump): Pair<String, Boolean> {
         val ident = jump.identifier
-        val label = jump.generatedLabel
         val addr = jump.address
         return when {
             ident!=null -> {
@@ -972,7 +1025,6 @@ $repeatLabel""")
                 else
                     Pair(asmSymbolName(ident), false)
             }
-            label!=null -> Pair(label, false)
             addr!=null -> Pair(addr.toHex(), false)
             else -> Pair("????", false)
         }
@@ -1026,6 +1078,15 @@ $repeatLabel""")
         val label = "_prog8_breakpoint_${breakpointLabels.size+1}"
         breakpointLabels.add(label)
         out(label)
+        if(options.breakpointCpuInstruction) {
+            val instruction =
+                when(options.compTarget.machine.cpu) {
+                    CpuType.CPU6502 -> "brk"
+                    CpuType.CPU65c02 -> "stp"
+                    else -> throw AssemblyError("invalid cpu type")
+                }
+            out("  $instruction")
+        }
     }
 
     internal fun signExtendAYlsb(valueDt: DataType) {
@@ -1213,7 +1274,6 @@ $repeatLabel""")
         val rightConstVal = right as? PtNumber
 
         val label = when {
-            jump.generatedLabel!=null -> jump.generatedLabel!!
             jump.identifier!=null -> asmSymbolName(jump.identifier!!)
             jump.address!=null -> jump.address!!.toHex()
             else -> throw AssemblyError("weird jump")
@@ -2997,30 +3057,16 @@ $repeatLabel""")
         }
     }
 
-    internal fun popCpuStack(dt: DataType, target: IPtVariable, scope: IPtSubroutine?) {
-        // note: because A is pushed first so popped last, saving A is often not required here.
-        val targetAsmSub = (target as PtNode).definingAsmSub()
-        if(targetAsmSub != null) {
-            val parameter = targetAsmSub.parameters.first { it.second.name==target.name }
-            popCpuStack(targetAsmSub, parameter.second, parameter.first)
-            return
-        }
-        val scopedName = when(target) {
-            is PtConstant -> target.scopedName
-            is PtMemMapped -> target.scopedName
-            is PtVariable -> target.scopedName
-            else -> throw AssemblyError("weird target var")
-        }
-        val tgt = AsmAssignTarget(TargetStorageKind.VARIABLE, this, target.type, scope, target.position, variableAsmName = asmVariableName(scopedName))
+    internal fun popCpuStack(dt: DataType) {
         if (dt in ByteDatatypes) {
             out("  pla")
-            assignRegister(RegisterOrPair.A, tgt)
-        } else {
+        } else if (dt in WordDatatypes) {
             if (isTargetCpu(CpuType.CPU65c02))
                 out("  ply |  pla")
             else
                 out("  pla |  tay |  pla")
-            assignRegister(RegisterOrPair.AY, tgt)
+        } else {
+            throw AssemblyError("can't pop type $dt")
         }
     }
 
@@ -3072,9 +3118,20 @@ $repeatLabel""")
 
     internal fun makeLabel(postfix: String): String {
         generatedLabelSequenceNumber++
-        return "prog8_label_asm_${generatedLabelSequenceNumber}_$postfix"
+        return "label_asm_${generatedLabelSequenceNumber}_$postfix"
     }
 
+    fun assignConstFloatToPointerAY(number: PtNumber) {
+        val floatConst = allocator.getFloatAsmConst(number.number)
+        out("""
+            pha
+            lda  #<$floatConst
+            sta  P8ZP_SCRATCH_W1
+            lda  #>$floatConst
+            sta  P8ZP_SCRATCH_W1+1
+            pla
+            jsr  floats.copy_float""")
+    }
 }
 
 /**

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

@@ -381,15 +381,9 @@ private fun optimizeStoreLoadSame(
         } else if(first=="phx" && second=="pla") {
             mods.add(Modification(lines[1].index, true, null))
             mods.add(Modification(lines[2].index, false, "  txa"))
-        } else if(first=="phx" && second=="ply") {
-            mods.add(Modification(lines[1].index, true, null))
-            mods.add(Modification(lines[2].index, false, "  txy"))
         } else if(first=="phy" && second=="pla") {
             mods.add(Modification(lines[1].index, true, null))
             mods.add(Modification(lines[2].index, false, "  tya"))
-        } else if(first=="phy" && second=="plx") {
-            mods.add(Modification(lines[1].index, true, null))
-            mods.add(Modification(lines[2].index, false, "  tyx"))
         }
     }
     return mods

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

@@ -1,9 +1,8 @@
 package prog8.codegen.cpu6502
 
-import prog8.code.ast.*
+import prog8.code.ast.PtAsmSub
 import prog8.code.core.Cx16VirtualRegisters
 import prog8.code.core.RegisterOrPair
-import prog8.code.core.RegisterOrStatusflag
 
 
 fun asmsub6502ArgsEvalOrder(sub: PtAsmSub): List<Int> {
@@ -11,50 +10,19 @@ fun asmsub6502ArgsEvalOrder(sub: PtAsmSub): List<Int> {
     // order is:
     //  1) cx16 virtual word registers,
     //  2) paired CPU registers,
-    //  3) single CPU registers (X last), except A,
+    //  3) single CPU registers (order Y,X,A),
     //  4) CPU Carry status flag
-    //  5) the A register itself last   (so everything before it can use the accumulator without having to save its value)
     val args = sub.parameters.withIndex()
     val (cx16regs, args2) = args.partition { it.value.first.registerOrPair in Cx16VirtualRegisters }
     val pairedRegisters = arrayOf(RegisterOrPair.AX, RegisterOrPair.AY, RegisterOrPair.XY)
     val (pairedRegs , args3) = args2.partition { it.value.first.registerOrPair in pairedRegisters }
-    val (regsWithoutA, args4) = args3.partition { it.value.first.registerOrPair != RegisterOrPair.A }
-    val (regA, rest) = args4.partition { it.value.first.registerOrPair != null }
+    val (singleRegs, rest) = args3.partition { it.value.first.registerOrPair != null }
 
     cx16regs.forEach { order += it.index }
     pairedRegs.forEach { order += it.index }
-    regsWithoutA.forEach {
-        if(it.value.first.registerOrPair != RegisterOrPair.X)
-            order += it.index
-    }
-    regsWithoutA.firstOrNull { it.value.first.registerOrPair==RegisterOrPair.X } ?.let { order += it.index}
+    singleRegs.sortedBy { it.value.first.registerOrPair!!.asCpuRegister() }.asReversed().forEach { order += it.index }
+    require(rest.all { it.value.first.registerOrPair==null && it.value.first.statusflag!=null})
     rest.forEach { order += it.index }
-    regA.forEach { order += it.index }
     require(order.size==sub.parameters.size)
     return order
 }
-
-fun asmsub6502ArgsHaveRegisterClobberRisk(
-    args: List<PtExpression>,
-    params: List<Pair<RegisterOrStatusflag, PtSubroutineParameter>>
-): Boolean {
-    fun isClobberRisk(expr: PtExpression): Boolean {
-        when (expr) {
-            is PtArrayIndexer -> {
-                return params.any {
-                    it.first.registerOrPair in listOf(RegisterOrPair.Y, RegisterOrPair.AY, RegisterOrPair.XY)
-                }
-            }
-            is PtBuiltinFunctionCall -> {
-                if (expr.name == "lsb" || expr.name == "msb")
-                    return isClobberRisk(expr.args[0])
-                if (expr.name == "mkword")
-                    return isClobberRisk(expr.args[0]) && isClobberRisk(expr.args[1])
-                return !expr.isSimple()
-            }
-            else -> return !expr.isSimple()
-        }
-    }
-
-    return args.size>1 && args.any { isClobberRisk(it) }
-}

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

@@ -1,5 +1,6 @@
 package prog8.codegen.cpu6502
 
+import prog8.code.StMemVar
 import prog8.code.StStaticVariable
 import prog8.code.ast.*
 import prog8.code.core.*
@@ -47,51 +48,51 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             "reverse" -> funcReverse(fcall)
             "memory" -> funcMemory(fcall, discardResult, resultRegister)
             "peekw" -> funcPeekW(fcall, resultRegister)
+            "peekf" -> funcPeekF(fcall, resultRegister)
             "peek" -> throw AssemblyError("peek() should have been replaced by @()")
             "pokew" -> funcPokeW(fcall)
-            "pokemon" -> { /* meme function */ }
+            "pokef" -> funcPokeF(fcall)
+            "pokemon" -> {
+                val memread = PtMemoryByte(fcall.position)
+                memread.add(fcall.args[0])
+                memread.parent = fcall
+                asmgen.assignExpressionToRegister(memread, RegisterOrPair.A, false)
+                asmgen.out("  pha")
+                val memtarget = AsmAssignTarget(TargetStorageKind.MEMORY, asmgen, DataType.UBYTE, fcall.definingISub(), fcall.position, memory=memread)
+                asmgen.assignExpressionTo(fcall.args[1], memtarget)
+                asmgen.out("  pla")
+            }
             "poke" -> throw AssemblyError("poke() should have been replaced by @()")
-            "push" -> asmgen.pushCpuStack(DataType.UBYTE, fcall.args[0])
-            "pushw" -> asmgen.pushCpuStack(DataType.UWORD, fcall.args[0])
-            "pop" -> {
-                require(fcall.args[0] is PtIdentifier) {
-                    "attempt to pop a value into a differently typed variable, or in something else that isn't supported ${fcall.position}"
-                }
-                val symbol = asmgen.symbolTable.lookup((fcall.args[0] as PtIdentifier).name)
-                val target = symbol!!.astNode as IPtVariable
-                asmgen.popCpuStack(DataType.UBYTE, target, fcall.definingISub())
-            }
-            "popw" -> {
-                require(fcall.args[0] is PtIdentifier) {
-                    "attempt to pop a value into a differently typed variable, or in something else that isn't supported ${fcall.position}"
-                }
-                val symbol = asmgen.symbolTable.lookup((fcall.args[0] as PtIdentifier).name)
-                val target = symbol!!.astNode as IPtVariable
-                asmgen.popCpuStack(DataType.UWORD, target, fcall.definingISub())
-            }
             "rsave" -> funcRsave()
             "rrestore" -> funcRrestore()
             "cmp" -> funcCmp(fcall)
-            "callfar" -> funcCallFar(fcall)
-            "prog8_lib_stringcompare" -> funcStringCompare(fcall)
-            "prog8_lib_square_byte" -> funcSquare(fcall, DataType.UBYTE)
-            "prog8_lib_square_word" -> funcSquare(fcall, DataType.UWORD)
+            "callfar" -> funcCallFar(fcall, resultRegister)
+            "call" -> funcCall(fcall)
+            "prog8_lib_stringcompare" -> funcStringCompare(fcall, resultRegister)
+            "prog8_lib_square_byte" -> funcSquare(fcall, DataType.UBYTE, resultRegister)
+            "prog8_lib_square_word" -> funcSquare(fcall, DataType.UWORD, resultRegister)
             else -> throw AssemblyError("missing asmgen for builtin func ${fcall.name}")
         }
 
         return BuiltinFunctions.getValue(fcall.name).returnType
     }
 
-    private fun funcSquare(fcall: PtBuiltinFunctionCall, resultType: DataType) {
+    private fun funcSquare(fcall: PtBuiltinFunctionCall, resultType: DataType, resultRegister: RegisterOrPair?) {
         // square of word value is faster with dedicated routine, square of byte just use the regular multiplication routine.
         when (resultType) {
             DataType.UBYTE -> {
                 asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.A)
                 asmgen.out("  tay  |  jsr  math.multiply_bytes")
+                if(resultRegister!=null)  {
+                    assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister, false, fcall.position, null, asmgen), CpuRegister.A, false, false)
+                }
             }
             DataType.UWORD -> {
                 asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.AY)
                 asmgen.out("  jsr  math.square")
+                if(resultRegister!=null)  {
+                    assignAsmGen.assignRegisterpairWord(AsmAssignTarget.fromRegisters(resultRegister, false, fcall.position, null, asmgen), RegisterOrPair.AY)
+                }
             }
             else -> {
                 throw AssemblyError("optimized square only for integer types")
@@ -110,8 +111,8 @@ 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, false)
-        assignAsmGen.assignRegisterByte(divisionTarget, CpuRegister.Y, false)
+        assignAsmGen.assignRegisterByte(remainderTarget, CpuRegister.A, false, false)
+        assignAsmGen.assignRegisterByte(divisionTarget, CpuRegister.Y, false, false)
     }
 
     private fun funcDivmodW(fcall: PtBuiltinFunctionCall) {
@@ -131,9 +132,12 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             sty  $remainderVar+1""")
     }
 
-    private fun funcStringCompare(fcall: PtBuiltinFunctionCall) {
+    private fun funcStringCompare(fcall: PtBuiltinFunctionCall, resultRegister: RegisterOrPair?) {
         asmgen.assignWordOperandsToAYAndVar(fcall.args[0], fcall.args[1], "P8ZP_SCRATCH_W2")
         asmgen.out("  jsr  prog8_lib.strcmp_mem")
+        if(resultRegister!=null)  {
+            assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister, false, fcall.position, null, asmgen), CpuRegister.A, false, false)
+        }
     }
 
     private fun funcRsave() {
@@ -173,7 +177,22 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                 plp""")
     }
 
-    private fun funcCallFar(fcall: PtBuiltinFunctionCall) {
+    private fun funcCall(fcall: PtBuiltinFunctionCall) {
+        val constAddr = fcall.args[0].asConstInteger()
+        if(constAddr!=null) {
+            asmgen.out("  jsr  ${constAddr.toHex()}")
+        } else {
+            asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.AY)     // jump address
+            asmgen.out("""
+              sta  (+)+1
+              sty  (+)+2
++             jsr  0       ; modified""")
+        }
+
+        // note: the routine can return a word value (in AY)
+    }
+
+    private fun funcCallFar(fcall: PtBuiltinFunctionCall, resultRegister: RegisterOrPair?) {
         if(asmgen.options.compTarget.name != "cx16")
             throw AssemblyError("callfar only works on cx16 target at this time")
 
@@ -187,6 +206,9 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
 +           .word  0
 +           .byte  0""")
         // note that by convention the values in A+Y registers are now the return value of the call.
+        if(resultRegister!=null)  {
+            assignAsmGen.assignRegisterpairWord(AsmAssignTarget.fromRegisters(resultRegister, false, fcall.position, null, asmgen), RegisterOrPair.AY)
+        }
     }
 
     private fun funcCmp(fcall: PtBuiltinFunctionCall) {
@@ -273,11 +295,11 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         when(fcall.args[0].type) {
             DataType.UBYTE -> {
                 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)
+                assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, false, false)
             }
             DataType.UWORD -> {
                 asmgen.out("  jsr  prog8_lib.func_sqrt16_into_A")
-                assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, false)
+                assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, false, false)
             }
             DataType.FLOAT -> {
                 asmgen.out("  jsr  floats.func_sqrt_into_FAC1")
@@ -288,106 +310,112 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
     }
 
     private fun funcReverse(fcall: PtBuiltinFunctionCall) {
-        val variable = fcall.args.single()
-        if (variable is PtIdentifier) {
-            val symbol = asmgen.symbolTable.lookup(variable.name)
-            val decl = symbol!!.astNode as IPtVariable
-            val numElements = when(decl) {
-                is PtConstant -> throw AssemblyError("cannot reverse a constant")
-                is PtMemMapped -> decl.arraySize
-                is PtVariable -> decl.arraySize
+        val variable = fcall.args.single() as PtIdentifier
+        val symbol = asmgen.symbolTable.lookup(variable.name)
+        val (dt, numElements) = when(symbol) {
+            is StStaticVariable  -> symbol.dt to symbol.length!!
+            is StMemVar -> symbol.dt to symbol.length!!
+            else -> DataType.UNDEFINED to 0
+        }
+        val varName = asmgen.asmVariableName(variable)
+        when (dt) {
+            DataType.ARRAY_UB, DataType.ARRAY_B -> {
+                asmgen.out("""
+                    lda  #<$varName
+                    ldy  #>$varName
+                    sta  P8ZP_SCRATCH_W1
+                    sty  P8ZP_SCRATCH_W1+1
+                    lda  #$numElements
+                    jsr  prog8_lib.func_reverse_b""")
             }
-            val varName = asmgen.asmVariableName(variable)
-            when (decl.type) {
-                DataType.ARRAY_UB, DataType.ARRAY_B -> {
-                    asmgen.out("""
-                        lda  #<$varName
-                        ldy  #>$varName
-                        sta  P8ZP_SCRATCH_W1
-                        sty  P8ZP_SCRATCH_W1+1
-                        lda  #$numElements
-                        jsr  prog8_lib.func_reverse_b""")
-                }
-                DataType.ARRAY_UW, DataType.ARRAY_W -> {
-                    asmgen.out("""
-                        lda  #<$varName
-                        ldy  #>$varName
-                        sta  P8ZP_SCRATCH_W1
-                        sty  P8ZP_SCRATCH_W1+1
-                        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  #<$varName
-                        ldy  #>$varName
-                        sta  P8ZP_SCRATCH_W1
-                        sty  P8ZP_SCRATCH_W1+1
-                        lda  #$numElements
-                        jsr  floats.func_reverse_f""")
-                }
-                in SplitWordArrayTypes -> TODO("split word reverse")
-                else -> throw AssemblyError("weird type")
+            DataType.ARRAY_UW, DataType.ARRAY_W -> {
+                asmgen.out("""
+                    lda  #<$varName
+                    ldy  #>$varName
+                    sta  P8ZP_SCRATCH_W1
+                    sty  P8ZP_SCRATCH_W1+1
+                    lda  #$numElements
+                    jsr  prog8_lib.func_reverse_w""")
             }
+            DataType.STR -> {
+                asmgen.out("""
+                    lda  #<$varName
+                    ldy  #>$varName
+                    sta  P8ZP_SCRATCH_W1
+                    sty  P8ZP_SCRATCH_W1+1
+                    lda  #${numElements-1}
+                    jsr  prog8_lib.func_reverse_b""")
+            }
+            DataType.ARRAY_F -> {
+                asmgen.out("""
+                    lda  #<$varName
+                    ldy  #>$varName
+                    sta  P8ZP_SCRATCH_W1
+                    sty  P8ZP_SCRATCH_W1+1
+                    lda  #$numElements
+                    jsr  floats.func_reverse_f""")
+            }
+            in SplitWordArrayTypes -> {
+                // reverse the lsb and msb arrays both, independently
+                asmgen.out("""
+                    lda  #<${varName}_lsb
+                    ldy  #>${varName}_lsb
+                    sta  P8ZP_SCRATCH_W1
+                    sty  P8ZP_SCRATCH_W1+1
+                    lda  #$numElements
+                    jsr  prog8_lib.func_reverse_b
+                    lda  #<${varName}_msb
+                    ldy  #>${varName}_msb
+                    sta  P8ZP_SCRATCH_W1
+                    sty  P8ZP_SCRATCH_W1+1
+                    lda  #$numElements
+                    jsr  prog8_lib.func_reverse_b""")
+            }
+            else -> throw AssemblyError("weird type")
         }
     }
 
     private fun funcSort(fcall: PtBuiltinFunctionCall) {
-        val variable = fcall.args.single()
-        if (variable is PtIdentifier) {
-            val symbol = asmgen.symbolTable.lookup(variable.name)
-            val decl = symbol!!.astNode as IPtVariable
-            val varName = asmgen.asmVariableName(variable)
-            val numElements = when(decl) {
-                is PtConstant -> throw AssemblyError("cannot sort a constant")
-                is PtMemMapped -> decl.arraySize
-                is PtVariable -> decl.arraySize
+        val variable = fcall.args.single() as PtIdentifier
+        val symbol = asmgen.symbolTable.lookup(variable.name)
+        val varName = asmgen.asmVariableName(variable)
+        val (dt, numElements) = when(symbol) {
+            is StStaticVariable  -> symbol.dt to symbol.length!!
+            is StMemVar -> symbol.dt to symbol.length!!
+            else -> DataType.UNDEFINED to 0
+        }
+        when (dt) {
+            DataType.ARRAY_UB, DataType.ARRAY_B -> {
+                asmgen.out("""
+                    lda  #<$varName
+                    ldy  #>$varName
+                    sta  P8ZP_SCRATCH_W1
+                    sty  P8ZP_SCRATCH_W1+1
+                    lda  #$numElements""")
+                asmgen.out(if (dt == DataType.ARRAY_UB) "  jsr  prog8_lib.func_sort_ub" else "  jsr  prog8_lib.func_sort_b")
             }
-            when (decl.type) {
-                DataType.ARRAY_UB, DataType.ARRAY_B -> {
-                    asmgen.out("""
-                        lda  #<$varName
-                        ldy  #>$varName
-                        sta  P8ZP_SCRATCH_W1
-                        sty  P8ZP_SCRATCH_W1+1
-                        lda  #$numElements""")
-                    asmgen.out(if (decl.type == DataType.ARRAY_UB) "  jsr  prog8_lib.func_sort_ub" else "  jsr  prog8_lib.func_sort_b")
-                }
-                DataType.ARRAY_UW, DataType.ARRAY_W -> {
-                    asmgen.out("""
-                        lda  #<$varName
-                        ldy  #>$varName
-                        sta  P8ZP_SCRATCH_W1
-                        sty  P8ZP_SCRATCH_W1+1
-                        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")
+            DataType.ARRAY_UW, DataType.ARRAY_W -> {
+                asmgen.out("""
+                    lda  #<$varName
+                    ldy  #>$varName
+                    sta  P8ZP_SCRATCH_W1
+                    sty  P8ZP_SCRATCH_W1+1
+                    lda  #$numElements""")
+                asmgen.out(if (dt == DataType.ARRAY_UW) "  jsr  prog8_lib.func_sort_uw" else "  jsr  prog8_lib.func_sort_w")
             }
-        } else
-            throw AssemblyError("weird type")
+            DataType.STR -> {
+                asmgen.out("""
+                    lda  #<$varName
+                    ldy  #>$varName
+                    sta  P8ZP_SCRATCH_W1
+                    sty  P8ZP_SCRATCH_W1+1
+                    lda  #${numElements-1}
+                    jsr  prog8_lib.func_sort_ub""")
+            }
+            DataType.ARRAY_F -> throw AssemblyError("sorting of floating point array is not supported")
+            in SplitWordArrayTypes -> TODO("split words sort")
+            else -> throw AssemblyError("weird type")
+        }
     }
 
     private fun funcRor2(fcall: PtBuiltinFunctionCall) {
@@ -396,8 +424,9 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UBYTE -> {
                 when (what) {
                     is PtArrayIndexer -> {
-                        translateRolRorArrayArgs(what.variable, what, "ror2", 'b')
-                        asmgen.out("  jsr  prog8_lib.ror2_array_ub")
+                        asmgen.loadScaledArrayIndexIntoRegister(what, CpuRegister.X)
+                        val varname = asmgen.asmVariableName(what.variable)
+                        asmgen.out("  lda  ${varname},x |  lsr  a |  bcc  + |  ora  #\$80 |+  |  sta  ${varname},x")
                     }
                     is PtMemoryByte -> {
                         if (what.address is PtNumber) {
@@ -418,10 +447,12 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UWORD -> {
                 when (what) {
                     is PtArrayIndexer -> {
+                        asmgen.loadScaledArrayIndexIntoRegister(what, CpuRegister.X)
+                        val varname = asmgen.asmVariableName(what.variable)
                         if(what.splitWords)
-                            TODO("ror2 split words ${what.position}")
-                        translateRolRorArrayArgs(what.variable, what, "ror2", 'w')
-                        asmgen.out("  jsr  prog8_lib.ror2_array_uw")
+                            asmgen.out("  lsr  ${varname}_msb,x |  ror  ${varname}_lsb,x |  bcc  + |  lda  ${varname}_msb,x |  ora  #\$80 |  sta  ${varname}_msb,x |+ ")
+                        else
+                            asmgen.out("  lsr  ${varname}+1,x |  ror  ${varname},x |  bcc  +  |  lda  ${varname}+1,x  |  ora  #\$80 |  sta  ${varname}+1,x |+ ")
                     }
                     is PtIdentifier -> {
                         val variable = asmgen.asmVariableName(what)
@@ -440,8 +471,11 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UBYTE -> {
                 when (what) {
                     is PtArrayIndexer -> {
-                        translateRolRorArrayArgs(what.variable, what, "ror", 'b')
-                        asmgen.out("  jsr  prog8_lib.ror_array_ub")
+                        if(!what.index.isSimple()) asmgen.out("  php")   // save Carry
+                        asmgen.loadScaledArrayIndexIntoRegister(what, CpuRegister.X)
+                        if(!what.index.isSimple()) asmgen.out("  plp")
+                        val varname = asmgen.asmVariableName(what.variable)
+                        asmgen.out("  ror  ${varname},x")
                     }
                     is PtMemoryByte -> {
                         if (what.address is PtNumber) {
@@ -450,15 +484,19 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                         } else {
                             val ptrAndIndex = asmgen.pointerViaIndexRegisterPossible(what.address)
                             if(ptrAndIndex!=null) {
+                                asmgen.out("  php")
                                 asmgen.assignExpressionToRegister(ptrAndIndex.second, RegisterOrPair.A)
                                 asmgen.saveRegisterStack(CpuRegister.A, true)
                                 asmgen.assignExpressionToRegister(ptrAndIndex.first, RegisterOrPair.AY)
                                 asmgen.out("  sta  (+) + 1 |  sty  (+) + 2")
                                 asmgen.restoreRegisterStack(CpuRegister.X, false)
                                 asmgen.out("""
+                                    plp
 +                                   ror  ${'$'}ffff,x           ; modified""")
                             } else {
+                                if(!what.address.isSimple()) asmgen.out("  php")   // save Carry
                                 asmgen.assignExpressionToRegister(what.address, RegisterOrPair.AY)
+                                if(!what.address.isSimple()) asmgen.out("  plp")
                                 asmgen.out("""
                                     sta  (+) + 1
                                     sty  (+) + 2
@@ -476,10 +514,14 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UWORD -> {
                 when (what) {
                     is PtArrayIndexer -> {
+                        if(!what.index.isSimple()) asmgen.out("  php")   // save Carry
+                        asmgen.loadScaledArrayIndexIntoRegister(what, CpuRegister.X)
+                        if(!what.index.isSimple()) asmgen.out("  plp")
+                        val varname = asmgen.asmVariableName(what.variable)
                         if(what.splitWords)
-                            TODO("ror split words ${what.position}")
-                        translateRolRorArrayArgs(what.variable, what, "ror", 'w')
-                        asmgen.out("  jsr  prog8_lib.ror_array_uw")
+                            asmgen.out("  ror  ${varname}_msb,x |  ror  ${varname}_lsb,x")
+                        else
+                            asmgen.out("  ror  ${varname}+1,x |  ror  ${varname},x")
                     }
                     is PtIdentifier -> {
                         val variable = asmgen.asmVariableName(what)
@@ -498,8 +540,9 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UBYTE -> {
                 when (what) {
                     is PtArrayIndexer -> {
-                        translateRolRorArrayArgs(what.variable, what, "rol2", 'b')
-                        asmgen.out("  jsr  prog8_lib.rol2_array_ub")
+                        asmgen.loadScaledArrayIndexIntoRegister(what, CpuRegister.X)
+                        val varname = asmgen.asmVariableName(what.variable)
+                        asmgen.out("  lda  ${varname},x |  cmp  #\$80 |  rol  a |  sta  ${varname},x")
                     }
                     is PtMemoryByte -> {
                         if (what.address is PtNumber) {
@@ -520,10 +563,12 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UWORD -> {
                 when (what) {
                     is PtArrayIndexer -> {
+                        asmgen.loadScaledArrayIndexIntoRegister(what, CpuRegister.X)
+                        val varname = asmgen.asmVariableName(what.variable)
                         if(what.splitWords)
-                            TODO("rol2 split words ${what.position}")
-                        translateRolRorArrayArgs(what.variable, what, "rol2", 'w')
-                        asmgen.out("  jsr  prog8_lib.rol2_array_uw")
+                            asmgen.out("  asl  ${varname}_lsb,x |  rol  ${varname}_msb,x |  bcc  + |  inc  ${varname}_lsb |+")
+                        else
+                            asmgen.out("  asl  ${varname},x |  rol  ${varname}+1,x |  bcc  + |  inc  ${varname},x |+  ")
                     }
                     is PtIdentifier -> {
                         val variable = asmgen.asmVariableName(what)
@@ -542,8 +587,11 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UBYTE -> {
                 when (what) {
                     is PtArrayIndexer -> {
-                        translateRolRorArrayArgs(what.variable, what, "rol", 'b')
-                        asmgen.out("  jsr  prog8_lib.rol_array_ub")
+                        if(!what.index.isSimple()) asmgen.out("  php")   // save Carry
+                        asmgen.loadScaledArrayIndexIntoRegister(what, CpuRegister.X)
+                        if(!what.index.isSimple()) asmgen.out("  plp")
+                        val varname = asmgen.asmVariableName(what.variable)
+                        asmgen.out("  rol  ${varname},x")
                     }
                     is PtMemoryByte -> {
                         if (what.address is PtNumber) {
@@ -552,15 +600,19 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                         } else {
                             val ptrAndIndex = asmgen.pointerViaIndexRegisterPossible(what.address)
                             if(ptrAndIndex!=null) {
+                                asmgen.out("  php")
                                 asmgen.assignExpressionToRegister(ptrAndIndex.second, RegisterOrPair.A)
                                 asmgen.saveRegisterStack(CpuRegister.A, true)
                                 asmgen.assignExpressionToRegister(ptrAndIndex.first, RegisterOrPair.AY)
                                 asmgen.out("  sta  (+) + 1 |  sty  (+) + 2")
                                 asmgen.restoreRegisterStack(CpuRegister.X, false)
                                 asmgen.out("""
+                                    plp
 +                                   rol  ${'$'}ffff,x           ; modified""")
                             } else {
+                                if(!what.address.isSimple()) asmgen.out("  php")   // save Carry
                                 asmgen.assignExpressionToRegister(what.address, RegisterOrPair.AY)
+                                if(!what.address.isSimple()) asmgen.out("  plp")
                                 asmgen.out("""
                                     sta  (+) + 1
                                     sty  (+) + 2
@@ -578,10 +630,14 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             DataType.UWORD -> {
                 when (what) {
                     is PtArrayIndexer -> {
+                        if(!what.index.isSimple()) asmgen.out("  php")   // save Carry
+                        asmgen.loadScaledArrayIndexIntoRegister(what, CpuRegister.X)
+                        if(!what.index.isSimple()) asmgen.out("  plp")
+                        val varname = asmgen.asmVariableName(what.variable)
                         if(what.splitWords)
-                            TODO("rol split words ${what.position}")
-                        translateRolRorArrayArgs(what.variable, what, "rol", 'w')
-                        asmgen.out("  jsr  prog8_lib.rol_array_uw")
+                            asmgen.out("  rol  ${varname}_lsb,x |  rol  ${varname}_msb,x")
+                        else
+                            asmgen.out("  rol  ${varname},x |  rol  ${varname}+1,x")
                     }
                     is PtIdentifier -> {
                         val variable = asmgen.asmVariableName(what)
@@ -594,23 +650,6 @@ 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${dt}._arg_target", DataType.UWORD)
-        } else {
-            val p = arrayvar.parent
-            val addressOf = PtAddressOf(arrayvar.position)
-            addressOf.add(arrayvar)
-            addressOf.parent = p
-            asmgen.assignExpressionToVariable(addressOf, "prog8_lib.${operation}_array_u${dt}._arg_target", DataType.UWORD)
-        }
-        asmgen.assignExpressionToVariable(indexer.index, "prog8_lib.${operation}_array_u${dt}._arg_index", DataType.UBYTE)
-    }
-
     private fun funcSetLsbMsb(fcall: PtBuiltinFunctionCall, msb: Boolean) {
         val target: AsmAssignTarget
         when(fcall.args[0]) {
@@ -686,7 +725,7 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
             assignAsmGen.assignConstantByte(target, 0)
         } else {
             asmgen.assignExpressionToRegister(fcall.args[1], RegisterOrPair.A, false)
-            assignAsmGen.assignRegisterByte(target, CpuRegister.A, false)
+            assignAsmGen.assignRegisterByte(target, CpuRegister.A, false, false)
         }
     }
 
@@ -701,20 +740,64 @@ 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, true)
+        assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, true, true)
     }
 
     private fun funcAnyAll(fcall: PtBuiltinFunctionCall, resultRegister: RegisterOrPair?, scope: IPtSubroutine?) {
-        outputAddressAndLenghtOfArray(fcall.args[0])
         val dt = fcall.args.single().type
+        val array = fcall.args[0] as PtIdentifier
         when (dt) {
-            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_${fcall.name}_f_into_A |  ldy  #0")
-            in SplitWordArrayTypes -> TODO("split word any/all")
+            DataType.ARRAY_B, DataType.ARRAY_UB, DataType.STR -> {
+                outputAddressAndLengthOfArray(array)
+                asmgen.out("  jsr  prog8_lib.func_${fcall.name}_b_into_A")
+            }
+            DataType.ARRAY_UW, DataType.ARRAY_W -> {
+                outputAddressAndLengthOfArray(array)
+                asmgen.out("  jsr  prog8_lib.func_${fcall.name}_w_into_A")
+            }
+            DataType.ARRAY_F -> {
+                outputAddressAndLengthOfArray(array)
+                asmgen.out("  jsr  floats.func_${fcall.name}_f_into_A")
+            }
+            in SplitWordArrayTypes -> {
+                val numElements = (asmgen.symbolTable.lookup(array.name) as StStaticVariable).length
+                when(fcall.name) {
+                    "any" -> {
+                        // any(lsb-array) or any(msb-array)
+                        val arrayName = asmgen.asmVariableName(array)
+                        asmgen.out("""
+                            lda  #<${arrayName}_lsb
+                            ldy  #>${arrayName}_lsb
+                            sta  P8ZP_SCRATCH_W1
+                            sty  P8ZP_SCRATCH_W1+1
+                            lda  #$numElements
+                        """)
+                        asmgen.out("  jsr  prog8_lib.func_${fcall.name}_b_into_A")
+                        asmgen.out("  bne  +")      // shortcircuit
+                        asmgen.out("""
+                            pha
+                            lda  #<${arrayName}_msb
+                            ldy  #>${arrayName}_msb
+                            sta  P8ZP_SCRATCH_W1
+                            sty  P8ZP_SCRATCH_W1+1
+                            lda  #$numElements
+                        """)
+                        asmgen.out("  jsr  prog8_lib.func_${fcall.name}_b_into_A")
+                        asmgen.out("""
+                            sta  P8ZP_SCRATCH_REG
+                            pla
+                            ora  P8ZP_SCRATCH_REG
++""")
+                    }
+                    "all" -> {
+                        TODO("split words all")
+                    }
+                    else -> throw AssemblyError("weird call")
+                }
+            }
             else -> throw AssemblyError("weird type $dt")
         }
-        assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, dt in SignedDatatypes)
+        assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A, dt in SignedDatatypes, true)
     }
 
     private fun funcAbs(fcall: PtBuiltinFunctionCall, resultRegister: RegisterOrPair?, scope: IPtSubroutine?) {
@@ -723,7 +806,7 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         when (dt) {
             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)
+                assignAsmGen.assignRegisterByte(AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, false, fcall.position, scope, asmgen), CpuRegister.A,false, true)
             }
             DataType.WORD -> {
                 asmgen.out("  jsr  prog8_lib.abs_w_into_AY")
@@ -743,6 +826,41 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         }
     }
 
+    private fun funcPokeF(fcall: PtBuiltinFunctionCall) {
+        when(val number = fcall.args[1]) {
+            is PtIdentifier -> {
+                val varName = asmgen.asmVariableName(number)
+                asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.AY)
+                asmgen.out("""
+                    pha
+                    lda  #<$varName
+                    sta  P8ZP_SCRATCH_W1
+                    lda  #>$varName
+                    sta  P8ZP_SCRATCH_W1+1
+                    pla
+                    jsr  floats.copy_float""")
+            }
+            is PtNumber -> {
+                asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.AY)
+                asmgen.assignConstFloatToPointerAY(number)
+            }
+            else -> {
+                val tempvar = asmgen.getTempVarName(DataType.FLOAT)
+                asmgen.assignExpressionTo(fcall.args[1],
+                    AsmAssignTarget(TargetStorageKind.VARIABLE, asmgen, DataType.FLOAT, fcall.definingISub(), fcall.position, tempvar, null, null, null, null))
+                asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.AY)
+                asmgen.out("""
+                    pha
+                    lda  #<$tempvar
+                    sta  P8ZP_SCRATCH_W1
+                    lda  #>$tempvar
+                    sta  P8ZP_SCRATCH_W1+1
+                    pla
+                    jsr  floats.copy_float""")
+            }
+        }
+    }
+
     private fun funcPokeW(fcall: PtBuiltinFunctionCall) {
         when(val addrExpr = fcall.args[0]) {
             is PtNumber -> {
@@ -799,6 +917,15 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         asmgen.out("  jsr  prog8_lib.func_pokew")
     }
 
+    private fun funcPeekF(fcall: PtBuiltinFunctionCall, resultRegister: RegisterOrPair?) {
+        asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.AY)
+        asmgen.out("  jsr  floats.MOVFM")
+        if(resultRegister!=null) {
+            assignAsmGen.assignFAC1float(
+                AsmAssignTarget(TargetStorageKind.REGISTER, asmgen, DataType.FLOAT, fcall.definingISub(), fcall.position, null, null, null, resultRegister, null))
+        }
+    }
+
     private fun funcPeekW(fcall: PtBuiltinFunctionCall, resultRegister: RegisterOrPair?) {
         fun fallback() {
             asmgen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.AY)
@@ -871,7 +998,7 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                 assignAsmGen.assignExpressionToRegister(fcall.args[0], RegisterOrPair.A, signed)    // value
                 asmgen.out("  jsr  prog8_lib.func_clamp_${fcall.type.toString().lowercase()}")
                 val targetReg = AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, signed, fcall.position, fcall.definingISub(), asmgen)
-                assignAsmGen.assignRegisterByte(targetReg, CpuRegister.A, signed)
+                assignAsmGen.assignRegisterByte(targetReg, CpuRegister.A, signed, true)
             }
             in WordDatatypes -> {
                 assignAsmGen.assignExpressionToVariable(fcall.args[1], "P8ZP_SCRATCH_W1", fcall.args[1].type)  // minimum
@@ -887,55 +1014,59 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
 
     private fun funcMin(fcall: PtBuiltinFunctionCall, 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
-+""")
-            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) {
+        when (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_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
-+""")
+                    lda  P8ZP_SCRATCH_B1
+    +""")
+                val targetReg = AsmAssignTarget.fromRegisters(resultRegister ?: RegisterOrPair.A, signed, fcall.position, fcall.definingISub(), asmgen)
+                asmgen.assignRegister(RegisterOrPair.A, targetReg)
+            }
+            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
+    +""")
+                }
+                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")
             }
-            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")
         }
     }
 
@@ -1111,7 +1242,25 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
                 else -> throw AssemblyError("invalid reg")
             }
         } else {
-            when(resultRegister) {
+            if(arg is PtArrayIndexer && resultRegister in setOf(null, RegisterOrPair.A, RegisterOrPair.Y, RegisterOrPair.X)) {
+                // just read the lsb byte out of the word array
+                val arrayVar = if(arg.splitWords) asmgen.asmVariableName(arg.variable)+"_lsb" else asmgen.asmVariableName(arg.variable)
+                when(resultRegister) {
+                    null, RegisterOrPair.A -> {
+                        asmgen.loadScaledArrayIndexIntoRegister(arg, CpuRegister.Y)
+                        asmgen.out("  lda  $arrayVar,y")
+                    }
+                    RegisterOrPair.Y -> {
+                        asmgen.loadScaledArrayIndexIntoRegister(arg, CpuRegister.X)
+                        asmgen.out("  lda  $arrayVar,x")
+                    }
+                    RegisterOrPair.X -> {
+                        asmgen.loadScaledArrayIndexIntoRegister(arg, CpuRegister.Y)
+                        asmgen.out("  ldx  $arrayVar,y")
+                    }
+                    else -> throw AssemblyError("invalid reg")
+                }
+            } else  when(resultRegister) {
                 null, RegisterOrPair.A -> {
                     asmgen.assignExpressionToRegister(arg, RegisterOrPair.AY)
                     // NOTE: we rely on the fact that the above assignment to AY, assigns the Lsb to A as the last instruction.
@@ -1152,16 +1301,14 @@ internal class BuiltinFunctionsAsmGen(private val program: PtProgram,
         }
     }
 
-    private fun outputAddressAndLenghtOfArray(arg: PtExpression) {
-        // address in P8ZP_SCRATCH_W1,  number of elements in A
-        arg as PtIdentifier
+    private fun outputAddressAndLengthOfArray(arg: PtIdentifier) {
+        // address goes in P8ZP_SCRATCH_W1,  number of elements in A
         val symbol = asmgen.symbolTable.lookup(arg.name)
-        val arrayVar = symbol!!.astNode as IPtVariable
-        val numElements = when(arrayVar) {
-            is PtConstant -> null
-            is PtMemMapped -> arrayVar.arraySize
-            is PtVariable -> arrayVar.arraySize
-        } ?: throw AssemblyError("length of non-array requested")
+        val numElements = when(symbol) {
+            is StStaticVariable -> symbol.length!!
+            is StMemVar -> symbol.length!!
+            else -> 0
+        }
         val identifierName = asmgen.asmVariableName(arg)
         asmgen.out("""
                     lda  #<$identifierName

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

@@ -1,7 +1,11 @@
 package prog8.codegen.cpu6502
 
 import com.github.michaelbull.result.fold
-import prog8.code.ast.*
+import prog8.code.StMemVar
+import prog8.code.StStaticVariable
+import prog8.code.ast.PtForLoop
+import prog8.code.ast.PtIdentifier
+import prog8.code.ast.PtRange
 import prog8.code.core.*
 import kotlin.math.absoluteValue
 
@@ -332,11 +336,10 @@ $endLabel""")
         asmgen.loopEndLabels.push(endLabel)
         val iterableName = asmgen.asmVariableName(ident)
         val symbol = asmgen.symbolTable.lookup(ident.name)
-        val decl = symbol!!.astNode as IPtVariable
-        val numElements = when(decl) {
-            is PtConstant -> throw AssemblyError("length of non-array requested")
-            is PtMemMapped -> decl.arraySize
-            is PtVariable -> decl.arraySize
+        val numElements = when(symbol) {
+            is StStaticVariable -> symbol.length!!
+            is StMemVar -> symbol.length!!
+            else -> 0
         }
         when(iterableDt) {
             DataType.STR -> {
@@ -364,7 +367,7 @@ $loopLabel          sty  $indexVar
                     lda  $iterableName,y
                     sta  ${asmgen.asmVariableName(stmt.variable)}""")
                 asmgen.translate(stmt.statements)
-                if(numElements!!<=255u) {
+                if(numElements<=255) {
                     asmgen.out("""
                         ldy  $indexVar
                         iny
@@ -379,7 +382,7 @@ $loopLabel          sty  $indexVar
                         bne  $loopLabel
                         beq  $endLabel""")
                 }
-                if(numElements>=16u) {
+                if(numElements>=16) {
                     // allocate index var on ZP if possible
                     val result = zeropage.allocate(indexVar, DataType.UBYTE, null, stmt.position, asmgen.errors)
                     result.fold(
@@ -392,7 +395,7 @@ $loopLabel          sty  $indexVar
                 asmgen.out(endLabel)
             }
             DataType.ARRAY_W, DataType.ARRAY_UW -> {
-                val length = numElements!! * 2u
+                val length = numElements * 2
                 val indexVar = asmgen.makeLabel("for_index")
                 val loopvarName = asmgen.asmVariableName(stmt.variable)
                 asmgen.out("""
@@ -403,7 +406,7 @@ $loopLabel          sty  $indexVar
                     lda  $iterableName+1,y
                     sta  $loopvarName+1""")
                 asmgen.translate(stmt.statements)
-                if(length<=127u) {
+                if(length<=127) {
                     asmgen.out("""
                         ldy  $indexVar
                         iny
@@ -420,7 +423,7 @@ $loopLabel          sty  $indexVar
                         bne  $loopLabel
                         beq  $endLabel""")
                 }
-                if(length>=16u) {
+                if(length>=16) {
                     // allocate index var on ZP if possible
                     val result = zeropage.allocate(indexVar, DataType.UBYTE, null, stmt.position, asmgen.errors)
                     result.fold(
@@ -433,7 +436,6 @@ $loopLabel          sty  $indexVar
                 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("""
@@ -444,7 +446,7 @@ $loopLabel          sty  $indexVar
                     lda  ${iterableName}_msb,y
                     sta  $loopvarName+1""")
                 asmgen.translate(stmt.statements)
-                if(numElements<=255u) {
+                if(numElements<=255) {
                     asmgen.out("""
                         ldy  $indexVar
                         iny
@@ -459,7 +461,7 @@ $loopLabel          sty  $indexVar
                         bne  $loopLabel
                         beq  $endLabel""")
                 }
-                if(numElements>=16u) {
+                if(numElements>=16) {
                     // allocate index var on ZP if possible
                     val result = zeropage.allocate(indexVar, DataType.UBYTE, null, stmt.position, asmgen.errors)
                     result.fold(

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

@@ -26,7 +26,7 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
         //       (you can use subroutine.shouldSaveX() and saveX()/restoreX() routines as a help for this)
 
         val symbol = asmgen.symbolTable.lookup(call.name)
-        val sub = symbol!!.astNode as IPtSubroutine
+        val sub = symbol?.astNode as IPtSubroutine
         val subAsmName = asmgen.asmSymbolName(call.name)
 
         if(sub is PtAsmSub) {
@@ -69,37 +69,59 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
         // remember: dealing with the X register and/or dealing with return values is the responsibility of the caller
     }
 
-    private fun argumentsViaRegisters(sub: PtAsmSub, call: PtFunctionCall) {
-        if(sub.parameters.size==1) {
-            argumentViaRegister(sub, IndexedValue(0, sub.parameters.single().second), call.args[0])
-        } else {
-            if(asmsub6502ArgsHaveRegisterClobberRisk(call.args, sub.parameters)) {
-                registerArgsViaCpuStackEvaluation(call, sub)
-            } else {
-                asmsub6502ArgsEvalOrder(sub).forEach {
-                    val param = sub.parameters[it]
-                    val arg = call.args[it]
-                    argumentViaRegister(sub, IndexedValue(it, param.second), arg)
-                }
+
+    private fun usesOtherRegistersWhileEvaluating(arg: PtExpression): Boolean {
+        return when(arg) {
+            is PtBuiltinFunctionCall -> {
+                if (arg.name == "lsb" || arg.name == "msb")
+                    return usesOtherRegistersWhileEvaluating(arg.args[0])
+                if (arg.name == "mkword")
+                    return usesOtherRegistersWhileEvaluating(arg.args[0]) || usesOtherRegistersWhileEvaluating(arg.args[1])
+                return !arg.isSimple()
             }
+            is PtAddressOf -> false
+            is PtIdentifier -> false
+            is PtMachineRegister -> false
+            is PtMemoryByte -> false
+            is PtNumber -> false
+            else -> true
         }
     }
 
-    private fun registerArgsViaCpuStackEvaluation(call: PtFunctionCall, callee: PtAsmSub) {
-        // this is called when one or more of the arguments are 'complex' and
-        // cannot be assigned to a register easily or risk clobbering other registers.
+    private fun argumentsViaRegisters(sub: PtAsmSub, call: PtFunctionCall) {
+        val registersUsed = mutableListOf<RegisterOrStatusflag>();
 
-        if(callee.parameters.isEmpty())
-            return
+        fun usedA() = registersUsed.any {it.registerOrPair==RegisterOrPair.A || it.registerOrPair==RegisterOrPair.AX || it.registerOrPair==RegisterOrPair.AY}
+        fun usedX() = registersUsed.any {it.registerOrPair==RegisterOrPair.X || it.registerOrPair==RegisterOrPair.AX || it.registerOrPair==RegisterOrPair.XY}
+        fun usedY() = registersUsed.any {it.registerOrPair==RegisterOrPair.Y || it.registerOrPair==RegisterOrPair.AY || it.registerOrPair==RegisterOrPair.XY}
 
-        // use the cpu hardware stack as intermediate storage for the arguments.
-        val argOrder = asmsub6502ArgsEvalOrder(callee)
-        argOrder.reversed().forEach {
-            asmgen.pushCpuStack(callee.parameters[it].second.type, call.args[it])
-        }
-        argOrder.forEach {
-            val param = callee.parameters[it]
-            asmgen.popCpuStack(callee, param.second, param.first)
+        if(sub.parameters.size==1) {
+            argumentViaRegister(sub, IndexedValue(0, sub.parameters.single().second), call.args[0])
+        } else {
+            val optimalEvalOrder = asmsub6502ArgsEvalOrder(sub)
+            optimalEvalOrder.forEach {
+                val param = sub.parameters[it]
+                val arg = call.args[it]
+                registersUsed += if(usesOtherRegistersWhileEvaluating(arg)) {
+                    if(!registersUsed.any{it.statusflag!=null || it.registerOrPair in CpuRegisters})
+                        argumentViaRegister(sub, IndexedValue(it, param.second), arg)
+                    else if(registersUsed.any {it.statusflag!=null}) {
+                        throw AssemblyError("call argument evaluation problem: can't save cpu statusregister parameter ${call.position}")
+                    }
+                    else {
+                        if(usedX()) asmgen.saveRegisterStack(CpuRegister.X, false)
+                        if(usedY()) asmgen.saveRegisterStack(CpuRegister.Y, false)
+                        if(usedA()) asmgen.saveRegisterStack(CpuRegister.A, false)
+                        val used = argumentViaRegister(sub, IndexedValue(it, param.second), arg)
+                        if(usedA()) asmgen.restoreRegisterStack(CpuRegister.A, false)
+                        if(usedY()) asmgen.restoreRegisterStack(CpuRegister.Y, true)
+                        if(usedX()) asmgen.restoreRegisterStack(CpuRegister.X, true)
+                        used
+                    }
+                } else {
+                    argumentViaRegister(sub, IndexedValue(it, param.second), arg)
+                }
+            }
         }
     }
 
@@ -112,7 +134,7 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
         asmgen.assignExpressionToVariable(value, varName, parameter.type)
     }
 
-    private fun argumentViaRegister(sub: IPtSubroutine, parameter: IndexedValue<PtSubroutineParameter>, value: PtExpression, registerOverride: RegisterOrPair? = null) {
+    private fun argumentViaRegister(sub: IPtSubroutine, parameter: IndexedValue<PtSubroutineParameter>, value: PtExpression, registerOverride: RegisterOrPair? = null): RegisterOrStatusflag {
         // pass argument via a register parameter
         if(!isArgumentTypeCompatible(value.type, parameter.value.type))
             throw AssemblyError("argument type incompatible")
@@ -156,6 +178,7 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
                     }
                 }
             } else throw AssemblyError("can only use Carry as status flag parameter")
+            return RegisterOrStatusflag(null, statusflag)
         }
         else {
             // via register or register pair
@@ -188,6 +211,7 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
                 }
                 asmgen.translateNormalAssignment(AsmAssignment(src, target, program.memsizer, Position.DUMMY), scope)
             }
+            return RegisterOrStatusflag(register, null)
         }
     }
 
@@ -209,3 +233,5 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
         return false
     }
 }
+
+

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

@@ -75,7 +75,7 @@ internal class PostIncrDecrAsmGen(private val program: PtProgram, private val as
         dec  ${asmArrayvarname}_msb+$constIndex
 +       dec  ${asmArrayvarname}_lsb+$constIndex""")
                     } else {
-                        asmgen.loadScaledArrayIndexIntoRegister(targetArrayIdx, elementDt, CpuRegister.X)
+                        asmgen.loadScaledArrayIndexIntoRegister(targetArrayIdx, CpuRegister.X)
                         if(incr)
                             asmgen.out(" inc  ${asmArrayvarname}_lsb,x |  bne  + |  inc  ${asmArrayvarname}_msb,x |+")
                         else
@@ -127,7 +127,7 @@ internal class PostIncrDecrAsmGen(private val program: PtProgram, private val as
                 }
                 else
                 {
-                    asmgen.loadScaledArrayIndexIntoRegister(targetArrayIdx, elementDt, CpuRegister.X)
+                    asmgen.loadScaledArrayIndexIntoRegister(targetArrayIdx, CpuRegister.X)
                     when(elementDt) {
                         in ByteDatatypes -> {
                             if(targetArrayIdx.usesPointerVariable) {

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

@@ -34,8 +34,8 @@ internal class ProgramAndVarsGen(
         header()
         val allBlocks = program.allBlocks()
 
-        if(allBlocks.first().name != "p8_main" && allBlocks.first().name != "main")
-            throw AssemblyError("first block should be 'main' or 'p8_main'")
+        if(allBlocks.first().name != "p8b_main" && allBlocks.first().name != "main")
+            throw AssemblyError("first block should be 'main' or 'p8b_main'")
 
         if(errors.noErrors())  {
             program.allBlocks().forEach { block2asm(it) }
@@ -133,29 +133,22 @@ internal class ProgramAndVarsGen(
                 pha""")
         }
 
-        // make sure that on the cx16 and c64, basic rom is banked in again when we exit the program
         when(compTarget.name) {
             "cx16" -> {
                 if(options.floats)
                     asmgen.out("  lda  #4 |  sta  $01")    // to use floats, make sure Basic rom is banked in
-                asmgen.out("  jsr  p8_main.p8_start")
+                asmgen.out("  jsr  p8b_main.p8s_start")
                 asmgen.out("  jmp  sys.cleanup_at_exit")
             }
             "c64" -> {
-                asmgen.out("  jsr  p8_main.p8_start |  lda  #31 |  sta  $01")
-                if(!options.noSysInit)
-                    asmgen.out("  jmp  sys.cleanup_at_exit")
-                else
-                    asmgen.out("  rts")
+                asmgen.out("  jsr  p8b_main.p8s_start |  lda  #31 |  sta  $01")
+                asmgen.out("  jmp  sys.cleanup_at_exit")
             }
             "c128" -> {
-                asmgen.out("  jsr  p8_main.p8_start |  lda  #0 |  sta ${"$"}ff00")
-                if(!options.noSysInit)
-                    asmgen.out("  jmp  sys.cleanup_at_exit")
-                else
-                    asmgen.out("  rts")
+                asmgen.out("  jsr  p8b_main.p8s_start |  lda  #0 |  sta ${"$"}ff00")
+                asmgen.out("  jmp  sys.cleanup_at_exit")
             }
-            else -> asmgen.jmp("p8_main.p8_start")
+            else -> asmgen.jmp("p8b_main.p8s_start")
         }
     }
 
@@ -173,47 +166,98 @@ internal class ProgramAndVarsGen(
     }
 
     private fun footer() {
+        var relocateBssVars = false
+        var relocateBssSlabs = false
+        var relocatedBssStart = 0u
+        var relocatedBssEnd = 0u
+
+        if(options.varsGolden) {
+            if(options.compTarget.machine.BSSGOLDENRAM_START == 0u ||
+                options.compTarget.machine.BSSGOLDENRAM_END == 0u ||
+                options.compTarget.machine.BSSGOLDENRAM_END <= options.compTarget.machine.BSSGOLDENRAM_START) {
+                throw AssemblyError("current compilation target hasn't got the golden ram area properly defined or it is simply not available")
+            }
+            relocateBssVars = true
+            relocatedBssStart = options.compTarget.machine.BSSGOLDENRAM_START
+            relocatedBssEnd = options.compTarget.machine.BSSGOLDENRAM_END
+        }
+        else if(options.varsHighBank!=null) {
+            if(options.compTarget.machine.BSSHIGHRAM_START == 0u ||
+                options.compTarget.machine.BSSHIGHRAM_END == 0u ||
+                options.compTarget.machine.BSSHIGHRAM_END <= options.compTarget.machine.BSSHIGHRAM_START) {
+                throw AssemblyError("current compilation target hasn't got the high ram area properly defined or it is simply not available")
+            }
+            if(options.slabsHighBank!=null && options.varsHighBank!=options.slabsHighBank)
+                throw AssemblyError("slabs and vars high bank must be the same")
+            relocateBssVars = true
+            relocatedBssStart = options.compTarget.machine.BSSHIGHRAM_START
+            relocatedBssEnd = options.compTarget.machine.BSSHIGHRAM_END
+        }
+
+        if(options.slabsGolden) {
+            if(options.compTarget.machine.BSSGOLDENRAM_START == 0u ||
+                options.compTarget.machine.BSSGOLDENRAM_END == 0u ||
+                options.compTarget.machine.BSSGOLDENRAM_END <= options.compTarget.machine.BSSGOLDENRAM_START) {
+                throw AssemblyError("current compilation target hasn't got the golden ram area properly defined or it is simply not available")
+            }
+            relocateBssSlabs = true
+            relocatedBssStart = options.compTarget.machine.BSSGOLDENRAM_START
+            relocatedBssEnd = options.compTarget.machine.BSSGOLDENRAM_END
+        }
+        else if(options.slabsHighBank!=null) {
+            if(options.compTarget.machine.BSSHIGHRAM_START == 0u ||
+                options.compTarget.machine.BSSHIGHRAM_END == 0u ||
+                options.compTarget.machine.BSSHIGHRAM_END <= options.compTarget.machine.BSSHIGHRAM_START) {
+                throw AssemblyError("current compilation target hasn't got the high ram area properly defined or it is simply not available")
+            }
+            if(options.varsHighBank!=null && options.varsHighBank!=options.slabsHighBank)
+                throw AssemblyError("slabs and vars high bank must be the same")
+            relocateBssSlabs = true
+            relocatedBssStart = options.compTarget.machine.BSSHIGHRAM_START
+            relocatedBssEnd = options.compTarget.machine.BSSHIGHRAM_END
+        }
+
         asmgen.out("; bss sections")
         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")
-            }
-            // BSS vars in high ram area, memory() slabs just concatenated at the end of the program.
-            if(symboltable.allMemorySlabs.isNotEmpty()) {
+        if(relocateBssVars) {
+            if(!relocateBssSlabs)
                 asmgen.out("  .dsection slabs_BSS")
-            }
             asmgen.out("prog8_program_end\t; end of program label for progend()")
-            asmgen.out("  * = ${options.compTarget.machine.BSSHIGHRAM_START.toHex()}")
+            asmgen.out("  * = ${relocatedBssStart.toHex()}")
             asmgen.out("prog8_bss_section_start")
             asmgen.out("  .dsection BSS")
-            asmgen.out("  .cerror * >= ${options.compTarget.machine.BSSHIGHRAM_END.toHex()}, \"too many variables for BSS section\"")
+            if(relocateBssSlabs)
+                asmgen.out("  .dsection slabs_BSS")
+            asmgen.out("  .cerror * > ${relocatedBssEnd.toHex()}, \"too many variables/data for BSS section\"")
             asmgen.out("prog8_bss_section_size = * - prog8_bss_section_start")
         } else {
-            // BSS vars followed by memory() slabs, concatenated at the end of the program.
             asmgen.out("prog8_bss_section_start")
             asmgen.out("  .dsection BSS")
             asmgen.out("prog8_bss_section_size = * - prog8_bss_section_start")
-            if(symboltable.allMemorySlabs.isNotEmpty()) {
+            if(!relocateBssSlabs)
                 asmgen.out("  .dsection slabs_BSS")
-            }
             asmgen.out("prog8_program_end\t; end of program label for progend()")
+            if(relocateBssSlabs) {
+                asmgen.out("  * = ${relocatedBssStart.toHex()}")
+                asmgen.out("  .dsection slabs_BSS")
+                asmgen.out("  .cerror * > ${relocatedBssEnd.toHex()}, \"too many data for slabs_BSS section\"")
+            }
         }
     }
 
     private fun block2asm(block: PtBlock) {
         asmgen.out("")
         asmgen.out("; ---- block: '${block.name}' ----")
-        if(block.address!=null)
-            asmgen.out("* = ${block.address!!.toHex()}")
+        if(block.options.address!=null)
+            asmgen.out("* = ${block.options.address!!.toHex()}")
         else {
-            if(block.alignment==PtBlock.BlockAlignment.WORD)
+            if(block.options.alignment==PtBlock.BlockAlignment.WORD)
                 asmgen.out("\t.align 2")
-            else if(block.alignment==PtBlock.BlockAlignment.PAGE)
+            else if(block.options.alignment==PtBlock.BlockAlignment.PAGE)
                 asmgen.out("\t.align $100")
         }
 
-        asmgen.out("${block.name}\t" + (if(block.forceOutput) ".block" else ".proc"))
+        asmgen.out("${block.name}\t" + (if(block.options.forceOutput) ".block" else ".proc"))
         asmgen.outputSourceLine(block)
 
         createBlockVariables(block)
@@ -236,7 +280,7 @@ internal class ProgramAndVarsGen(
             asmgen.out("  rts\n  .bend")
         }
 
-        asmgen.out(if(block.forceOutput) "\n\t.bend" else "\n\t.pend")
+        asmgen.out(if(block.options.forceOutput) "\n\t.bend" else "\n\t.pend")
     }
 
     private fun getVars(scope: StNode): Map<String, StNode> =
@@ -276,7 +320,7 @@ internal class ProgramAndVarsGen(
 
         val asmStartScope: String
         val asmEndScope: String
-        if(sub.definingBlock()!!.forceOutput) {
+        if(sub.definingBlock()!!.options.forceOutput) {
             asmStartScope = ".block"
             asmEndScope = ".bend"
         } else {
@@ -299,7 +343,7 @@ internal class ProgramAndVarsGen(
 
         val asmStartScope: String
         val asmEndScope: String
-        if(sub.definingBlock()!!.forceOutput) {
+        if(sub.definingBlock()!!.options.forceOutput) {
             asmStartScope = ".block"
             asmEndScope = ".bend"
         } else {
@@ -331,7 +375,7 @@ 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.name=="p8_start") && (sub.definingBlock()!!.name=="main" || sub.definingBlock()!!.name=="p8_main"))
+        if((sub.name=="start" || sub.name=="p8s_start") && (sub.definingBlock()!!.name=="main" || sub.definingBlock()!!.name=="p8b_main"))
             entrypointInitialization()
 
         if(functioncallAsmGen.optimizeIntArgsViaRegisters(sub)) {
@@ -497,7 +541,15 @@ internal class ProgramAndVarsGen(
         for ((scopedName, zpvar) in zpVariables) {
             if (scopedName.startsWith("cx16.r"))
                 continue        // The 16 virtual registers of the cx16 are not actual variables in zp, they're memory mapped
-            asmgen.out("${scopedName.substringAfterLast('.')} \t= ${zpvar.address} \t; zp ${zpvar.dt}")
+            val variable = symboltable.flat.getValue(scopedName) as StStaticVariable
+            if(variable.dt in SplitWordArrayTypes) {
+                val lsbAddr = zpvar.address
+                val msbAddr = zpvar.address + (zpvar.size/2).toUInt()
+                asmgen.out("${scopedName.substringAfterLast('.')}_lsb \t= $lsbAddr \t; zp ${zpvar.dt} (lsbs)")
+                asmgen.out("${scopedName.substringAfterLast('.')}_msb \t= $msbAddr \t; zp ${zpvar.dt} (msbs)")
+            } else {
+                asmgen.out("${scopedName.substringAfterLast('.')} \t= ${zpvar.address} \t; zp ${zpvar.dt}")
+            }
         }
     }
 

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

@@ -23,13 +23,13 @@ internal class AnyExprAsmGen(
                     return assignByteBinExpr(expr, assign)
                 if (expr.left.type in WordDatatypes && expr.right.type in WordDatatypes) {
                     require(expr.operator in ComparisonOperators)
-                    TODO("words operands comparison -> byte")
+                    throw AssemblyError("words operands comparison -> byte, should have been handled by assignOptimizedComparisonWords()")
                 }
                 if (expr.left.type==DataType.FLOAT && expr.right.type==DataType.FLOAT) {
                     require(expr.operator in ComparisonOperators)
                     return assignFloatBinExpr(expr, assign)
                 }
-                TODO("weird expr operand types")
+                throw AssemblyError("weird expr operand types: ${expr.left.type} and ${expr.right.type}")
             }
             in WordDatatypes -> {
                 require(expr.left.type in WordDatatypes && expr.right.type in WordDatatypes) {
@@ -49,54 +49,24 @@ internal class AnyExprAsmGen(
 
     private fun assignWordBinExpr(expr: PtBinaryExpression): Boolean {
         when(expr.operator) {
-            "+" -> {
-                TODO("word + at ${expr.position}")
-            }
-            "-" -> {
-                TODO("word - at ${expr.position}")
-            }
-            "*" -> {
-                TODO("word * at ${expr.position}")
-            }
-            "/" -> {
-                TODO("word / at ${expr.position}")
-            }
-            "<<" -> {
-                TODO("word << at ${expr.position}")
-            }
-            ">>" -> {
-                TODO("word >> at ${expr.position}")
-            }
-            "%" -> {
-                TODO("word % at ${expr.position}")
-            }
-            "&", "and" -> {
-                TODO("word and at ${expr.position}")
-            }
-            "|", "or" -> {
-                TODO("word or at ${expr.position}")
-            }
-            "^", "xor" -> {
-                TODO("word xor at ${expr.position}")
-            }
-            "==" -> {
-                TODO("word == at ${expr.position}")
-            }
-            "!=" -> {
-                TODO("word != at ${expr.position}")
-            }
-            "<" -> {
-                TODO("word < at ${expr.position}")
-            }
-            "<=" -> {
-                TODO("word <= at ${expr.position}")
-            }
-            ">" -> {
-                TODO("word > at ${expr.position}")
-            }
-            ">=" -> {
-                TODO("word >= at ${expr.position}")
-            }
+            "+" -> TODO("word + at ${expr.position}")
+            "-" -> TODO("word - at ${expr.position}")
+            "*" -> TODO("word * at ${expr.position}")
+            "/" -> TODO("word / at ${expr.position}")
+            "<<" -> TODO("word << at ${expr.position}")
+            ">>" -> TODO("word >> at ${expr.position}")
+            "%" -> TODO("word % at ${expr.position}")
+            "and" -> TODO("word logical and (with optional shortcircuit) ${expr.position}")
+            "or" -> TODO("word logical or (with optional shortcircuit) ${expr.position}")
+            "&" -> TODO("word and at ${expr.position}")
+            "|" -> TODO("word or at ${expr.position}")
+            "^", "xor" -> TODO("word xor at ${expr.position}")
+            "==" -> TODO("word == at ${expr.position}")
+            "!=" -> TODO("word != at ${expr.position}")
+            "<" -> TODO("word < at ${expr.position}")
+            "<=" -> TODO("word <= at ${expr.position}")
+            ">" -> TODO("word > at ${expr.position}")
+            ">=" -> TODO("word >= at ${expr.position}")
             else -> return false
         }
     }
@@ -134,7 +104,9 @@ internal class AnyExprAsmGen(
             "%" -> {
                 TODO("byte % at ${expr.position}")
             }
-            "&", "and" -> {
+            "and" -> TODO("logical and (with optional shortcircuit) ${expr.position}")
+            "or" -> TODO("logical or (with optional shortcircuit) ${expr.position}")
+            "&" -> {
                 asmgen.assignExpressionToRegister(expr.left, RegisterOrPair.A, false)
                 asmgen.out("  pha")
                 asmgen.assignExpressionToVariable(expr.right, "P8ZP_SCRATCH_B1", DataType.UBYTE)
@@ -142,7 +114,7 @@ internal class AnyExprAsmGen(
                 asmgen.assignRegister(RegisterOrPair.A, assign.target)
                 return true
             }
-            "|", "or" -> {
+            "|" -> {
                 asmgen.assignExpressionToRegister(expr.left, RegisterOrPair.A, false)
                 asmgen.out("  pha")
                 asmgen.assignExpressionToVariable(expr.right, "P8ZP_SCRATCH_B1", DataType.UBYTE)

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

@@ -37,8 +37,10 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             "*=" -> inplaceModification(assign.target, "*", assign.source)
             "/=" -> inplaceModification(assign.target, "/", assign.source)
             "|=" -> inplaceModification(assign.target, "|", assign.source)
+            "or=" -> inplaceModification(assign.target, "or", assign.source)
             "&=" -> inplaceModification(assign.target, "&", assign.source)
-            "^=" -> inplaceModification(assign.target, "^", assign.source)
+            "and=" -> inplaceModification(assign.target, "and", assign.source)
+            "^=", "xor=" -> inplaceModification(assign.target, "^", assign.source)
             "<<=" -> inplaceModification(assign.target, "<<", assign.source)
             ">>=" -> inplaceModification(assign.target, ">>", assign.source)
             "%=" -> inplaceModification(assign.target, "%", assign.source)
@@ -205,239 +207,251 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             }
             TargetStorageKind.ARRAY -> {
                 val indexNum = target.array!!.index as? PtNumber
-                val indexVar = target.array.index as? PtIdentifier
-                when {
-                    indexNum!=null -> {
-                        val targetVarName = if(target.array.splitWords)
-                            "${target.asmVarname} + ${indexNum.number.toInt()}"
-                        else
-                            "${target.asmVarname} + ${indexNum.number.toInt()*program.memsizer.memorySize(target.datatype)}"
-                        when (target.datatype) {
-                            in ByteDatatypes -> {
-                                when(value.kind) {
-                                    SourceStorageKind.LITERALNUMBER -> inplacemodificationByteVariableWithLiteralval(targetVarName, target.datatype, operator, value.number!!.number.toInt())
-                                    SourceStorageKind.VARIABLE -> inplacemodificationByteVariableWithVariable(targetVarName, target.datatype, operator, value.asmVarname)
-                                    SourceStorageKind.REGISTER -> inplacemodificationByteVariableWithVariable(targetVarName, target.datatype, operator, regName(value))
-                                    SourceStorageKind.MEMORY -> inplacemodificationByteVariableWithValue(targetVarName, target.datatype, operator, value.memory!!)
-                                    SourceStorageKind.ARRAY -> inplacemodificationByteVariableWithValue(targetVarName, target.datatype, operator, value.array!!)
-                                    SourceStorageKind.EXPRESSION -> {
-                                        if(value.expression is PtTypeCast) {
-                                            if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator)) return
-                                            inplacemodificationByteVariableWithValue(targetVarName, target.datatype, operator, value.expression)
-                                        } else {
-                                            inplacemodificationByteVariableWithValue(targetVarName, target.datatype, operator, value.expression!!)
-                                        }
+                if (indexNum!=null) {
+                    val targetVarName = if(target.array.splitWords)
+                        "${target.asmVarname} + ${indexNum.number.toInt()}"
+                    else
+                        "${target.asmVarname} + ${indexNum.number.toInt()*program.memsizer.memorySize(target.datatype)}"
+                    when (target.datatype) {
+                        in ByteDatatypes -> {
+                            when(value.kind) {
+                                SourceStorageKind.LITERALNUMBER -> inplacemodificationByteVariableWithLiteralval(targetVarName, target.datatype, operator, value.number!!.number.toInt())
+                                SourceStorageKind.VARIABLE -> inplacemodificationByteVariableWithVariable(targetVarName, target.datatype, operator, value.asmVarname)
+                                SourceStorageKind.REGISTER -> inplacemodificationByteVariableWithVariable(targetVarName, target.datatype, operator, regName(value))
+                                SourceStorageKind.MEMORY -> inplacemodificationByteVariableWithValue(targetVarName, target.datatype, operator, value.memory!!)
+                                SourceStorageKind.ARRAY -> inplacemodificationByteVariableWithValue(targetVarName, target.datatype, operator, value.array!!)
+                                SourceStorageKind.EXPRESSION -> {
+                                    if(value.expression is PtTypeCast) {
+                                        if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator)) return
+                                        inplacemodificationByteVariableWithValue(targetVarName, target.datatype, operator, value.expression)
+                                    } else {
+                                        inplacemodificationByteVariableWithValue(targetVarName, target.datatype, operator, value.expression!!)
                                     }
                                 }
                             }
-                            in WordDatatypes -> {
-                                val block = target.origAstTarget?.definingBlock()
-                                when(value.kind) {
-                                    SourceStorageKind.LITERALNUMBER -> inplacemodificationWordWithLiteralval(targetVarName, target.datatype, operator, value.number!!.number.toInt(), block)
-                                    SourceStorageKind.VARIABLE -> inplacemodificationWordWithVariable(targetVarName, target.datatype, operator, value.asmVarname, value.datatype, block)
-                                    SourceStorageKind.REGISTER -> inplacemodificationWordWithVariable(targetVarName, target.datatype, operator, regName(value), value.datatype, block)
-                                    SourceStorageKind.MEMORY -> inplacemodificationWordWithMemread(targetVarName, target.datatype, operator, value.memory!!)
-                                    SourceStorageKind.ARRAY -> inplacemodificationWordWithValue(targetVarName, target.datatype, operator, value.array!!, block)
-                                    SourceStorageKind.EXPRESSION -> {
-                                        if(value.expression is PtTypeCast) {
-                                            if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator)) return
-                                            inplacemodificationWordWithValue(targetVarName, target.datatype, operator, value.expression, block)
-                                        } else {
-                                            inplacemodificationWordWithValue(targetVarName, target.datatype, operator, value.expression!!, block)
-                                        }
-                                    }
-                                }
-                            }
-                            DataType.FLOAT -> {
-                                when(value.kind) {
-                                    SourceStorageKind.LITERALNUMBER -> inplacemodificationFloatWithLiteralval(targetVarName, operator, value.number!!.number)
-                                    SourceStorageKind.VARIABLE -> inplacemodificationFloatWithVariable(targetVarName, operator, value.asmVarname)
-                                    SourceStorageKind.REGISTER -> inplacemodificationFloatWithVariable(targetVarName, operator, regName(value))
-                                    SourceStorageKind.MEMORY -> TODO("memread into float array")
-                                    SourceStorageKind.ARRAY -> inplacemodificationFloatWithValue(targetVarName, operator, value.array!!)
-                                    SourceStorageKind.EXPRESSION -> {
-                                        if(value.expression is PtTypeCast) {
-                                            if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator)) return
-                                            inplacemodificationFloatWithValue(targetVarName, operator, value.expression)
-                                        } else {
-                                            inplacemodificationFloatWithValue(targetVarName, operator, value.expression!!)
-                                        }
-                                    }
-                                }
-                            }
-                            else -> throw AssemblyError("weird type to do in-place modification on ${target.datatype}")
                         }
+
+                        in WordDatatypes -> {
+                            val block = target.origAstTarget?.definingBlock()
+                            when(value.kind) {
+                                SourceStorageKind.LITERALNUMBER -> inplacemodificationWordWithLiteralval(targetVarName, target.datatype, operator, value.number!!.number.toInt(), block)
+                                SourceStorageKind.VARIABLE -> inplacemodificationWordWithVariable(targetVarName, target.datatype, operator, value.asmVarname, value.datatype, block)
+                                SourceStorageKind.REGISTER -> inplacemodificationWordWithVariable(targetVarName, target.datatype, operator, regName(value), value.datatype, block)
+                                SourceStorageKind.MEMORY -> inplacemodificationWordWithMemread(targetVarName, target.datatype, operator, value.memory!!)
+                                SourceStorageKind.ARRAY -> inplacemodificationWordWithValue(targetVarName, target.datatype, operator, value.array!!, block)
+                                SourceStorageKind.EXPRESSION -> {
+                                    if(value.expression is PtTypeCast) {
+                                        if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator)) return
+                                        inplacemodificationWordWithValue(targetVarName, target.datatype, operator, value.expression, block)
+                                    } else {
+                                        inplacemodificationWordWithValue(targetVarName, target.datatype, operator, value.expression!!, block)
+                                    }
+                                }
+                            }
+                        }
+
+                        DataType.FLOAT -> {
+                            when(value.kind) {
+                                SourceStorageKind.LITERALNUMBER -> inplacemodificationFloatWithLiteralval(targetVarName, operator, value.number!!.number)
+                                SourceStorageKind.VARIABLE -> inplacemodificationFloatWithVariable(targetVarName, operator, value.asmVarname)
+                                SourceStorageKind.REGISTER -> inplacemodificationFloatWithVariable(targetVarName, operator, regName(value))
+                                SourceStorageKind.MEMORY -> TODO("memread into float array")
+                                SourceStorageKind.ARRAY -> inplacemodificationFloatWithValue(targetVarName, operator, value.array!!)
+                                SourceStorageKind.EXPRESSION -> {
+                                    if(value.expression is PtTypeCast) {
+                                        if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator)) return
+                                        inplacemodificationFloatWithValue(targetVarName, operator, value.expression)
+                                    } else {
+                                        inplacemodificationFloatWithValue(targetVarName, operator, value.expression!!)
+                                    }
+                                }
+                            }
+                        }
+
+                        else -> throw AssemblyError("weird type to do in-place modification on ${target.datatype}")
                     }
-                    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.saveRegisterStack(CpuRegister.Y, false)
+                }
+                else {
+                    when (target.datatype) {
+                        in ByteDatatypes -> {
+                            if(value.kind==SourceStorageKind.EXPRESSION
+                                && value.expression is PtTypeCast
+                                && tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator))
+                                return
+                            asmgen.loadScaledArrayIndexIntoRegister(target.array, CpuRegister.Y)
+                            asmgen.saveRegisterStack(CpuRegister.Y, false)
+                            asmgen.out("  lda  ${target.array.variable.name},y")
+                            when(value.kind) {
+                                SourceStorageKind.LITERALNUMBER -> {
+                                    inplacemodificationRegisterAwithVariable(operator, "#${value.number!!.number.toInt()}", target.datatype in SignedDatatypes)
+                                    asmgen.restoreRegisterStack(CpuRegister.Y, true)
+                                }
+
+                                SourceStorageKind.VARIABLE -> {
+                                    inplacemodificationRegisterAwithVariable(operator, value.asmVarname, target.datatype in SignedDatatypes)
+                                    asmgen.restoreRegisterStack(CpuRegister.Y, true)
+                                }
+
+                                SourceStorageKind.REGISTER -> {
+                                    inplacemodificationRegisterAwithVariable(operator, regName(value), target.datatype in SignedDatatypes)
+                                    asmgen.restoreRegisterStack(CpuRegister.Y, true)
+                                }
+
+                                SourceStorageKind.MEMORY -> {
+                                    asmgen.out("  sta  P8ZP_SCRATCH_B1")
+                                    inplacemodificationByteVariableWithValue("P8ZP_SCRATCH_B1", target.datatype, operator, value.memory!!)
+                                    asmgen.restoreRegisterStack(CpuRegister.Y, false)
+                                    asmgen.out("  lda  P8ZP_SCRATCH_B1")
+                                }
+
+                                SourceStorageKind.ARRAY -> {
+                                    asmgen.out("  sta  P8ZP_SCRATCH_B1")
+                                    inplacemodificationByteVariableWithValue("P8ZP_SCRATCH_B1", target.datatype, operator, value.array!!)
+                                    asmgen.restoreRegisterStack(CpuRegister.Y, false)
+                                    asmgen.out("  lda  P8ZP_SCRATCH_B1")
+                                }
+
+                                SourceStorageKind.EXPRESSION -> {
+                                    asmgen.out("  sta  P8ZP_SCRATCH_B1")
+                                    if(value.expression is PtTypeCast)
+                                        inplacemodificationByteVariableWithValue("P8ZP_SCRATCH_B1", target.datatype, operator, value.expression)
+                                    else
+                                        inplacemodificationByteVariableWithValue("P8ZP_SCRATCH_B1", target.datatype, operator, value.expression!!)
+                                    asmgen.restoreRegisterStack(CpuRegister.Y, false)
+                                    asmgen.out("  lda  P8ZP_SCRATCH_B1")
+                                }
+                            }
+                            asmgen.out("  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, CpuRegister.Y)
+                            asmgen.saveRegisterStack(CpuRegister.Y, false)
+                            if(target.array.splitWords) {
+                                asmgen.out("  lda  ${target.array.variable.name}_lsb,y")
+                                asmgen.out("  ldx  ${target.array.variable.name}_msb,y")
+                            } else {
                                 asmgen.out("  lda  ${target.array.variable.name},y")
-                                when(value.kind) {
-                                    SourceStorageKind.LITERALNUMBER -> {
-                                        inplacemodificationRegisterAwithVariable(operator, "#${value.number!!.number.toInt()}", target.datatype in SignedDatatypes)
-                                        asmgen.restoreRegisterStack(CpuRegister.Y, true)
-                                    }
-                                    SourceStorageKind.VARIABLE -> {
-                                        inplacemodificationRegisterAwithVariable(operator, value.asmVarname, target.datatype in SignedDatatypes)
-                                        asmgen.restoreRegisterStack(CpuRegister.Y, true)
-                                    }
-                                    SourceStorageKind.REGISTER -> {
-                                        inplacemodificationRegisterAwithVariable(operator, regName(value), target.datatype in SignedDatatypes)
-                                        asmgen.restoreRegisterStack(CpuRegister.Y, true)
-                                    }
-                                    SourceStorageKind.MEMORY -> {
-                                        asmgen.out("  sta  P8ZP_SCRATCH_B1")
-                                        inplacemodificationByteVariableWithValue("P8ZP_SCRATCH_B1", target.datatype, operator, value.memory!!)
-                                        asmgen.restoreRegisterStack(CpuRegister.Y, false)
-                                        asmgen.out("  lda  P8ZP_SCRATCH_B1")
-                                    }
-                                    SourceStorageKind.ARRAY -> {
-                                        asmgen.out("  sta  P8ZP_SCRATCH_B1")
-                                        inplacemodificationByteVariableWithValue("P8ZP_SCRATCH_B1", target.datatype, operator, value.array!!)
-                                        asmgen.restoreRegisterStack(CpuRegister.Y, false)
-                                        asmgen.out("  lda  P8ZP_SCRATCH_B1")
-                                    }
-                                    SourceStorageKind.EXPRESSION -> {
-                                        asmgen.out("  sta  P8ZP_SCRATCH_B1")
-                                        if(value.expression is PtTypeCast)
-                                            inplacemodificationByteVariableWithValue("P8ZP_SCRATCH_B1", target.datatype, operator, value.expression)
-                                        else
-                                            inplacemodificationByteVariableWithValue("P8ZP_SCRATCH_B1", target.datatype, operator, value.expression!!)
-                                        asmgen.restoreRegisterStack(CpuRegister.Y, false)
-                                        asmgen.out("  lda  P8ZP_SCRATCH_B1")
-                                    }
-                                }
-                                asmgen.out("  sta  ${target.array.variable.name},y")
+                                asmgen.out("  ldx  ${target.array.variable.name}+1,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.saveRegisterStack(CpuRegister.Y, false)
-                                if(target.array.splitWords) {
-                                    asmgen.out("  lda  ${target.array.variable.name}_lsb,y")
-                                    asmgen.out("  ldx  ${target.array.variable.name}_msb,y")
-                                } else {
-                                    asmgen.out("  lda  ${target.array.variable.name},y")
-                                    asmgen.out("  ldx  ${target.array.variable.name}+1,y")
-                                }
-                                val block = target.origAstTarget?.definingBlock()
-                                when(value.kind) {
-                                    SourceStorageKind.LITERALNUMBER -> {
-                                        val number = value.number!!.number.toInt()
-                                        if(!inplacemodificationRegisterAXwithLiteralval(operator, number)) {
-                                            asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
-                                            inplacemodificationWordWithLiteralval("P8ZP_SCRATCH_W1", target.datatype, operator, number, block)
-                                            asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
-                                        }
-                                    }
-                                    SourceStorageKind.VARIABLE -> {
-                                        if(!inplacemodificationRegisterAXwithVariable(
-                                                operator,
-                                                value.asmVarname,
-                                                value.datatype
-                                            )) {
-                                            asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
-                                            inplacemodificationWordWithVariable("P8ZP_SCRATCH_W1", target.datatype, operator, value.asmVarname, value.datatype, block)
-                                            asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
-                                        }
-                                    }
-                                    SourceStorageKind.REGISTER -> {
-                                        if(!inplacemodificationRegisterAXwithVariable(
-                                                operator,
-                                                regName(value),
-                                                value.datatype
-                                            )) {
-                                            asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
-                                            inplacemodificationWordWithVariable("P8ZP_SCRATCH_W1", target.datatype, operator, regName(value), value.datatype, block)
-                                            asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
-                                        }
-                                    }
-                                    SourceStorageKind.MEMORY -> {
+                            val block = target.origAstTarget?.definingBlock()
+                            when(value.kind) {
+                                SourceStorageKind.LITERALNUMBER -> {
+                                    val number = value.number!!.number.toInt()
+                                    if(!inplacemodificationRegisterAXwithLiteralval(operator, number)) {
                                         asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
-                                        inplacemodificationWordWithMemread("P8ZP_SCRATCH_W1", target.datatype, operator, value.memory!!)
+                                        inplacemodificationWordWithLiteralval("P8ZP_SCRATCH_W1", target.datatype, operator, number, block)
                                         asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
                                     }
-                                    SourceStorageKind.ARRAY -> {
-                                        asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
-                                        inplacemodificationWordWithValue("P8ZP_SCRATCH_W1", target.datatype, operator, value.array!!, block)
-                                        asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
-                                    }
-                                    SourceStorageKind.EXPRESSION -> {
-                                        asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
-                                        if(value.expression is PtTypeCast)
-                                            inplacemodificationWordWithValue("P8ZP_SCRATCH_W1", target.datatype, operator, value.expression, block)
-                                        else
-                                            inplacemodificationWordWithValue("P8ZP_SCRATCH_W1", target.datatype, operator, value.expression!!, block)
-                                        asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
-                                    }
-                                }
-                                asmgen.restoreRegisterStack(CpuRegister.Y, true)
-                                if(target.array.splitWords)
-                                    asmgen.out("  sta  ${target.array.variable.name}_lsb,y |  txa |  sta  ${target.array.variable.name}_msb,y")
-                                else
-                                    asmgen.out("  sta  ${target.array.variable.name},y |  txa |  sta  ${target.array.variable.name}+1,y")
-                            }
-                            DataType.FLOAT -> {
-                                // copy array value into tempvar
-                                val tempvar = asmgen.getTempVarName(DataType.FLOAT)
-                                asmgen.loadScaledArrayIndexIntoRegister(target.array, DataType.FLOAT, CpuRegister.A)
-                                asmgen.out("""
-                                    ldy  #>${target.asmVarname}
-                                    clc
-                                    adc  #<${target.asmVarname}
-                                    bcc  +
-                                    iny
-+                                   sta  P8ZP_SCRATCH_W1
-                                    sty  P8ZP_SCRATCH_W1+1
-                                    pha  ; save array ptr lsb
-                                    tya
-                                    pha  ; save array ptr msb
-                                    lda  #<$tempvar
-                                    ldy  #>$tempvar
-                                    jsr  floats.copy_float""")
-
-                                // calculate on tempvar
-                                when(value.kind) {
-                                    SourceStorageKind.LITERALNUMBER -> inplacemodificationFloatWithLiteralval(tempvar, operator, value.number!!.number)
-                                    SourceStorageKind.VARIABLE -> inplacemodificationFloatWithVariable(tempvar, operator, value.asmVarname)
-                                    SourceStorageKind.REGISTER -> inplacemodificationFloatWithVariable(tempvar, operator, regName(value))
-                                    SourceStorageKind.MEMORY -> TODO("memread into float")
-                                    SourceStorageKind.ARRAY -> inplacemodificationFloatWithValue(tempvar, operator, value.array!!)
-                                    SourceStorageKind.EXPRESSION -> {
-                                        if(value.expression is PtTypeCast) {
-                                            if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator))
-                                                return
-                                            inplacemodificationFloatWithValue(tempvar, operator, value.expression)
-                                        } else {
-                                            inplacemodificationFloatWithValue(tempvar, operator, value.expression!!)
-                                        }
-                                    }
                                 }
 
-                                // copy tempvar back into array
-                                asmgen.out("""
-                                    lda  #<$tempvar
-                                    ldy  #>$tempvar
-                                    sta  P8ZP_SCRATCH_W1
-                                    sty  P8ZP_SCRATCH_W1+1
-                                    pla  ; restore array ptr msb
-                                    tay
-                                    pla  ; restore array ptr lsb
-                                    jsr  floats.copy_float""")
+                                SourceStorageKind.VARIABLE -> {
+                                    if(!inplacemodificationRegisterAXwithVariable(
+                                            operator,
+                                            value.asmVarname,
+                                            value.datatype
+                                        )) {
+                                        asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
+                                        inplacemodificationWordWithVariable("P8ZP_SCRATCH_W1", target.datatype, operator, value.asmVarname, value.datatype, block)
+                                        asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
+                                    }
+                                }
+
+                                SourceStorageKind.REGISTER -> {
+                                    if(!inplacemodificationRegisterAXwithVariable(
+                                            operator,
+                                            regName(value),
+                                            value.datatype
+                                        )) {
+                                        asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
+                                        inplacemodificationWordWithVariable("P8ZP_SCRATCH_W1", target.datatype, operator, regName(value), value.datatype, block)
+                                        asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
+                                    }
+                                }
+
+                                SourceStorageKind.MEMORY -> {
+                                    asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
+                                    inplacemodificationWordWithMemread("P8ZP_SCRATCH_W1", target.datatype, operator, value.memory!!)
+                                    asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
+                                }
+
+                                SourceStorageKind.ARRAY -> {
+                                    asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
+                                    inplacemodificationWordWithValue("P8ZP_SCRATCH_W1", target.datatype, operator, value.array!!, block)
+                                    asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
+                                }
+
+                                SourceStorageKind.EXPRESSION -> {
+                                    asmgen.out("  sta  P8ZP_SCRATCH_W1 |  stx  P8ZP_SCRATCH_W1+1")
+                                    if(value.expression is PtTypeCast)
+                                        inplacemodificationWordWithValue("P8ZP_SCRATCH_W1", target.datatype, operator, value.expression, block)
+                                    else
+                                        inplacemodificationWordWithValue("P8ZP_SCRATCH_W1", target.datatype, operator, value.expression!!, block)
+                                    asmgen.out("  lda  P8ZP_SCRATCH_W1 |  ldx  P8ZP_SCRATCH_W1+1")
+                                }
                             }
-                            else -> throw AssemblyError("weird type to do in-place modification on ${target.datatype}")
+                            asmgen.restoreRegisterStack(CpuRegister.Y, true)
+                            if(target.array.splitWords)
+                                asmgen.out("  sta  ${target.array.variable.name}_lsb,y |  txa |  sta  ${target.array.variable.name}_msb,y")
+                            else
+                                asmgen.out("  sta  ${target.array.variable.name},y |  txa |  sta  ${target.array.variable.name}+1,y")
                         }
+
+                        DataType.FLOAT -> {
+                            // copy array value into tempvar
+                            val tempvar = asmgen.getTempVarName(DataType.FLOAT)
+                            asmgen.loadScaledArrayIndexIntoRegister(target.array, CpuRegister.A)
+                            asmgen.out("""
+                                                ldy  #>${target.asmVarname}
+                                                clc
+                                                adc  #<${target.asmVarname}
+                                                bcc  +
+                                                iny
+            +                                   sta  P8ZP_SCRATCH_W1
+                                                sty  P8ZP_SCRATCH_W1+1
+                                                pha  ; save array ptr lsb
+                                                tya
+                                                pha  ; save array ptr msb
+                                                lda  #<$tempvar
+                                                ldy  #>$tempvar
+                                                jsr  floats.copy_float""")
+
+                            // calculate on tempvar
+                            when(value.kind) {
+                                SourceStorageKind.LITERALNUMBER -> inplacemodificationFloatWithLiteralval(tempvar, operator, value.number!!.number)
+                                SourceStorageKind.VARIABLE -> inplacemodificationFloatWithVariable(tempvar, operator, value.asmVarname)
+                                SourceStorageKind.REGISTER -> inplacemodificationFloatWithVariable(tempvar, operator, regName(value))
+                                SourceStorageKind.MEMORY -> TODO("memread into float")
+                                SourceStorageKind.ARRAY -> inplacemodificationFloatWithValue(tempvar, operator, value.array!!)
+                                SourceStorageKind.EXPRESSION -> {
+                                    if(value.expression is PtTypeCast) {
+                                        if (tryInplaceModifyWithRemovedRedundantCast(value.expression, target, operator))
+                                            return
+                                        inplacemodificationFloatWithValue(tempvar, operator, value.expression)
+                                    } else {
+                                        inplacemodificationFloatWithValue(tempvar, operator, value.expression!!)
+                                    }
+                                }
+                            }
+
+                            // copy tempvar back into array
+                            asmgen.out("""
+                                                lda  #<$tempvar
+                                                ldy  #>$tempvar
+                                                sta  P8ZP_SCRATCH_W1
+                                                sty  P8ZP_SCRATCH_W1+1
+                                                pla  ; restore array ptr msb
+                                                tay
+                                                pla  ; restore array ptr lsb
+                                                jsr  floats.copy_float""")
+                        }
+
+                        else -> throw AssemblyError("weird type to do in-place modification on ${target.datatype}")
                     }
-                    else -> throw AssemblyError("indexer expression should have been replaced by auto indexer var")
                 }
             }
             TargetStorageKind.REGISTER -> throw AssemblyError("no asm gen for reg in-place modification")
@@ -445,7 +459,6 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
     }
 
     private fun inplacemodificationRegisterAXwithVariable(operator: String, variable: String, varDt: DataType): Boolean {
-        // note: we only optimize addition and subtraction, and these are the same for unsigned or signed.
         when(operator) {
             "+" -> {
                 return if(varDt in WordDatatypes) {
@@ -495,12 +508,44 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                     true
                 }
             }
-            else -> return false        // TODO optimize more operators, such as the bitwise logical ones? Might need to know if signed
+            "|" -> {
+                asmgen.out("""
+                    ora  $variable
+                    tay
+                    txa
+                    ora  $variable+1
+                    tax
+                    tya
+                """)
+                return true
+            }
+            "&" -> {
+                asmgen.out("""
+                    and  $variable
+                    tay
+                    txa
+                    and  $variable+1
+                    tax
+                    tya
+                """)
+                return true
+            }
+            "^" -> {
+                asmgen.out("""
+                    eor  $variable
+                    tay
+                    txa
+                    eor  $variable+1
+                    tax
+                    tya
+                """)
+                return true
+            }
+            else -> return false
         }
     }
 
     private fun inplacemodificationRegisterAXwithLiteralval(operator: String, number: Int): Boolean {
-        // note: we only optimize addition and subtraction, and these are the same for unsigned or signed.
         when(operator) {
             "+" -> {
                 return if(number in -128..255) {
@@ -544,7 +589,40 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                     true
                 }
             }
-            else -> return false        // TODO optimize more operators, such as the bitwise logical ones? Might need to know if signed
+            "|" -> {
+                asmgen.out("""
+                    ora  #<$number
+                    tay
+                    txa
+                    ora  #>$number
+                    tax
+                    tya
+                """)
+                return true
+            }
+            "&" -> {
+                asmgen.out("""
+                    and  #<$number
+                    tay
+                    txa
+                    and  #>$number
+                    tax
+                    tya
+                """)
+                return true
+            }
+            "^" -> {
+                asmgen.out("""
+                    eor  #<$number
+                    tay
+                    txa
+                    eor  #>$number
+                    tax
+                    tya
+                """)
+                return true
+            }
+            else -> return false
         }
     }
 
@@ -715,12 +793,38 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
     }
 
     private fun inplacemodificationByteVariableWithValue(name: String, dt: DataType, operator: String, value: PtExpression) {
+        val shortcutLabel = asmgen.makeLabel("shortcut")
+        when (operator) {
+            "and" -> {
+                // short-circuit  LEFT and RIGHT  -->  if LEFT then RIGHT else LEFT   (== if !LEFT then LEFT else RIGHT)
+                asmgen.out("  lda  $name |  beq  $shortcutLabel")
+                asmgen.assignExpressionToRegister(value, RegisterOrPair.A, dt in SignedDatatypes)
+                asmgen.out("""
+                     and  $name
+                     sta  $name
+$shortcutLabel:""")
+                return
+            }
+            "or" -> {
+                // short-circuit  LEFT or RIGHT  -->  if LEFT then LEFT else RIGHT
+                asmgen.out("  lda  $name |  bne  $shortcutLabel")
+                asmgen.assignExpressionToRegister(value, RegisterOrPair.A, dt in SignedDatatypes)
+                asmgen.out("""
+                     ora  $name
+                     sta  $name
+$shortcutLabel:""")
+                return
+            }
+        }
+
+        // normal evaluation
         asmgen.assignExpressionToRegister(value, RegisterOrPair.A, dt in SignedDatatypes)
         inplacemodificationRegisterAwithVariableWithSwappedOperands(operator, name, dt in SignedDatatypes)
         asmgen.out("  sta  $name")
     }
 
     private fun inplacemodificationByteVariableWithVariable(name: String, dt: DataType, operator: String, otherName: String) {
+        // note: no logical and/or shortcut here, not worth it due to simple right operand
         asmgen.out("  lda  $name")
         inplacemodificationRegisterAwithVariable(operator, otherName, dt in SignedDatatypes)
         asmgen.out("  sta  $name")
@@ -774,9 +878,9 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
 +""")
                 }
             }
-            "&" -> asmgen.out("  and  $variable")
-            "|" -> asmgen.out("  ora  $variable")
-            "^" -> asmgen.out("  eor  $variable")
+            "&", "and" -> asmgen.out("  and  $variable")
+            "|", "or" -> asmgen.out("  ora  $variable")
+            "^","xor" -> asmgen.out("  eor  $variable")
             "==" -> {
                 asmgen.out("""
                     cmp  $variable
@@ -918,8 +1022,9 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             "<<" -> {
                 asmgen.out("""
                     tay
-                    beq  +
                     lda  $variable
+                    cpy  #0
+                    beq  +
 -                   asl  a
                     dey
                     bne  -
@@ -929,8 +1034,9 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                 if(!signed) {
                     asmgen.out("""
                         tay
-                        beq  +
                         lda  $variable
+                        cpy  #0
+                        beq  +
 -                       lsr  a
                         dey
                         bne  -
@@ -938,8 +1044,9 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                 } else {
                     asmgen.out("""
                         tay
-                        beq  +
                         lda  $variable
+                        cpy  #0
+                        beq  +
                         sta  P8ZP_SCRATCH_B1
 -                       asl  a
                         ror  P8ZP_SCRATCH_B1
@@ -1051,6 +1158,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
 
     private fun inplacemodificationByteVariableWithLiteralval(name: String, dt: DataType, operator: String, value: Int) {
         // note: this contains special optimized cases because we know the exact value. Don't replace this with another routine.
+        // note: no logical and/or shortcut here, not worth it due to simple right operand
         when (operator) {
             "+" -> asmgen.out(" lda  $name |  clc |  adc  #$value |  sta  $name")
             "-" -> asmgen.out(" lda  $name |  sec |  sbc  #$value |  sta  $name")
@@ -1113,9 +1221,9 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                     }
                 }
             }
-            "&" -> asmgen.out(" lda  $name |  and  #$value |  sta  $name")
-            "|" -> asmgen.out(" lda  $name |  ora  #$value |  sta  $name")
-            "^" -> asmgen.out(" lda  $name |  eor  #$value |  sta  $name")
+            "&", "and" -> immediateAndInplace(name, value)
+            "|", "or" -> immediateOrInplace(name, value)
+            "^", "xor" -> asmgen.out(" lda  $name |  eor  #$value |  sta  $name")
             "==" -> {
                 asmgen.out("""
                     lda  $name
@@ -1235,6 +1343,22 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
         }
     }
 
+    private fun immediateAndInplace(name: String, value: Int) {
+        if(asmgen.isTargetCpu(CpuType.CPU65c02)) {
+            asmgen.out(" lda  #${value xor 255} |  trb  $name")     // reset bit
+        } else  {
+            asmgen.out(" lda  $name |  and  #$value |  sta  $name")
+        }
+    }
+
+    private fun immediateOrInplace(name: String, value: Int) {
+        if(asmgen.isTargetCpu(CpuType.CPU65c02) && ((value and (value-1))==0)) {
+            asmgen.out(" lda  #$value |  tsb  $name")       // set bit
+        } else  {
+            asmgen.out(" lda  $name |  ora  #$value |  sta  $name")
+        }
+    }
+
     private fun inplacemodificationWordWithMemread(name: String, dt: DataType, operator: String, memread: PtMemoryByte) {
         when (operator) {
             "+" -> {
@@ -1345,7 +1469,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                 if(value in asmgen.optimizedWordMultiplications) {
                     asmgen.out("  lda  $name |  ldy  $name+1 |  jsr  math.mul_word_$value |  sta  $name |  sty  $name+1")
                 } else {
-                    if(block?.veraFxMuls==true)
+                    if(block?.options?.veraFxMuls==true)
                         // cx16 verafx hardware mul
                         asmgen.out("""
                             lda  $name
@@ -1532,7 +1656,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                         asmgen.out("  lda  $name+1 |  and  #>$value |  sta  $name+1")
                     }
                     value < 0x0100 -> {
-                        asmgen.out("  lda  $name |  and  #$value |  sta  $name")
+                        immediateAndInplace(name, value)
                         if(asmgen.isTargetCpu(CpuType.CPU65c02))
                             asmgen.out("  stz  $name+1")
                         else
@@ -1545,7 +1669,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                 when {
                     value == 0 -> {}
                     value and 255 == 0 -> asmgen.out("  lda  $name+1 |  ora  #>$value |  sta  $name+1")
-                    value < 0x0100 -> asmgen.out("  lda  $name |  ora  #$value |  sta  $name")
+                    value < 0x0100 -> immediateOrInplace(name, value)
                     else -> asmgen.out("  lda  $name |  ora  #<$value |  sta  $name |  lda  $name+1 |  ora  #>$value |  sta  $name+1")
                 }
             }
@@ -1812,7 +1936,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                                 sta  $name+1""")
                     }
                     "*" -> {
-                        if(block?.veraFxMuls==true) {
+                        if(block?.options?.veraFxMuls==true) {
                             // cx16 verafx hardware muls
                             if(valueDt==DataType.UBYTE) {
                                 asmgen.out("  lda  $otherName |  sta  cx16.r1")
@@ -1899,12 +2023,12 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                     }
                     "<<" -> {
                         asmgen.out("""
-                        ldy  $otherName
-                        beq  +
--                       asl  $name
-                        rol  $name+1
-                        dey
-                        bne  -
+                            ldy  $otherName
+                            beq  +
+-                           asl  $name
+                            rol  $name+1
+                            dey
+                            bne  -
 +""")
                     }
                     ">>" -> {
@@ -1980,7 +2104,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
                     "+" -> asmgen.out("  lda  $name |  clc |  adc  $otherName |  sta  $name |  lda  $name+1 |  adc  $otherName+1 |  sta  $name+1")
                     "-" -> asmgen.out("  lda  $name |  sec |  sbc  $otherName |  sta  $name |  lda  $name+1 |  sbc  $otherName+1 |  sta  $name+1")
                     "*" -> {
-                        if(block?.veraFxMuls==true)
+                        if(block?.options?.veraFxMuls==true)
                             // cx16 verafx hardware muls
                             asmgen.out("""
                                 lda  $name
@@ -2184,7 +2308,7 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
 
     private fun inplacemodificationWordWithValue(name: String, dt: DataType, operator: String, value: PtExpression, block: PtBlock?) {
         fun multiplyVarByWordInAY() {
-            if(block?.veraFxMuls==true)
+            if(block?.options?.veraFxMuls==true)
                 // cx16 verafx hardware muls
                 asmgen.out("""
                     sta  cx16.r1
@@ -2649,16 +2773,23 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
         val constValueName = allocator.getFloatAsmConst(value)
         when (operator) {
             "+" -> {
-                if (value == 0.0)
-                    return
-                asmgen.out("""
-                    lda  #<$name
-                    ldy  #>$name
-                    jsr  floats.MOVFM
-                    lda  #<$constValueName
-                    ldy  #>$constValueName
-                    jsr  floats.FADD
-                """)
+                when (value) {
+                    0.0 -> return
+                    0.5 -> asmgen.out("""
+                        lda  #<$name
+                        ldy  #>$name
+                        jsr  floats.MOVFM
+                        jsr  floats.FADDH
+                    """)
+                    else -> asmgen.out("""
+                        lda  #<$name
+                        ldy  #>$name
+                        jsr  floats.MOVFM
+                        lda  #<$constValueName
+                        ldy  #>$constValueName
+                        jsr  floats.FADD
+                    """)
+                }
             }
             "-" -> {
                 if (value == 0.0)
@@ -2674,14 +2805,23 @@ internal class AugmentableAssignmentAsmGen(private val program: PtProgram,
             }
             "*" -> {
                 // assume that code optimization is already done on the AST level for special cases such as 0, 1, 2...
-                asmgen.out("""
-                    lda  #<$name
-                    ldy  #>$name
-                    jsr  floats.MOVFM
-                    lda  #<$constValueName
-                    ldy  #>$constValueName
-                    jsr  floats.FMULT
-                """)
+                if(value==10.0) {
+                    asmgen.out("""
+                        lda  #<$name
+                        ldy  #>$name
+                        jsr  floats.MOVFM
+                        jsr  floats.MUL10
+                    """)
+                } else {
+                    asmgen.out("""
+                        lda  #<$name
+                        ldy  #>$name
+                        jsr  floats.MOVFM
+                        lda  #<$constValueName
+                        ldy  #>$constValueName
+                        jsr  floats.FMULT
+                    """)
+                }
             }
             "/" -> {
                 if (value == 0.0)

codeGenCpu6502/test/Dummies.kt

@@ -18,6 +18,8 @@ internal object DummyMemsizer : IMemSizer {
 }
 
 internal object DummyStringEncoder : IStringEncoding {
+    override val defaultEncoding: Encoding = Encoding.ISO
+
     override fun encodeString(str: String, encoding: Encoding): List<UByte> {
         return emptyList()
     }
@@ -27,11 +29,11 @@ internal object DummyStringEncoder : IStringEncoding {
     }
 }
 
-internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors: Boolean=true, private val keepMessagesAfterReporting: Boolean=false):
-    IErrorReporter {
+internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors: Boolean=true, private val keepMessagesAfterReporting: Boolean=false): IErrorReporter {
 
     val errors = mutableListOf<String>()
     val warnings = mutableListOf<String>()
+    val infos = mutableListOf<String>()
 
     override fun err(msg: String, position: Position) {
         val text = "${position.toClickableStr()} $msg"
@@ -45,6 +47,12 @@ internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors:
             warnings.add(text)
     }
 
+    override fun info(msg: String, position: Position) {
+        val text = "${position.toClickableStr()} $msg"
+        if(text !in infos)
+            infos.add(text)
+    }
+
     override fun undefined(symbol: List<String>, position: Position) {
         err("undefined symbol: ${symbol.joinToString(".")}", position)
     }
@@ -52,10 +60,11 @@ internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors:
     override fun noErrors(): Boolean  = errors.isEmpty()
 
     override fun report() {
+        infos.forEach { println("UNITTEST COMPILATION REPORT: INFO: $it") }
         warnings.forEach { println("UNITTEST COMPILATION REPORT: WARNING: $it") }
         errors.forEach { println("UNITTEST COMPILATION REPORT: ERROR: $it") }
         if(throwExceptionAtReportIfErrors)
-            finalizeNumErrors(errors.size, warnings.size)
+            finalizeNumErrors(errors.size, warnings.size, infos.size)
         if(!keepMessagesAfterReporting) {
             clear()
         }
@@ -64,5 +73,6 @@ internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors:
     fun clear() {
         errors.clear()
         warnings.clear()
+        infos.clear()
     }
 }

codeGenCpu6502/test/TestCodegen.kt

@@ -22,6 +22,7 @@ class TestCodegen: FunSpec({
             CbmPrgLauncherType.NONE,
             ZeropageType.DONTUSE,
             zpReserved = emptyList(),
+            zpAllowed = CompilationOptions.AllZeropageAllowed,
             floats = true,
             noSysInit = false,
             compTarget = target,
@@ -44,7 +45,7 @@ class TestCodegen: FunSpec({
 //}
         val codegen = AsmGen6502(prefixSymbols = false)
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main",false, SourceCode.Generated("test"), PtBlock.Options(), 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))
@@ -94,7 +95,7 @@ class TestCodegen: FunSpec({
         program.add(block)
 
         // define the "cx16.r0" virtual register
-        val cx16block = PtBlock("cx16", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val cx16block = PtBlock("cx16", false, SourceCode.Generated("test"), PtBlock.Options(), Position.DUMMY)
         cx16block.add(PtMemMapped("r0", DataType.UWORD, 100u, null, Position.DUMMY))
         program.add(cx16block)
 

codeGenIntermediate/build.gradle

@@ -30,7 +30,7 @@ dependencies {
     // implementation "org.jetbrains.kotlin:kotlin-reflect"
     implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18"
 
-    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.6.2'
+    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.8.0'
 }
 
 sourceSets {

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

@@ -50,25 +50,26 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
         assignment: PtAugmentedAssign
     ): IRCodeChunks {
         val value = assignment.value
-        val vmDt = irType(value.type)
+        val targetDt = irType(assignment.target.type)
+        val signed = assignment.target.type in SignedDatatypes
         return when(assignment.operator) {
-            "+=" -> expressionEval.operatorPlusInplace(address, null, vmDt, value)
-            "-=" -> expressionEval.operatorMinusInplace(address, null, vmDt, value)
-            "*=" -> expressionEval.operatorMultiplyInplace(address, null, vmDt, value)
-            "/=" -> expressionEval.operatorDivideInplace(address, null, vmDt, value.type in SignedDatatypes, value)
-            "|=" -> expressionEval.operatorOrInplace(address, null, vmDt, value)
-            "&=" -> expressionEval.operatorAndInplace(address, null, vmDt, value)
-            "^=" -> expressionEval.operatorXorInplace(address, null, vmDt, value)
-            "<<=" -> expressionEval.operatorShiftLeftInplace(address, null, vmDt, value)
-            ">>=" -> expressionEval.operatorShiftRightInplace(address, null, vmDt, value.type in SignedDatatypes, value)
-            "%=" -> expressionEval.operatorModuloInplace(address, null, vmDt, value)
-            "==" -> expressionEval.operatorEqualsInplace(address, null, vmDt, value)
-            "!=" -> expressionEval.operatorNotEqualsInplace(address, null, vmDt, value)
-            "<" -> expressionEval.operatorLessInplace(address, null, vmDt, value.type in SignedDatatypes, value)
-            ">" -> expressionEval.operatorGreaterInplace(address, null, vmDt, value.type in SignedDatatypes, value)
-            "<=" -> expressionEval.operatorLessEqualInplace(address, null, vmDt, value.type in SignedDatatypes, value)
-            ">=" -> expressionEval.operatorGreaterEqualInplace(address, null, vmDt, value.type in SignedDatatypes, value)
-            in PrefixOperators -> inplacePrefix(assignment.operator, vmDt, address, null)
+            "+=" -> expressionEval.operatorPlusInplace(address, null, targetDt, value)
+            "-=" -> expressionEval.operatorMinusInplace(address, null, targetDt, value)
+            "*=" -> expressionEval.operatorMultiplyInplace(address, null, targetDt, value)
+            "/=" -> expressionEval.operatorDivideInplace(address, null, targetDt, signed, value)
+            "|=" -> expressionEval.operatorOrInplace(address, null, targetDt, value)
+            "&=" -> expressionEval.operatorAndInplace(address, null, targetDt, value)
+            "^=" -> expressionEval.operatorXorInplace(address, null, targetDt, value)
+            "<<=" -> expressionEval.operatorShiftLeftInplace(address, null, targetDt, value)
+            ">>=" -> expressionEval.operatorShiftRightInplace(address, null, targetDt, signed, value)
+            "%=" -> expressionEval.operatorModuloInplace(address, null, targetDt, value)
+            "==" -> expressionEval.operatorEqualsInplace(address, null, targetDt, value)
+            "!=" -> expressionEval.operatorNotEqualsInplace(address, null, targetDt, value)
+            "<" -> expressionEval.operatorLessInplace(address, null, targetDt, signed, value)
+            ">" -> expressionEval.operatorGreaterInplace(address, null, targetDt, signed, value)
+            "<=" -> expressionEval.operatorLessEqualInplace(address, null, targetDt, signed, value)
+            ">=" -> expressionEval.operatorGreaterEqualInplace(address, null, targetDt, signed, value)
+            in PrefixOperators -> inplacePrefix(assignment.operator, targetDt, address, null)
 
             else -> throw AssemblyError("invalid augmented assign operator ${assignment.operator}")
         }
@@ -76,24 +77,27 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
 
     private fun assignVarAugmented(symbol: String, assignment: PtAugmentedAssign): IRCodeChunks {
         val value = assignment.value
+        val signed = assignment.target.type in SignedDatatypes
         val targetDt = irType(assignment.target.type)
-        return when (assignment.operator) {
+        return when(assignment.operator) {
             "+=" -> expressionEval.operatorPlusInplace(null, symbol, targetDt, value)
             "-=" -> expressionEval.operatorMinusInplace(null, symbol, targetDt, value)
             "*=" -> expressionEval.operatorMultiplyInplace(null, symbol, targetDt, value)
-            "/=" -> expressionEval.operatorDivideInplace(null, symbol, targetDt, value.type in SignedDatatypes, value)
+            "/=" -> expressionEval.operatorDivideInplace(null, symbol, targetDt, signed, value)
             "|=" -> expressionEval.operatorOrInplace(null, symbol, targetDt, value)
+            "or=" -> expressionEval.operatorLogicalOrInplace(null, symbol, targetDt, value)
             "&=" -> expressionEval.operatorAndInplace(null, symbol, targetDt, value)
-            "^=" -> expressionEval.operatorXorInplace(null, symbol, targetDt, value)
+            "and=" -> expressionEval.operatorLogicalAndInplace(null, symbol, targetDt, value)
+            "^=", "xor=" -> expressionEval.operatorXorInplace(null, symbol, targetDt, value)
             "<<=" -> expressionEval.operatorShiftLeftInplace(null, symbol, targetDt, value)
-            ">>=" -> expressionEval.operatorShiftRightInplace(null, symbol, targetDt, value.type in SignedDatatypes, value)
+            ">>=" -> expressionEval.operatorShiftRightInplace(null, symbol, targetDt, signed, value)
             "%=" -> expressionEval.operatorModuloInplace(null, symbol, targetDt, value)
             "==" -> expressionEval.operatorEqualsInplace(null, symbol, targetDt, value)
             "!=" -> expressionEval.operatorNotEqualsInplace(null, symbol, targetDt, value)
-            "<" -> expressionEval.operatorLessInplace(null, symbol, targetDt, value.type in SignedDatatypes, value)
-            ">" -> expressionEval.operatorGreaterInplace(null, symbol, targetDt, value.type in SignedDatatypes, value)
-            "<=" -> expressionEval.operatorLessEqualInplace(null, symbol, targetDt, value.type in SignedDatatypes, value)
-            ">=" -> expressionEval.operatorGreaterEqualInplace(null, symbol, targetDt, value.type in SignedDatatypes, value)
+            "<" -> expressionEval.operatorLessInplace(null, symbol, targetDt, signed, value)
+            ">" -> expressionEval.operatorGreaterInplace(null, symbol, targetDt, signed, value)
+            "<=" -> expressionEval.operatorLessEqualInplace(null, symbol, targetDt, signed, value)
+            ">=" -> expressionEval.operatorGreaterEqualInplace(null, symbol, targetDt, signed, value)
             in PrefixOperators -> inplacePrefix(assignment.operator, targetDt, null, symbol)
             else -> throw AssemblyError("invalid augmented assign operator ${assignment.operator}")
         }
@@ -102,11 +106,11 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
     private fun fallbackAssign(origAssign: PtAugmentedAssign): IRCodeChunks {
         val value: PtExpression
         if(origAssign.operator in PrefixOperators) {
-            value = PtPrefix(origAssign.operator, origAssign.value.type, origAssign.value.position)
+            value = PtPrefix(origAssign.operator, origAssign.target.type, origAssign.value.position)
             value.add(origAssign.value)
         } else {
             require(origAssign.operator.endsWith('='))
-            value = PtBinaryExpression(origAssign.operator.dropLast(1), origAssign.value.type, origAssign.value.position)
+            value = PtBinaryExpression(origAssign.operator.dropLast(1), origAssign.target.type, origAssign.value.position)
             val left: PtExpression = origAssign.target.children.single() as PtExpression
             value.add(left)
             value.add(origAssign.value)
@@ -227,11 +231,11 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
                 if(fixedIndex!=null) {
                     val chunk = IRCodeChunk(null, null).also {
                         if(targetArray.splitWords) {
-                            it += IRInstruction(Opcode.STOREZM, IRDataType.BYTE, immediate = arrayLength, labelSymbol = "${variable}_lsb+$fixedIndex")
-                            it += IRInstruction(Opcode.STOREZM, IRDataType.BYTE, immediate = arrayLength, labelSymbol = "${variable}_msb+$fixedIndex")
+                            it += IRInstruction(Opcode.STOREZM, IRDataType.BYTE, immediate = arrayLength, labelSymbol = "${variable}_lsb", symbolOffset = fixedIndex)
+                            it += IRInstruction(Opcode.STOREZM, IRDataType.BYTE, immediate = arrayLength, labelSymbol = "${variable}_msb", symbolOffset = fixedIndex)
                         }
                         else
-                            it += IRInstruction(Opcode.STOREZM, targetDt, labelSymbol = "$variable+${fixedIndex*itemsize}")
+                            it += IRInstruction(Opcode.STOREZM, targetDt, labelSymbol = variable, symbolOffset = fixedIndex*itemsize)
                     }
                     result += chunk
                 } else {
@@ -251,7 +255,7 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
                     if(fixedIndex!=null) {
                         val offset = fixedIndex*itemsize
                         val chunk = IRCodeChunk(null, null).also {
-                            it += IRInstruction(Opcode.STOREM, targetDt, fpReg1 = valueFpRegister, labelSymbol = "$variable+$offset")
+                            it += IRInstruction(Opcode.STOREM, targetDt, fpReg1 = valueFpRegister, labelSymbol = variable, symbolOffset = offset)
                         }
                         result += chunk
                     } else {
@@ -266,12 +270,12 @@ internal class AssignmentGen(private val codeGen: IRCodeGen, private val express
                         val chunk = IRCodeChunk(null, null).also {
                             if(targetArray.splitWords) {
                                 val msbReg = codeGen.registers.nextFree()
-                                it += IRInstruction(Opcode.STOREM, IRDataType.BYTE, reg1 = valueRegister, immediate = arrayLength, labelSymbol = "${variable}_lsb+$fixedIndex")
+                                it += IRInstruction(Opcode.STOREM, IRDataType.BYTE, reg1 = valueRegister, immediate = arrayLength, labelSymbol = "${variable}_lsb", symbolOffset = 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")
+                                it += IRInstruction(Opcode.STOREM, IRDataType.BYTE, reg1 = msbReg, immediate = arrayLength, labelSymbol = "${variable}_msb", symbolOffset = fixedIndex)
                             }
                             else
-                                it += IRInstruction(Opcode.STOREM, targetDt, reg1 = valueRegister, labelSymbol = "$variable+${fixedIndex*itemsize}")
+                                it += IRInstruction(Opcode.STOREM, targetDt, reg1 = valueRegister, labelSymbol = variable, symbolOffset = fixedIndex*itemsize)
                         }
                         result += chunk
                     } else {

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

@@ -20,20 +20,19 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             "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)
-            "pushw" -> funcPushw(call)
             "rsave", "rrestore" -> ExpressionCodeResult.EMPTY  // vm doesn't have registers to save/restore
             "callfar" -> funcCallfar(call)
+            "call" -> funcCall(call)
             "msb" -> funcMsb(call)
             "lsb" -> funcLsb(call)
             "memory" -> funcMemory(call)
-            "peek" -> funcPeek(call)
-            "peekw" -> funcPeekW(call)
-            "poke" -> funcPoke(call)
-            "pokew" -> funcPokeW(call)
-            "pokemon" -> ExpressionCodeResult.EMPTY     // easter egg function
+            "peek" -> funcPeek(call, IRDataType.BYTE)
+            "peekw" -> funcPeek(call, IRDataType.WORD)
+            "peekf" -> funcPeek(call, IRDataType.FLOAT)
+            "poke" -> funcPoke(call, IRDataType.BYTE)
+            "pokew" -> funcPoke(call, IRDataType.WORD)
+            "pokef" -> funcPoke(call, IRDataType.FLOAT)
+            "pokemon" -> funcPokemon(call)
             "mkword" -> funcMkword(call)
             "clamp__byte", "clamp__ubyte", "clamp__word", "clamp__uword" -> funcClamp(call)
             "min__byte", "min__ubyte", "min__word", "min__uword" -> funcMin(call)
@@ -42,10 +41,10 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             "reverse" -> funcReverse(call)
             "setlsb" -> funcSetLsbMsb(call, false)
             "setmsb" -> funcSetLsbMsb(call, true)
-            "rol" -> funcRolRor(Opcode.ROXL, call)
-            "ror" -> funcRolRor(Opcode.ROXR, call)
-            "rol2" -> funcRolRor(Opcode.ROL, call)
-            "ror2" -> funcRolRor(Opcode.ROR, call)
+            "rol" -> funcRolRor(call)
+            "ror" -> funcRolRor(call)
+            "rol2" -> funcRolRor(call)
+            "ror2" -> funcRolRor(call)
             "prog8_lib_stringcompare" -> funcStringCompare(call)
             "prog8_lib_square_byte" -> funcSquare(call, IRDataType.BYTE)
             "prog8_lib_square_word" -> funcSquare(call, IRDataType.WORD)
@@ -69,6 +68,17 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         }
     }
 
+    private fun funcCall(call: PtBuiltinFunctionCall): ExpressionCodeResult {
+        val result = mutableListOf<IRCodeChunkBase>()
+        val addressTr = exprGen.translateExpression(call.args[0])
+        addToResult(result, addressTr, addressTr.resultReg, -1)
+        addInstr(result, IRInstruction(Opcode.CALLI, reg1 = addressTr.resultReg), null)
+        if(call.void)
+            return ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
+        else
+            return ExpressionCodeResult(result, IRDataType.WORD, codeGen.registers.nextFree(), -1)      // TODO actually the result is returned in CPU registers AY...
+    }
+
     private fun funcCallfar(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
         addInstr(result, IRInstruction(Opcode.PREPARECALL, immediate = 3), null)
@@ -138,6 +148,31 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
     private fun funcAny(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val arrayName = call.args[0] as PtIdentifier
         val arrayLength = codeGen.symbolTable.getLength(arrayName.name)
+        val result = mutableListOf<IRCodeChunkBase>()
+        val lengthReg = codeGen.registers.nextFree()
+
+        if(arrayName.type in SplitWordArrayTypes) {
+            // any(lsb-array) or any(msb-array)
+            addInstr(result, IRInstruction(Opcode.PREPARECALL, immediate = 2), null)
+            val trLsb = exprGen.translateExpression(PtIdentifier(arrayName.name+"_lsb", DataType.ARRAY_UB, call.position))
+            addToResult(result, trLsb, trLsb.resultReg, -1)
+            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = lengthReg, immediate = arrayLength), null)
+            result += codeGen.makeSyscall(IMSyscall.ANY_BYTE, listOf(IRDataType.WORD to trLsb.resultReg, IRDataType.BYTE to lengthReg), IRDataType.BYTE to trLsb.resultReg)
+            val shortcircuitLabel = codeGen.createLabelName()
+            result += IRCodeChunk(null, null).also {
+                it += IRInstruction(Opcode.CMPI, IRDataType.BYTE, reg1 = trLsb.resultReg, immediate = 0)
+                it += IRInstruction(Opcode.BSTNE, labelSymbol = shortcircuitLabel)
+                it += IRInstruction(Opcode.PREPARECALL, immediate = 2)
+            }
+            val trMsb = exprGen.translateExpression(PtIdentifier(arrayName.name+"_msb", DataType.ARRAY_UB, call.position))
+            addToResult(result, trMsb, trMsb.resultReg, -1)
+            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = lengthReg, immediate = arrayLength), null)
+            result += codeGen.makeSyscall(IMSyscall.ANY_BYTE, listOf(IRDataType.WORD to trMsb.resultReg, IRDataType.BYTE to lengthReg), IRDataType.BYTE to trMsb.resultReg)
+            addInstr(result, IRInstruction(Opcode.ORR, IRDataType.BYTE, reg1=trLsb.resultReg, reg2=trMsb.resultReg), null)
+            result += IRCodeChunk(shortcircuitLabel, null)
+            return ExpressionCodeResult(result, IRDataType.BYTE, trLsb.resultReg, -1)
+        }
+
         val syscall =
             when (arrayName.type) {
                 DataType.ARRAY_UB,
@@ -147,11 +182,9 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
                 DataType.ARRAY_F -> IMSyscall.ANY_FLOAT
                 else -> throw IllegalArgumentException("weird type")
             }
-        val result = mutableListOf<IRCodeChunkBase>()
         addInstr(result, IRInstruction(Opcode.PREPARECALL, immediate = 2), null)
-        val tr = exprGen.translateExpression(call.args[0])
+        val tr = exprGen.translateExpression(arrayName)
         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)
@@ -160,6 +193,12 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
     private fun funcAll(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val arrayName = call.args[0] as PtIdentifier
         val arrayLength = codeGen.symbolTable.getLength(arrayName.name)
+
+        if(arrayName.type in SplitWordArrayTypes) {
+            // this is a bit complicated to calculate.... have to check all recombined (lsb,msb) words for $0000
+            TODO("all(split words $arrayName)")
+        }
+
         val syscall =
             when(arrayName.type) {
                 DataType.ARRAY_UB,
@@ -171,7 +210,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             }
         val result = mutableListOf<IRCodeChunkBase>()
         addInstr(result, IRInstruction(Opcode.PREPARECALL, immediate = 2), null)
-        val tr = exprGen.translateExpression(call.args[0])
+        val tr = exprGen.translateExpression(arrayName)
         addToResult(result, tr, tr.resultReg, -1)
         val lengthReg = codeGen.registers.nextFree()
         addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = lengthReg, immediate = arrayLength), null)
@@ -221,14 +260,21 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
 
     private fun funcSgn(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
-        val vmDt = irType(call.type)
         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.SGN, vmDt, reg1 = resultReg, reg2 = tr.resultReg)
+
+        if(tr.dt==IRDataType.FLOAT) {
+            addToResult(result, tr, -1, tr.resultFpReg)
+            result += IRCodeChunk(null, null).also {
+                it += IRInstruction(Opcode.SGN, tr.dt, reg1 = resultReg, fpReg1 = tr.resultFpReg)
+            }
+        } else {
+            addToResult(result, tr, tr.resultReg, -1)
+            result += IRCodeChunk(null, null).also {
+                it += IRInstruction(Opcode.SGN, tr.dt, reg1 = resultReg, reg2 = tr.resultReg)
+            }
         }
-        return ExpressionCodeResult(result, vmDt, resultReg, -1)
+        return ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
     }
 
     private fun funcSqrt(call: PtBuiltinFunctionCall): ExpressionCodeResult {
@@ -260,64 +306,40 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
                 }
                 return ExpressionCodeResult(result, IRDataType.FLOAT, -1, resultFpReg)
             }
-            else -> throw AssemblyError("invalid dt for sqrt")
+            else -> throw AssemblyError("invalid dt")
         }
     }
 
-    private fun funcPop(call: PtBuiltinFunctionCall): ExpressionCodeResult {
-        val code = IRCodeChunk(null, null)
-        val reg = codeGen.registers.nextFree()
-        code += IRInstruction(Opcode.POP, IRDataType.BYTE, reg1=reg)
-        val result = mutableListOf<IRCodeChunkBase>(code)
-        result += assignRegisterTo(call.args.single(), reg)
-        return ExpressionCodeResult(result, IRDataType.BYTE, reg, -1)
-    }
-
-    private fun funcPopw(call: PtBuiltinFunctionCall): ExpressionCodeResult {
-        val code = IRCodeChunk(null, null)
-        val reg = codeGen.registers.nextFree()
-        code += IRInstruction(Opcode.POP, IRDataType.WORD, reg1=reg)
-        val result = mutableListOf<IRCodeChunkBase>(code)
-        result += assignRegisterTo(call.args.single(), reg)
-        return ExpressionCodeResult(result, IRDataType.WORD, reg, -1)
-    }
-
-    private fun funcPush(call: PtBuiltinFunctionCall): ExpressionCodeResult {
-        val result = mutableListOf<IRCodeChunkBase>()
-        val tr = exprGen.translateExpression(call.args.single())
-        addToResult(result, tr, tr.resultReg, -1)
-        result += IRCodeChunk(null, null).also {
-            it += IRInstruction(Opcode.PUSH, IRDataType.BYTE, reg1=tr.resultReg)
-        }
-        return ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
-    }
-
-    private fun funcPushw(call: PtBuiltinFunctionCall): ExpressionCodeResult {
-        val result = mutableListOf<IRCodeChunkBase>()
-        val tr = exprGen.translateExpression(call.args.single())
-        addToResult(result, tr, tr.resultReg, -1)
-        result += IRCodeChunk(null, null).also {
-            it += IRInstruction(Opcode.PUSH, IRDataType.WORD, reg1 = tr.resultReg)
-        }
-        return ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
-    }
-
     private fun funcReverse(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val arrayName = call.args[0] as PtIdentifier
         val arrayLength = codeGen.symbolTable.getLength(arrayName.name)
+        val lengthReg = codeGen.registers.nextFree()
+        val result = mutableListOf<IRCodeChunkBase>()
+
+        if(arrayName.type in SplitWordArrayTypes) {
+            // reverse the lsb and msb arrays both, independently
+            addInstr(result, IRInstruction(Opcode.PREPARECALL, immediate = 2), null)
+            val trLsb = exprGen.translateExpression(PtIdentifier(arrayName.name+"_lsb", DataType.ARRAY_UB, call.position))
+            addToResult(result, trLsb, trLsb.resultReg, -1)
+            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = lengthReg, immediate = if(arrayName.type==DataType.STR) arrayLength!!-1 else arrayLength), null)
+            result += codeGen.makeSyscall(IMSyscall.REVERSE_BYTES, listOf(IRDataType.WORD to trLsb.resultReg, IRDataType.BYTE to lengthReg), null)
+            val trMsb = exprGen.translateExpression(PtIdentifier(arrayName.name+"_msb", DataType.ARRAY_UB, call.position))
+            addToResult(result, trMsb, trMsb.resultReg, -1)
+            addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = lengthReg, immediate = if(arrayName.type==DataType.STR) arrayLength!!-1 else arrayLength), null)
+            result += codeGen.makeSyscall(IMSyscall.REVERSE_BYTES, listOf(IRDataType.WORD to trMsb.resultReg, IRDataType.BYTE to lengthReg), null)
+            return ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
+        }
+
         val syscall =
             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>()
         addInstr(result, IRInstruction(Opcode.PREPARECALL, immediate = 2), null)
-        val tr = exprGen.translateExpression(call.args[0])
+        val tr = exprGen.translateExpression(arrayName)
         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)
@@ -339,7 +361,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
             }
         val result = mutableListOf<IRCodeChunkBase>()
         addInstr(result, IRInstruction(Opcode.PREPARECALL, immediate = 2), null)
-        val tr = exprGen.translateExpression(call.args[0])
+        val tr = exprGen.translateExpression(arrayName)
         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)
@@ -349,13 +371,18 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
 
     private fun funcMkword(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
-        val msbTr = exprGen.translateExpression(call.args[0])
-        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=resultReg, reg2 = lsbTr.resultReg, reg3 = msbTr.resultReg)
+        if((call.args[0] as? PtNumber)?.number == 0.0) {
+            // msb is 0, use EXT
+            val lsbTr = exprGen.translateExpression(call.args[1])
+            addToResult(result, lsbTr, lsbTr.resultReg, -1)
+            addInstr(result, IRInstruction(Opcode.EXT, IRDataType.BYTE, reg1=resultReg, reg2 = lsbTr.resultReg), null)
+        } else {
+            val msbTr = exprGen.translateExpression(call.args[0])
+            addToResult(result, msbTr, msbTr.resultReg, -1)
+            val lsbTr = exprGen.translateExpression(call.args[1])
+            addToResult(result, lsbTr, lsbTr.resultReg, -1)
+            addInstr(result, IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=resultReg, reg2 = msbTr.resultReg, reg3 = lsbTr.resultReg), null)
         }
         return ExpressionCodeResult(result, IRDataType.WORD, resultReg, -1)
     }
@@ -435,118 +462,148 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         return ExpressionCodeResult(result, type, leftTr.resultReg, -1)
     }
 
-    private fun funcPokeW(call: PtBuiltinFunctionCall): ExpressionCodeResult {
+    private fun funcPoke(call: PtBuiltinFunctionCall, dt: IRDataType): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
         if(codeGen.isZero(call.args[1])) {
             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, address = address)
+                    it += IRInstruction(Opcode.STOREZM, dt, address = address)
                 }
             } else {
                 val tr = exprGen.translateExpression(call.args[0])
                 addToResult(result, tr, tr.resultReg, -1)
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.STOREZI, IRDataType.WORD, reg1 = tr.resultReg)
+                    it += IRInstruction(Opcode.STOREZI, dt, reg1 = tr.resultReg)
                 }
             }
         } else {
             if (call.args[0] is PtNumber) {
                 val address = (call.args[0] as PtNumber).number.toInt()
                 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, address = address)
+                if(dt==IRDataType.FLOAT) {
+                    addToResult(result, tr, -1, tr.resultFpReg)
+                    result += IRCodeChunk(null, null).also {
+                        it += IRInstruction(Opcode.STOREM, dt, fpReg1 = tr.resultFpReg, address = address)
+                    }
+                } else {
+                    addToResult(result, tr, tr.resultReg, -1)
+                    result += IRCodeChunk(null, null).also {
+                        it += IRInstruction(Opcode.STOREM, dt, reg1 = tr.resultReg, address = address)
+                    }
                 }
             } else {
                 val addressTr = exprGen.translateExpression(call.args[0])
                 addToResult(result, addressTr, addressTr.resultReg, -1)
                 val valueTr = exprGen.translateExpression(call.args[1])
-                addToResult(result, valueTr, valueTr.resultReg, -1)
-                result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.STOREI, IRDataType.WORD, reg1 = valueTr.resultReg, reg2 = addressTr.resultReg)
+                if(dt==IRDataType.FLOAT) {
+                    addToResult(result, valueTr, -1, valueTr.resultFpReg)
+                    result += IRCodeChunk(null, null).also {
+                        it += IRInstruction(Opcode.STOREI, IRDataType.FLOAT, reg1 = addressTr.resultReg, fpReg1 = valueTr.resultFpReg)
+                    }
+                } else {
+                    addToResult(result, valueTr, valueTr.resultReg, -1)
+                    result += IRCodeChunk(null, null).also {
+                        it += IRInstruction(Opcode.STOREI, dt, reg1 = valueTr.resultReg, reg2 = addressTr.resultReg)
+                    }
                 }
             }
         }
         return ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
     }
 
-    private fun funcPoke(call: PtBuiltinFunctionCall): ExpressionCodeResult {
+    private fun funcPeek(call: PtBuiltinFunctionCall, dt: IRDataType): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
-        if(codeGen.isZero(call.args[1])) {
-            if (call.args[0] is PtNumber) {
+        return if(dt==IRDataType.FLOAT) {
+            if(call.args[0] is PtNumber) {
+                val resultFpRegister = codeGen.registers.nextFreeFloat()
                 val address = (call.args[0] as PtNumber).number.toInt()
+                result += IRCodeChunk(null, null).also {
+                    it += IRInstruction(Opcode.LOADM, IRDataType.FLOAT, fpReg1 = resultFpRegister, address = address)
+                }
+                ExpressionCodeResult(result, IRDataType.FLOAT, -1, resultFpRegister)
+            } else {
+                val tr = exprGen.translateExpression(call.args.single())
+                addToResult(result, tr, tr.resultReg, -1)
+                val resultFpReg = codeGen.registers.nextFreeFloat()
+                result += IRCodeChunk(null, null).also {
+                    it += IRInstruction(Opcode.LOADI, IRDataType.FLOAT, reg1 = tr.resultReg, fpReg1 = resultFpReg)
+                }
+                ExpressionCodeResult(result, IRDataType.FLOAT, -1, resultFpReg)
+            }
+        } else {
+            if (call.args[0] is PtNumber) {
+                val resultRegister = codeGen.registers.nextFree()
+                val address = (call.args[0] as PtNumber).number.toInt()
+                result += IRCodeChunk(null, null).also {
+                    it += IRInstruction(Opcode.LOADM, dt, reg1 = resultRegister, address = address)
+                }
+                ExpressionCodeResult(result, dt, resultRegister, -1)
+            } else {
+                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.LOADI, dt, reg1 = resultReg, reg2 = tr.resultReg)
+                }
+                ExpressionCodeResult(result, dt, resultReg, -1)
+            }
+        }
+    }
+
+    private fun funcPokemon(call: PtBuiltinFunctionCall): ExpressionCodeResult {
+        val result = mutableListOf<IRCodeChunkBase>()
+        val address = call.args[0]
+
+        fun pokeM(result: MutableList<IRCodeChunkBase>, address: Int, value: PtExpression) {
+            if(codeGen.isZero(value)) {
                 result += IRCodeChunk(null, null).also {
                     it += IRInstruction(Opcode.STOREZM, IRDataType.BYTE, address = address)
                 }
             } else {
-                val tr = exprGen.translateExpression(call.args[0])
-                addToResult(result, tr, tr.resultReg, -1)
-                result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.STOREZI, IRDataType.BYTE, reg1 = tr.resultReg)
-                }
-            }
-        } else {
-            if (call.args[0] is PtNumber) {
-                val address = (call.args[0] as PtNumber).number.toInt()
-                val tr = exprGen.translateExpression(call.args[1])
+                val tr = exprGen.translateExpression(value)
                 addToResult(result, tr, tr.resultReg, -1)
                 result += IRCodeChunk(null, null).also {
                     it += IRInstruction(Opcode.STOREM, IRDataType.BYTE, reg1 = tr.resultReg, address = address)
                 }
+            }
+        }
+
+        fun pokeI(result: MutableList<IRCodeChunkBase>, register: Int, value: PtExpression) {
+            if(codeGen.isZero(value)) {
+                result += IRCodeChunk(null, null).also {
+                    it += IRInstruction(Opcode.STOREZI, IRDataType.BYTE, reg1 = register)
+                }
             } else {
-                val addressTr = exprGen.translateExpression(call.args[0])
-                addToResult(result, addressTr, addressTr.resultReg, -1)
-                val valueTr = exprGen.translateExpression(call.args[1])
+                val valueTr = exprGen.translateExpression(value)
                 addToResult(result, valueTr, valueTr.resultReg, -1)
                 result += IRCodeChunk(null, null).also {
-                    it += IRInstruction(Opcode.STOREI, IRDataType.BYTE, reg1 = valueTr.resultReg, reg2 = addressTr.resultReg)
+                    it += IRInstruction(Opcode.STOREI, IRDataType.BYTE, reg1 = valueTr.resultReg, reg2 = register)
                 }
             }
         }
-        return ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
-    }
 
-    private fun funcPeekW(call: PtBuiltinFunctionCall): ExpressionCodeResult {
-        val result = mutableListOf<IRCodeChunkBase>()
-        return if(call.args[0] is PtNumber) {
+        return if(address is PtNumber) {
             val resultRegister = codeGen.registers.nextFree()
-            val address = (call.args[0] as PtNumber).number.toInt()
+            val addressNum = address.number.toInt()
             result += IRCodeChunk(null, null).also {
-                it += IRInstruction(Opcode.LOADM, IRDataType.WORD, reg1 = resultRegister, address = address)
+                it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1 = resultRegister, address = addressNum)
             }
+            pokeM(result, addressNum, call.args[1])
             ExpressionCodeResult(result, IRDataType.BYTE, resultRegister, -1)
         } else {
-            val tr = exprGen.translateExpression(call.args.single())
-            addToResult(result, tr, tr.resultReg, -1)
+            val addressTr = exprGen.translateExpression(address)
+            addToResult(result, addressTr, addressTr.resultReg, -1)
             val resultReg = codeGen.registers.nextFree()
             result += IRCodeChunk(null, null).also {
-                it += IRInstruction(Opcode.LOADI, IRDataType.WORD, reg1 = resultReg, reg2 = tr.resultReg)
-            }
-            ExpressionCodeResult(result, IRDataType.WORD, resultReg, -1)
-        }
-    }
-
-    private fun funcPeek(call: PtBuiltinFunctionCall): ExpressionCodeResult {
-        val result = mutableListOf<IRCodeChunkBase>()
-        return if(call.args[0] is PtNumber) {
-            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, address = address)
-            }
-            ExpressionCodeResult(result, IRDataType.BYTE, resultRegister, -1)
-        } else {
-            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.LOADI, IRDataType.BYTE, reg1 = resultReg, reg2 = tr.resultReg)
+                it += IRInstruction(Opcode.LOADI, IRDataType.BYTE, reg1 = resultReg, reg2 = addressTr.resultReg)
             }
+            pokeI(result, addressTr.resultReg, call.args[1])
             ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
         }
     }
 
+
     private fun funcMemory(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val name = (call.args[0] as PtString).value
         val code = IRCodeChunk(null, null)
@@ -558,7 +615,7 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
     private fun funcLsb(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         return exprGen.translateExpression(call.args.single())
         // note: if a word result is needed, the upper byte is cleared by the typecast that follows. No need to do it here.
-        // TODO to be more strict, maybe we *should* introduce a new result register that is of type .b?
+        // ....To be more strict, maybe we should introduce a new result register that is of type .b?
     }
 
     private fun funcMsb(call: PtBuiltinFunctionCall): ExpressionCodeResult {
@@ -566,22 +623,58 @@ internal class BuiltinFuncGen(private val codeGen: IRCodeGen, private val exprGe
         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.MSIG, IRDataType.BYTE, reg1 = resultReg, reg2 = tr.resultReg)
-        }
+        addInstr(result, IRInstruction(Opcode.MSIG, IRDataType.BYTE, reg1 = resultReg, reg2 = tr.resultReg), null)
         // note: if a word result is needed, the upper byte is cleared by the typecast that follows. No need to do it here.
         return ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
     }
 
-    private fun funcRolRor(opcode: Opcode, call: PtBuiltinFunctionCall): ExpressionCodeResult {
-        val vmDt = irType(call.args[0].type)
+    private fun funcRolRor(call: PtBuiltinFunctionCall): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
-        val tr = exprGen.translateExpression(call.args[0])
-        addToResult(result, tr, tr.resultReg, -1)
-        result += IRCodeChunk(null, null).also {
-            it += IRInstruction(opcode, vmDt, reg1 = tr.resultReg)
+        val arg = call.args[0]
+        val vmDt = irType(arg.type)
+        val opcodeMemAndReg = when(call.name) {
+            "rol" -> Opcode.ROXLM to Opcode.ROXL
+            "ror" -> Opcode.ROXRM to Opcode.ROXR
+            "rol2" -> Opcode.ROLM to Opcode.ROL
+            "ror2" -> Opcode.RORM to Opcode.ROR
+            else -> throw AssemblyError("wrong func")
         }
-        result += assignRegisterTo(call.args[0], tr.resultReg)
+
+        val ident = arg as? PtIdentifier
+        if(ident!=null) {
+            addInstr(result, IRInstruction(opcodeMemAndReg.first, vmDt, labelSymbol = ident.name), null)
+            return ExpressionCodeResult(result, vmDt, -1, -1)
+        }
+
+        val memAddr  = (arg as? PtMemoryByte)?.address?.asConstInteger()
+        if(memAddr!=null) {
+            addInstr(result, IRInstruction(opcodeMemAndReg.first, vmDt, address = memAddr), null)
+            return ExpressionCodeResult(result, vmDt, -1, -1)
+        }
+
+        val arr = (arg as? PtArrayIndexer)
+        val index = arr?.index?.asConstInteger()
+        if(arr!=null && index!=null) {
+            val variable = arr.variable.name
+            val itemsize = codeGen.program.memsizer.memorySize(arr.type)
+            addInstr(result, IRInstruction(opcodeMemAndReg.first, vmDt, labelSymbol = variable, symbolOffset = index*itemsize), null)
+            return ExpressionCodeResult(result, vmDt, -1, -1)
+        }
+
+        val opcode = opcodeMemAndReg.second
+        val saveCarry = opcode in OpcodesThatDependOnCarry && !arg.isSimple()
+        if(saveCarry)
+            addInstr(result, IRInstruction(Opcode.PUSHST), null)    // save Carry
+        val tr = exprGen.translateExpression(arg)
+        addToResult(result, tr, tr.resultReg, -1)
+        if(saveCarry)
+            addInstr(result, IRInstruction(Opcode.POPST), null)
+        addInstr(result, IRInstruction(opcode, vmDt, reg1 = tr.resultReg), null)
+        if(saveCarry)
+            addInstr(result, IRInstruction(Opcode.PUSHST), null)    // save Carry
+        result += assignRegisterTo(arg, tr.resultReg)
+        if(saveCarry)
+            addInstr(result, IRInstruction(Opcode.POPST), null)
         return ExpressionCodeResult(result, vmDt, -1, -1)
     }
 

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

@@ -31,7 +31,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 val code = IRCodeChunk(null, null)
                 if(vmDt==IRDataType.FLOAT) {
                     val resultFpRegister = codeGen.registers.nextFreeFloat()
-                    code += IRInstruction(Opcode.LOAD, vmDt, fpReg1 = resultFpRegister, immediateFp = expr.number.toFloat())
+                    code += IRInstruction(Opcode.LOAD, vmDt, fpReg1 = resultFpRegister, immediateFp = expr.number)
                     ExpressionCodeResult(code, vmDt,-1, resultFpRegister)
                 }
                 else {
@@ -154,7 +154,19 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 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")
+            DataType.ARRAY_F -> {
+                addInstr(result, IRInstruction(Opcode.PREPARECALL, immediate = 3), null)
+                val elementTr = translateExpression(check.element)
+                addToResult(result, elementTr, -1, elementTr.resultFpReg)
+                val iterableTr = translateExpression(check.iterable)
+                addToResult(result, iterableTr, iterableTr.resultReg, -1)
+                val lengthReg = codeGen.registers.nextFree()
+                val resultReg = 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.FLOATARRAY_CONTAINS, listOf(IRDataType.FLOAT to elementTr.resultFpReg, IRDataType.WORD to iterableTr.resultReg, IRDataType.BYTE to lengthReg), IRDataType.BYTE to resultReg)
+                return ExpressionCodeResult(result, IRDataType.BYTE, resultReg, -1)
+            }
             else -> throw AssemblyError("weird iterable dt ${check.iterable.type} for ${check.iterable.name}")
         }
     }
@@ -168,7 +180,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         if(arrayIx.usesPointerVariable) {
             if(eltSize!=1)
                 throw AssemblyError("non-array var indexing requires bytes dt")
-            if(arrayIx.index.type!=DataType.UBYTE)
+            if(arrayIx.index.type !in ByteDatatypes)
                 throw AssemblyError("non-array var indexing requires bytes index")
             val tr = translateExpression(arrayIx.index)
             addToResult(result, tr, tr.resultReg, -1)
@@ -188,18 +200,18 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 val memOffset = (arrayIx.index as PtNumber).number.toInt()
                 result += IRCodeChunk(null, null).also {
                     val tmpRegMsb = codeGen.registers.nextFree()
-                    it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=resultRegister, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_lsb+$memOffset")
-                    it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=tmpRegMsb, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_msb+$memOffset")
-                    it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=finalResultReg, reg2=resultRegister, reg3=tmpRegMsb)
+                    it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=tmpRegMsb, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_msb", symbolOffset = memOffset)
+                    it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=resultRegister, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_lsb", symbolOffset = memOffset)
+                    it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=finalResultReg, reg2=tmpRegMsb, reg3=resultRegister)
                 }
             } else {
                 val tr = translateExpression(arrayIx.index)
                 addToResult(result, tr, tr.resultReg, -1)
                 result += IRCodeChunk(null, null).also {
                     val tmpRegMsb = codeGen.registers.nextFree()
-                    it += IRInstruction(Opcode.LOADX, IRDataType.BYTE, reg1=resultRegister, reg2 = tr.resultReg, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_lsb")
                     it += IRInstruction(Opcode.LOADX, IRDataType.BYTE, reg1=tmpRegMsb, reg2 = tr.resultReg, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_msb")
-                    it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=finalResultReg, reg2=resultRegister, reg3=tmpRegMsb)
+                    it += IRInstruction(Opcode.LOADX, IRDataType.BYTE, reg1=resultRegister, reg2 = tr.resultReg, immediate = arrayLength, labelSymbol= "${arrayVarSymbol}_lsb")
+                    it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=finalResultReg, reg2=tmpRegMsb, reg3=resultRegister)
                 }
             }
             return ExpressionCodeResult(result, vmDt, finalResultReg, -1)
@@ -207,14 +219,14 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
 
         var resultFpRegister = -1
         if(arrayIx.index is PtNumber) {
-            val memOffset = ((arrayIx.index as PtNumber).number.toInt() * eltSize).toString()
+            val memOffset = ((arrayIx.index as PtNumber).number.toInt() * eltSize)
             if(vmDt==IRDataType.FLOAT) {
                 resultFpRegister = codeGen.registers.nextFreeFloat()
-                addInstr(result, IRInstruction(Opcode.LOADM, IRDataType.FLOAT, fpReg1=resultFpRegister, labelSymbol = "$arrayVarSymbol+$memOffset"), null)
+                addInstr(result, IRInstruction(Opcode.LOADM, IRDataType.FLOAT, fpReg1=resultFpRegister, labelSymbol = arrayVarSymbol, symbolOffset = memOffset), null)
             }
             else {
                 resultRegister = codeGen.registers.nextFree()
-                addInstr(result, IRInstruction(Opcode.LOADM, vmDt, reg1=resultRegister, labelSymbol = "$arrayVarSymbol+$memOffset"), null)
+                addInstr(result, IRInstruction(Opcode.LOADM, vmDt, reg1=resultRegister, labelSymbol = arrayVarSymbol, symbolOffset = memOffset), null)
             }
         } else {
             val tr = translateExpression(arrayIx.index)
@@ -365,9 +377,11 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             "*" -> operatorMultiply(binExpr, vmDt)
             "/" -> operatorDivide(binExpr, vmDt, signed)
             "%" -> operatorModulo(binExpr, vmDt)
-            "|" -> operatorOr(binExpr, vmDt)
-            "&" -> operatorAnd(binExpr, vmDt)
-            "^" -> operatorXor(binExpr, vmDt)
+            "|" -> operatorOr(binExpr, vmDt, true)
+            "&" -> operatorAnd(binExpr, vmDt, true)
+            "^", "xor" -> operatorXor(binExpr, vmDt)
+            "or" -> operatorOr(binExpr, vmDt, false)
+            "and" -> operatorAnd(binExpr, vmDt, false)
             "<<" -> operatorShiftLeft(binExpr, vmDt)
             ">>" -> operatorShiftRight(binExpr, vmDt, signed)
             "==" -> operatorEquals(binExpr, vmDt, false)
@@ -428,6 +442,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                     result += tr.chunks
                 }
                 // return value
+                var statusFlagResult: Statusflag? = null
                 val returnRegSpec = if(fcall.void) null else {
                     if(callTarget.returns.isEmpty())
                         null
@@ -436,8 +451,11 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                         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 {
+                            statusFlagResult = returns.register.statusflag
+                            val returnRegister = if(statusFlagResult==null) codeGen.registers.nextFree() else -1
+                            FunctionCallArgs.RegSpec(returnIrType, returnRegister, returns.register)
+                        }
                     } else {
                         // 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 }
@@ -455,12 +473,48 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                     else
                         IRInstruction(Opcode.CALL, address = callTarget.address!!.toInt(), fcallArgs = FunctionCallArgs(argRegisters, returnRegSpec))
                 addInstr(result, call, null)
+                var finalReturnRegister = returnRegSpec?.registerNum ?: -1
+
+                if(fcall.parent is PtAssignment || fcall.parent is PtTypeCast) {
+                    // look if the status flag bit should actually be returned as a 0/1 byte value in a result register (so it can be assigned)
+                    if(statusFlagResult!=null && returnRegSpec!=null) {
+                        // assign status flag bit to the return value register
+                        finalReturnRegister = returnRegSpec.registerNum
+                        if(finalReturnRegister<0)
+                            finalReturnRegister = codeGen.registers.nextFree()
+                        when(statusFlagResult) {
+                            Statusflag.Pc -> {
+                                addInstr(result, IRInstruction(Opcode.SCS, returnRegSpec.dt, reg1=finalReturnRegister), null)
+                            }
+                            else -> {
+                                val branchOpcode = when(statusFlagResult) {
+                                    Statusflag.Pc -> throw AssemblyError("carry should be treated separately")
+                                    Statusflag.Pz -> Opcode.BSTEQ
+                                    Statusflag.Pv -> Opcode.BSTVS
+                                    Statusflag.Pn -> Opcode.BSTNEG
+                                }
+                                val setLabel = codeGen.createLabelName()
+                                val endLabel = codeGen.createLabelName()
+                                result += IRCodeChunk(null, null).also {
+                                    it += IRInstruction(branchOpcode, labelSymbol = setLabel)
+                                    it += IRInstruction(Opcode.LOAD, returnRegSpec.dt, reg1=finalReturnRegister, immediate = 0)
+                                    it += IRInstruction(Opcode.JUMP, labelSymbol = endLabel)
+                                }
+                                result += IRCodeChunk(setLabel, null).also {
+                                    it += IRInstruction(Opcode.LOAD, returnRegSpec.dt, reg1=finalReturnRegister, immediate = 1)
+                                }
+                                result += IRCodeChunk(endLabel, null)
+                            }
+                        }
+                    }
+                }
+
                 return if(fcall.void)
                     ExpressionCodeResult(result, IRDataType.BYTE, -1, -1)
                 else if(fcall.type==DataType.FLOAT)
-                    ExpressionCodeResult(result, returnRegSpec!!.dt, -1, returnRegSpec.registerNum)
+                    ExpressionCodeResult(result, returnRegSpec!!.dt, -1, finalReturnRegister)
                 else
-                    ExpressionCodeResult(result, returnRegSpec!!.dt, returnRegSpec.registerNum, -1)
+                    ExpressionCodeResult(result, returnRegSpec!!.dt, finalReturnRegister, -1)
             }
             else -> throw AssemblyError("invalid node type")
         }
@@ -652,37 +706,61 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         }
     }
 
-    private fun operatorAnd(binExpr: PtBinaryExpression, vmDt: IRDataType): ExpressionCodeResult {
+    private fun operatorAnd(binExpr: PtBinaryExpression, vmDt: IRDataType, bitwise: Boolean): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
-        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, immediate = (binExpr.right as PtNumber).number.toInt()), null)
-            ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
-        } else {
+        if(!bitwise && !binExpr.right.isSimple()) {
+            // short-circuit  LEFT and RIGHT  -->  if LEFT then RIGHT else LEFT   (== if !LEFT then LEFT else RIGHT)
             val leftTr = translateExpression(binExpr.left)
             addToResult(result, leftTr, leftTr.resultReg, -1)
+            val shortcutLabel = codeGen.createLabelName()
+            addInstr(result, IRInstruction(Opcode.BSTEQ, labelSymbol = shortcutLabel), null)
             val rightTr = translateExpression(binExpr.right)
-            addToResult(result, rightTr, rightTr.resultReg, -1)
-            addInstr(result, IRInstruction(Opcode.ANDR, vmDt, reg1 = leftTr.resultReg, reg2 = rightTr.resultReg), null)
-            ExpressionCodeResult(result, vmDt, leftTr.resultReg, -1)
+            addToResult(result, rightTr, leftTr.resultReg, -1)
+            result += IRCodeChunk(shortcutLabel, null)
+            return ExpressionCodeResult(result, vmDt, leftTr.resultReg, -1)
+        } else {
+            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, immediate = (binExpr.right as PtNumber).number.toInt()), null)
+                ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
+            } else {
+                val leftTr = translateExpression(binExpr.left)
+                addToResult(result, leftTr, leftTr.resultReg, -1)
+                val rightTr = translateExpression(binExpr.right)
+                addToResult(result, rightTr, rightTr.resultReg, -1)
+                addInstr(result, IRInstruction(Opcode.ANDR, vmDt, reg1 = leftTr.resultReg, reg2 = rightTr.resultReg), null)
+                ExpressionCodeResult(result, vmDt, leftTr.resultReg, -1)
+            }
         }
     }
 
-    private fun operatorOr(binExpr: PtBinaryExpression, vmDt: IRDataType): ExpressionCodeResult {
+    private fun operatorOr(binExpr: PtBinaryExpression, vmDt: IRDataType, bitwise: Boolean): ExpressionCodeResult {
         val result = mutableListOf<IRCodeChunkBase>()
-        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, immediate = (binExpr.right as PtNumber).number.toInt()), null)
-            ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
-        } else {
+        if(!bitwise && !binExpr.right.isSimple()) {
+            // short-circuit  LEFT or RIGHT  -->  if LEFT then LEFT else RIGHT
             val leftTr = translateExpression(binExpr.left)
             addToResult(result, leftTr, leftTr.resultReg, -1)
+            val shortcutLabel = codeGen.createLabelName()
+            addInstr(result, IRInstruction(Opcode.BSTNE, labelSymbol = shortcutLabel), null)
             val rightTr = translateExpression(binExpr.right)
-            addToResult(result, rightTr, rightTr.resultReg, -1)
-            addInstr(result, IRInstruction(Opcode.ORR, vmDt, reg1 = leftTr.resultReg, reg2 = rightTr.resultReg), null)
-            ExpressionCodeResult(result, vmDt, leftTr.resultReg, -1)
+            addToResult(result, rightTr, leftTr.resultReg, -1)
+            result += IRCodeChunk(shortcutLabel, null)
+            return ExpressionCodeResult(result, vmDt, leftTr.resultReg, -1)
+        } else {
+            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, immediate = (binExpr.right as PtNumber).number.toInt()), null)
+                ExpressionCodeResult(result, vmDt, tr.resultReg, -1)
+            } else {
+                val leftTr = translateExpression(binExpr.left)
+                addToResult(result, leftTr, leftTr.resultReg, -1)
+                val rightTr = translateExpression(binExpr.right)
+                addToResult(result, rightTr, rightTr.resultReg, -1)
+                addInstr(result, IRInstruction(Opcode.ORR, vmDt, reg1 = leftTr.resultReg, reg2 = rightTr.resultReg), null)
+                ExpressionCodeResult(result, vmDt, leftTr.resultReg, -1)
+            }
         }
     }
 
@@ -711,7 +789,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             if(constFactorRight!=null && constFactorRight.type!=DataType.FLOAT) {
                 val tr = translateExpression(binExpr.left)
                 addToResult(result, tr, -1, tr.resultFpReg)
-                val factor = constFactorRight.number.toFloat()
+                val factor = constFactorRight.number
                 result += codeGen.divideByConstFloat(tr.resultFpReg, factor)
                 return ExpressionCodeResult(result, vmDt, -1, tr.resultFpReg)
             } else {
@@ -767,13 +845,13 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
             return if(constFactorLeft!=null) {
                 val tr = translateExpression(binExpr.right)
                 addToResult(result, tr, -1, tr.resultFpReg)
-                val factor = constFactorLeft.number.toFloat()
+                val factor = constFactorLeft.number
                 result += codeGen.multiplyByConstFloat(tr.resultFpReg, factor)
                 ExpressionCodeResult(result, vmDt, -1, tr.resultFpReg)
             } else if(constFactorRight!=null) {
                 val tr = translateExpression(binExpr.left)
                 addToResult(result, tr, -1, tr.resultFpReg)
-                val factor = constFactorRight.number.toFloat()
+                val factor = constFactorRight.number
                 result += codeGen.multiplyByConstFloat(tr.resultFpReg, factor)
                 ExpressionCodeResult(result, vmDt, -1, tr.resultFpReg)
             } else {
@@ -821,7 +899,7 @@ 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, immediateFp = (binExpr.right as PtNumber).number.toFloat()), null)
+                    addInstr(result, IRInstruction(Opcode.SUB, vmDt, fpReg1 = tr.resultFpReg, immediateFp = (binExpr.right as PtNumber).number), null)
                     ExpressionCodeResult(result, vmDt, -1, tr.resultFpReg)
                 } else {
                     val leftTr = translateExpression(binExpr.left)
@@ -876,7 +954,7 @@ 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, immediateFp = (binExpr.right as PtNumber).number.toFloat()), null)
+                    addInstr(result, IRInstruction(Opcode.ADD, vmDt, fpReg1 = tr.resultFpReg, immediateFp = (binExpr.right as PtNumber).number), null)
                     ExpressionCodeResult(result, vmDt, -1, tr.resultFpReg)
                 } else {
                     val leftTr = translateExpression(binExpr.left)
@@ -926,8 +1004,38 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         addInstr(result, if(knownAddress!=null)
             IRInstruction(Opcode.ANDM, vmDt, reg1=tr.resultReg, address = knownAddress)
         else
-            IRInstruction(Opcode.ANDM, vmDt, reg1=tr.resultReg, labelSymbol = symbol)
-            ,null)
+            IRInstruction(Opcode.ANDM, vmDt, reg1=tr.resultReg, labelSymbol = symbol),null)
+        return result
+    }
+
+    internal fun operatorLogicalAndInplace(knownAddress: Int?, symbol: String?, vmDt: IRDataType, operand: PtExpression): IRCodeChunks {
+        val result = mutableListOf<IRCodeChunkBase>()
+        val tr = translateExpression(operand)
+        if(!operand.isSimple()) {
+            // short-circuit  LEFT and RIGHT  -->  if LEFT then RIGHT else LEFT   (== if !LEFT then LEFT else RIGHT)
+            val inplaceReg = codeGen.registers.nextFree()
+            val shortcutLabel = codeGen.createLabelName()
+            result += IRCodeChunk(null, null).also {
+                it += if(knownAddress!=null)
+                    IRInstruction(Opcode.LOADM, vmDt, reg1=inplaceReg, address = knownAddress)
+                else
+                    IRInstruction(Opcode.LOADM, vmDt, reg1=inplaceReg, labelSymbol = symbol)
+                it += IRInstruction(Opcode.BSTEQ, labelSymbol = shortcutLabel)
+            }
+            addToResult(result, tr, tr.resultReg, -1)
+            addInstr(result, if(knownAddress!=null)
+                IRInstruction(Opcode.STOREM, vmDt, reg1=tr.resultReg, address = knownAddress)
+            else
+                IRInstruction(Opcode.STOREM, vmDt, reg1=tr.resultReg, labelSymbol = symbol), null)
+            result += IRCodeChunk(shortcutLabel, null)
+        } else {
+            // normal evaluation, it is *likely* shorter and faster because of the simple operands.
+            addToResult(result, tr, tr.resultReg, -1)
+            addInstr(result, if(knownAddress!=null)
+                IRInstruction(Opcode.ANDM, vmDt, reg1=tr.resultReg, address = knownAddress)
+            else
+                IRInstruction(Opcode.ANDM, vmDt, reg1=tr.resultReg, labelSymbol = symbol),null)
+        }
         return result
     }
 
@@ -938,8 +1046,38 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         addInstr(result, if(knownAddress!=null)
             IRInstruction(Opcode.ORM, vmDt, reg1=tr.resultReg, address = knownAddress)
         else
-            IRInstruction(Opcode.ORM, vmDt, reg1=tr.resultReg, labelSymbol = symbol)
-            , null)
+            IRInstruction(Opcode.ORM, vmDt, reg1=tr.resultReg, labelSymbol = symbol), null)
+        return result
+    }
+
+    internal fun operatorLogicalOrInplace(knownAddress: Int?, symbol: String?, vmDt: IRDataType, operand: PtExpression): IRCodeChunks {
+        val result = mutableListOf<IRCodeChunkBase>()
+        val tr = translateExpression(operand)
+        if(!operand.isSimple()) {
+            // short-circuit  LEFT or RIGHT  -->  if LEFT then LEFT else RIGHT
+            val inplaceReg = codeGen.registers.nextFree()
+            val shortcutLabel = codeGen.createLabelName()
+            result += IRCodeChunk(null, null).also {
+                it += if(knownAddress!=null)
+                    IRInstruction(Opcode.LOADM, vmDt, reg1=inplaceReg, address = knownAddress)
+                else
+                    IRInstruction(Opcode.LOADM, vmDt, reg1=inplaceReg, labelSymbol = symbol)
+                it += IRInstruction(Opcode.BSTNE, labelSymbol = shortcutLabel)
+            }
+            addToResult(result, tr, tr.resultReg, -1)
+            addInstr(result, if(knownAddress!=null)
+                IRInstruction(Opcode.STOREM, vmDt, reg1=tr.resultReg, address = knownAddress)
+            else
+                IRInstruction(Opcode.STOREM, vmDt, reg1=tr.resultReg, labelSymbol = symbol), null)
+            result += IRCodeChunk(shortcutLabel, null)
+        } else {
+            // normal evaluation, it is *likely* shorter and faster because of the simple operands.
+            addToResult(result, tr, tr.resultReg, -1)
+            addInstr(result, if(knownAddress!=null)
+                IRInstruction(Opcode.ORM, vmDt, reg1=tr.resultReg, address = knownAddress)
+            else
+                IRInstruction(Opcode.ORM, vmDt, reg1=tr.resultReg, labelSymbol = symbol), null)
+        }
         return result
     }
 
@@ -948,7 +1086,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         val constFactorRight = operand as? PtNumber
         if(vmDt==IRDataType.FLOAT) {
             if(constFactorRight!=null && constFactorRight.type!=DataType.FLOAT) {
-                val factor = constFactorRight.number.toFloat()
+                val factor = constFactorRight.number
                 result += codeGen.divideByConstFloatInplace(knownAddress, symbol, factor)
             } else {
                 val tr = translateExpression(operand)
@@ -997,7 +1135,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
         val constFactorRight = operand as? PtNumber
         if(vmDt==IRDataType.FLOAT) {
             if(constFactorRight!=null) {
-                val factor = constFactorRight.number.toFloat()
+                val factor = constFactorRight.number
                 result += codeGen.multiplyByConstFloatInplace(knownAddress, symbol, factor)
             } else {
                 val tr = translateExpression(operand)
@@ -1322,7 +1460,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 // in-place modify a memory location
                 result += IRCodeChunk(null, null).also {
                     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.LOAD, IRDataType.FLOAT, fpReg1 = numberReg, immediateFp = operand.number)
                     it += IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg, fpReg2 = numberReg)
                     it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, immediate = 0)
                     it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpResultReg, reg2=cmpReg, reg3=zeroReg)
@@ -1333,7 +1471,7 @@ internal class ExpressionGen(private val codeGen: IRCodeGen) {
                 // 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, immediateFp = operand.number.toFloat())
+                    it += IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = numberReg, immediateFp = operand.number)
                     it += IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=cmpReg, fpReg1 = valueReg, fpReg2 = numberReg)
                     it += IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1=zeroReg, immediate = 0)
                     it += IRInstruction(compareAndSetOpcode, IRDataType.BYTE, reg1=cmpResultReg, reg2=cmpReg, reg3=zeroReg)

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

@@ -18,7 +18,6 @@ class IRCodeGen(
     private val expressionEval = ExpressionGen(this)
     private val builtinFuncGen = BuiltinFuncGen(this, expressionEval)
     private val assignmentGen = AssignmentGen(this, expressionEval)
-    private var irSymbolTable: IRSymbolTable = IRSymbolTable(null)
     internal val registers = RegisterPool()
 
     fun generate(): IRProgram {
@@ -26,7 +25,7 @@ class IRCodeGen(
         moveAllNestedSubroutinesToBlockScope()
         verifyNameScoping(program, symbolTable)
 
-        irSymbolTable = IRSymbolTable(symbolTable)
+        val irSymbolTable = IRSymbolTable.fromStDuringCodegen(symbolTable)
         val irProg = IRProgram(program.name, irSymbolTable, options, program.encoding)
 
         // collect global variables initializers
@@ -144,19 +143,10 @@ class IRCodeGen(
                 (idx, instr) ->
                     val symbolExpr = instr.labelSymbol
                     if(symbolExpr!=null) {
-                        val symbol: String
-                        val index: UInt
-                        if('+' in symbolExpr) {
-                            val operands = symbolExpr.split('+', )
-                            symbol = operands[0]
-                            index = operands[1].toUInt()
-                        } else {
-                            symbol = symbolExpr
-                            index = 0u
-                        }
-                        val target = symbolTable.flat[symbol]
+                        val index = instr.labelSymbolOffset ?: 0
+                        val target = symbolTable.flat[symbolExpr]
                         if (target is StMemVar) {
-                            replacements.add(Triple(chunk, idx, target.address+index))
+                            replacements.add(Triple(chunk, idx, target.address+index.toUInt()))
                         }
                     }
                 }
@@ -283,15 +273,10 @@ class IRCodeGen(
             else -> TODO("missing codegen for $node")
         }
 
-        chunks.forEach { chunk ->
-            require(chunk.isNotEmpty() || chunk.label != null) {
-                "chunk should have instructions and/or a label"
-            }
-        }
+        val nonEmptyChunks = chunks.filter { it.isNotEmpty() || it.label != null }
+        nonEmptyChunks.filterIsInstance<IRCodeChunk>().firstOrNull()?.appendSrcPosition(node.position)
 
-        chunks.filterIsInstance<IRCodeChunk>().firstOrNull()?.appendSrcPosition(node.position)
-
-        return chunks
+        return nonEmptyChunks
     }
 
     private fun readBinaryData(node: PtIncludeBinary): Collection<UByte> {
@@ -321,7 +306,7 @@ class IRCodeGen(
                 }
                 addInstr(result, branchIns, null)
             } else {
-                val label = if(goto.generatedLabel!=null) goto.generatedLabel else goto.identifier!!.name
+                val label = goto.identifier!!.name
                 val branchIns = when(branch.condition) {
                     BranchCondition.CS -> IRInstruction(Opcode.BSTCS, labelSymbol = label)
                     BranchCondition.CC -> IRInstruction(Opcode.BSTCC, labelSymbol = label)
@@ -470,7 +455,6 @@ class IRCodeGen(
                         addInstr(result, IRInstruction(Opcode.LOAD, IRDataType.BYTE, reg1 = indexReg, immediate = 0), null)
                         result += IRCodeChunk(loopLabel, null).also {
                             it += IRInstruction(Opcode.LOADX, IRDataType.BYTE, reg1 = tmpReg, reg2 = indexReg, labelSymbol = iterable.name)
-                            it += IRInstruction(Opcode.CMPI, IRDataType.BYTE, reg1 = tmpReg, immediate = 0)
                             it += IRInstruction(Opcode.BSTEQ, labelSymbol = endLabel)
                             it += IRInstruction(Opcode.STOREM, IRDataType.BYTE, reg1 = tmpReg, labelSymbol = loopvarSymbol)
                         }
@@ -491,15 +475,18 @@ class IRCodeGen(
                             val tmpRegLsb = registers.nextFree()
                             val tmpRegMsb = registers.nextFree()
                             val concatReg = registers.nextFree()
-                            it += IRInstruction(Opcode.LOADX, IRDataType.BYTE, reg1=tmpRegLsb, reg2=indexReg, immediate = iterableLength, labelSymbol=iterable.name+"_lsb")
                             it += IRInstruction(Opcode.LOADX, IRDataType.BYTE, reg1=tmpRegMsb, reg2=indexReg, immediate = iterableLength, labelSymbol=iterable.name+"_msb")
-                            it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=concatReg, reg2=tmpRegLsb, reg3=tmpRegMsb)
+                            it += IRInstruction(Opcode.LOADX, IRDataType.BYTE, reg1=tmpRegLsb, reg2=indexReg, immediate = iterableLength, labelSymbol=iterable.name+"_lsb")
+                            it += IRInstruction(Opcode.CONCAT, IRDataType.BYTE, reg1=concatReg, reg2=tmpRegMsb, reg3=tmpRegLsb)
                             it += IRInstruction(Opcode.STOREM, irType(elementDt), reg1=concatReg, labelSymbol = loopvarSymbol)
                         }
                         result += translateNode(forLoop.statements)
                         result += IRCodeChunk(null, null).also {
                             it += IRInstruction(Opcode.INC, IRDataType.BYTE, reg1=indexReg)
-                            it += IRInstruction(Opcode.CMPI, IRDataType.BYTE, reg1=indexReg, immediate = if(iterableLength==256) 0 else iterableLength)
+                            if(iterableLength!=256) {
+                                // for length 256, the compare is actually against 0, which doesn't require a separate CMP instruction
+                                it += IRInstruction(Opcode.CMPI, IRDataType.BYTE, reg1=indexReg, immediate = iterableLength)
+                            }
                             it += IRInstruction(Opcode.BSTNE, labelSymbol = loopLabel)
                         }
                     }
@@ -516,7 +503,10 @@ class IRCodeGen(
                         result += translateNode(forLoop.statements)
                         result += addConstReg(IRDataType.BYTE, indexReg, elementSize)
                         result += IRCodeChunk(null, null).also {
-                            it += IRInstruction(Opcode.CMPI, IRDataType.BYTE, reg1=indexReg, immediate = if(lengthBytes==256) 0 else lengthBytes)
+                            if(lengthBytes!=256) {
+                                // for length 256, the compare is actually against 0, which doesn't require a separate CMP instruction
+                                it += IRInstruction(Opcode.CMPI, IRDataType.BYTE, reg1=indexReg, immediate = lengthBytes)
+                            }
                             it += IRInstruction(Opcode.BSTNE, labelSymbol = loopLabel)
                         }
                     }
@@ -705,23 +695,23 @@ class IRCodeGen(
         return code
     }
 
-    internal fun multiplyByConstFloat(fpReg: Int, factor: Float): IRCodeChunk {
+    internal fun multiplyByConstFloat(fpReg: Int, factor: Double): IRCodeChunk {
         val code = IRCodeChunk(null, null)
-        if(factor==1f)
+        if(factor==1.0)
             return code
-        code += if(factor==0f) {
-            IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = fpReg, immediateFp = 0f)
+        code += if(factor==0.0) {
+            IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = fpReg, immediateFp = 0.0)
         } else {
             IRInstruction(Opcode.MUL, IRDataType.FLOAT, fpReg1 = fpReg, immediateFp = factor)
         }
         return code
     }
 
-    internal fun multiplyByConstFloatInplace(knownAddress: Int?, symbol: String?, factor: Float): IRCodeChunk {
+    internal fun multiplyByConstFloatInplace(knownAddress: Int?, symbol: String?, factor: Double): IRCodeChunk {
         val code = IRCodeChunk(null, null)
-        if(factor==1f)
+        if(factor==1.0)
             return code
-        if(factor==0f) {
+        if(factor==0.0) {
             code += if(knownAddress!=null)
                 IRInstruction(Opcode.STOREZM, IRDataType.FLOAT, address = knownAddress)
             else
@@ -802,25 +792,25 @@ class IRCodeGen(
         return code
     }
 
-    internal fun divideByConstFloat(fpReg: Int, factor: Float): IRCodeChunk {
+    internal fun divideByConstFloat(fpReg: Int, factor: Double): IRCodeChunk {
         val code = IRCodeChunk(null, null)
-        if(factor==1f)
+        if(factor==1.0)
             return code
-        code += if(factor==0f) {
-            IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = fpReg, immediateFp = Float.MAX_VALUE)
+        code += if(factor==0.0) {
+            IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = fpReg, immediateFp = Double.MAX_VALUE)
         } else {
             IRInstruction(Opcode.DIVS, IRDataType.FLOAT, fpReg1 = fpReg, immediateFp = factor)
         }
         return code
     }
 
-    internal fun divideByConstFloatInplace(knownAddress: Int?, symbol: String?, factor: Float): IRCodeChunk {
+    internal fun divideByConstFloatInplace(knownAddress: Int?, symbol: String?, factor: Double): IRCodeChunk {
         val code = IRCodeChunk(null, null)
-        if(factor==1f)
+        if(factor==1.0)
             return code
-        if(factor==0f) {
+        if(factor==0.0) {
             val maxvalueReg = registers.nextFreeFloat()
-            code += IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = maxvalueReg, immediateFp = Float.MAX_VALUE)
+            code += IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = maxvalueReg, immediateFp = Double.MAX_VALUE)
             code += if(knownAddress!=null)
                 IRInstruction(Opcode.STOREM, IRDataType.FLOAT, fpReg1 = maxvalueReg, address = knownAddress)
             else
@@ -976,7 +966,6 @@ class IRCodeGen(
                         fpReg2 = rightTr.resultFpReg
                     )
                     when(condition.operator) {
-                        // TODO: the converted list of operators
                         "==" -> {
                             it += IRInstruction(Opcode.CMPI, IRDataType.BYTE, reg1 = compResultReg, immediate = 0)
                             it += branchInstr(goto, Opcode.BSTEQ)
@@ -986,7 +975,6 @@ class IRCodeGen(
                             it += branchInstr(goto, Opcode.BSTNE)
                         }
                         else -> {
-                            // TODO: the old list of operators, still to be converted
                             val gotoOpcode = when (condition.operator) {
                                 "<" -> Opcode.BLTS
                                 ">" -> Opcode.BGTS
@@ -996,8 +984,6 @@ class IRCodeGen(
                             }
                             it += if (goto.address != null)
                                 IRInstruction(gotoOpcode, IRDataType.BYTE, reg1 = compResultReg, immediate = 0, address = goto.address?.toInt())
-                            else if (goto.generatedLabel != null)
-                                IRInstruction(gotoOpcode, IRDataType.BYTE, reg1 = compResultReg, immediate = 0, labelSymbol = goto.generatedLabel)
                             else
                                 IRInstruction(gotoOpcode, IRDataType.BYTE, reg1 = compResultReg, immediate = 0, labelSymbol = goto.identifier!!.name)
                         }
@@ -1016,11 +1002,8 @@ class IRCodeGen(
         }
     }
 
-    // TODO use this everywhere
     private fun branchInstr(goto: PtJump, branchOpcode: Opcode) = if (goto.address != null)
         IRInstruction(branchOpcode, address = goto.address?.toInt())
-    else if (goto.generatedLabel != null)
-        IRInstruction(branchOpcode, labelSymbol = goto.generatedLabel)
     else
         IRInstruction(branchOpcode, labelSymbol = goto.identifier!!.name)
 
@@ -1034,18 +1017,20 @@ class IRCodeGen(
         val condition = ifElse.condition as PtBinaryExpression
         val leftTr = expressionEval.translateExpression(condition.left)
         addToResult(result, leftTr, leftTr.resultReg, -1)
+        val lastInstruction = leftTr.chunks.last().instructions.lastOrNull()
+        val requireCompareZero = lastInstruction?.opcode !in OpcodesThatSetStatusbits
         when(condition.operator) {
-            // TODO: converted list of operators
             "==" -> {
-                addInstr(result, IRInstruction(Opcode.CMPI, irDtLeft, reg1 = leftTr.resultReg, immediate = 0), null)
+                if(requireCompareZero)
+                    addInstr(result, IRInstruction(Opcode.CMPI, irDtLeft, reg1 = leftTr.resultReg, immediate = 0), null)
                 addInstr(result, branchInstr(goto, Opcode.BSTEQ), null)
             }
             "!=" -> {
-                addInstr(result, IRInstruction(Opcode.CMPI, irDtLeft, reg1 = leftTr.resultReg, immediate = 0), null)
+                if(requireCompareZero)
+                    addInstr(result, IRInstruction(Opcode.CMPI, irDtLeft, reg1 = leftTr.resultReg, immediate = 0), null)
                 addInstr(result, branchInstr(goto, Opcode.BSTNE), null)
             }
             else -> {
-                // TODO: to-be converted operators
                 val opcode = when (condition.operator) {
                     "<" -> if (signed) Opcode.BLTS else Opcode.BLT
                     ">" -> if (signed) Opcode.BGTS else Opcode.BGT
@@ -1055,8 +1040,6 @@ class IRCodeGen(
                 }
                 if (goto.address != null)
                     addInstr(result, IRInstruction(opcode, irDtLeft, reg1 = leftTr.resultReg, immediate = 0, address = goto.address?.toInt()), null)
-                else if (goto.generatedLabel != null)
-                    addInstr(result, IRInstruction(opcode, irDtLeft, reg1 = leftTr.resultReg, immediate = 0, labelSymbol = goto.generatedLabel), null)
                 else
                     addInstr(result, IRInstruction(opcode, irDtLeft, reg1 = leftTr.resultReg, immediate = 0, labelSymbol = goto.identifier!!.name), null)
             }
@@ -1072,12 +1055,13 @@ class IRCodeGen(
     ) {
         val condition = ifElse.condition as? PtBinaryExpression
         if(condition==null) {
-            val tr = expressionEval.translateExpression(ifElse.condition)
-            result += tr.chunks
-            result += IRCodeChunk(null, null).also {
-                it += IRInstruction(Opcode.CMPI, irDtLeft, reg1 = tr.resultReg, immediate = 0)
-                it += branchInstr(goto, Opcode.BSTNE)
-            }
+            throw AssemblyError("expected condition")
+//            val tr = expressionEval.translateExpression(ifElse.condition)
+//            result += tr.chunks
+//            result += IRCodeChunk(null, null).also {
+//                it += IRInstruction(Opcode.CMPI, irDtLeft, reg1 = tr.resultReg, immediate = 0)      // was redundant CMP most likely
+//                it += branchInstr(goto, Opcode.BSTNE)
+//            }
         } else {
             val leftTr = expressionEval.translateExpression(condition.left)
             addToResult(result, leftTr, leftTr.resultReg, -1)
@@ -1085,7 +1069,6 @@ class IRCodeGen(
             if(number!=null) {
                 val firstReg = leftTr.resultReg
                 when(condition.operator) {
-                    // TODO: the converted operators
                     "==" -> {
                         addInstr(result, IRInstruction(Opcode.CMPI, irDtLeft, reg1 = firstReg, immediate = number), null)
                         addInstr(result, branchInstr(goto, Opcode.BSTEQ), null)
@@ -1095,7 +1078,6 @@ class IRCodeGen(
                         addInstr(result, branchInstr(goto, Opcode.BSTNE), null)
                     }
                     else -> {
-                        // TODO: to-be converted operators
                         val opcode = when (condition.operator) {
                             "<" -> if(signed) Opcode.BLTS else Opcode.BLT
                             ">" -> if(signed) Opcode.BGTS else Opcode.BGT
@@ -1105,8 +1087,6 @@ class IRCodeGen(
                         }
                         if (goto.address != null)
                             addInstr(result, IRInstruction(opcode, irDtLeft, reg1 = firstReg, immediate = number, address = goto.address?.toInt()), null)
-                        else if (goto.generatedLabel != null)
-                            addInstr(result, IRInstruction(opcode, irDtLeft, reg1 = firstReg, immediate = number, labelSymbol = goto.generatedLabel), null)
                         else
                             addInstr(result, IRInstruction(opcode, irDtLeft, reg1 = firstReg, immediate = number, labelSymbol = goto.identifier!!.name), null)
                     }
@@ -1164,8 +1144,6 @@ class IRCodeGen(
                 } else {
                     if (goto.address != null)
                         addInstr(result, IRInstruction(opcode, irDtLeft, reg1 = firstReg, reg2 = secondReg, address = goto.address?.toInt()), null)
-                    else if (goto.generatedLabel != null)
-                        addInstr(result, IRInstruction(opcode, irDtLeft, reg1 = firstReg, reg2 = secondReg, labelSymbol = goto.generatedLabel), null)
                     else
                         addInstr(result, IRInstruction(opcode, irDtLeft, reg1 = firstReg, reg2 = secondReg, labelSymbol = goto.identifier!!.name), null)
                 }
@@ -1187,7 +1165,7 @@ class IRCodeGen(
             addToResult(result, leftTr, -1, leftTr.resultFpReg)
             result += IRCodeChunk(null, null).also {
                 val rightFpReg = registers.nextFreeFloat()
-                it += IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = rightFpReg, immediateFp = 0f)
+                it += IRInstruction(Opcode.LOAD, IRDataType.FLOAT, fpReg1 = rightFpReg, immediateFp = 0.0)
                 it += IRInstruction(Opcode.FCOMP, IRDataType.FLOAT, reg1=compResultReg, fpReg1 = leftTr.resultFpReg, fpReg2 = rightFpReg)
             }
             when (condition.operator) {
@@ -1267,33 +1245,6 @@ class IRCodeGen(
         val condition = ifElse.condition as? PtBinaryExpression
         if(condition==null) {
             throw AssemblyError("if-else condition is not a binaryexpression, should have been converted?")
-//            if(irDtLeft==IRDataType.FLOAT)
-//                throw AssemblyError("condition value should not be float")
-//            val tr = expressionEval.translateExpression(ifElse.condition)
-//            result += tr.chunks
-//            if(ifElse.elseScope.children.isNotEmpty()) {
-//                val elseLabel = createLabelName()
-//                result += IRCodeChunk(null, null).also {
-//                    it += IRInstruction(Opcode.CMPI, irDtLeft, reg1=tr.resultReg, immediate = 0)
-//                    it += IRInstruction(Opcode.BSTEQ, labelSymbol = elseLabel)
-//                    TODO("test this")
-//                }
-//                result += translateNode(ifElse.ifScope)
-//                val afterIfLabel = createLabelName()
-//                addInstr(result, IRInstruction(Opcode.JUMP, labelSymbol = afterIfLabel), null)
-//                result += labelFirstChunk(translateNode(ifElse.elseScope), elseLabel)
-//                result += IRCodeChunk(afterIfLabel, null)
-//            } else {
-//                val afterIfLabel = createLabelName()
-//                result += IRCodeChunk(null, null).also {
-//                    it += IRInstruction(Opcode.CMPI, irDtLeft, reg1=tr.resultReg, immediate = 0)
-//                    it += IRInstruction(Opcode.BSTEQ, labelSymbol = afterIfLabel)
-//                    TODO("test this")
-//                }
-//                result += translateNode(ifElse.ifScope)
-//                result += IRCodeChunk(afterIfLabel, null)
-//            }
-//            return result
         } else {
             if (irDtLeft == IRDataType.FLOAT) {
                 val leftTr = expressionEval.translateExpression(condition.left)
@@ -1491,21 +1442,21 @@ class IRCodeGen(
                 when(postIncrDecr.operator) {
                     "++" -> {
                         result += IRCodeChunk(null, null).also {
-                            it += IRInstruction(Opcode.INCM, IRDataType.BYTE, labelSymbol = "${variable}_lsb+$fixedIndex")
+                            it += IRInstruction(Opcode.INCM, IRDataType.BYTE, labelSymbol = "${variable}_lsb", symbolOffset = fixedIndex)
                             it += IRInstruction(Opcode.BSTNE, labelSymbol = skipLabel)
-                            it += IRInstruction(Opcode.INCM, IRDataType.BYTE, labelSymbol = "${variable}_msb+$fixedIndex")
+                            it += IRInstruction(Opcode.INCM, IRDataType.BYTE, labelSymbol = "${variable}_msb", symbolOffset = fixedIndex)
                         }
                         result += IRCodeChunk(skipLabel, null)
                     }
                     "--" -> {
                         val valueReg=registers.nextFree()
                         result += IRCodeChunk(null, null).also {
-                            it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=valueReg, labelSymbol = "${variable}_lsb+$fixedIndex")
+                            it += IRInstruction(Opcode.LOADM, IRDataType.BYTE, reg1=valueReg, labelSymbol = "${variable}_lsb", symbolOffset = fixedIndex)
                             it += IRInstruction(Opcode.BSTNE, labelSymbol = skipLabel)
-                            it += IRInstruction(Opcode.DECM, IRDataType.BYTE, labelSymbol = "${variable}_msb+$fixedIndex")
+                            it += IRInstruction(Opcode.DECM, IRDataType.BYTE, labelSymbol = "${variable}_msb", symbolOffset = fixedIndex)
                         }
                         result += IRCodeChunk(skipLabel, null).also {
-                            it += IRInstruction(Opcode.DECM, IRDataType.BYTE, labelSymbol = "${variable}_lsb+$fixedIndex")
+                            it += IRInstruction(Opcode.DECM, IRDataType.BYTE, labelSymbol = "${variable}_lsb", symbolOffset = fixedIndex)
                         }
                     }
                     else -> throw AssemblyError("weird operator")
@@ -1596,7 +1547,7 @@ class IRCodeGen(
                             it += IRInstruction(Opcode.STOREIX, irDt, reg1 = dataReg, reg2 = indexReg, labelSymbol = variable)
                         }
                     } else {
-                        addInstr(result, IRInstruction(operationMem, irDt, labelSymbol = "$variable+$offset"), null)
+                        addInstr(result, IRInstruction(operationMem, irDt, labelSymbol = variable, symbolOffset = offset), null)
                     }
                 } else {
                     val indexTr = expressionEval.translateExpression(array.index)
@@ -1641,10 +1592,7 @@ class IRCodeGen(
         addToResult(result, countTr, countTr.resultReg, -1)
         if(constIntValue(repeat.count)==null) {
             // check if the counter is already zero
-            result += IRCodeChunk(null, null).also {
-                it += IRInstruction(Opcode.CMPI, irDt, reg1=countTr.resultReg, immediate = 0)
-                it += IRInstruction(Opcode.BSTEQ, labelSymbol = skipRepeatLabel)
-            }
+            addInstr(result, IRInstruction(Opcode.BSTEQ, labelSymbol = skipRepeatLabel), null)
         }
         result += labelFirstChunk(translateNode(repeat.statements), repeatLabel)
         result += IRCodeChunk(null, null).also {
@@ -1661,9 +1609,7 @@ class IRCodeGen(
         if(jump.address!=null) {
             chunk += IRInstruction(Opcode.JUMP, address = jump.address!!.toInt())
         } else {
-            if (jump.generatedLabel != null)
-                chunk += IRInstruction(Opcode.JUMP, labelSymbol = jump.generatedLabel!!)
-            else if (jump.identifier != null) {
+            if (jump.identifier != null) {
                 val symbol = symbolTable.lookup(jump.identifier!!.name)
                 if(symbol?.type==StNodeType.MEMVAR || symbol?.type==StNodeType.STATICVAR) {
                     val jumpReg = registers.nextFree()
@@ -1707,7 +1653,15 @@ class IRCodeGen(
     }
 
     private fun translate(block: PtBlock): IRBlock {
-        val irBlock = IRBlock(block.name, block.address, block.library, block.forceOutput, translate(block.alignment), block.position)   // no use for other attributes yet?
+        val irBlock = IRBlock(block.name, block.library,
+            IRBlock.Options(
+                block.options.address,
+                block.options.forceOutput,
+                block.options.noSymbolPrefixing,
+                block.options.veraFxMuls,
+                block.options.ignoreUnused,
+                translate(block.options.alignment)
+            ), block.position)
         for (child in block.children) {
             when(child) {
                 is PtNop -> { /* nothing */ }
@@ -1755,7 +1709,7 @@ class IRCodeGen(
                 is PtLabel -> {
                     irBlock += IRCodeChunk(child.name, null)
                 }
-                else -> TODO("weird child node $child")
+                else -> TODO("weird block child node $child")
             }
         }
         return irBlock
@@ -1779,7 +1733,7 @@ class IRCodeGen(
     private var labelSequenceNumber = 0
     internal fun createLabelName(): String {
         labelSequenceNumber++
-        return "prog8_label_gen_$labelSequenceNumber"
+        return "label_gen_$labelSequenceNumber"
     }
 
     internal fun translateBuiltinFunc(call: PtBuiltinFunctionCall): ExpressionCodeResult

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

@@ -45,7 +45,8 @@ class IRPeepholeOptimizer(private val irprog: IRProgram) {
                         val indexedInstructions = chunk1.instructions.withIndex()
                             .map { IndexedValue(it.index, it.value) }
                         val changed = removeNops(chunk1, indexedInstructions)
-                                || removeDoubleLoadsAndStores(chunk1, indexedInstructions)       // TODO not yet implemented
+                                || replaceConcatZeroMsbWithExt(chunk1, indexedInstructions)
+                                || removeDoubleLoadsAndStores(chunk1, indexedInstructions)
                                 || removeUselessArithmetic(chunk1, indexedInstructions)
                                 || removeNeedlessCompares(chunk1, indexedInstructions)
                                 || removeWeirdBranches(chunk1, chunk2, indexedInstructions)
@@ -63,6 +64,22 @@ class IRPeepholeOptimizer(private val irprog: IRProgram) {
         irprog.linkChunks()  // re-link
     }
 
+    private fun replaceConcatZeroMsbWithExt(chunk: IRCodeChunk, indexedInstructions: List<IndexedValue<IRInstruction>>): Boolean {
+        var changed = false
+        indexedInstructions.reversed().forEach { (idx, ins) ->
+            if (ins.opcode == Opcode.CONCAT) {
+                // if the previous instruction loads a zero in the msb, this can be turned into EXT.B instead
+                val prev = indexedInstructions[idx-1].value
+                if(prev.opcode==Opcode.LOAD && prev.immediate==0 && prev.reg1==ins.reg2) {
+                    chunk.instructions[idx] = IRInstruction(Opcode.EXT, IRDataType.BYTE, reg1 = ins.reg1, reg2 = ins.reg3)
+                    chunk.instructions.removeAt(idx-1)
+                    changed = true
+                }
+            }
+        }
+        return changed
+    }
+
     private fun removeEmptyChunks(sub: IRSubroutine) {
         if(sub.chunks.isEmpty())
             return
@@ -70,7 +87,7 @@ 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: merge both labels into 1)
+            If next chunk has label -> label name should be the same, remove original. Otherwise 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.
@@ -79,6 +96,7 @@ class IRPeepholeOptimizer(private val irprog: IRProgram) {
 
         val relabelChunks = mutableListOf<Pair<Int, String>>()
         val removeChunks = mutableListOf<Int>()
+        val replaceLabels = mutableMapOf<String, String>()
 
         sub.chunks.withIndex().forEach { (index, chunk) ->
             if(chunk is IRCodeChunk && chunk.instructions.isEmpty()) {
@@ -92,10 +110,18 @@ class IRPeepholeOptimizer(private val irprog: IRProgram) {
                             relabelChunks += Pair(index + 1, chunk.label!!)
                             removeChunks += index
                         } else {
-                            if (chunk.label == nextchunk.label)
-                                removeChunks += index
-                            else {
-                                // TODO: merge labels on same chunk
+                            // merge both labels into 1 except if this is the label chunk at the start of the subroutine
+                            if(index>0) {
+                                if (chunk.label == nextchunk.label)
+                                    removeChunks += index
+                                else {
+                                    removeChunks += index
+                                    replaceLabels[chunk.label!!] = nextchunk.label!!
+                                    replaceLabels.entries.forEach { (key, value) ->
+                                        if (value == chunk.label)
+                                            replaceLabels[key] = nextchunk.label!!
+                                    }
+                                }
                             }
                         }
                     }
@@ -112,6 +138,25 @@ class IRPeepholeOptimizer(private val irprog: IRProgram) {
             sub.chunks[index] = chunk
         }
         removeChunks.reversed().forEach { index -> sub.chunks.removeAt(index) }
+
+        sub.chunks.forEach { chunk ->
+            chunk.instructions.withIndex().forEach { (idx, instr) ->
+                instr.labelSymbol?.let {
+                    if(instr.opcode in OpcodesThatBranch) {
+                        replaceLabels.forEach { (from, to) ->
+                            if (it == from) {
+                                chunk.instructions[idx] = instr.copy(labelSymbol = to)
+                            }
+                            else {
+                                val actualPrefix = "$from."
+                                if (it.startsWith(actualPrefix))
+                                    chunk.instructions[idx] = instr.copy(labelSymbol = "$to.")
+                            }
+                        }
+                    }
+                }
+            }
+        }
     }
 
     private fun joinChunks(sub: IRSubroutine) {
@@ -279,14 +324,16 @@ class IRPeepholeOptimizer(private val irprog: IRProgram) {
         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)) {
+                if(previous.reg1==ins.reg1) {
+                    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 OpcodesThatDependOnCarry) {
+                            chunk.instructions.removeAt(idx)
+                            changed = true
+                        }
                     }
                 }
             }
@@ -363,18 +410,21 @@ class IRPeepholeOptimizer(private val irprog: IRProgram) {
     }
 
     private fun removeDoubleLoadsAndStores(chunk: IRCodeChunk, indexedInstructions: List<IndexedValue<IRInstruction>>): Boolean {
+        return false
+
+/*
         var changed = false
         indexedInstructions.forEach { (idx, ins) ->
 
             // TODO: detect multiple loads to the same target registers, only keep first (if source is not I/O memory)
             // TODO: detect multiple stores to the same target, only keep first (if target is not I/O memory)
             // TODO: detect multiple float ffrom/fto to the same target, only keep first
-            // TODO: detect multiple sequential rnd with same reg1, only keep one
             // TODO: detect subsequent same xors/nots/negs, remove the pairs completely as they cancel out
             // TODO: detect multiple same ands, ors; only keep first
             // TODO: (hard) detect multiple registers being assigned the same value (and not changed) - use only 1 of them
             // ...
         }
         return changed
+*/
     }
 }

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

@@ -1,7 +1,6 @@
 package prog8.codegen.intermediate
 
 import prog8.code.core.IErrorReporter
-import prog8.code.core.SourceCode.Companion.LIBRARYFILEPREFIX
 import prog8.intermediate.*
 
 
@@ -37,9 +36,9 @@ class IRUnusedCodeRemover(
                 }
             }
 
-            irprog.st.allVariables().forEach { stVar->
-                val initValue = stVar.onetimeInitializationArrayValue
-                if(initValue!=null && !initValue.isEmpty()) {
+            irprog.st.allVariables().forEach { variable ->
+                val initValue = variable.onetimeInitializationArrayValue
+                if(!initValue.isNullOrEmpty()) {
                     if(initValue.any {
                         it.addressOfSymbol?.startsWith(blockLabel)==true
                     })
@@ -61,8 +60,8 @@ class IRUnusedCodeRemover(
         irprog.blocks.forEach { block ->
             block.children.filterIsInstance<IRSubroutine>().reversed().forEach { sub ->
                 if(sub.isEmpty()) {
-                    if(!sub.position.file.startsWith(LIBRARYFILEPREFIX)) {
-                        errors.warn("unused subroutine ${sub.label}", sub.position)
+                    if(!block.options.ignoreUnused) {
+                        errors.info("unused subroutine '${sub.label}'", sub.position)
                     }
                     block.children.remove(sub)
                     irprog.st.removeTree(sub.label)
@@ -82,8 +81,8 @@ class IRUnusedCodeRemover(
         irprog.blocks.forEach { block ->
             block.children.filterIsInstance<IRAsmSubroutine>().reversed().forEach { sub ->
                 if(sub.isEmpty()) {
-                    if(!sub.position.file.startsWith(LIBRARYFILEPREFIX)) {
-                        errors.warn("unused asmsubroutine ${sub.label}", sub.position)
+                    if(!block.options.ignoreUnused) {
+                        errors.info("unused subroutine '${sub.label}'", sub.position)
                     }
                     block.children.remove(sub)
                     irprog.st.removeTree(sub.label)
@@ -103,7 +102,7 @@ class IRUnusedCodeRemover(
         // check if asmsub is called from another asmsub
         irprog.blocks.asSequence().forEach { block ->
             block.children.filterIsInstance<IRAsmSubroutine>().forEach { sub ->
-                if (block.forceOutput || block.library)
+                if (block.options.forceOutput || block.library)
                     linkedAsmSubs += sub
                 if (sub.asmChunk.isNotEmpty()) {
                     allSubs.forEach { (label, asmsub) ->
@@ -168,7 +167,13 @@ class IRUnusedCodeRemover(
                 it.next?.let { next -> new += next }
                 it.instructions.forEach { instr ->
                     if (instr.branchTarget == null)
-                        instr.labelSymbol?.let { label -> allLabeledChunks[label]?.let { chunk -> new += chunk } }
+                        instr.labelSymbol?.let { label ->
+                            val chunk = allLabeledChunks[label.substringBeforeLast('.')]
+                            if(chunk!=null)
+                                new+=chunk
+                            else
+                                allLabeledChunks[label]?.let { new += it }
+                        }
                     else
                         new += instr.branchTarget!!
                 }
@@ -215,6 +220,16 @@ class IRUnusedCodeRemover(
                 linkedChunks += chunk
         }
 
+        // make sure that chunks that are only used as a prefix of a label, are also marked as linked
+        linkedChunks.toList().forEach { chunk ->
+            chunk.instructions.forEach {
+                if(it.labelSymbol!=null) {
+                    val chunkName = it.labelSymbol!!.substringBeforeLast('.')
+                    allLabeledChunks[chunkName]?.let { linkedChunks+=it }
+                }
+            }
+        }
+
         return removeUnlinkedChunks(linkedChunks)
     }
 

codeGenIntermediate/test/Dummies.kt

@@ -16,6 +16,8 @@ internal object DummyMemsizer : IMemSizer {
 }
 
 internal object DummyStringEncoder : IStringEncoding {
+    override val defaultEncoding: Encoding = Encoding.ISO
+
     override fun encodeString(str: String, encoding: Encoding): List<UByte> {
         return emptyList()
     }
@@ -25,11 +27,11 @@ internal object DummyStringEncoder : IStringEncoding {
     }
 }
 
-internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors: Boolean=true, private val keepMessagesAfterReporting: Boolean=false):
-    IErrorReporter {
+internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors: Boolean=true, private val keepMessagesAfterReporting: Boolean=false): IErrorReporter {
 
     val errors = mutableListOf<String>()
     val warnings = mutableListOf<String>()
+    val infos = mutableListOf<String>()
 
     override fun err(msg: String, position: Position) {
         val text = "${position.toClickableStr()} $msg"
@@ -43,6 +45,12 @@ internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors:
             warnings.add(text)
     }
 
+    override fun info(msg: String, position: Position) {
+        val text = "${position.toClickableStr()} $msg"
+        if(text !in infos)
+            infos.add(text)
+    }
+
     override fun undefined(symbol: List<String>, position: Position) {
         err("undefined symbol: ${symbol.joinToString(".")}", position)
     }
@@ -50,10 +58,11 @@ internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors:
     override fun noErrors(): Boolean  = errors.isEmpty()
 
     override fun report() {
+        infos.forEach { println("UNITTEST COMPILATION REPORT: INFO: $it") }
         warnings.forEach { println("UNITTEST COMPILATION REPORT: WARNING: $it") }
         errors.forEach { println("UNITTEST COMPILATION REPORT: ERROR: $it") }
         if(throwExceptionAtReportIfErrors)
-            finalizeNumErrors(errors.size, warnings.size)
+            finalizeNumErrors(errors.size, warnings.size, infos.size)
         if(!keepMessagesAfterReporting) {
             clear()
         }
@@ -62,5 +71,6 @@ internal class ErrorReporterForTests(private val throwExceptionAtReportIfErrors:
     fun clear() {
         errors.clear()
         warnings.clear()
+        infos.clear()
     }
 }

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, false, false, IRBlock.BlockAlignment.NONE, Position.DUMMY)
+        val block = IRBlock("main", false, IRBlock.Options(), Position.DUMMY)
         val sub = IRSubroutine("main.start", emptyList(), null, Position.DUMMY)
         chunks.forEach { sub += it }
         block += sub
@@ -18,12 +18,13 @@ class TestIRPeepholeOpt: FunSpec({
             CbmPrgLauncherType.NONE,
             ZeropageType.DONTUSE,
             emptyList(),
+            CompilationOptions.AllZeropageAllowed,
             floats = false,
             noSysInit = true,
             compTarget = target,
             loadAddress = target.machine.PROGRAM_LOAD_ADDRESS
         )
-        val prog = IRProgram("test", IRSymbolTable(null), options, target)
+        val prog = IRProgram("test", IRSymbolTable(), options, target)
         prog.addBlock(block)
         prog.linkChunks()
         prog.validate()
@@ -52,9 +53,9 @@ class TestIRPeepholeOpt: FunSpec({
 
     test("remove jmp to label below") {
         val c1 = IRCodeChunk("main.start", null)
-        c1 += IRInstruction(Opcode.JUMP, labelSymbol = "label")  // removed, but chunk stays because of label
+        c1 += IRInstruction(Opcode.JUMP, labelSymbol = "label")
         val c2 = IRCodeChunk("label", null)
-        c2 += IRInstruction(Opcode.JUMP, labelSymbol = "label2") // removed, but chunk stays because of label
+        c2 += IRInstruction(Opcode.JUMP, labelSymbol = "label2")
         c2 += IRInstruction(Opcode.NOP)  // removed
         val c3 = IRCodeChunk("label2", null)
         c3 += IRInstruction(Opcode.JUMP, labelSymbol = "label3")
@@ -66,15 +67,13 @@ class TestIRPeepholeOpt: FunSpec({
         irProg.chunks().flatMap { it.instructions }.size shouldBe 5
         val opt = IRPeepholeOptimizer(irProg)
         opt.optimize(true, ErrorReporterForTests())
-        irProg.chunks().size shouldBe 4
+        irProg.chunks().size shouldBe 3
         irProg.chunks()[0].label shouldBe "main.start"
-        irProg.chunks()[1].label shouldBe "label"
-        irProg.chunks()[2].label shouldBe "label2"
-        irProg.chunks()[3].label shouldBe "label3"
+        irProg.chunks()[1].label shouldBe "label2"
+        irProg.chunks()[2].label shouldBe "label3"
         irProg.chunks()[0].isEmpty() shouldBe true
-        irProg.chunks()[1].isEmpty() shouldBe true
-        irProg.chunks()[2].isEmpty() shouldBe false
-        irProg.chunks()[3].isEmpty() shouldBe true
+        irProg.chunks()[1].isEmpty() shouldBe false
+        irProg.chunks()[2].isEmpty() shouldBe true
         val instr = irProg.chunks().flatMap { it.instructions }
         instr.size shouldBe 2
         instr[0].opcode shouldBe Opcode.JUMP

codeGenIntermediate/test/TestVmCodeGen.kt

@@ -19,6 +19,7 @@ class TestVmCodeGen: FunSpec({
             CbmPrgLauncherType.NONE,
             ZeropageType.DONTUSE,
             zpReserved = emptyList(),
+            zpAllowed = CompilationOptions.AllZeropageAllowed,
             floats = true,
             noSysInit = false,
             compTarget = target,
@@ -41,7 +42,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", false, SourceCode.Generated("test"), PtBlock.Options(), 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 +92,7 @@ class TestVmCodeGen: FunSpec({
         program.add(block)
 
         // define the "cx16.r0" virtual register
-        val cx16block = PtBlock("cx16", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val cx16block = PtBlock("cx16", false, SourceCode.Generated("test"), PtBlock.Options(), Position.DUMMY)
         cx16block.add(PtMemMapped("r0", DataType.UWORD, 100u, null, Position.DUMMY))
         program.add(cx16block)
 
@@ -120,7 +121,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", false, SourceCode.Generated("test"), PtBlock.Options(), 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)
@@ -183,7 +184,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", false, SourceCode.Generated("test"), PtBlock.Options(), 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)
@@ -242,7 +243,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", false, SourceCode.Generated("test"), PtBlock.Options(), 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)
@@ -250,7 +251,7 @@ class TestVmCodeGen: FunSpec({
         cmp1.add(PtIdentifier("main.start.f1", DataType.FLOAT, Position.DUMMY))
         cmp1.add(PtNumber(DataType.FLOAT, 42.0, Position.DUMMY))
         if1.add(cmp1)
-        if1.add(PtNodeGroup().also { it.add(PtJump(null, 0xc000u, null, Position.DUMMY)) })
+        if1.add(PtNodeGroup().also { it.add(PtJump(null, 0xc000u, Position.DUMMY)) })
         if1.add(PtNodeGroup())
         sub.add(if1)
         val if2 = PtIfElse(Position.DUMMY)
@@ -258,7 +259,7 @@ class TestVmCodeGen: FunSpec({
         cmp2.add(PtIdentifier("main.start.f1", DataType.FLOAT, Position.DUMMY))
         cmp2.add(PtNumber(DataType.FLOAT, 42.0, Position.DUMMY))
         if2.add(cmp2)
-        if2.add(PtNodeGroup().also { it.add(PtJump(null, 0xc000u, null, Position.DUMMY)) })
+        if2.add(PtNodeGroup().also { it.add(PtJump(null, 0xc000u, Position.DUMMY)) })
         if2.add(PtNodeGroup())
         sub.add(if2)
         block.add(sub)
@@ -289,7 +290,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", false, SourceCode.Generated("test"), PtBlock.Options(), 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)
@@ -352,7 +353,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", false, SourceCode.Generated("test"), PtBlock.Options(), 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)
@@ -411,7 +412,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", false, SourceCode.Generated("test"), PtBlock.Options(), 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)
@@ -419,7 +420,7 @@ class TestVmCodeGen: FunSpec({
         cmp1.add(PtIdentifier("main.start.ub1", DataType.UBYTE, Position.DUMMY))
         cmp1.add(PtNumber(DataType.UBYTE, 42.0, Position.DUMMY))
         if1.add(cmp1)
-        if1.add(PtNodeGroup().also { it.add(PtJump(null, 0xc000u, null, Position.DUMMY)) })
+        if1.add(PtNodeGroup().also { it.add(PtJump(null, 0xc000u, Position.DUMMY)) })
         if1.add(PtNodeGroup())
         sub.add(if1)
         val if2 = PtIfElse(Position.DUMMY)
@@ -427,7 +428,7 @@ class TestVmCodeGen: FunSpec({
         cmp2.add(PtIdentifier("main.start.ub1", DataType.UBYTE, Position.DUMMY))
         cmp2.add(PtNumber(DataType.UBYTE, 42.0, Position.DUMMY))
         if2.add(cmp2)
-        if2.add(PtNodeGroup().also { it.add(PtJump(null, 0xc000u, null, Position.DUMMY)) })
+        if2.add(PtNodeGroup().also { it.add(PtJump(null, 0xc000u, Position.DUMMY)) })
         if2.add(PtNodeGroup())
         sub.add(if2)
         block.add(sub)
@@ -451,7 +452,7 @@ class TestVmCodeGen: FunSpec({
 //}
         val codegen = VmCodeGen()
         val program = PtProgram("test", DummyMemsizer, DummyStringEncoder)
-        val block = PtBlock("main", null, false, false, false, false, PtBlock.BlockAlignment.NONE, SourceCode.Generated("test"), Position.DUMMY)
+        val block = PtBlock("main", false, SourceCode.Generated("test"), PtBlock.Options(), 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)

codeOptimizers/build.gradle

@@ -27,6 +27,7 @@ dependencies {
     implementation project(':codeCore')
     implementation project(':compilerAst')
     implementation "org.jetbrains.kotlin:kotlin-stdlib-jdk8"
+    implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18"
     // implementation "org.jetbrains.kotlin:kotlin-reflect"
 }
 

codeOptimizers/codeOptimizers.iml

@@ -11,5 +11,6 @@
     <orderEntry type="library" name="KotlinJavaRuntime" level="project" />
     <orderEntry type="module" module-name="codeCore" />
     <orderEntry type="module" module-name="compilerAst" />
+    <orderEntry type="library" name="michael.bull.kotlin.result.jvm" level="project" />
   </component>
 </module>

codeOptimizers/src/prog8/optimizer/ConstExprEvaluator.kt

@@ -1,17 +1,19 @@
 package prog8.optimizer
 
+import prog8.ast.Program
 import prog8.ast.base.ExpressionError
 import prog8.ast.base.FatalAstException
-import prog8.ast.expressions.Expression
+import prog8.ast.expressions.FunctionCallExpression
 import prog8.ast.expressions.NumericLiteral
 import prog8.code.core.DataType
 import prog8.code.core.IntegerDatatypes
 import prog8.code.core.Position
+import kotlin.math.*
 
 
 class ConstExprEvaluator {
 
-    fun evaluate(left: NumericLiteral, operator: String, right: NumericLiteral): Expression {
+    fun evaluate(left: NumericLiteral, operator: String, right: NumericLiteral): NumericLiteral {
         try {
             return when(operator) {
                 "+" -> plus(left, right)
@@ -19,9 +21,9 @@ class ConstExprEvaluator {
                 "*" -> multiply(left, right)
                 "/" -> divide(left, right)
                 "%" -> remainder(left, right)
-                "&" -> bitwiseand(left, right)
-                "|" -> bitwiseor(left, right)
-                "^" -> bitwisexor(left, right)
+                "&" -> bitwiseAnd(left, right)
+                "|" -> bitwiseOr(left, right)
+                "^" -> bitwiseXor(left, right)
                 "<" -> NumericLiteral.fromBoolean(left < right, left.position)
                 ">" -> NumericLiteral.fromBoolean(left > right, left.position)
                 "<=" -> NumericLiteral.fromBoolean(left <= right, left.position)
@@ -30,6 +32,9 @@ class ConstExprEvaluator {
                 "!=" -> NumericLiteral.fromBoolean(left != right, left.position)
                 "<<" -> shiftedleft(left, right)
                 ">>" -> shiftedright(left, right)
+                "and" -> logicalAnd(left, right)
+                "or" -> logicalOr(left, right)
+                "xor" -> logicalXor(left, right)
                 else -> throw FatalAstException("const evaluation for invalid operator $operator")
             }
         } catch (ax: FatalAstException) {
@@ -37,7 +42,7 @@ class ConstExprEvaluator {
         }
     }
 
-    private fun shiftedright(left: NumericLiteral, amount: NumericLiteral): Expression {
+    private fun shiftedright(left: NumericLiteral, amount: NumericLiteral): NumericLiteral {
         if(left.type !in IntegerDatatypes || amount.type !in IntegerDatatypes)
             throw ExpressionError("cannot compute $left >> $amount", left.position)
         val result =
@@ -48,52 +53,81 @@ class ConstExprEvaluator {
         return NumericLiteral(left.type, result.toDouble(), left.position)
     }
 
-    private fun shiftedleft(left: NumericLiteral, amount: NumericLiteral): Expression {
+    private fun shiftedleft(left: NumericLiteral, amount: NumericLiteral): NumericLiteral {
         if(left.type !in IntegerDatatypes || amount.type !in IntegerDatatypes)
             throw ExpressionError("cannot compute $left << $amount", left.position)
         val result = left.number.toInt().shl(amount.number.toInt())
-        return NumericLiteral(left.type, result.toDouble(), left.position)
+//        when(left.type) {
+//            DataType.BOOL -> result = result and 1
+//            DataType.UBYTE -> result = result and 255
+//            DataType.BYTE -> result = result.toByte().toInt()
+//            DataType.UWORD -> result = result and 65535
+//            DataType.WORD -> result = result.toShort().toInt()
+//            else -> { /* keep as it is */ }
+//        }
+        return NumericLiteral.optimalNumeric(result.toDouble(), left.position)
     }
 
-    private fun bitwisexor(left: NumericLiteral, right: NumericLiteral): NumericLiteral {
+    private fun bitwiseXor(left: NumericLiteral, right: NumericLiteral): NumericLiteral {
         if(left.type== DataType.UBYTE) {
             if(right.type in IntegerDatatypes) {
                 return NumericLiteral(DataType.UBYTE, (left.number.toInt() xor (right.number.toInt() and 255)).toDouble(), left.position)
             }
         } else if(left.type== DataType.UWORD) {
             if(right.type in IntegerDatatypes) {
-                return NumericLiteral(DataType.UWORD, (left.number.toInt() xor right.number.toInt()).toDouble(), left.position)
+                return NumericLiteral(DataType.UWORD, (left.number.toInt() xor right.number.toInt() and 65535).toDouble(), left.position)
+            }
+        } else if(left.type== DataType.LONG) {
+            if(right.type in IntegerDatatypes) {
+                return NumericLiteral.optimalNumeric((left.number.toInt() xor right.number.toInt()).toDouble(), left.position)
             }
         }
         throw ExpressionError("cannot calculate $left ^ $right", left.position)
     }
 
-    private fun bitwiseor(left: NumericLiteral, right: NumericLiteral): NumericLiteral {
+    private fun bitwiseOr(left: NumericLiteral, right: NumericLiteral): NumericLiteral {
         if(left.type== DataType.UBYTE) {
             if(right.type in IntegerDatatypes) {
                 return NumericLiteral(DataType.UBYTE, (left.number.toInt() or (right.number.toInt() and 255)).toDouble(), left.position)
             }
         } else if(left.type== DataType.UWORD) {
             if(right.type in IntegerDatatypes) {
-                return NumericLiteral(DataType.UWORD, (left.number.toInt() or right.number.toInt()).toDouble(), left.position)
+                return NumericLiteral(DataType.UWORD, (left.number.toInt() or right.number.toInt() and 65535).toDouble(), left.position)
+            }
+        } else if(left.type== DataType.LONG) {
+            if(right.type in IntegerDatatypes) {
+                return NumericLiteral.optimalNumeric((left.number.toInt() or right.number.toInt()).toDouble(), left.position)
             }
         }
         throw ExpressionError("cannot calculate $left | $right", left.position)
     }
 
-    private fun bitwiseand(left: NumericLiteral, right: NumericLiteral): NumericLiteral {
+    private fun bitwiseAnd(left: NumericLiteral, right: NumericLiteral): NumericLiteral {
         if(left.type== DataType.UBYTE) {
             if(right.type in IntegerDatatypes) {
                 return NumericLiteral(DataType.UBYTE, (left.number.toInt() and (right.number.toInt() and 255)).toDouble(), left.position)
             }
         } else if(left.type== DataType.UWORD) {
             if(right.type in IntegerDatatypes) {
-                return NumericLiteral(DataType.UWORD, (left.number.toInt() and right.number.toInt()).toDouble(), left.position)
+                return NumericLiteral(DataType.UWORD, (left.number.toInt() and right.number.toInt() and 65535).toDouble(), left.position)
+            }
+        } else if(left.type== DataType.LONG) {
+            if(right.type in IntegerDatatypes) {
+                return NumericLiteral.optimalNumeric((left.number.toInt() and right.number.toInt()).toDouble(), left.position)
             }
         }
         throw ExpressionError("cannot calculate $left & $right", left.position)
     }
 
+    private fun logicalAnd(left: NumericLiteral, right: NumericLiteral): NumericLiteral =
+        NumericLiteral.fromBoolean(left.asBooleanValue and right.asBooleanValue, left.position)
+
+    private fun logicalOr(left: NumericLiteral, right: NumericLiteral): NumericLiteral =
+        NumericLiteral.fromBoolean(left.asBooleanValue or right.asBooleanValue, left.position)
+
+    private fun logicalXor(left: NumericLiteral, right: NumericLiteral): NumericLiteral =
+        NumericLiteral.fromBoolean(left.asBooleanValue xor right.asBooleanValue, left.position)
+
     private fun plus(left: NumericLiteral, right: NumericLiteral): NumericLiteral {
         val error = "cannot add $left and $right"
         return when (left.type) {
@@ -206,4 +240,154 @@ class ConstExprEvaluator {
             else -> throw ExpressionError(error, left.position)
         }
     }
+
+    fun evaluate(call: FunctionCallExpression, program: Program): NumericLiteral? {
+        if(call.target.nameInSource.size!=2)
+            return null  // likely a builtin function, or user function, these get evaluated elsewhere
+        val constArgs = call.args.mapNotNull { it.constValue(program) }
+        if(constArgs.size!=call.args.size)
+            return null
+
+        return when(call.target.nameInSource[0]) {
+            "math" -> evalMath(call, constArgs)
+            "floats" -> evalFloats(call, constArgs)
+            "string" -> evalString(call, constArgs)
+            else -> null
+        }
+    }
+
+    private fun evalFloats(func: FunctionCallExpression, args: List<NumericLiteral>): NumericLiteral? {
+        val result= when(func.target.nameInSource[1]) {
+            "pow" -> args[0].number.pow(args[1].number)
+            "sin" -> sin(args[0].number)
+            "cos" -> cos(args[0].number)
+            "tan" -> tan(args[0].number)
+            "atan" -> atan(args[0].number)
+            "ln" -> ln(args[0].number)
+            "log2" -> log2(args[0].number)
+            "rad" -> args[0].number/360.0 * 2 * PI
+            "deg" -> args[0].number/ 2 / PI * 360.0
+            "round" -> round(args[0].number)
+            "floor" -> floor(args[0].number)
+            "ceil" -> ceil(args[0].number)
+            "minf", "min" -> min(args[0].number, args[1].number)
+            "maxf", "max" -> max(args[0].number, args[1].number)
+            "clampf", "clamp" -> {
+                var value = args[0].number
+                val minimum = args[1].number
+                val maximum = args[2].number
+                if(value<minimum)
+                    value=minimum
+                if(value<maximum)
+                    value
+                else
+                    maximum
+            }
+            else -> null
+        }
+
+        return if(result==null)
+            null
+        else
+            NumericLiteral(DataType.FLOAT, result, func.position)
+    }
+
+    private fun evalMath(func: FunctionCallExpression, args: List<NumericLiteral>): NumericLiteral? {
+        return when(func.target.nameInSource[1]) {
+            "sin8u" -> {
+                val value = truncate(128.0 + 127.5 * sin(args.single().number / 256.0 * 2 * PI))
+                NumericLiteral(DataType.UBYTE, value, func.position)
+            }
+            "cos8u" -> {
+                val value = truncate(128.0 + 127.5 * cos(args.single().number / 256.0 * 2 * PI))
+                NumericLiteral(DataType.UBYTE, value, func.position)
+            }
+            "sin8" -> {
+                val value = truncate(127.0  * sin(args.single().number / 256.0 * 2 * PI))
+                NumericLiteral(DataType.BYTE, value, func.position)
+            }
+            "cos8" -> {
+                val value = truncate(127.0  * cos(args.single().number / 256.0 * 2 * PI))
+                NumericLiteral(DataType.BYTE, value, func.position)
+            }
+            "sinr8u" -> {
+                val value = truncate(128.0 + 127.5 * sin(args.single().number / 180.0 * 2 * PI))
+                NumericLiteral(DataType.UBYTE, value, func.position)
+            }
+            "cosr8u" -> {
+                val value = truncate(128.0 + 127.5 * cos(args.single().number / 180.0 * 2 * PI))
+                NumericLiteral(DataType.UBYTE, value, func.position)
+            }
+            "sinr8" -> {
+                val value = truncate(127.0  * sin(args.single().number / 180.0 * 2 * PI))
+                NumericLiteral(DataType.BYTE, value, func.position)
+            }
+            "cosr8" -> {
+                val value = truncate(127.0  * cos(args.single().number / 180.0 * 2 * PI))
+                NumericLiteral(DataType.BYTE, value, func.position)
+            }
+            "log2" -> {
+                val value = truncate(log2(args.single().number))
+                NumericLiteral(DataType.UBYTE, value, func.position)
+            }
+            "log2w" -> {
+                val value = truncate(log2(args.single().number))
+                NumericLiteral(DataType.UWORD, value, func.position)
+            }
+            "atan2" -> {
+                val x1f = args[0].number
+                val y1f = args[1].number
+                val x2f = args[2].number
+                val y2f = args[3].number
+                var radians = atan2(y2f-y1f, x2f-x1f)
+                if(radians<0)
+                    radians+=2*PI
+                NumericLiteral(DataType.UWORD, floor(radians/2.0/PI*256.0), func.position)
+            }
+            "diff" -> {
+                val n1 = args[0].number
+                val n2 = args[1].number
+                val value = if(n1>n2) n1-n2 else n2-n1
+                NumericLiteral(DataType.UBYTE, value, func.position)
+            }
+            "diffw" -> {
+                val n1 = args[0].number
+                val n2 = args[1].number
+                val value = if(n1>n2) n1-n2 else n2-n1
+                NumericLiteral(DataType.UWORD, value, func.position)
+            }
+            else -> null
+        }
+    }
+
+    private fun evalString(func: FunctionCallExpression, args: List<NumericLiteral>): NumericLiteral? {
+        return when(func.target.nameInSource[1]) {
+            "isdigit" -> {
+                val char = args[0].number.toInt()
+                NumericLiteral.fromBoolean(char in 48..57, func.position)
+            }
+            "isupper" -> {
+                // shifted petscii has 2 ranges that contain the upper case letters... 97-122 and 193-218
+                val char = args[0].number.toInt()
+                NumericLiteral.fromBoolean(char in 97..122 || char in 193..218, func.position)
+            }
+            "islower" -> {
+                val char = args[0].number.toInt()
+                NumericLiteral.fromBoolean(char in 65..90, func.position)
+            }
+            "isletter" -> {
+                val char = args[0].number.toInt()
+                NumericLiteral.fromBoolean(char in 65..90 || char in 97..122 || char in 193..218, func.position)
+            }
+            "isspace" -> {
+                val char = args[0].number.toInt()
+                NumericLiteral.fromBoolean(char in arrayOf(32, 13, 9, 10, 141, 160), func.position)
+            }
+            "isprint" -> {
+                val char = args[0].number.toInt()
+                NumericLiteral.fromBoolean(char in 32..127 || char>=160, func.position)
+            }
+            else -> null
+        }
+    }
 }

codeOptimizers/src/prog8/optimizer/ConstantFoldingOptimizer.kt

@@ -2,7 +2,6 @@ package prog8.optimizer
 
 import prog8.ast.Node
 import prog8.ast.Program
-import prog8.ast.base.FatalAstException
 import prog8.ast.expressions.*
 import prog8.ast.maySwapOperandOrder
 import prog8.ast.statements.ForLoop
@@ -13,10 +12,18 @@ import prog8.ast.walk.AstWalker
 import prog8.ast.walk.IAstModification
 import prog8.code.core.AssociativeOperators
 import prog8.code.core.DataType
+import prog8.code.core.IErrorReporter
 import kotlin.math.floor
 
 
-class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
+class ConstantFoldingOptimizer(private val program: Program, private val errors: IErrorReporter) : AstWalker() {
+
+    private val evaluator = ConstExprEvaluator()
+
+    override fun after(addressOf: AddressOf, parent: Node): Iterable<IAstModification> {
+        val constAddr = addressOf.constValue(program) ?: return noModifications
+        return listOf(IAstModification.ReplaceNode(addressOf, constAddr, parent))
+    }
 
     override fun before(memread: DirectMemoryRead, parent: Node): Iterable<IAstModification> {
         // @( &thing )  -->  thing  (but only if thing is a byte type!)
@@ -80,6 +87,8 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
             else if(expr.operator=="*" && rightconst!=null && expr.left is StringLiteral) {
                 // mutiply a string.
                 val part = expr.left as StringLiteral
+                if(part.value.isEmpty())
+                    errors.warn("resulting string has length zero", part.position)
                 val newStr = StringLiteral(part.value.repeat(rightconst.number.toInt()), part.encoding, expr.position)
                 return listOf(IAstModification.ReplaceNode(expr, newStr, parent))
             }
@@ -87,7 +96,8 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
 
         if(expr.operator=="==" && rightconst!=null) {
             val leftExpr = expr.left as? BinaryExpression
-            if(leftExpr!=null) {
+            // only do this shuffling when the LHS is not a constant itself (otherwise problematic nested replacements)
+            if(leftExpr?.constValue(program) != null) {
                 val leftRightConst = leftExpr.right.constValue(program)
                 if(leftRightConst!=null) {
                     when (leftExpr.operator) {
@@ -134,9 +144,8 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
             }
         }
 
-        val evaluator = ConstExprEvaluator()
-
-        // const fold when both operands are a const
+        // const fold when both operands are a const.
+        // if in a chained comparison, that one has to be desugared first though.
         if(leftconst != null && rightconst != null) {
             val result = evaluator.evaluate(leftconst, expr.operator, rightconst)
             modifications += IAstModification.ReplaceNode(expr, result, parent)
@@ -184,10 +193,12 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
                         val newExpr = BinaryExpression(leftBinExpr.left, "*", constants, expr.position)
                         return listOf(IAstModification.ReplaceNode(expr, newExpr, parent))
                     } else if (leftBinExpr.operator=="/") {
-                        //  (X / C2) * rightConst   -->  X * (rightConst/C2)
-                        val constants = BinaryExpression(rightconst, "/", c2, c2.position)
-                        val newExpr = BinaryExpression(leftBinExpr.left, "*", constants, expr.position)
-                        return listOf(IAstModification.ReplaceNode(expr, newExpr, parent))
+                        if(expr.inferType(program).istype(DataType.FLOAT)) {
+                            //  (X / C2) * rightConst   -->  X * (rightConst/C2)    only valid for floating point
+                            val constants = BinaryExpression(rightconst, "/", c2, c2.position)
+                            val newExpr = BinaryExpression(leftBinExpr.left, "*", constants, expr.position)
+                            return listOf(IAstModification.ReplaceNode(expr, newExpr, parent))
+                        }
                     }
                 }
             }
@@ -195,7 +206,7 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
                 val c2 = leftBinExpr.right.constValue(program)
                 if(c2!=null && leftBinExpr.operator=="/") {
                     // (X / C1) / C2   -->  X / (C1*C2)
-                    // NOTE: do not optimize  (X * C1) / C2   -->  X * (C1/C2)  because this causes precision loss on integers
+                    // NOTE: do not optimize  (X * C1) / C2  for integers,  -->  X * (C1/C2)  because this causes precision loss on integers
                     val constants = BinaryExpression(c2, "*", rightconst, c2.position)
                     val newExpr = BinaryExpression(leftBinExpr.left, "/", constants, expr.position)
                     return listOf(IAstModification.ReplaceNode(expr, newExpr, parent))
@@ -227,10 +238,14 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
                 }
                 else if(leftBinExpr.operator=="*" && rightBinExpr.operator=="*"){
                     if (c1 != null && c2 != null && c1==c2) {
-                        //(X * C) <plusmin> (Y * C)  =>  (X <plusmin> Y) * C
-                        val xwithy = BinaryExpression(leftBinExpr.left, expr.operator, rightBinExpr.left, expr.position)
-                        val newExpr = BinaryExpression(xwithy, "*", c1, expr.position)
-                        modifications += IAstModification.ReplaceNode(expr, newExpr, parent)
+                        //(X * C) <plusmin> (Y * C)  =>  (X <plusmin> Y) * C    (only if types of X and Y are the same!)
+                        val xDt = leftBinExpr.left.inferType(program)
+                        val yDt = rightBinExpr.left.inferType(program)
+                        if(xDt==yDt) {
+                            val xwithy = BinaryExpression(leftBinExpr.left, expr.operator, rightBinExpr.left, expr.position)
+                            val newExpr = BinaryExpression(xwithy, "*", c1, expr.position)
+                            modifications += IAstModification.ReplaceNode(expr, newExpr, parent)
+                        }
                     }
                 }
             }
@@ -311,8 +326,13 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
         val constvalue = functionCallExpr.constValue(program)
         return if(constvalue!=null)
             listOf(IAstModification.ReplaceNode(functionCallExpr, constvalue, parent))
-        else
-            noModifications
+        else {
+            val const2 = evaluator.evaluate(functionCallExpr, program)
+            return if(const2!=null)
+                listOf(IAstModification.ReplaceNode(functionCallExpr, const2, parent))
+            else
+                noModifications
+        }
     }
 
     override fun after(bfc: BuiltinFunctionCall, parent: Node): Iterable<IAstModification> {
@@ -386,7 +406,7 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
                         return listOf(IAstModification.ReplaceNode(forLoop.iterable, newIter, forLoop))
                 }
             }
-            else -> throw FatalAstException("invalid loopvar datatype $loopvar")
+            else -> { /* nothing for floats, these are not allowed in for loops and will give an error elsewhere */ }
         }
 
         return noModifications
@@ -417,6 +437,11 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
         return noModifications
     }
 
+    override fun after(typecast: TypecastExpression, parent: Node): Iterable<IAstModification> {
+        val constValue = typecast.constValue(program) ?: return noModifications
+        return listOf(IAstModification.ReplaceNode(typecast, constValue, parent))
+    }
+
     private class ShuffleOperands(val expr: BinaryExpression,
                                   val exprOperator: String?,
                                   val subExpr: BinaryExpression,
@@ -526,7 +551,8 @@ class ConstantFoldingOptimizer(private val program: Program) : AstWalker() {
                     return IAstModification.ReplaceNode(expr, change, expr.parent)
                 }
             }
-            else if(expr.operator=="*" && subExpr.operator=="/") {
+            else if(expr.operator=="*" && subExpr.operator=="/" && subExpr.inferType(program).istype(DataType.FLOAT)) {
+                // division optimizations only valid for floats
                 if(leftIsConst) {
                     val change = if(subleftIsConst) {
                         // C1*(C2/V) -> (C1*C2)/V

codeOptimizers/src/prog8/optimizer/ConstantIdentifierReplacer.kt

@@ -1,5 +1,7 @@
 package prog8.optimizer
 
+import prog8.ast.IFunctionCall
+import prog8.ast.IStatementContainer
 import prog8.ast.Node
 import prog8.ast.Program
 import prog8.ast.base.FatalAstException
@@ -9,10 +11,17 @@ import prog8.ast.statements.*
 import prog8.ast.walk.AstWalker
 import prog8.ast.walk.IAstModification
 import prog8.code.core.*
+import prog8.compiler.CallGraph
 
 // Fix up the literal value's type to match that of the vardecl
 //   (also check range literal operands types before they get expanded into arrays for instance)
-class VarConstantValueTypeAdjuster(private val program: Program, private val errors: IErrorReporter) : AstWalker() {
+class VarConstantValueTypeAdjuster(
+    private val program: Program,
+    private val options: CompilationOptions,
+    private val errors: IErrorReporter
+) : AstWalker() {
+
+    private val callGraph by lazy { CallGraph(program) }
 
     override fun after(decl: VarDecl, parent: Node): Iterable<IAstModification> {
 
@@ -25,9 +34,10 @@ class VarConstantValueTypeAdjuster(private val program: Program, private val err
                 && declConstValue.type != decl.datatype) {
                 // avoid silent float roundings
                 if(decl.datatype in IntegerDatatypes && declConstValue.type == DataType.FLOAT) {
-                    errors.err("refused rounding of float to avoid loss of precision", decl.value!!.position)
+                    errors.err("refused truncating of float to avoid loss of precision", decl.value!!.position)
                 } else {
                     // cast the numeric literal to the appropriate datatype of the variable
+                    declConstValue.linkParents(decl)
                     val cast = declConstValue.cast(decl.datatype)
                     if (cast.isValid)
                         return listOf(IAstModification.ReplaceNode(decl.value!!, cast.valueOrZero(), decl))
@@ -37,6 +47,73 @@ class VarConstantValueTypeAdjuster(private val program: Program, private val err
             errors.err(x.message, x.position)
         }
 
+        // replace variables by constants, if possible
+        if(options.optimize) {
+            if (decl.sharedWithAsm || decl.type != VarDeclType.VAR || decl.origin != VarDeclOrigin.USERCODE || decl.datatype !in NumericDatatypes)
+                return noModifications
+            if (decl.value != null && decl.value!!.constValue(program) == null)
+                return noModifications
+            val usages = callGraph.usages(decl)
+            val (writes, reads) = usages
+                .partition {
+                    it is InlineAssembly  // can't really tell if it's written to or only read, assume the worst
+                            || it.parent is AssignTarget
+                            || it.parent is ForLoop
+                            || it.parent is AddressOf
+                            || (it.parent as? IFunctionCall)?.target?.nameInSource?.singleOrNull() in InplaceModifyingBuiltinFunctions
+                }
+            val singleAssignment =
+                writes.singleOrNull()?.parent?.parent as? Assignment ?: writes.singleOrNull()?.parent as? Assignment
+            if (singleAssignment == null) {
+                if (writes.isEmpty()) {
+                    if(reads.isEmpty()) {
+                        // variable is never used AT ALL so we just remove it altogether
+                        if("ignore_unused" !in decl.definingBlock.options())
+                            errors.info("removing unused variable '${decl.name}'", decl.position)
+                        return listOf(IAstModification.Remove(decl, parent as IStatementContainer))
+                    }
+                    val declValue = decl.value?.constValue(program)
+                    if (declValue != null) {
+                        // variable is never written to, so it can be replaced with a constant, IF the value is a constant
+                        errors.info("variable is never written to and was replaced by a constant", decl.position)
+                        val const = VarDecl(VarDeclType.CONST, decl.origin, decl.datatype, decl.zeropage, decl.arraysize, decl.name, decl.names, declValue, decl.sharedWithAsm, decl.splitArray, decl.position)
+                        return listOf(
+                            IAstModification.ReplaceNode(decl, const, parent),
+                        )
+                    }
+                }
+            } else {
+                if (singleAssignment.origin == AssignmentOrigin.VARINIT && singleAssignment.value.constValue(program) != null) {
+                    if(reads.isEmpty()) {
+                        // variable is never used AT ALL so we just remove it altogether, including the single assignment
+                        if("ignore_unused" !in decl.definingBlock.options())
+                            errors.info("removing unused variable '${decl.name}'", decl.position)
+                        return listOf(
+                            IAstModification.Remove(decl, parent as IStatementContainer),
+                            IAstModification.Remove(singleAssignment, singleAssignment.parent as IStatementContainer)
+                        )
+                    }
+                    // variable only has a single write and it is the initialization value, so it can be replaced with a constant, IF the value is a constant
+                    errors.info("variable is never written to and was replaced by a constant", decl.position)
+                    val const = VarDecl(VarDeclType.CONST, decl.origin, decl.datatype, decl.zeropage, decl.arraysize, decl.name, decl.names, singleAssignment.value, decl.sharedWithAsm, decl.splitArray, decl.position)
+                    return listOf(
+                        IAstModification.ReplaceNode(decl, const, parent),
+                        IAstModification.Remove(singleAssignment, singleAssignment.parent as IStatementContainer)
+                    )
+                }
+            }
+            /*
+        TODO: need to check if there are no variable usages between the declaration and the assignment (because these rely on the original initialization value)
+        if(writes.size==2) {
+            val firstAssignment = writes[0].parent as? Assignment
+            val secondAssignment = writes[1].parent as? Assignment
+            if(firstAssignment?.origin==AssignmentOrigin.VARINIT && secondAssignment?.value?.constValue(program)!=null) {
+                errors.warn("variable is only assigned once here, consider using this as the initialization value in the declaration instead", secondAssignment.position)
+            }
+        }
+        */
+        }
+
         return noModifications
     }
 
@@ -195,6 +272,14 @@ class VarConstantValueTypeAdjuster(private val program: Program, private val err
 // This is needed because further constant optimizations depend on those.
 internal class ConstantIdentifierReplacer(private val program: Program, private val errors: IErrorReporter, private val compTarget: ICompilationTarget) : AstWalker() {
 
+    override fun before(addressOf: AddressOf, parent: Node): Iterable<IAstModification> {
+        val constValue = addressOf.constValue(program)
+        if(constValue!=null) {
+            return listOf(IAstModification.ReplaceNode(addressOf, constValue, parent))
+        }
+        return noModifications
+    }
+
     override fun after(identifier: IdentifierReference, parent: Node): Iterable<IAstModification> {
         // replace identifiers that refer to const value, with the value itself
         // if it's a simple type and if it's not a left hand side variable
@@ -227,16 +312,20 @@ internal class ConstantIdentifierReplacer(private val program: Program, private
                     listOf(IAstModification.ReplaceNode(arrayIdx, memread, arrayIdx.parent))
                 }
             }
-            return when (cval.type) {
-                in NumericDatatypes -> listOf(
-                    IAstModification.ReplaceNode(
-                        identifier,
-                        NumericLiteral(cval.type, cval.number, identifier.position),
-                        identifier.parent
+            when (cval.type) {
+                in NumericDatatypes -> {
+                    if(parent is AddressOf)
+                        return noModifications      // cannot replace the identifier INSIDE the addr-of here, let's do it later.
+                    return listOf(
+                        IAstModification.ReplaceNode(
+                            identifier,
+                            NumericLiteral(cval.type, cval.number, identifier.position),
+                            identifier.parent
+                        )
                     )
-                )
+                }
                 in PassByReferenceDatatypes -> throw InternalCompilerException("pass-by-reference type should not be considered a constant")
-                else -> noModifications
+                else -> return noModifications
             }
         } catch (x: UndefinedSymbolError) {
             errors.err(x.message, x.position)
@@ -244,14 +333,23 @@ internal class ConstantIdentifierReplacer(private val program: Program, private
         }
     }
 
-    override fun before(decl: VarDecl, parent: Node): Iterable<IAstModification> {
+    override fun after(decl: VarDecl, parent: Node): Iterable<IAstModification> {
         // the initializer value can't refer to the variable itself (recursive definition)
         if(decl.value?.referencesIdentifier(listOf(decl.name)) == true || decl.arraysize?.indexExpr?.referencesIdentifier(listOf(decl.name)) == true) {
             errors.err("recursive var declaration", decl.position)
             return noModifications
         }
 
-        if(decl.type== VarDeclType.CONST || decl.type== VarDeclType.VAR) {
+        if(decl.isArray && decl.type==VarDeclType.MEMORY && decl.value !is IdentifierReference) {
+            val memaddr = decl.value?.constValue(program)
+            if(memaddr!=null && memaddr !== decl.value) {
+                return listOf(IAstModification.SetExpression(
+                    { decl.value = it }, memaddr, decl
+                ))
+            }
+        }
+
+        if(decl.type==VarDeclType.CONST || decl.type==VarDeclType.VAR) {
             if(decl.isArray){
                 val arraysize = decl.arraysize
                 if(arraysize==null) {
@@ -291,8 +389,11 @@ internal class ConstantIdentifierReplacer(private val program: Program, private
     }
 
     private fun createConstArrayInitializerValue(decl: VarDecl): ArrayLiteral? {
-        // convert the initializer range expression from a range or int, to an actual array.
 
+        if(decl.type==VarDeclType.MEMORY)
+            return null     // memory mapped arrays can never have an initializer value other than the address where they're mapped.
+
+        // convert the initializer range expression from a range or int, to an actual array.
         when(decl.datatype) {
             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

codeOptimizers/src/prog8/optimizer/ExpressionSimplifier.kt

@@ -13,7 +13,6 @@ import prog8.ast.statements.Jump
 import prog8.ast.walk.AstWalker
 import prog8.ast.walk.IAstModification
 import prog8.code.core.*
-import prog8.code.target.VMTarget
 import kotlin.math.abs
 import kotlin.math.log2
 import kotlin.math.pow
@@ -66,30 +65,36 @@ class ExpressionSimplifier(private val program: Program,
                 )
             }
             if(elsepart.statements.singleOrNull() is Jump) {
-                val invertedCondition = invertCondition(ifElse.condition)
-                if(invertedCondition!=null) {
-                    return listOf(
-                        IAstModification.ReplaceNode(ifElse.condition, invertedCondition, ifElse),
-                        IAstModification.InsertAfter(ifElse, truepart, parent as IStatementContainer),
-                        IAstModification.ReplaceNode(elsepart, AnonymousScope(mutableListOf(), elsepart.position), ifElse),
-                        IAstModification.ReplaceNode(truepart, elsepart, ifElse)
-                    )
-                }
+                val invertedCondition = invertCondition(ifElse.condition, program)
+                return listOf(
+                    IAstModification.ReplaceNode(ifElse.condition, invertedCondition, ifElse),
+                    IAstModification.InsertAfter(ifElse, truepart, parent as IStatementContainer),
+                    IAstModification.ReplaceNode(elsepart, AnonymousScope(mutableListOf(), elsepart.position), ifElse),
+                    IAstModification.ReplaceNode(truepart, elsepart, ifElse)
+                )
             }
         }
 
-        if(compTarget.name!=VMTarget.NAME) {
-            val booleanCondition = ifElse.condition as? BinaryExpression
-            if(booleanCondition!=null && booleanCondition.operator=="&") {
-                // special optimization of WORD & $ff00  ->  just and the msb of WORD with $ff
-                val rightNum = booleanCondition.right as? NumericLiteral
-                if(rightNum!=null && rightNum.type==DataType.UWORD && (rightNum.number.toInt() and 0x00ff)==0) {
+        val booleanCondition = ifElse.condition as? BinaryExpression
+        if(booleanCondition!=null && booleanCondition.operator=="&") {
+            val rightNum = booleanCondition.right as? NumericLiteral
+            if (rightNum!=null && rightNum.type==DataType.UWORD) {
+                if ((rightNum.number.toInt() and 0x00ff) == 0) {
+                    // if WORD & $xx00  ->  if msb(WORD) & $xx
                     val msb = BuiltinFunctionCall(IdentifierReference(listOf("msb"), booleanCondition.left.position), mutableListOf(booleanCondition.left), booleanCondition.left.position)
                     val bytevalue = NumericLiteral(DataType.UBYTE, (rightNum.number.toInt() shr 8).toDouble(), booleanCondition.right.position)
                     return listOf(
                         IAstModification.ReplaceNode(booleanCondition.left, msb, booleanCondition),
                         IAstModification.ReplaceNode(booleanCondition.right, bytevalue, booleanCondition))
                 }
+                else if ((rightNum.number.toInt() and 0xff00) == 0) {
+                    // if WORD & $00ff  ->  if lsb(WORD) & $ff
+                    val lsb = BuiltinFunctionCall(IdentifierReference(listOf("lsb"), booleanCondition.left.position), mutableListOf(booleanCondition.left), booleanCondition.left.position)
+                    val bytevalue = NumericLiteral(DataType.UBYTE, rightNum.number, booleanCondition.right.position)
+                    return listOf(
+                        IAstModification.ReplaceNode(booleanCondition.left, lsb, booleanCondition),
+                        IAstModification.ReplaceNode(booleanCondition.right, bytevalue, booleanCondition))
+                }
             }
         }
 
@@ -97,6 +102,10 @@ class ExpressionSimplifier(private val program: Program,
     }
 
     override fun after(expr: BinaryExpression, parent: Node): Iterable<IAstModification> {
+        val newExpr = applyAbsorptionLaws(expr)
+        if(newExpr!=null)
+            return listOf(IAstModification.ReplaceNode(expr, newExpr, parent))
+
         val leftVal = expr.left.constValue(program)
         val rightVal = expr.right.constValue(program)
 
@@ -158,8 +167,8 @@ class ExpressionSimplifier(private val program: Program,
                     val y = determineY(x, leftBinExpr)
                     if (y != null) {
                         val yPlus1 = BinaryExpression(y, "+", NumericLiteral(leftDt, 1.0, y.position), y.position)
-                        val newExpr = BinaryExpression(x, "*", yPlus1, x.position)
-                        return listOf(IAstModification.ReplaceNode(expr, newExpr, parent))
+                        val replacement = BinaryExpression(x, "*", yPlus1, x.position)
+                        return listOf(IAstModification.ReplaceNode(expr, replacement, parent))
                     }
                 } else {
                     // Y*X - X  ->  X*(Y - 1)
@@ -168,8 +177,8 @@ class ExpressionSimplifier(private val program: Program,
                     val y = determineY(x, leftBinExpr)
                     if (y != null) {
                         val yMinus1 = BinaryExpression(y, "-", NumericLiteral(leftDt, 1.0, y.position), y.position)
-                        val newExpr = BinaryExpression(x, "*", yMinus1, x.position)
-                        return listOf(IAstModification.ReplaceNode(expr, newExpr, parent))
+                        val replacement = BinaryExpression(x, "*", yMinus1, x.position)
+                        return listOf(IAstModification.ReplaceNode(expr, replacement, parent))
                     }
                 }
             } else if (rightBinExpr?.operator == "*") {
@@ -180,8 +189,8 @@ class ExpressionSimplifier(private val program: Program,
                     val y = determineY(x, rightBinExpr)
                     if (y != null) {
                         val yPlus1 = BinaryExpression(y, "+", NumericLiteral.optimalInteger(1, y.position), y.position)
-                        val newExpr = BinaryExpression(x, "*", yPlus1, x.position)
-                        return listOf(IAstModification.ReplaceNode(expr, newExpr, parent))
+                        val replacement = BinaryExpression(x, "*", yPlus1, x.position)
+                        return listOf(IAstModification.ReplaceNode(expr, replacement, parent))
                     }
                 }
             }
@@ -270,7 +279,7 @@ class ExpressionSimplifier(private val program: Program,
 
         // simplify when a term is constant and directly determines the outcome
         val constFalse = NumericLiteral.fromBoolean(false, expr.position)
-        val newExpr: Expression? = when (expr.operator) {
+        val newExpr2 = when (expr.operator) {
             "|" -> {
                 when {
                     leftVal?.number==0.0 -> expr.right
@@ -315,33 +324,102 @@ class ExpressionSimplifier(private val program: Program,
         }
 
         if(rightVal!=null && leftDt==DataType.BOOL) {
-            // see if we can replace comparison against true/1 with simpler comparison against zero
-            if (expr.operator == "==") {
-                if (rightVal.number == 1.0) {
-                    val zero = NumericLiteral(DataType.UBYTE, 0.0, expr.right.position)
-                    return listOf(
-                        IAstModification.SetExpression({expr.operator="!="}, expr, parent),
-                        IAstModification.ReplaceNode(expr.right, zero, expr)
-                    )
-                }
-            }
-            if (expr.operator == "!=") {
-                if (rightVal.number == 1.0) {
-                    val zero = NumericLiteral(DataType.UBYTE, 0.0, expr.right.position)
-                    return listOf(
-                        IAstModification.SetExpression({expr.operator="=="}, expr, parent),
-                        IAstModification.ReplaceNode(expr.right, zero, expr)
-                    )
+            // boolean compare against a number -> just keep the boolean, no compare
+            if(expr.operator=="==" || expr.operator=="!=") {
+                val test = if (expr.operator == "==") rightVal.asBooleanValue else !rightVal.asBooleanValue
+                return if (test) {
+                    listOf(IAstModification.ReplaceNode(expr, expr.left, parent))
+                } else {
+                    listOf(IAstModification.ReplaceNode(expr, invertCondition(expr.left, program), parent))
                 }
             }
         }
 
-        if(newExpr != null)
-            return listOf(IAstModification.ReplaceNode(expr, newExpr, parent))
+        if(newExpr2 != null)
+            return listOf(IAstModification.ReplaceNode(expr, newExpr2, parent))
+
+        if (rightVal!=null && (expr.operator == "==" || expr.operator == "!=")) {
+            val bitwise = expr.left as? BinaryExpression
+            if(bitwise!=null && bitwise.operator=="&" && bitwise.inferType(program).isWords) {
+                val andNum = (bitwise.right as? NumericLiteral)?.number?.toInt()
+                if (andNum!=null) {
+                    if ((andNum and 0x00ff) == 0) {
+                        // (WORD & $xx00)==y  ->  (msb(WORD) & $xx)==y
+                        val msb = BuiltinFunctionCall(IdentifierReference(listOf("msb"), bitwise.left.position), mutableListOf(bitwise.left), bitwise.left.position)
+                        val bytevalue = NumericLiteral(DataType.UBYTE, (andNum shr 8).toDouble(), bitwise.right.position)
+                        val rightvalByte = NumericLiteral(DataType.UBYTE, (rightVal.number.toInt() shr 8).toDouble(), rightVal.position)
+                        return listOf(
+                            IAstModification.ReplaceNode(bitwise.left, msb, bitwise),
+                            IAstModification.ReplaceNode(bitwise.right, bytevalue, bitwise),
+                            IAstModification.ReplaceNode(expr.right, rightvalByte, expr)
+                        )
+                    }
+                    else if((andNum and 0xff00) == 0) {
+                        // (WORD & $00xx)==y  ->  (lsb(WORD) & $xx)==y
+                        val lsb = BuiltinFunctionCall(IdentifierReference(listOf("lsb"), bitwise.left.position), mutableListOf(bitwise.left), bitwise.left.position)
+                        val bytevalue = NumericLiteral(DataType.UBYTE, andNum.toDouble(), bitwise.right.position)
+                        val rightvalByte = NumericLiteral(DataType.UBYTE, (rightVal.number.toInt() and 255).toDouble(), rightVal.position)
+                        return listOf(
+                            IAstModification.ReplaceNode(bitwise.left, lsb, bitwise),
+                            IAstModification.ReplaceNode(bitwise.right, bytevalue, bitwise),
+                            IAstModification.ReplaceNode(expr.right, rightvalByte, expr)
+                        )
+                    }
+                }
+            }
+        }
 
         return noModifications
     }
 
+    override fun after(expr: PrefixExpression, parent: Node): Iterable<IAstModification> {
+        if(expr.operator=="not") {
+            // not  X <compare> Y  ->   X <invertedcompare> Y
+            val binExpr = expr.expression as? BinaryExpression
+            if(binExpr!=null) {
+                val invertedOperator = invertedComparisonOperator(binExpr.operator)
+                if(invertedOperator!=null) {
+                    val inverted = BinaryExpression(binExpr.left, invertedOperator, binExpr.right, binExpr.position)
+                    return listOf(IAstModification.ReplaceNode(expr, inverted, parent))
+                }
+            }
+        }
+        return noModifications
+    }
+
+
+    private fun applyAbsorptionLaws(expr: BinaryExpression): Expression? {
+        val rightB = expr.right as? BinaryExpression
+        if(rightB!=null) {
+            // absorption laws:  a or (a and b) --> a,  a and (a or b) --> a
+            if(expr.operator=="or" && rightB.operator=="and") {
+                if(expr.left isSameAs rightB.left || expr.left isSameAs rightB.right) {
+                    return expr.left
+                }
+            }
+            else if(expr.operator=="and" && rightB.operator=="or") {
+                if(expr.left isSameAs rightB.left || expr.left isSameAs rightB.right) {
+                    return expr.left
+                }
+            }
+        }
+        val leftB = expr.left as? BinaryExpression
+        if(leftB!=null) {
+            // absorption laws:  (a and b) or a --> a,  (a or b) and a --> a
+            if(expr.operator=="or" && leftB.operator=="and") {
+                if(expr.right isSameAs leftB.left || expr.left isSameAs leftB.right) {
+                    return expr.right
+                }
+            }
+            else if(expr.operator=="and" && leftB.operator=="or") {
+                if(expr.right isSameAs leftB.left || expr.right isSameAs leftB.right) {
+                    return expr.right
+                }
+            }
+        }
+        return null
+    }
+
     override fun after(functionCallExpr: FunctionCallExpression, parent: Node): Iterable<IAstModification> {
         if(functionCallExpr.target.nameInSource == listOf("lsb")) {
             if(functionCallExpr.args.isEmpty())
@@ -349,10 +427,15 @@ class ExpressionSimplifier(private val program: Program,
             val arg = functionCallExpr.args[0]
             if(arg is TypecastExpression) {
                 val valueDt = arg.expression.inferType(program)
-                if (valueDt istype DataType.BYTE || valueDt istype DataType.UBYTE) {
-                    // useless lsb() of byte value that was typecasted to word
+                if (valueDt istype DataType.UBYTE) {
+                    // useless lsb() of ubyte value
                     return listOf(IAstModification.ReplaceNode(functionCallExpr, arg.expression, parent))
                 }
+                else if (valueDt istype DataType.BYTE) {
+                    // useless lsb() of byte value, but as lsb() returns unsigned, we have to cast now.
+                    val cast = TypecastExpression(arg.expression, DataType.UBYTE, true, arg.position)
+                    return listOf(IAstModification.ReplaceNode(functionCallExpr, cast, parent))
+                }
             } else {
                 if(arg is IdentifierReference && arg.nameInSource.size==2
                     && arg.nameInSource[0]=="cx16" && arg.nameInSource[1].uppercase() in RegisterOrPair.names) {
@@ -361,10 +444,15 @@ class ExpressionSimplifier(private val program: Program,
                     return listOf(IAstModification.ReplaceNode(functionCallExpr, highReg, parent))
                 }
                 val argDt = arg.inferType(program)
-                if (argDt istype DataType.BYTE || argDt istype DataType.UBYTE) {
+                if (argDt istype DataType.UBYTE) {
                     // useless lsb() of byte value
                     return listOf(IAstModification.ReplaceNode(functionCallExpr, arg, parent))
                 }
+                else if (argDt istype DataType.BYTE) {
+                    // useless lsb() of byte value, but as lsb() returns unsigned, we have to cast now.
+                    val cast = TypecastExpression(arg, DataType.UBYTE, true, arg.position)
+                    return listOf(IAstModification.ReplaceNode(functionCallExpr, cast, parent))
+                }
             }
         }
         else if(functionCallExpr.target.nameInSource == listOf("msb")) {
@@ -405,6 +493,14 @@ class ExpressionSimplifier(private val program: Program,
                 return listOf(IAstModification.ReplaceNode(functionCallExpr, cast, parent))
             }
         }
+        else if(functionCallExpr.target.nameInSource == listOf("string", "contains")) {
+            val target = (functionCallExpr.args[0] as? IdentifierReference)?.targetVarDecl(program)
+            if(target?.value is StringLiteral) {
+                errors.info("for actual strings, use a regular containment check instead: 'char in string'", functionCallExpr.position)
+                val contains = ContainmentCheck(functionCallExpr.args[1], functionCallExpr.args[0], functionCallExpr.position)
+                return listOf(IAstModification.ReplaceNode(functionCallExpr as Node, contains, parent))
+            }
+        }
 
         return noModifications
     }
@@ -540,10 +636,25 @@ class ExpressionSimplifier(private val program: Program,
                         return expr.left
                     }
                 }
+                256.0 -> {
+                    when(leftDt) {
+                        DataType.UBYTE -> return NumericLiteral(DataType.UBYTE, 0.0, expr.position)
+                        DataType.BYTE -> return null        // is either 0 or -1 we cannot tell here
+                        DataType.UWORD, DataType.WORD -> {
+                            // just use:  msb(value) as type
+                            val msb = BuiltinFunctionCall(IdentifierReference(listOf("msb"), expr.position), mutableListOf(expr.left), expr.position)
+                            return if(leftDt==DataType.WORD)
+                                TypecastExpression(msb, DataType.BYTE, true, expr.position)
+                            else
+                                TypecastExpression(msb, DataType.UWORD, true, expr.position)
+                        }
+                        else -> return null
+                    }
+                }
                 in powersOfTwo -> {
                     if (leftDt==DataType.UBYTE || leftDt==DataType.UWORD) {
                         // Unsigned number divided by a power of two => shift right
-                        // Signed number can't simply be bitshifted in this case (due to rounding issues for negative values),
+                        // Signed number can't simply be bitshifted in this case (due to rounding issues for negative values),  TODO is this correct???
                         // so we leave that as is and let the code generator deal with it.
                         val numshifts = log2(cv).toInt()
                         return BinaryExpression(expr.left, ">>", NumericLiteral.optimalInteger(numshifts, expr.position), expr.position)
@@ -705,8 +816,8 @@ class ExpressionSimplifier(private val program: Program,
             }
             DataType.UWORD -> {
                 if (amount >= 16) {
-                    errors.warn("shift always results in 0", expr.position)
-                    return NumericLiteral.optimalInteger(0, expr.position)
+                    errors.err("useless to shift by more than 15 bits", expr.position)
+                    return null
                 }
                 else if(amount==8) {
                     // shift right by 8 bits is just a byte operation: msb(X) as uword
@@ -720,11 +831,21 @@ class ExpressionSimplifier(private val program: Program,
                 }
             }
             DataType.WORD -> {
-                // bit-shifting a signed value shouldn't be allowed by the compiler but here we go...
-                if (amount > 16) {
-                    expr.right = NumericLiteral.optimalInteger(16, expr.right.position)
+                if (amount >= 16) {
+                    errors.err("useless to shift by more than 15 bits", expr.position)
                     return null
                 }
+                else if(amount == 8) {
+                    // shift right by 8 bits is just a byte operation: msb(X) as byte  (will get converted to word later)
+                    val msb = FunctionCallExpression(IdentifierReference(listOf("msb"), expr.position), mutableListOf(expr.left), expr.position)
+                    return TypecastExpression(msb, DataType.BYTE, true, expr.position)
+                }
+                else if(amount > 8) {
+                    // same as above but with residual shifts. Take care to do signed shift.
+                    val msb = FunctionCallExpression(IdentifierReference(listOf("msb"), expr.position), mutableListOf(expr.left), expr.position)
+                    val signed = TypecastExpression(msb, DataType.BYTE, true, expr.position)
+                    return BinaryExpression(signed, ">>", NumericLiteral.optimalInteger(amount - 8, expr.position), expr.position)
+                }
             }
             else -> {
             }

codeOptimizers/src/prog8/optimizer/Extensions.kt

@@ -7,13 +7,13 @@ import prog8.code.core.ICompilationTarget
 import prog8.code.core.IErrorReporter
 
 
-fun Program.constantFold(errors: IErrorReporter, compTarget: ICompilationTarget) {
-    val valuetypefixer = VarConstantValueTypeAdjuster(this, errors)
+fun Program.constantFold(errors: IErrorReporter, options: CompilationOptions) {
+    val valuetypefixer = VarConstantValueTypeAdjuster(this, options, errors)
     valuetypefixer.visit(this)
     if(errors.noErrors()) {
         valuetypefixer.applyModifications()
 
-        val replacer = ConstantIdentifierReplacer(this, errors, compTarget)
+        val replacer = ConstantIdentifierReplacer(this, errors, options.compTarget)
         replacer.visit(this)
         if (errors.noErrors()) {
             replacer.applyModifications()
@@ -22,7 +22,7 @@ fun Program.constantFold(errors: IErrorReporter, compTarget: ICompilationTarget)
             if(errors.noErrors()) {
                 valuetypefixer.applyModifications()
 
-                val optimizer = ConstantFoldingOptimizer(this)
+                val optimizer = ConstantFoldingOptimizer(this, errors)
                 optimizer.visit(this)
                 while (errors.noErrors() && optimizer.applyModifications() > 0) {
                     optimizer.visit(this)

codeOptimizers/src/prog8/optimizer/Inliner.kt

@@ -34,85 +34,100 @@ class Inliner(private val program: Program, private val options: CompilationOpti
                 val containsSubsOrVariables = subroutine.statements.any { it is VarDecl || it is Subroutine}
                 if(!containsSubsOrVariables) {
                     if(subroutine.statements.size==1 || (subroutine.statements.size==2 && isEmptyReturn(subroutine.statements[1]))) {
-                        // subroutine is possible candidate to be inlined
-                        subroutine.inline =
-                            when(val stmt=subroutine.statements[0]) {
-                                is Return -> {
-                                    if(stmt.value is NumericLiteral)
-                                        true
-                                    else if(stmt.value==null)
-                                        true
-                                    else if (stmt.value is IdentifierReference) {
-                                        makeFullyScoped(stmt.value as IdentifierReference)
-                                        true
-                                    } else if(stmt.value!! is IFunctionCall && (stmt.value as IFunctionCall).args.size<=1 && (stmt.value as IFunctionCall).args.all { it is NumericLiteral || it is IdentifierReference }) {
-                                        when (stmt.value) {
-                                            is BuiltinFunctionCall -> {
-                                                makeFullyScoped(stmt.value as BuiltinFunctionCall)
-                                                true
+                        if(subroutine !== program.entrypoint) {
+                            // subroutine is possible candidate to be inlined
+                            subroutine.inline =
+                                when (val stmt = subroutine.statements[0]) {
+                                    is Return -> {
+                                        if (stmt.value is NumericLiteral)
+                                            true
+                                        else if (stmt.value == null)
+                                            true
+                                        else if (stmt.value is IdentifierReference) {
+                                            makeFullyScoped(stmt.value as IdentifierReference)
+                                            true
+                                        } else if (stmt.value!! is IFunctionCall && (stmt.value as IFunctionCall).args.size <= 1 && (stmt.value as IFunctionCall).args.all { it is NumericLiteral || it is IdentifierReference }) {
+                                            when (stmt.value) {
+                                                is BuiltinFunctionCall -> {
+                                                    makeFullyScoped(stmt.value as BuiltinFunctionCall)
+                                                    true
+                                                }
+
+                                                is FunctionCallExpression -> {
+                                                    makeFullyScoped(stmt.value as FunctionCallExpression)
+                                                    true
+                                                }
+
+                                                else -> false
                                             }
-                                            is FunctionCallExpression -> {
-                                                makeFullyScoped(stmt.value as FunctionCallExpression)
-                                                true
-                                            }
-                                            else -> false
-                                        }
-                                    } else
-                                        false
-                                }
-                                is Assignment -> {
-                                    if(stmt.value.isSimple) {
-                                        val targetInline =
-                                            if(stmt.target.identifier!=null) {
-                                                makeFullyScoped(stmt.target.identifier!!)
-                                                true
-                                            } else if(stmt.target.memoryAddress?.addressExpression is NumericLiteral || stmt.target.memoryAddress?.addressExpression is IdentifierReference) {
-                                                if(stmt.target.memoryAddress?.addressExpression is IdentifierReference)
-                                                    makeFullyScoped(stmt.target.memoryAddress?.addressExpression as IdentifierReference)
-                                                true
-                                            } else
-                                                false
-                                        val valueInline =
-                                            if(stmt.value is IdentifierReference) {
-                                                makeFullyScoped(stmt.value as IdentifierReference)
-                                                true
-                                            } else if((stmt.value as? DirectMemoryRead)?.addressExpression is NumericLiteral || (stmt.value as? DirectMemoryRead)?.addressExpression is IdentifierReference) {
-                                                if((stmt.value as? DirectMemoryRead)?.addressExpression is IdentifierReference)
-                                                    makeFullyScoped((stmt.value as? DirectMemoryRead)?.addressExpression as IdentifierReference)
-                                                true
-                                            } else
-                                                false
-                                        targetInline || valueInline
-                                    } else
-                                        false
-                                }
-                                is BuiltinFunctionCallStatement -> {
-                                    val inline = stmt.args.size<=1 && stmt.args.all { it is NumericLiteral || it is IdentifierReference }
-                                    if(inline)
-                                        makeFullyScoped(stmt)
-                                    inline
-                                }
-                                is FunctionCallStatement -> {
-                                    val inline = stmt.args.size<=1 && stmt.args.all { it is NumericLiteral || it is IdentifierReference }
-                                    if(inline)
-                                        makeFullyScoped(stmt)
-                                    inline
-                                }
-                                is PostIncrDecr -> {
-                                    if(stmt.target.identifier!=null) {
-                                        makeFullyScoped(stmt.target.identifier!!)
-                                        true
+                                        } else
+                                            false
                                     }
-                                    else if(stmt.target.memoryAddress?.addressExpression is NumericLiteral || stmt.target.memoryAddress?.addressExpression is IdentifierReference) {
-                                        if(stmt.target.memoryAddress?.addressExpression is IdentifierReference)
-                                            makeFullyScoped(stmt.target.memoryAddress?.addressExpression as IdentifierReference)
-                                        true
-                                    } else
-                                        false
+
+                                    is Assignment -> {
+                                        if (stmt.value.isSimple) {
+                                            val targetInline =
+                                                if (stmt.target.identifier != null) {
+                                                    makeFullyScoped(stmt.target.identifier!!)
+                                                    true
+                                                } else if (stmt.target.memoryAddress?.addressExpression is NumericLiteral || stmt.target.memoryAddress?.addressExpression is IdentifierReference) {
+                                                    if (stmt.target.memoryAddress?.addressExpression is IdentifierReference)
+                                                        makeFullyScoped(stmt.target.memoryAddress?.addressExpression as IdentifierReference)
+                                                    true
+                                                } else
+                                                    false
+                                            val valueInline =
+                                                if (stmt.value is IdentifierReference) {
+                                                    makeFullyScoped(stmt.value as IdentifierReference)
+                                                    true
+                                                } else if ((stmt.value as? DirectMemoryRead)?.addressExpression is NumericLiteral || (stmt.value as? DirectMemoryRead)?.addressExpression is IdentifierReference) {
+                                                    if ((stmt.value as? DirectMemoryRead)?.addressExpression is IdentifierReference)
+                                                        makeFullyScoped((stmt.value as? DirectMemoryRead)?.addressExpression as IdentifierReference)
+                                                    true
+                                                } else
+                                                    false
+                                            targetInline || valueInline
+                                        } else
+                                            false
+                                    }
+
+                                    is BuiltinFunctionCallStatement -> {
+                                        val inline =
+                                            stmt.args.size <= 1 && stmt.args.all { it is NumericLiteral || it is IdentifierReference }
+                                        if (inline)
+                                            makeFullyScoped(stmt)
+                                        inline
+                                    }
+
+                                    is FunctionCallStatement -> {
+                                        val inline =
+                                            stmt.args.size <= 1 && stmt.args.all { it is NumericLiteral || it is IdentifierReference }
+                                        if (inline)
+                                            makeFullyScoped(stmt)
+                                        inline
+                                    }
+
+                                    is PostIncrDecr -> {
+                                        if (stmt.target.identifier != null) {
+                                            makeFullyScoped(stmt.target.identifier!!)
+                                            true
+                                        } else if (stmt.target.memoryAddress?.addressExpression is NumericLiteral || stmt.target.memoryAddress?.addressExpression is IdentifierReference) {
+                                            if (stmt.target.memoryAddress?.addressExpression is IdentifierReference)
+                                                makeFullyScoped(stmt.target.memoryAddress?.addressExpression as IdentifierReference)
+                                            true
+                                        } else
+                                            false
+                                    }
+
+                                    is Jump -> true
+                                    else -> false
                                 }
-                                is Jump -> true
-                                else -> false
-                            }
+                        }
+                    }
+
+                    if(subroutine.inline && subroutine.statements.size>1) {
+                        require(subroutine.statements.size==2 && isEmptyReturn(subroutine.statements[1]))
+                        subroutine.statements.removeLast()      // get rid of the Return, to be able to inline the (single) statement preceding it.
                     }
                 }
             }
@@ -129,16 +144,20 @@ class Inliner(private val program: Program, private val options: CompilationOpti
 
         private fun makeFullyScoped(call: BuiltinFunctionCallStatement) {
             val scopedArgs = makeScopedArgs(call.args)
-            val scopedCall = BuiltinFunctionCallStatement(call.target.copy(), scopedArgs.toMutableList(), call.position)
-            modifications += IAstModification.ReplaceNode(call, scopedCall, call.parent)
+            if(scopedArgs.any()) {
+                val scopedCall = BuiltinFunctionCallStatement(call.target.copy(), scopedArgs.toMutableList(), call.position)
+                modifications += IAstModification.ReplaceNode(call, scopedCall, call.parent)
+            }
         }
 
         private fun makeFullyScoped(call: FunctionCallStatement) {
             call.target.targetSubroutine(program)?.let { sub ->
                 val scopedName = IdentifierReference(sub.scopedName, call.target.position)
                 val scopedArgs = makeScopedArgs(call.args)
-                val scopedCall = FunctionCallStatement(scopedName, scopedArgs.toMutableList(), call.void, call.position)
-                modifications += IAstModification.ReplaceNode(call, scopedCall, call.parent)
+                if(scopedArgs.any()) {
+                    val scopedCall = FunctionCallStatement(scopedName, scopedArgs.toMutableList(), call.void, call.position)
+                    modifications += IAstModification.ReplaceNode(call, scopedCall, call.parent)
+                }
             }
         }
 
@@ -146,8 +165,10 @@ class Inliner(private val program: Program, private val options: CompilationOpti
             call.target.targetSubroutine(program)?.let { sub ->
                 val scopedName = IdentifierReference(sub.scopedName, call.target.position)
                 val scopedArgs = makeScopedArgs(call.args)
-                val scopedCall = BuiltinFunctionCall(scopedName, scopedArgs.toMutableList(), call.position)
-                modifications += IAstModification.ReplaceNode(call, scopedCall, call.parent)
+                if(scopedArgs.any()) {
+                    val scopedCall = BuiltinFunctionCall(scopedName, scopedArgs.toMutableList(), call.position)
+                    modifications += IAstModification.ReplaceNode(call, scopedCall, call.parent)
+                }
             }
         }
 
@@ -155,8 +176,10 @@ class Inliner(private val program: Program, private val options: CompilationOpti
             call.target.targetSubroutine(program)?.let { sub ->
                 val scopedName = IdentifierReference(sub.scopedName, call.target.position)
                 val scopedArgs = makeScopedArgs(call.args)
-                val scopedCall = FunctionCallExpression(scopedName, scopedArgs.toMutableList(), call.position)
-                modifications += IAstModification.ReplaceNode(call, scopedCall, call.parent)
+                if(scopedArgs.any()) {
+                    val scopedCall = FunctionCallExpression(scopedName, scopedArgs.toMutableList(), call.position)
+                    modifications += IAstModification.ReplaceNode(call, scopedCall, call.parent)
+                }
             }
         }
 
@@ -165,7 +188,8 @@ class Inliner(private val program: Program, private val options: CompilationOpti
                 when (it) {
                     is NumericLiteral -> it.copy()
                     is IdentifierReference -> {
-                        val scoped = (it.targetStatement(program)!! as INamedStatement).scopedName
+                        val target = it.targetStatement(program) ?: return emptyList()
+                        val scoped = (target as INamedStatement).scopedName
                         IdentifierReference(scoped, it.position)
                     }
                     else -> throw InternalCompilerException("expected only number or identifier arg, otherwise too complex")

codeOptimizers/src/prog8/optimizer/StatementOptimizer.kt

@@ -45,11 +45,7 @@ class StatementOptimizer(private val program: Program,
                             mutableListOf(NumericLiteral(DataType.UBYTE, firstCharEncoded.toDouble(), pos)),
                             functionCallStatement.void, pos
                         )
-                        val stringDecl = string.parent as VarDecl
-                        return listOf(
-                            IAstModification.ReplaceNode(functionCallStatement, chrout, parent),
-                            IAstModification.Remove(stringDecl, stringDecl.parent as IStatementContainer)
-                        )
+                        return listOf(IAstModification.ReplaceNode(functionCallStatement, chrout, parent))
                     } else if (string.value.length == 2) {
                         val firstTwoCharsEncoded = options.compTarget.encodeString(string.value.take(2), string.encoding)
                         val chrout1 = FunctionCallStatement(
@@ -62,11 +58,9 @@ class StatementOptimizer(private val program: Program,
                             mutableListOf(NumericLiteral(DataType.UBYTE, firstTwoCharsEncoded[1].toDouble(), pos)),
                             functionCallStatement.void, pos
                         )
-                        val stringDecl = string.parent as VarDecl
                         return listOf(
                             IAstModification.InsertBefore(functionCallStatement, chrout1, parent as IStatementContainer),
-                            IAstModification.ReplaceNode(functionCallStatement, chrout2, parent),
-                            IAstModification.Remove(stringDecl, stringDecl.parent as IStatementContainer)
+                            IAstModification.ReplaceNode(functionCallStatement, chrout2, parent)
                         )
                     }
                 }
@@ -77,6 +71,11 @@ class StatementOptimizer(private val program: Program,
     }
 
     override fun after(ifElse: IfElse, parent: Node): Iterable<IAstModification> {
+        val constvalue = ifElse.condition.constValue(program)
+        if(constvalue!=null) {
+            errors.warn("condition is always ${constvalue.asBooleanValue}", ifElse.condition.position)
+        }
+
         // remove empty if statements
         if(ifElse.truepart.isEmpty() && ifElse.elsepart.isEmpty())
             return listOf(IAstModification.Remove(ifElse, parent as IStatementContainer))
@@ -93,17 +92,12 @@ class StatementOptimizer(private val program: Program,
             )
         }
 
-        val constvalue = ifElse.condition.constValue(program)
         if(constvalue!=null) {
             return if(constvalue.asBooleanValue){
                 // always true -> keep only if-part
-                if(!ifElse.definingModule.isLibrary)
-                    errors.warn("condition is always true", ifElse.condition.position)
                 listOf(IAstModification.ReplaceNode(ifElse, ifElse.truepart, parent))
             } else {
                 // always false -> keep only else-part
-                if(!ifElse.definingModule.isLibrary)
-                    errors.warn("condition is always false", ifElse.condition.position)
                 listOf(IAstModification.ReplaceNode(ifElse, ifElse.elsepart, parent))
             }
         }
@@ -116,12 +110,27 @@ class StatementOptimizer(private val program: Program,
                 IAstModification.InsertAfter(ifElse, elsePart, parent as IStatementContainer)
             )
         }
+
+        // switch if/else around if the else is just a jump or branch
+        if(ifElse.elsepart.isNotEmpty() && ifElse.elsepart.statements.size==1) {
+            val jump = ifElse.elsepart.statements[0]
+            if(jump is Jump) {
+                val newTruePart = AnonymousScope(mutableListOf(jump), ifElse.elsepart.position)
+                val newElsePart = AnonymousScope(ifElse.truepart.statements, ifElse.truepart.position)
+                return listOf(
+                    IAstModification.ReplaceNode(ifElse.elsepart, newElsePart, ifElse),
+                    IAstModification.ReplaceNode(ifElse.truepart, newTruePart, ifElse),
+                    IAstModification.ReplaceNode(ifElse.condition, invertCondition(ifElse.condition, program), ifElse)
+                )
+            }
+        }
+
         return noModifications
     }
 
     override fun after(forLoop: ForLoop, parent: Node): Iterable<IAstModification> {
         if(forLoop.body.isEmpty()) {
-            errors.warn("removing empty for loop", forLoop.position)
+            errors.info("removing empty for loop", forLoop.position)
             return listOf(IAstModification.Remove(forLoop, parent as IStatementContainer))
         } else if(forLoop.body.statements.size==1) {
             val loopvar = forLoop.body.statements[0] as? VarDecl
@@ -157,7 +166,7 @@ class StatementOptimizer(private val program: Program,
                     return listOf(IAstModification.ReplaceNode(forLoop, scope, parent))
                 }
             }
-            else if(iterable.datatype in ArrayDatatypes) {
+            else if(iterable.isArray) {
                 val size = iterable.arraysize!!.constIndex()
                 if(size==1) {
                     // loop over array of length 1 -> just assign the single value
@@ -174,6 +183,70 @@ class StatementOptimizer(private val program: Program,
             }
         }
 
+        val loopvarDt = forLoop.loopVarDt(program)
+        if(loopvarDt.istype(DataType.UWORD) || loopvarDt.istype(DataType.UBYTE)) {
+            if (range != null && range.from.constValue(program)?.number == 0.0 && range.step.constValue(program)?.number==1.0) {
+                val toBinExpr = range.to as? BinaryExpression
+                if(toBinExpr!=null && toBinExpr.operator=="-" && toBinExpr.right.constValue(program)?.number==1.0) {
+                    // FOR var IN 0 TO X-1 .... ---> var=0, DO {... , var++} UNTIL var==X
+                    val pos = forLoop.position
+                    val condition = BinaryExpression(forLoop.loopVar.copy(), "==", toBinExpr.left, pos)
+                    val incOne = PostIncrDecr(AssignTarget(forLoop.loopVar.copy(), null, null, pos), "++", pos)
+                    forLoop.body.statements.add(incOne)
+                    val replacement = AnonymousScope(mutableListOf(
+                        Assignment(AssignTarget(forLoop.loopVar.copy(), null, null, pos),
+                            NumericLiteral.optimalNumeric(0.0, pos),
+                            AssignmentOrigin.OPTIMIZER, pos),
+                        UntilLoop(forLoop.body, condition, pos)
+                    ), pos)
+                    return listOf(IAstModification.ReplaceNode(forLoop, replacement, parent))
+                }
+
+                if(options.compTarget.name!=VMTarget.NAME) {
+                    // this optimization is not effective for the VM target.
+                    val toConst = range.to.constValue(program)
+                    if (toConst == null) {
+                        // FOR var in 0 TO X ... --->  var=0, REPEAT { ... , IF var==X break , var++ }
+                        val pos = forLoop.position
+                        val incOne = PostIncrDecr(AssignTarget(forLoop.loopVar.copy(), null, null, pos), "++", pos)
+                        val breakCondition = IfElse(
+                            BinaryExpression(forLoop.loopVar, "==", range.to, pos),
+                            AnonymousScope(mutableListOf(Break(pos)), pos),
+                            AnonymousScope(mutableListOf(), pos),
+                            pos
+                        )
+                        forLoop.body.statements.add(breakCondition)
+                        forLoop.body.statements.add(incOne)
+                        val replacement = AnonymousScope(mutableListOf(
+                            Assignment(AssignTarget(forLoop.loopVar.copy(), null, null, pos),
+                                NumericLiteral.optimalNumeric(0.0, pos),
+                                AssignmentOrigin.OPTIMIZER, pos),
+                            RepeatLoop(null, forLoop.body, pos)
+                        ), pos)
+                        return listOf(IAstModification.ReplaceNode(forLoop, replacement, parent))
+                    }
+                }
+            }
+
+            if (range != null && range.to.constValue(program)?.number == 0.0 && range.step.constValue(program)?.number==-1.0) {
+                val fromExpr = range.from
+                if(fromExpr.constValue(program)==null) {
+                    // FOR X = something DOWNTO 0 {...} -->  X=something,  DO { ... , X-- } UNTIL X=255 (or 65535 if uword)
+                    val pos = forLoop.position
+                    val checkValue = NumericLiteral(loopvarDt.getOr(DataType.UNDEFINED), if(loopvarDt.istype(DataType.UBYTE)) 255.0 else 65535.0, pos)
+                    val condition = BinaryExpression(forLoop.loopVar.copy(), "==", checkValue, pos)
+                    val decOne = PostIncrDecr(AssignTarget(forLoop.loopVar.copy(), null, null, pos), "--", pos)
+                    forLoop.body.statements.add(decOne)
+                    val replacement = AnonymousScope(mutableListOf(
+                        Assignment(AssignTarget(forLoop.loopVar.copy(), null, null, pos),
+                            fromExpr, AssignmentOrigin.OPTIMIZER, pos),
+                        UntilLoop(forLoop.body, condition, pos)
+                    ), pos)
+                    return listOf(IAstModification.ReplaceNode(forLoop, replacement, parent))
+                }
+            }
+        }
+
         return noModifications
     }
 
@@ -213,7 +286,7 @@ class StatementOptimizer(private val program: Program,
         val iter = repeatLoop.iterations
         if(iter!=null) {
             if(repeatLoop.body.isEmpty()) {
-                errors.warn("empty loop removed", repeatLoop.position)
+                errors.info("empty loop removed", repeatLoop.position)
                 return listOf(IAstModification.Remove(repeatLoop, parent as IStatementContainer))
             }
             val iterations = iter.constValue(program)?.number?.toInt()
@@ -395,14 +468,14 @@ class StatementOptimizer(private val program: Program,
 
         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)
+            errors.info("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.
+                    // 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)
@@ -430,6 +503,27 @@ class StatementOptimizer(private val program: Program,
                 }
             }
         }
+
+        val constantValue = whenStmt.condition.constValue(program)?.number
+        if(constantValue!=null) {
+            // when condition is a constant
+            var matchingChoice: WhenChoice? = null
+            loop@ for(choice in whenStmt.choices) {
+                for(value in choice.values ?: emptyList()) {
+                    if(value.constValue(program)?.number == constantValue) {
+                        matchingChoice = choice
+                        break@loop
+                    }
+                }
+            }
+            if(matchingChoice==null)
+                matchingChoice = whenStmt.choices.singleOrNull { it.values==null }
+            if(matchingChoice!=null) {
+                // get rid of the whole when-statement and just leave the matching choice
+                return listOf(IAstModification.ReplaceNode(whenStmt, matchingChoice.statements, parent))
+            }
+        }
+
         return noModifications
     }
 

codeOptimizers/src/prog8/optimizer/UnusedCodeRemover.kt

@@ -66,14 +66,14 @@ class UnusedCodeRemover(private val program: Program,
             if (block.containsNoCodeNorVars) {
                 if(block.name != internedStringsModuleName) {
                     if(!block.statements.any { it is Subroutine && it.hasBeenInlined })
-                        errors.warn("removing unused block '${block.name}'", block.position)
+                        errors.info("removing unused block '${block.name}'", block.position)
                 }
                 return listOf(IAstModification.Remove(block, parent as IStatementContainer))
             }
             if(callgraph.unused(block)) {
                 if(block.statements.any{ it !is VarDecl || it.type== VarDeclType.VAR}) {
                     if(!block.statements.any { it is Subroutine && it.hasBeenInlined })
-                        errors.warn("removing unused block '${block.name}'", block.position)
+                        errors.info("removing unused block '${block.name}'", block.position)
                 }
                 if(!block.statements.any { it is Subroutine && it.hasBeenInlined }) {
                     program.removeInternedStringsFromRemovedBlock(block)
@@ -90,8 +90,8 @@ class UnusedCodeRemover(private val program: Program,
         if (subroutine !== program.entrypoint && !forceOutput && !subroutine.isAsmSubroutine) {
             if(callgraph.unused(subroutine)) {
                 if(subroutine.containsNoCodeNorVars) {
-                    if(!subroutine.definingModule.isLibrary)
-                        errors.warn("removing empty subroutine '${subroutine.name}'", subroutine.position)
+                    if("ignore_unused" !in subroutine.definingBlock.options())
+                        errors.info("removing empty subroutine '${subroutine.name}'", subroutine.position)
                     val removals = mutableListOf(IAstModification.Remove(subroutine, parent as IStatementContainer))
                     callgraph.calledBy[subroutine]?.let {
                         for(node in it)
@@ -99,8 +99,8 @@ class UnusedCodeRemover(private val program: Program,
                     }
                     return removals
                 }
-                if(!subroutine.definingModule.isLibrary && !subroutine.hasBeenInlined) {
-                    errors.warn("unused subroutine '${subroutine.name}'", subroutine.position)
+                if(!subroutine.hasBeenInlined && "ignore_unused" !in subroutine.definingBlock.options()) {
+                    errors.info("unused subroutine '${subroutine.name}'", subroutine.position)
                 }
                 if(!subroutine.inline) {
                     program.removeInternedStringsFromRemovedSubroutine(subroutine)
@@ -119,8 +119,8 @@ class UnusedCodeRemover(private val program: Program,
             if (!forceOutput && decl.origin==VarDeclOrigin.USERCODE && !decl.sharedWithAsm) {
                 val usages = callgraph.usages(decl)
                 if (usages.isEmpty()) {
-                    if(!decl.definingModule.isLibrary)
-                        errors.warn("removing unused variable '${decl.name}'", decl.position)
+                    if("ignore_unused" !in decl.definingBlock.options())
+                        errors.info("removing unused variable '${decl.name}'", decl.position)
                     return listOf(IAstModification.Remove(decl, parent as IStatementContainer))
                 }
                 else {
@@ -131,23 +131,30 @@ class UnusedCodeRemover(private val program: Program,
                             if(assignment!=null && assignment.origin==AssignmentOrigin.VARINIT) {
                                 if(assignment.value.isSimple) {
                                     // remove the vardecl
-                                    if(!decl.definingModule.isLibrary)
-                                        errors.warn("removing unused variable '${decl.name}'", decl.position)
+                                    if("ignore_unused" !in decl.definingBlock.options())
+                                        errors.info("removing unused variable '${decl.name}'", decl.position)
                                     return listOf(
                                         IAstModification.Remove(decl, parent as IStatementContainer),
                                         IAstModification.Remove(assignment, assignment.parent as IStatementContainer)
                                     )
                                 } else if(assignment.value is IFunctionCall) {
                                     // replace the unused variable's initializer function call by a void
-                                    errors.warn("replaced unused variable '${decl.name}' with void call, maybe this can be removed altogether", decl.position)
-                                    val fcall = assignment.value as IFunctionCall
-                                    val voidCall = FunctionCallStatement(fcall.target, fcall.args, true, fcall.position)
-                                    return listOf(
-                                        IAstModification.ReplaceNode(decl, voidCall, parent),
-                                        IAstModification.Remove(assignment, assignment.parent as IStatementContainer)
-                                    )
+                                    // but only if the vardecl immediately precedes it!
+                                    if(singleUse.parent.parent === parent) {
+                                        val declIndex = (parent as IStatementContainer).statements.indexOf(decl)
+                                        val singleUseIndex = (parent as IStatementContainer).statements.indexOf(singleUse.parent)
+                                        if(declIndex==singleUseIndex-1) {
+                                            errors.info("replaced unused variable '${decl.name}' with void call, maybe this can be removed altogether", decl.position)
+                                            val fcall = assignment.value as IFunctionCall
+                                            val voidCall = FunctionCallStatement(fcall.target, fcall.args, true, fcall.position)
+                                            return listOf(
+                                                IAstModification.ReplaceNode(decl, voidCall, parent),
+                                                IAstModification.Remove(assignment, assignment.parent as IStatementContainer)
+                                            )
+                                        }
+                                    }
                                 } else {
-                                    errors.warn("variable '${decl.name}' is unused but has non-trivial initialization assignment. Leaving this in but maybe it can be removed altogether", decl.position)
+                                    errors.info("variable '${decl.name}' is unused but has non-trivial initialization assignment. Leaving this in but maybe it can be removed altogether", decl.position)
                                 }
                             }
                         }

compiler/build.gradle

@@ -40,7 +40,7 @@ dependencies {
     implementation "com.michael-bull.kotlin-result:kotlin-result-jvm:1.1.18"
 
     testImplementation project(':intermediate')
-    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.6.2'
+    testImplementation 'io.kotest:kotest-runner-junit5-jvm:5.8.0'
 }
 
 configurations.all {

compiler/res/prog8lib/atari/syslib.p8

@@ -1,20 +1,26 @@
 ; Prog8 definitions for the Atari800XL
 ; Including memory registers, I/O registers, Basic and Kernal subroutines.
 
+%option no_symbol_prefixing, ignore_unused
+
 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
 
+        &uword COLCRS = 85
+        &ubyte ROWCRS = 84
+
+    romsub $F24A = getchar() -> ubyte @A
+    romsub $F2B0 = outchar(ubyte character @ A)
+    romsub $F2FD = waitkey() -> ubyte @A
+
 }
 
 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()  {
@@ -198,9 +204,57 @@ _longcopy
         }}
     }
 
-    inline asmsub exit(ubyte returnvalue @A) {
+    sub disable_caseswitch() {
+        ; no-op
+    }
+
+    sub enable_caseswitch() {
+        ; no-op
+    }
+
+    asmsub save_prog8_internals() {
+        %asm {{
+            lda  P8ZP_SCRATCH_B1
+            sta  save_SCRATCH_ZPB1
+            lda  P8ZP_SCRATCH_REG
+            sta  save_SCRATCH_ZPREG
+            lda  P8ZP_SCRATCH_W1
+            sta  save_SCRATCH_ZPWORD1
+            lda  P8ZP_SCRATCH_W1+1
+            sta  save_SCRATCH_ZPWORD1+1
+            lda  P8ZP_SCRATCH_W2
+            sta  save_SCRATCH_ZPWORD2
+            lda  P8ZP_SCRATCH_W2+1
+            sta  save_SCRATCH_ZPWORD2+1
+            rts
+save_SCRATCH_ZPB1	.byte  0
+save_SCRATCH_ZPREG	.byte  0
+save_SCRATCH_ZPWORD1	.word  0
+save_SCRATCH_ZPWORD2	.word  0
+        }}
+    }
+
+    asmsub restore_prog8_internals() {
+        %asm {{
+            lda  save_prog8_internals.save_SCRATCH_ZPB1
+            sta  P8ZP_SCRATCH_B1
+            lda  save_prog8_internals.save_SCRATCH_ZPREG
+            sta  P8ZP_SCRATCH_REG
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD1
+            sta  P8ZP_SCRATCH_W1
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD1+1
+            sta  P8ZP_SCRATCH_W1+1
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD2
+            sta  P8ZP_SCRATCH_W2
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD2+1
+            sta  P8ZP_SCRATCH_W2+1
+            rts
+        }}
+    }
+
+    asmsub exit(ubyte returnvalue @A) {
         ; -- immediately exit the program with a return code in the A register
-        ;    TODO
+        ;    TODO where to store A as exit code?
         %asm {{
             ldx  prog8_lib.orig_stackpointer
             txs
@@ -215,11 +269,37 @@ _longcopy
         }}
     }
 
+    inline asmsub push(ubyte value @A) {
+        %asm {{
+            pha
+        }}
+    }
+
+    inline asmsub pushw(uword value @AY) {
+        %asm {{
+            pha
+            tya
+            pha
+        }}
+    }
+
+    inline asmsub pop() -> ubyte @A {
+        %asm {{
+            pla
+        }}
+    }
+
+    inline asmsub popw() -> uword @AY {
+        %asm {{
+            pla
+            tay
+            pla
+        }}
+    }
+
 }
 
 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
     ; TODO
@@ -351,4 +431,9 @@ cx16 {
         }}
     }
 
+    sub cpu_is_65816() -> bool {
+        ; Returns true when you have a 65816 cpu, false when it's a 6502.
+        return false
+    }
+
 }

compiler/res/prog8lib/atari/textio.p8

@@ -1,4 +1,5 @@
 ; Prog8 definitions for the Text I/O and Screen routines for the Atari 800XL
+; Reference:  https://www.atariarchives.org/mapping/appendix12.php
 
 %import syslib
 %import conv
@@ -6,7 +7,7 @@
 
 txt {
 
-    %option no_symbol_prefixing
+    %option no_symbol_prefixing, ignore_unused
 
 const ubyte DEFAULT_WIDTH = 40
 const ubyte DEFAULT_HEIGHT = 24
@@ -24,12 +25,22 @@ sub spc() {
     txt.chrout(' ')
 }
 
-asmsub column(ubyte col @A) clobbers(A, X, Y) {
+sub column(ubyte col) {
     ; ---- set the cursor on the given column (starting with 0) on the current line
-    ;      TODO
-    %asm {{
-        rts
-    }}
+    atari.COLCRS = col
+}
+
+sub get_column() -> ubyte {
+    return atari.COLCRS as ubyte
+}
+
+sub row(ubyte rownum) {
+    ; ---- set the cursor on the given row (starting with 0) on the current line
+    atari.ROWCRS = rownum
+}
+
+sub get_row() -> ubyte {
+    return atari.ROWCRS
 }
 
 asmsub  fill_screen (ubyte character @ A, ubyte color @ Y) clobbers(A)  {
@@ -114,9 +125,6 @@ asmsub  scroll_down  (bool alsocolors @ Pc) clobbers(A)  {
 }
 
 
-romsub $F2B0 = outchar(ubyte character @ A)
-romsub $F2Fd = waitkey()
-
 asmsub chrout(ubyte character @ A) {
 	%asm {{
 		sta  _tmp_outchar+1
@@ -126,7 +134,7 @@ asmsub chrout(ubyte character @ A) {
 		pha
 _tmp_outchar
 		lda  #0
-		jsr  outchar
+		jsr  atari.outchar
 		pla
 		tay
 		pla
@@ -372,12 +380,20 @@ sub  setcc  (ubyte col, ubyte row, ubyte char, ubyte charcolor)  {
         }}
 }
 
-asmsub  plot  (ubyte col @ Y, ubyte row @ A) {
-	; ---- set cursor at specific position
-	; TODO
-	%asm  {{
-	    rts
-	}}
+sub  plot(ubyte col, ubyte rownum) {
+    column(col)
+    row(rownum)
+}
+
+sub get_cursor(uword colptr, uword rowptr) {
+    @(colptr) = get_column()
+    @(rowptr) = get_row()
+}
+
+asmsub waitkey() -> ubyte @A {
+    %asm {{
+        jmp  atari.waitkey
+    }}
 }
 
 asmsub width() clobbers(X,Y) -> ubyte @A {

compiler/res/prog8lib/c128/syslib.p8

@@ -1,9 +1,10 @@
 ; Prog8 definitions for the Commodore-128
 ; Including memory registers, I/O registers, Basic and Kernal subroutines.
 
+%option no_symbol_prefixing, ignore_unused
+
 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
@@ -78,7 +79,7 @@ romsub $FFAB = UNTLK() clobbers(A)                              ; command serial
 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 $FFB7 = READST() -> ubyte @ A                            ; read I/O status word  (use CLEARST to reset it to 0)
 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
@@ -128,12 +129,19 @@ asmsub RDTIM16() clobbers(X) -> uword @AY {
     }}
 }
 
+sub CLEARST() {
+    ; -- Set the ST status variable back to 0. (there's no direct kernal call for this)
+    ;    Note: a drive error state (blinking led) isn't cleared! You can use diskio.status() to clear that.
+    SETNAM(0, $0000)
+    SETLFS(15, 3, 15)
+    void OPEN()
+    CLOSE(15)
+}
 
 }
 
 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
@@ -288,7 +296,6 @@ c64 {
 
 c128 {
 ; ---- C128 specific registers ----
-    %option no_symbol_prefixing
 
     &ubyte  VM1     = $0A2C         ; shadow for VUC $d018 in text mode
     &ubyte  VM2     = $0A2D         ; shadow for VIC $d018 in bitmap screen mode
@@ -325,7 +332,6 @@ asmsub  disable_basic() clobbers(A) {
 
 sys {
     ; ------- lowlevel system routines --------
-    %option no_symbol_prefixing
 
     const ubyte target = 128         ;  compilation target specifier.  64 = C64, 128 = C128,  16 = CommanderX16.
 
@@ -367,7 +373,13 @@ asmsub  init_system_phase2()  {
 asmsub  cleanup_at_exit() {
     ; executed when the main subroutine does rts
     %asm {{
-        jmp  sys.enable_runstop_and_charsetswitch
+        lda  #0
+        sta  $ff00          ; default bank 15
+        jsr  cbm.CLRCHN		; reset i/o channels
+        jsr  enable_runstop_and_charsetswitch
+_exitcode = *+1
+        lda  #0        ; exit code possibly modified in exit()
+        rts
     }}
 }
 
@@ -391,77 +403,78 @@ asmsub  enable_runstop_and_charsetswitch() clobbers(A) {
     }}
 }
 
-asmsub  set_irq(uword handler @AY, bool useKernal @Pc) clobbers(A)  {
+    asmsub save_prog8_internals() {
+        %asm {{
+            lda  P8ZP_SCRATCH_B1
+            sta  save_SCRATCH_ZPB1
+            lda  P8ZP_SCRATCH_REG
+            sta  save_SCRATCH_ZPREG
+            lda  P8ZP_SCRATCH_W1
+            sta  save_SCRATCH_ZPWORD1
+            lda  P8ZP_SCRATCH_W1+1
+            sta  save_SCRATCH_ZPWORD1+1
+            lda  P8ZP_SCRATCH_W2
+            sta  save_SCRATCH_ZPWORD2
+            lda  P8ZP_SCRATCH_W2+1
+            sta  save_SCRATCH_ZPWORD2+1
+            rts
+save_SCRATCH_ZPB1	.byte  0
+save_SCRATCH_ZPREG	.byte  0
+save_SCRATCH_ZPWORD1	.word  0
+save_SCRATCH_ZPWORD2	.word  0
+        }}
+    }
+
+    asmsub restore_prog8_internals() {
+        %asm {{
+            lda  save_prog8_internals.save_SCRATCH_ZPB1
+            sta  P8ZP_SCRATCH_B1
+            lda  save_prog8_internals.save_SCRATCH_ZPREG
+            sta  P8ZP_SCRATCH_REG
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD1
+            sta  P8ZP_SCRATCH_W1
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD1+1
+            sta  P8ZP_SCRATCH_W1+1
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD2
+            sta  P8ZP_SCRATCH_W2
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD2+1
+            sta  P8ZP_SCRATCH_W2+1
+            rts
+        }}
+    }
+
+asmsub  set_irq(uword handler @AY) clobbers(A)  {
 	%asm {{
-	        sta  _modified+1
-	        sty  _modified+2
-	        lda  #0
-	        rol  a
-	        sta  _use_kernal
 		sei
+        sta  _modified+1
+        sty  _modified+2
 		lda  #<_irq_handler
 		sta  cbm.CINV
 		lda  #>_irq_handler
 		sta  cbm.CINV+1
 		cli
 		rts
-_irq_handler    jsr  _irq_handler_init
-_modified	jsr  $ffff                      ; modified
-		jsr  _irq_handler_end
-		lda  _use_kernal
-		bne  +
-		lda  #$ff
+_irq_handler
+        jsr  sys.save_prog8_internals
+        cld
+_modified
+        jsr  $ffff                      ; modified
+        pha
+        jsr  sys.restore_prog8_internals
+        pla
+        beq  +
+		jmp  cbm.IRQDFRT		; continue with normal kernal irq routine
++		lda  #$ff
 		sta  c64.VICIRQ			; acknowledge raster irq
 		lda  c64.CIA1ICR		; acknowledge CIA1 interrupt
-		; end irq processing - don't use kernal's irq handling
 		pla
 		tay
 		pla
 		tax
 		pla
 		rti
-+		jmp  cbm.IRQDFRT		; continue with normal kernal irq routine
 
 _use_kernal     .byte  0
-
-_irq_handler_init
-		; save all zp scratch registers as these might be clobbered by the irq routine
-		lda  P8ZP_SCRATCH_B1
-		sta  IRQ_SCRATCH_ZPB1
-		lda  P8ZP_SCRATCH_REG
-		sta  IRQ_SCRATCH_ZPREG
-		lda  P8ZP_SCRATCH_W1
-		sta  IRQ_SCRATCH_ZPWORD1
-		lda  P8ZP_SCRATCH_W1+1
-		sta  IRQ_SCRATCH_ZPWORD1+1
-		lda  P8ZP_SCRATCH_W2
-		sta  IRQ_SCRATCH_ZPWORD2
-		lda  P8ZP_SCRATCH_W2+1
-		sta  IRQ_SCRATCH_ZPWORD2+1
-		cld
-		rts
-
-_irq_handler_end
-		; restore all zp scratch registers
-		lda  IRQ_SCRATCH_ZPB1
-		sta  P8ZP_SCRATCH_B1
-		lda  IRQ_SCRATCH_ZPREG
-		sta  P8ZP_SCRATCH_REG
-		lda  IRQ_SCRATCH_ZPWORD1
-		sta  P8ZP_SCRATCH_W1
-		lda  IRQ_SCRATCH_ZPWORD1+1
-		sta  P8ZP_SCRATCH_W1+1
-		lda  IRQ_SCRATCH_ZPWORD2
-		sta  P8ZP_SCRATCH_W2
-		lda  IRQ_SCRATCH_ZPWORD2+1
-		sta  P8ZP_SCRATCH_W2+1
-		rts
-
-IRQ_SCRATCH_ZPB1	.byte  0
-IRQ_SCRATCH_ZPREG	.byte  0
-IRQ_SCRATCH_ZPWORD1	.word  0
-IRQ_SCRATCH_ZPWORD2	.word  0
-
 		}}
 }
 
@@ -481,40 +494,39 @@ asmsub  restore_irq() clobbers(A) {
 	}}
 }
 
-asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0, bool useKernal @Pc) clobbers(A) {
+asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0) clobbers(A) {
 	%asm {{
-	        sta  _modified+1
-	        sty  _modified+2
-	        lda  #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  cbm.CINV
-		lda  #>_raster_irq_handler
-		sta  cbm.CINV+1
-		cli
-		rts
+        sei
+        sta  _modified+1
+        sty  _modified+2
+        lda  cx16.r0
+        ldy  cx16.r0+1
+        jsr  _setup_raster_irq
+        lda  #<_raster_irq_handler
+        sta  cbm.CINV
+        lda  #>_raster_irq_handler
+        sta  cbm.CINV+1
+        cli
+        rts
 
 _raster_irq_handler
-		jsr  set_irq._irq_handler_init
-_modified	jsr  $ffff              ; modified
-		jsr  set_irq._irq_handler_end
+		jsr  sys.save_prog8_internals
+		cld
+_modified
+        jsr  $ffff              ; modified
+        pha
+		jsr  sys.restore_prog8_internals
         lda  #$ff
         sta  c64.VICIRQ			; acknowledge raster irq
-		lda  set_irq._use_kernal
-		bne  +
-		; end irq processing - don't use kernal's irq handling
 		pla
+		beq  +
+		jmp  cbm.IRQDFRT        ; continue with kernal irq routine
++		pla
 		tay
 		pla
 		tax
 		pla
 		rti
-+		jmp  cbm.IRQDFRT                ; continue with kernal irq routine
 
 _setup_raster_irq
 		pha
@@ -734,16 +746,27 @@ _longcopy
         }}
     }
 
-    inline asmsub exit(ubyte returnvalue @A) {
-        ; -- immediately exit the program with a return code in the A register
+    inline asmsub disable_caseswitch() {
+        %asm {{
+            lda  #$80
+            sta  247
+        }}
+    }
+
+    inline asmsub enable_caseswitch() {
         %asm {{
             lda  #0
-            sta  $ff00          ; default bank 15
-            jsr  cbm.CLRCHN		; reset i/o channels
-            jsr  sys.enable_runstop_and_charsetswitch
+            sta  247
+        }}
+    }
+
+    asmsub exit(ubyte returnvalue @A) {
+        ; -- immediately exit the program with a return code in the A register
+        %asm {{
+            sta  cleanup_at_exit._exitcode
             ldx  prog8_lib.orig_stackpointer
             txs
-            rts		; return to original caller
+            jmp  cleanup_at_exit
         }}
     }
 
@@ -754,11 +777,37 @@ _longcopy
         }}
     }
 
+    inline asmsub push(ubyte value @A) {
+        %asm {{
+            pha
+        }}
+    }
+
+    inline asmsub pushw(uword value @AY) {
+        %asm {{
+            pha
+            tya
+            pha
+        }}
+    }
+
+    inline asmsub pop() -> ubyte @A {
+        %asm {{
+            pla
+        }}
+    }
+
+    inline asmsub popw() -> uword @AY {
+        %asm {{
+            pla
+            tay
+            pla
+        }}
+    }
+
 }
 
 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
     ; (because there's no room for them in the zeropage)
@@ -890,4 +939,10 @@ cx16 {
             rts
         }}
     }
+
+    sub cpu_is_65816() -> bool {
+        ; Returns true when you have a 65816 cpu, false when it's a 6502.
+        return false
+    }
+
 }

compiler/res/prog8lib/c128/textio.p8

@@ -1,11 +1,13 @@
 ; Prog8 definitions for the Text I/O and Screen routines for the Commodore-128
+; All routines work with Screencode character encoding, except `print`, `chrout` and `input_chars`,
+; these work with PETSCII encoding instead.
 
 %import syslib
 %import conv
 
 
 txt {
-    %option no_symbol_prefixing
+    %option no_symbol_prefixing, ignore_unused
 
 const ubyte DEFAULT_WIDTH = 40
 const ubyte DEFAULT_HEIGHT = 25
@@ -38,6 +40,45 @@ asmsub column(ubyte col @A) clobbers(A, X, Y) {
     }}
 }
 
+
+asmsub get_column() -> ubyte @Y {
+    %asm {{
+        sec
+        jmp cbm.PLOT
+    }}
+}
+
+asmsub row(ubyte rownum @A) clobbers(A, X, Y) {
+    ; ---- set the cursor on the given row (starting with 0) on the current line
+    %asm {{
+        sec
+        jsr  cbm.PLOT
+        tax
+        clc
+        jmp  cbm.PLOT
+    }}
+}
+
+asmsub get_row() -> ubyte @X {
+    %asm {{
+        sec
+        jmp cbm.PLOT
+    }}
+}
+
+sub get_cursor(uword colptr, uword rowptr) {
+    %asm {{
+        sec
+        jsr cbm.PLOT
+        tya
+        ldy #$00
+        sta (colptr),y
+        txa
+        sta (rowptr),y
+    }}
+}
+
+
 asmsub  fill_screen (ubyte char @ A, ubyte color @ Y) clobbers(A)  {
 	; ---- fill the character screen with the given fill character and character color.
 	;      (assumes screen and color matrix are at their default addresses)
@@ -238,10 +279,10 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use cbm.CHROUT directly ofcourse.
+romsub $FFD2 = chrout(ubyte char @ A)    ; for consistency. You can also use cbm.CHROUT directly ofcourse. Note: takes a PETSCII encoded character.
 
 asmsub  print (str text @ AY) clobbers(A,Y)  {
-	; ---- print null terminated string from A/Y
+	; ---- print null terminated string, in PETSCII encoding, 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 cbm.CHROUT of that single char.
@@ -427,7 +468,8 @@ asmsub  print_w  (word value @ AY) clobbers(A,X,Y)  {
 }
 
 asmsub  input_chars  (uword buffer @ AY) clobbers(A) -> ubyte @ Y  {
-	; ---- Input a string (max. 80 chars) from the keyboard. Returns length in Y. (string is terminated with a 0 byte as well)
+	; ---- Input a string (max. 80 chars) from the keyboard, in PETSCII encoding.
+	;      Returns length in Y. (string is terminated with a 0 byte as well)
 	;      It assumes the keyboard is selected as I/O channel!
 
 	%asm {{
@@ -584,4 +626,11 @@ asmsub height() clobbers(X, Y) -> ubyte @A {
     }}
 }
 
+asmsub waitkey() -> ubyte @A {
+    %asm {{
+-       jsr cbm.GETIN
+        beq -
+        rts
+    }}
+}
 }

compiler/res/prog8lib/c64/floats.p8

@@ -1,14 +1,13 @@
 ; Prog8 definitions for floating point handling on the Commodore-64
 
-%option enable_floats
+%option enable_floats, no_symbol_prefixing, ignore_unused
 %import floats_functions
 
 floats {
 	; ---- this block contains C-64 floating point related functions ----
-    %option no_symbol_prefixing
-
-        const float  PI     = 3.141592653589793
-        const float  TWOPI  = 6.283185307179586
+        const float  π      = 3.141592653589793
+        const float  PI     = π
+        const float  TWOPI  = 2*π
 
 
 ; ---- C64 basic and kernal ROM float constants and functions ----
@@ -46,13 +45,13 @@ romsub $b391 = GIVAYF(ubyte lo @ Y, ubyte hi @ A) clobbers(A,X,Y)
 
 romsub $b3a2 = FREADUY(ubyte value @ Y) clobbers(A,X,Y)     ; 8 bit unsigned Y -> float in fac1
 romsub $bc3c = FREADSA(byte value @ A) clobbers(A,X,Y)      ; 8 bit signed A -> float in fac1
-romsub $b7b5 = FREADSTR(ubyte length @ A) clobbers(A,X,Y)   ; str -> fac1, $22/23 must point to string, A=string length.  Also see parse_f()
+romsub $b7b5 = FREADSTR(ubyte length @ A) clobbers(A,X,Y)   ; str -> fac1, $22/23 must point to string, A=string length.  Also see parse()
 romsub $aabc = FPRINTLN() clobbers(A,X,Y)                   ; print string of fac1, on one line (= with newline) destroys fac1.  (consider FOUT + STROUT as well)
 romsub $bddd = FOUT() clobbers(X) -> uword @ AY             ; fac1 -> string, address returned in AY ($0100)
 
 romsub $b849 = FADDH() clobbers(A,X,Y)                      ; fac1 += 0.5, for rounding- call this before INT
 romsub $bae2 = MUL10() clobbers(A,X,Y)                      ; fac1 *= 10
-romsub $bafe = DIV10() clobbers(A,X,Y)                      ; fac1 /= 10 , CAUTION: result is always positive!
+romsub $bafe = DIV10() clobbers(A,X,Y)                      ; fac1 /= 10 , CAUTION: result is always positive! You have to fix sign manually!
 romsub $bc5b = 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 $b86a = FADDT() clobbers(A,X,Y)                      ; fac1 += fac2
@@ -77,6 +76,7 @@ romsub $bf71 = SQR() clobbers(A,X,Y)                        ; fac1 = SQRT(fac1)
 romsub $bf74 = SQRA() clobbers(A,X,Y)                       ; fac1 = SQRT(fac2)
 romsub $bfed = EXP() clobbers(A,X,Y)                        ; fac1 = EXP(fac1)  (e ** fac1)
 romsub $bfb4 = NEGOP() clobbers(A)                          ; switch the sign of fac1 (fac1 = -fac1)
+romsub $b8d7 = NORMAL() clobbers(A)                         ; normalize FAC1
 romsub $e097 = RND() clobbers(A,X,Y)                        ; fac1 = RND(fac1) float random number generator
 romsub $e264 = COS() clobbers(A,X,Y)                        ; fac1 = COS(fac1)
 romsub $e26b = SIN() clobbers(A,X,Y)                        ; fac1 = SIN(fac1)
@@ -157,7 +157,7 @@ asmsub  GETADRAY  () clobbers(X) -> uword @ AY  {
 
 &uword AYINT_facmo = $64      ; $64/$65 contain result of AYINT
 
-sub rndf() -> float {
+sub rnd() -> float {
     %asm {{
         lda  #1
         jsr  FREADSA
@@ -165,7 +165,7 @@ sub rndf() -> float {
     }}
 }
 
-asmsub parse_f(str value @AY) -> float @FAC1 {
+asmsub parse(str value @AY) -> float @FAC1 {
     %asm {{
         sta  $22
         sty  $23
@@ -175,6 +175,12 @@ asmsub parse_f(str value @AY) -> float @FAC1 {
     }}
 }
 
+asmsub normalize(float value @FAC1) -> float @ FAC1 {
+    %asm {{
+        jmp  floats.NORMAL
+    }}
+}
+
 
 %asminclude "library:c64/floats.asm"
 %asminclude "library:c64/floats_funcs.asm"

compiler/res/prog8lib/c64/floats_funcs.asm

@@ -147,3 +147,54 @@ func_sqrt_into_FAC1     .proc
         jsr  MOVFM
         jmp  SQR
 	.pend
+
+
+
+containment_floatarray    .proc
+	; -- check if a value exists in a float array.
+	;    parameters: FAC1: value to check, P8ZP_SCRATCH_W1: address of the word array, Y = length of array (>=1).
+	;    returns boolean 0/1 in A.
+	sty  P8ZP_SCRATCH_REG
+	ldx  #<floats.floats_temp_var
+	ldy  #>floats.floats_temp_var
+	jsr  floats.MOVMF
+	ldx  P8ZP_SCRATCH_REG
+	ldy  #0
+-       lda  floats.floats_temp_var
+	cmp  (P8ZP_SCRATCH_W1),y
+	bne  _firstmiss
+	iny
+	lda  floats.floats_temp_var+1
+	cmp  (P8ZP_SCRATCH_W1),y
+	bne  _secondmiss
+	iny
+	lda  floats.floats_temp_var+2
+	cmp  (P8ZP_SCRATCH_W1),y
+	bne  _thirdmiss
+	iny
+	lda  floats.floats_temp_var+3
+	cmp  (P8ZP_SCRATCH_W1),y
+	bne  _fourthmiss
+	iny
+	lda  floats.floats_temp_var+4
+	cmp  (P8ZP_SCRATCH_W1),y
+	bne  _fifthmiss
+	lda  #1
+	rts
+
+_firstmiss
+	iny
+_secondmiss
+	iny
+_thirdmiss
+	iny
+_fourthmiss
+	iny
+_fifthmiss
+        iny
+	dex
+	bne  -
+        lda  #0
+        rts
+
+	.pend

compiler/res/prog8lib/c64/graphics.p8

@@ -1,11 +1,14 @@
 ; bitmap pixel graphics module for the C64
 ; only black/white monochrome 320x200 for now
 ;
+; NOTE: For sake of speed, NO BOUNDS CHECKING is performed in most routines!
+;       You'll have to make sure yourself that you're not writing outside of bitmap boundaries!
+;
 ; Sets character matrix and bitmap screen memory at a higher memory location $5c00-$7fff
 ; so that the program itself can be larger without starting to overwrite the graphics memory.
 
 graphics {
-    %option no_symbol_prefixing
+    %option ignore_unused
 
     const uword WIDTH = 320
     const ubyte HEIGHT = 200
@@ -172,11 +175,11 @@ graphics {
 
         if separate_pixels {
             %asm {{
-                lda  pixaddr
+                lda  p8v_pixaddr
                 sta  P8ZP_SCRATCH_W1
-                lda  pixaddr+1
+                lda  p8v_pixaddr+1
                 sta  P8ZP_SCRATCH_W1+1
-                ldy  separate_pixels
+                ldy  p8v_separate_pixels
                 lda  hline_filled_right,y
                 eor  #255
                 ldy  #0
@@ -190,23 +193,23 @@ graphics {
 
         if length {
             %asm {{
-                lda  length
+                lda  p8v_length
                 and  #7
-                sta  separate_pixels
-                lsr  length+1
-                ror  length
-                lsr  length+1
-                ror  length
-                lsr  length+1
-                ror  length
-                lda  pixaddr
+                sta  p8v_separate_pixels
+                lsr  p8v_length+1
+                ror  p8v_length
+                lsr  p8v_length+1
+                ror  p8v_length
+                lsr  p8v_length+1
+                ror  p8v_length
+                lda  p8v_pixaddr
                 sta  _modified+1
-                lda  pixaddr+1
+                lda  p8v_pixaddr+1
                 sta  _modified+2
-                lda  length
-                ora  length+1
+                lda  p8v_length
+                ora  p8v_length+1
                 beq  _zero
-                ldy  length
+                ldy  p8v_length
                 ldx  #$ff
 _modified       stx  $ffff      ; modified
                 lda  _modified+1
@@ -218,7 +221,7 @@ _modified       stx  $ffff      ; modified
 +               dey
                 bne  _modified
 _zero
-                ldy  separate_pixels
+                ldy  p8v_separate_pixels
                 beq  hline_zero2
                 lda  _modified+1
                 sta  P8ZP_SCRATCH_W1
@@ -244,7 +247,8 @@ hline_zero2
     }
 
     sub circle(uword xcenter, ubyte ycenter, ubyte radius) {
-        ; Midpoint algorithm
+        ; Warning: NO BOUNDS CHECKS. Make sure circle fits in the screen.
+        ; Midpoint algorithm.
         if radius==0
             return
         ubyte @zp ploty
@@ -273,16 +277,17 @@ hline_zero2
             internal_plotx = xcenter - yy
             internal_plot(ploty)
             yy++
-            if decisionOver2<=0
-                decisionOver2 += (yy as word)*2+1
-            else {
+            if decisionOver2>=0 {
                 radius--
-                decisionOver2 += (yy as word -radius)*2+1
+                decisionOver2 -= radius*$0002
             }
+            decisionOver2 += yy*$0002
+            decisionOver2++
         }
     }
 
     sub disc(uword xcenter, ubyte ycenter, ubyte radius) {
+        ; Warning: NO BOUNDS CHECKS. Make sure circle fits in the screen.
         ; Midpoint algorithm, filled
         if radius==0
             return
@@ -295,12 +300,12 @@ hline_zero2
             horizontal_line(xcenter-yy, ycenter+radius, yy*2+1)
             horizontal_line(xcenter-yy, ycenter-radius, yy*2+1)
             yy++
-            if decisionOver2<=0
-                decisionOver2 += (yy as word)*2+1
-            else {
+            if decisionOver2>=0 {
                 radius--
-                decisionOver2 += (yy as word -radius)*2+1
+                decisionOver2 -= radius*$0002
             }
+            decisionOver2 += yy*$0002
+            decisionOver2++
         }
     }
 
@@ -314,9 +319,9 @@ hline_zero2
 
     inline asmsub  plot(uword plotx @XY, ubyte ploty @A) clobbers (A, X, Y) {
         %asm {{
-            stx  graphics.internal_plotx
-            sty  graphics.internal_plotx+1
-            jsr  graphics.internal_plot
+            stx  p8b_graphics.p8v_internal_plotx
+            sty  p8b_graphics.p8v_internal_plotx+1
+            jsr  p8b_graphics.p8s_internal_plot
         }}
     }
 
@@ -328,11 +333,11 @@ hline_zero2
     asmsub  internal_plot(ubyte ploty @A) clobbers (A, X, Y) {      ; internal_plotx is 16 bits 0 to 319... doesn't fit in a register
         %asm {{
         tay
-        lda  internal_plotx+1
+        lda  p8v_internal_plotx+1
         sta  P8ZP_SCRATCH_W2+1
         lsr  a            ; 0
         sta  P8ZP_SCRATCH_W2
-        lda  internal_plotx
+        lda  p8v_internal_plotx
         pha
         and  #7
         tax
@@ -360,7 +365,7 @@ _ormask     .byte 128, 64, 32, 16, 8, 4, 2, 1
 ; the y lookup tables encodes this formula:  BITMAP_ADDRESS + 320*(py>>3) + (py & 7)    (y from 0..199)
 ; We use the 64tass syntax for range expressions to calculate this table on assembly time.
 
-_plot_y_values := BITMAP_ADDRESS + 320*(range(200)>>3) + (range(200) & 7)
+_plot_y_values := p8c_BITMAP_ADDRESS + 320*(range(200)>>3) + (range(200) & 7)
 
 _y_lookup_lo    .byte  <_plot_y_values
 _y_lookup_hi    .byte  >_plot_y_values
@@ -370,9 +375,9 @@ _y_lookup_hi    .byte  >_plot_y_values
 
     asmsub get_y_lookup(ubyte yy @Y) -> uword @AY {
         %asm {{
-            lda  internal_plot._y_lookup_lo,y
+            lda  p8s_internal_plot._y_lookup_lo,y
             pha
-            lda  internal_plot._y_lookup_hi,y
+            lda  p8s_internal_plot._y_lookup_hi,y
             tay
             pla
             rts

compiler/res/prog8lib/c64/syslib.p8

@@ -1,13 +1,11 @@
 ; Prog8 definitions for the Commodore-64
 ; Including memory registers, I/O registers, Basic and Kernal subroutines.
 
+%option no_symbol_prefixing, ignore_unused
 
 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
@@ -82,7 +80,7 @@ romsub $FFAB = UNTLK() clobbers(A)                              ; command serial
 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 $FFB7 = READST() -> ubyte @ A                            ; read I/O status word  (use CLEARST to reset it to 0)
 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
@@ -128,13 +126,20 @@ asmsub RDTIM16() clobbers(X) -> uword @AY {
     }}
 }
 
+sub CLEARST() {
+    ; -- Set the ST status variable back to 0. (there's no direct kernal call for this)
+    ;    Note: a drive error state (blinking led) isn't cleared! You can use diskio.status() to clear that.
+    SETNAM(0, $0000)
+    SETLFS(15, 3, 15)
+    void OPEN()
+    CLOSE(15)
+}
+
 }
 
 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)
         ; (depending on the VIC bank and screen ram address selection these can be shifted around though,
         ; see the two routines after this for a dynamic way of determining the correct memory location)
@@ -324,8 +329,6 @@ c64 {
 sys {
     ; ------- lowlevel system routines --------
 
-    %option no_symbol_prefixing
-
     const ubyte target = 64         ;  compilation target specifier.  64 = C64, 128 = C128, 16 = CommanderX16.
 
 
@@ -369,7 +372,13 @@ asmsub  init_system_phase2()  {
 asmsub  cleanup_at_exit() {
     ; executed when the main subroutine does rts
     %asm {{
-        jmp  sys.enable_runstop_and_charsetswitch
+        lda  #31
+        sta  $01        ; bank the kernal in
+        jsr  cbm.CLRCHN		; reset i/o channels
+        jsr  enable_runstop_and_charsetswitch
+_exitcode = *+1
+        lda  #0        ; exit code possibly modified in exit()
+        rts
     }}
 }
 
@@ -393,78 +402,77 @@ asmsub  enable_runstop_and_charsetswitch() clobbers(A) {
     }}
 }
 
-asmsub  set_irq(uword handler @AY, bool useKernal @Pc) clobbers(A)  {
+    asmsub save_prog8_internals() {
+        %asm {{
+            lda  P8ZP_SCRATCH_B1
+            sta  save_SCRATCH_ZPB1
+            lda  P8ZP_SCRATCH_REG
+            sta  save_SCRATCH_ZPREG
+            lda  P8ZP_SCRATCH_W1
+            sta  save_SCRATCH_ZPWORD1
+            lda  P8ZP_SCRATCH_W1+1
+            sta  save_SCRATCH_ZPWORD1+1
+            lda  P8ZP_SCRATCH_W2
+            sta  save_SCRATCH_ZPWORD2
+            lda  P8ZP_SCRATCH_W2+1
+            sta  save_SCRATCH_ZPWORD2+1
+            rts
+save_SCRATCH_ZPB1	.byte  0
+save_SCRATCH_ZPREG	.byte  0
+save_SCRATCH_ZPWORD1	.word  0
+save_SCRATCH_ZPWORD2	.word  0
+        }}
+    }
+
+    asmsub restore_prog8_internals() {
+        %asm {{
+            lda  save_prog8_internals.save_SCRATCH_ZPB1
+            sta  P8ZP_SCRATCH_B1
+            lda  save_prog8_internals.save_SCRATCH_ZPREG
+            sta  P8ZP_SCRATCH_REG
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD1
+            sta  P8ZP_SCRATCH_W1
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD1+1
+            sta  P8ZP_SCRATCH_W1+1
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD2
+            sta  P8ZP_SCRATCH_W2
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD2+1
+            sta  P8ZP_SCRATCH_W2+1
+            rts
+        }}
+    }
+
+asmsub  set_irq(uword handler @AY) clobbers(A)  {
 	%asm {{
-	        sta  _modified+1
-	        sty  _modified+2
-	        lda  #0
-	        rol  a
-	        sta  _use_kernal
-		sei
+	    sei
+        sta  _modified+1
+        sty  _modified+2
 		lda  #<_irq_handler
 		sta  cbm.CINV
 		lda  #>_irq_handler
 		sta  cbm.CINV+1
 		cli
 		rts
-_irq_handler    jsr  _irq_handler_init
-_modified	jsr  $ffff                      ; modified
-		jsr  _irq_handler_end
-		lda  _use_kernal
-		bne  +
-		lda  #$ff
+_irq_handler
+        jsr  sys.save_prog8_internals
+        cld
+_modified
+        jsr  $ffff                      ; modified
+        pha
+		jsr  sys.restore_prog8_internals
+		pla
+		beq  +
+		jmp  cbm.IRQDFRT		; continue with normal kernal irq routine
++		lda  #$ff
 		sta  c64.VICIRQ			; acknowledge raster irq
 		lda  c64.CIA1ICR		; acknowledge CIA1 interrupt
-		; end irq processing - don't use kernal's irq handling
 		pla
 		tay
 		pla
 		tax
 		pla
 		rti
-+		jmp  cbm.IRQDFRT		; continue with normal kernal irq routine
-
-_use_kernal     .byte  0
-
-_irq_handler_init
-		; save all zp scratch registers as these might be clobbered by the irq routine
-		lda  P8ZP_SCRATCH_B1
-		sta  IRQ_SCRATCH_ZPB1
-		lda  P8ZP_SCRATCH_REG
-		sta  IRQ_SCRATCH_ZPREG
-		lda  P8ZP_SCRATCH_W1
-		sta  IRQ_SCRATCH_ZPWORD1
-		lda  P8ZP_SCRATCH_W1+1
-		sta  IRQ_SCRATCH_ZPWORD1+1
-		lda  P8ZP_SCRATCH_W2
-		sta  IRQ_SCRATCH_ZPWORD2
-		lda  P8ZP_SCRATCH_W2+1
-		sta  IRQ_SCRATCH_ZPWORD2+1
-		cld
-		rts
-
-_irq_handler_end
-		; restore all zp scratch registers
-		lda  IRQ_SCRATCH_ZPB1
-		sta  P8ZP_SCRATCH_B1
-		lda  IRQ_SCRATCH_ZPREG
-		sta  P8ZP_SCRATCH_REG
-		lda  IRQ_SCRATCH_ZPWORD1
-		sta  P8ZP_SCRATCH_W1
-		lda  IRQ_SCRATCH_ZPWORD1+1
-		sta  P8ZP_SCRATCH_W1+1
-		lda  IRQ_SCRATCH_ZPWORD2
-		sta  P8ZP_SCRATCH_W2
-		lda  IRQ_SCRATCH_ZPWORD2+1
-		sta  P8ZP_SCRATCH_W2+1
-		rts
-
-IRQ_SCRATCH_ZPB1	.byte  0
-IRQ_SCRATCH_ZPREG	.byte  0
-IRQ_SCRATCH_ZPWORD1	.word  0
-IRQ_SCRATCH_ZPWORD2	.word  0
-
-		}}
+    }}
 }
 
 asmsub  restore_irq() clobbers(A) {
@@ -483,40 +491,39 @@ asmsub  restore_irq() clobbers(A) {
 	}}
 }
 
-asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0, bool useKernal @Pc) clobbers(A) {
+asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0) clobbers(A) {
 	%asm {{
-	        sta  _modified+1
-	        sty  _modified+2
-	        lda  #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  cbm.CINV
-		lda  #>_raster_irq_handler
-		sta  cbm.CINV+1
-		cli
-		rts
+	    sei
+        sta  _modified+1
+        sty  _modified+2
+        lda  cx16.r0
+        ldy  cx16.r0+1
+        jsr  _setup_raster_irq
+        lda  #<_raster_irq_handler
+        sta  cbm.CINV
+        lda  #>_raster_irq_handler
+        sta  cbm.CINV+1
+        cli
+        rts
 
 _raster_irq_handler
-		jsr  set_irq._irq_handler_init
-_modified	jsr  $ffff              ; modified
-		jsr  set_irq._irq_handler_end
+		jsr  sys.save_prog8_internals
+		cld
+_modified
+        jsr  $ffff              ; modified
+        pha
+        jsr  sys.restore_prog8_internals
         lda  #$ff
         sta  c64.VICIRQ			; acknowledge raster irq
-		lda  set_irq._use_kernal
-		bne  +
-		; end irq processing - don't use kernal's irq handling
-		pla
+        pla
+        beq  +
+		jmp  cbm.IRQDFRT                ; continue with kernal irq routine
++		pla
 		tay
 		pla
 		tax
 		pla
 		rti
-+		jmp  cbm.IRQDFRT                ; continue with kernal irq routine
 
 _setup_raster_irq
 		pha
@@ -737,16 +744,27 @@ _longcopy
         }}
     }
 
-    inline asmsub exit(ubyte returnvalue @A) {
+    inline asmsub disable_caseswitch() {
+        %asm {{
+            lda  #$80
+            sta  657
+        }}
+    }
+
+    inline asmsub enable_caseswitch() {
+        %asm {{
+            lda  #0
+            sta  657
+        }}
+    }
+
+    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  cbm.CLRCHN		; reset i/o channels
-            jsr  sys.enable_runstop_and_charsetswitch
+            sta  cleanup_at_exit._exitcode
             ldx  prog8_lib.orig_stackpointer
             txs
-            rts		; return to original caller
+            jmp  cleanup_at_exit
         }}
     }
 
@@ -757,12 +775,36 @@ _longcopy
         }}
     }
 
+    inline asmsub push(ubyte value @A) {
+        %asm {{
+            pha
+        }}
+    }
+
+    inline asmsub pushw(uword value @AY) {
+        %asm {{
+            pha
+            tya
+            pha
+        }}
+    }
+
+    inline asmsub pop() -> ubyte @A {
+        %asm {{
+            pla
+        }}
+    }
+
+    inline asmsub popw() -> uword @AY {
+        %asm {{
+            pla
+            tay
+            pla
+        }}
+    }
 }
 
 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 in the default configuration)
@@ -896,4 +938,9 @@ cx16 {
         }}
     }
 
+    sub cpu_is_65816() -> bool {
+        ; Returns true when you have a 65816 cpu, false when it's a 6502.
+        return false
+    }
+
 }

compiler/res/prog8lib/c64/textio.p8

@@ -1,4 +1,6 @@
 ; Prog8 definitions for the Text I/O and Screen routines for the Commodore-64
+; All routines work with Screencode character encoding, except `print`, `chrout` and `input_chars`,
+; these work with PETSCII encoding instead.
 
 %import syslib
 %import conv
@@ -6,7 +8,7 @@
 
 txt {
 
-    %option no_symbol_prefixing
+    %option no_symbol_prefixing, ignore_unused
 
 const ubyte DEFAULT_WIDTH = 40
 const ubyte DEFAULT_HEIGHT = 25
@@ -39,6 +41,45 @@ asmsub column(ubyte col @A) clobbers(A, X, Y) {
     }}
 }
 
+
+asmsub get_column() -> ubyte @Y {
+    %asm {{
+        sec
+        jmp cbm.PLOT
+    }}
+}
+
+asmsub row(ubyte rownum @A) clobbers(A, X, Y) {
+    ; ---- set the cursor on the given row (starting with 0) on the current line
+    %asm {{
+        sec
+        jsr  cbm.PLOT
+        tax
+        clc
+        jmp  cbm.PLOT
+    }}
+}
+
+asmsub get_row() -> ubyte @X {
+    %asm {{
+        sec
+        jmp cbm.PLOT
+    }}
+}
+
+sub get_cursor(uword colptr, uword rowptr) {
+    %asm {{
+        sec
+        jsr cbm.PLOT
+        tya
+        ldy #$00
+        sta (colptr),y
+        txa
+        sta (rowptr),y
+    }}
+}
+
+
 asmsub  fill_screen (ubyte character @ A, ubyte color @ Y) clobbers(A)  {
 	; ---- fill the character screen with the given fill character and character color.
 	;      (assumes screen and color matrix are at their default addresses)
@@ -237,10 +278,10 @@ _scroll_screen  ; scroll only the screen memory
 	}}
 }
 
-romsub $FFD2 = chrout(ubyte character @ A)    ; for consistency. You can also use cbm.CHROUT directly ofcourse.
+romsub $FFD2 = chrout(ubyte character @ A)    ; for consistency. You can also use cbm.CHROUT directly ofcourse. Note: takes a PETSCII encoded character.
 
 asmsub  print (str text @ AY) clobbers(A,Y)  {
-	; ---- print null terminated string from A/Y
+	; ---- print null terminated string, in PESCII encoding, 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 cbm.CHROUT of that single char.
@@ -426,7 +467,8 @@ asmsub  print_w  (word value @ AY) clobbers(A,X,Y)  {
 }
 
 asmsub  input_chars  (uword buffer @ AY) clobbers(A) -> ubyte @ Y  {
-	; ---- Input a string (max. 80 chars) from the keyboard. Returns length in Y. (string is terminated with a 0 byte as well)
+	; ---- Input a string (max. 80 chars) from the keyboard, in PETSCII encoding.
+	;      Returns length in Y. (string is terminated with a 0 byte as well)
 	;      It assumes the keyboard is selected as I/O channel!
 
 	%asm {{
@@ -583,4 +625,11 @@ asmsub height() clobbers(X, Y) -> ubyte @A {
     }}
 }
 
+asmsub waitkey() -> ubyte @A {
+    %asm {{
+-       jsr cbm.GETIN
+        beq -
+        rts
+    }}
+}
 }

compiler/res/prog8lib/conv.p8

@@ -2,7 +2,7 @@
 
 conv {
 
-    %option no_symbol_prefixing
+    %option no_symbol_prefixing, ignore_unused
 
 ; ----- number conversions to decimal strings ----
 

compiler/res/prog8lib/cx16/bmx.p8

@@ -0,0 +1,362 @@
+; Routines to load and save "BMX" files (commander X16 bitmap format) Version 1.
+; Only uncompressed images are supported for now.
+; BMX Specification: https://cx16forum.com/forum/viewtopic.php?t=6945
+
+%import diskio
+%option ignore_unused
+
+bmx {
+
+    ubyte[32] header
+    str FILEID = petscii:"bmx"
+
+    ubyte bitsperpixel          ; consider using set_bpp() when setting this
+    ubyte vera_colordepth       ; consider using set_vera_colordepth() when setting this
+    uword width
+    uword height
+    ubyte border
+    uword palette_entries       ; 1-256
+    ubyte palette_start
+    ubyte compression
+    uword @shared palette_buffer_ptr = 0    ; should you want to load or save the palette into main memory instead of directly into vram
+
+    uword error_message             ; pointer to error message, or 0 if all ok
+    ubyte old_drivenumber
+
+    sub open(ubyte drivenumber, str filename) -> bool {
+        ; Open a BMX bitmap file and reads the header information.
+        ; Returns true if all is ok, false otherwise + error_message will be set.
+        error_message = 0
+        old_drivenumber = diskio.drivenumber
+        diskio.drivenumber = drivenumber
+        if diskio.f_open(filename) {
+            diskio.reset_read_channel()
+            if read_header() {
+                if parse_header() {
+                    if palette_entries>0 {
+                        if width<=640 {
+                            return true
+                        } else
+                            error_message = "image too large"
+                    } else
+                        error_message = "invalid bmx file"
+                } ; note: parse_header sets the error message by itself
+            } else
+                error_message = "invalid bmx file"
+        } else
+            error_message = diskio.status()
+
+        close()
+        return false
+    }
+
+    sub close() {
+        ; if you want to close the file before actually loading palette or bitmap data.
+        diskio.f_close()
+        diskio.drivenumber = old_drivenumber
+    }
+
+    sub continue_load(ubyte vbank, uword vaddr) -> bool {
+        ; Continues loading the palette and bitmap data from the opened BMX file.
+        ; Parameters: the vram bank and address where the bitmap data should go.
+        ; You can set palette_buffer_ptr if you want the palette buffered rather than directly into vram.
+        ; Note: does not change vera screen mode or colordepth! You have to do that yourself!
+        ; Returns true if all is ok, false otherwise + error_message will be set.
+        error_message = 0
+        diskio.reset_read_channel()
+        if width==320 or width==640 {
+            if compression==0 {
+                if read_palette() {
+                    if not read_bitmap(vbank, vaddr)
+                        error_message = "bitmap error"
+                } else
+                    error_message = "palette error"
+            } else
+                error_message = "compression not supported"
+        } else
+            error_message = "width not 320 or 640"      ; note: use continue_load_stamp() to read other sizes
+
+        close()
+        return error_message==0
+    }
+
+    sub continue_load_stamp(ubyte vbank, uword vaddr, uword screenwidth) -> bool {
+        ; Continues loading the palette and bitmap "stamp" data from the opened BMX file.
+        ; "Stamp" means: load an image that is smaller than the screen (so we need to pad around it)
+        ; Parameters:the vram bank and address where the bitmap data should go,
+        ; and the screen width that the stamp image is loaded into.
+        ; You can set palette_buffer_ptr if you want the palette buffered rather than directly into vram.
+        ; Note: does not change vera screen mode or colordepth! You have to do that yourself!
+        ; Returns true if all is ok, false otherwise + error_message will be set.
+        error_message = 0
+        diskio.reset_read_channel()
+        if compression==0 {
+            if read_palette() {
+                if not read_bitmap_padded(vbank, vaddr, screenwidth)
+                    error_message = "bitmap error"
+            } else
+                error_message = "palette error"
+        } else
+            error_message = "compression not supported"
+
+        close()
+        return error_message==0
+    }
+
+    sub continue_load_only_palette() -> bool {
+        ; Continues loading the palette but not the bitmap data from the opened BMX file.
+        ; You can set palette_buffer_ptr if you want the palette buffered rather than directly into vram.
+        ; Returns true if all is ok, false otherwise + error_message will be set.
+        ; Afterwards the file is closed and you can no longer read additional data from it!
+        error_message = 0
+        diskio.reset_read_channel()
+        if not read_palette()
+            error_message = "palette error"
+        close()
+        return error_message==0
+    }
+
+    sub save(ubyte drivenumber, str filename, ubyte vbank, uword vaddr, uword screenwidth) -> bool {
+        ; Save bitmap and palette data from vram into a BMX file.
+        ; First you must have set all bmx.* variables to the correct values! (like width, height..)
+        ; Parameters:
+        ; drive number and filename to save to,
+        ; vram bank and address of the bitmap data to save,
+        ; and optionally the screen width if you want to save a "stamp" that's smaller than the screen.
+        ; If you're saving the whole screen width, you can leave screenwidth at 0.
+        ; Returns: success status. If false, error_message points to the error message string.
+        error_message = 0
+        if compression {
+            error_message = "compression not supported"
+            return false
+        }
+        if screenwidth==0 {
+            if width!=320 and width!=640 {
+                error_message = "width not 320 or 640"
+                return false
+            }
+        } else {
+            if width>screenwidth {
+                error_message = "image too large"
+                return false
+            }
+        }
+
+        old_drivenumber = diskio.drivenumber
+        diskio.drivenumber = drivenumber
+        if diskio.f_open_w(filename) {
+            cx16.r0 = diskio.status()
+            if cx16.r0[0]!='0' {
+                error_message = cx16.r0
+                goto save_end
+            }
+            diskio.reset_write_channel()
+            if write_header() {
+                if write_palette() {
+                    if not write_bitmap(vbank, vaddr, screenwidth)
+                        error_message = "bitmap error"
+                } else
+                    error_message = "palette error"
+            } else
+                error_message = "header error"
+        } else
+             error_message = diskio.status()
+save_end:
+        diskio.f_close_w()
+        diskio.drivenumber = old_drivenumber
+        return error_message==0
+    }
+
+    sub set_bpp(ubyte bpp) {
+        ubyte[8] depths = [0,1,1,2,2,2,2,3]
+        vera_colordepth = depths[bpp-1]
+    }
+
+    sub set_vera_colordepth(ubyte depth) {
+        vera_colordepth = depth
+        bitsperpixel = 1 << depth
+    }
+
+; ------------------- helper routines, usually not needed to call yourself -------------------------
+
+    sub read_header() -> bool {
+        ; load a BMX header from the currently active input file
+        for cx16.r0L in 0 to sizeof(header)-1 {
+            header[cx16.r0L] = cbm.CHRIN()
+        }
+        return not cbm.READST()
+    }
+
+    sub read_palette() -> bool {
+        ; load palette data from the currently active input file
+        ; if palette_buffer_ptr is not 0, the palette data is read into that memory buffer,
+        ; otherwise it is read directly into the palette in vram.
+        cx16.vaddr(1, $fa00+palette_start*2, 0, 1)
+        cx16.r3 = palette_buffer_ptr
+        cx16.r2L = lsb(palette_entries)
+        do {
+            cx16.r4L = cbm.CHRIN()
+            cx16.r4H = cbm.CHRIN()
+            if cx16.r3 {
+                pokew(cx16.r3, cx16.r4)             ; into memory
+                cx16.r3+=2
+            } else {
+                cx16.VERA_DATA0 = cx16.r4L          ; into vram
+                cx16.VERA_DATA0 = cx16.r4H
+            }
+            cx16.r2L--
+        } until cx16.r2L==0
+        return cbm.READST()==0 or cbm.READST()&$40    ; no error or eof?
+    }
+
+    sub read_bitmap_padded(ubyte vbank, uword vaddr, uword screenwidth) -> bool {
+        ; load bitmap "stamp" into vram from the currently active input file
+        cx16.r3 = bytes_per_scanline(width)         ; num bytes per image scanline
+        cx16.r2 = bytes_per_scanline(screenwidth)   ; num bytes per screen scanline
+        repeat height {
+            cx16.vaddr(vbank, vaddr, 0, 1)
+            read_scanline(cx16.r3)
+            vaddr += cx16.r2
+            if_cs
+                vbank++
+        }
+        return cbm.READST()==0 or cbm.READST()&$40    ; no error or eof?
+    }
+
+    sub read_bitmap(ubyte vbank, uword vaddr) -> bool {
+        ; load contiguous bitmap into vram from the currently active input file
+        cx16.r3 = bytes_per_scanline(width)
+        cx16.vaddr(vbank, vaddr, 0, 1)
+        repeat height
+            read_scanline(cx16.r3)
+        return cbm.READST()==0 or cbm.READST()&$40    ; no error or eof?
+    }
+
+    sub read_scanline(uword size) {
+        while size {
+            cx16.r0 = cx16.MACPTR(min(255, size) as ubyte, &cx16.VERA_DATA0, true)
+            if_cs {
+                ; no MACPTR support
+                repeat size
+                    cx16.VERA_DATA0 = cbm.CHRIN()
+                return
+            }
+            size -= cx16.r0
+        }
+    }
+
+    sub write_header() -> bool {
+        ; save out the BMX header to the currently active output file
+        build_header()
+        for cx16.r0L in 0 to sizeof(header)-1 {
+            cbm.CHROUT(header[cx16.r0L])
+        }
+        return not cbm.READST()
+    }
+
+    sub write_palette() -> bool {
+        ; write palette data to the currently active input file
+        ; if palette_buffer_ptr is not 0, the palette data is read from that memory buffer,
+        ; otherwise it is read directly from the palette in vram.
+        cx16.r3 = palette_buffer_ptr
+        cx16.r2L = lsb(palette_entries)
+        cx16.vaddr(1, $fa00+palette_start*2, 0, 1)
+        do {
+            if cx16.r3 {
+                cbm.CHROUT(@(cx16.r3))      ; from memory
+                cx16.r3++
+                cbm.CHROUT(@(cx16.r3))
+                cx16.r3++
+            } else {
+                cbm.CHROUT(cx16.VERA_DATA0) ; from vram
+                cbm.CHROUT(cx16.VERA_DATA0)
+            }
+            cx16.r2L--
+        } until cx16.r2L==0
+        return not cbm.READST()
+    }
+
+    sub write_bitmap(ubyte vbank, uword vaddr, uword screenwidth) -> bool {
+        ; screenwidth=0: save contiguous bitmap from vram to the currently active output file
+        ; screenwidth>0: save "stamp" bitmap from vram to the currently active output file
+        cx16.vaddr(vbank, vaddr, 0, 1)
+        cx16.r3 = bytes_per_scanline(width)         ; num bytes per image scanline
+        cx16.r2 = 0
+        if screenwidth
+            cx16.r2 = bytes_per_scanline(screenwidth-width)   ; num bytes padding per screen scanline
+        repeat height {
+            write_scanline(cx16.r3)
+            repeat cx16.r2
+                cx16.r0L = cx16.VERA_DATA0
+        }
+        return not cbm.READST()
+
+        sub write_scanline(uword size) {
+            while size {
+                cx16.r0L = lsb(size)
+                if msb(size)
+                    cx16.r0L = 0        ; 256 bytes
+                cx16.r0 = cx16.MCIOUT(cx16.r0L, &cx16.VERA_DATA0, true)
+                if_cs {
+                    ; no MCIOUT support
+                    repeat size
+                        cbm.CHROUT(cx16.VERA_DATA0)
+                    return
+                }
+                size -= cx16.r0
+            }
+        }
+    }
+
+    sub bytes_per_scanline(uword w) -> uword {
+        ubyte[4] shifts = [3,2,1,0]
+        return w >> shifts[vera_colordepth]
+    }
+
+    sub parse_header() -> bool {
+        if header[0]==FILEID[0] and header[1]==FILEID[1] and header[2]==FILEID[2] {
+            if header[3]==1 {       ; only version 1 supported for now
+                bitsperpixel = header[4]
+                vera_colordepth = header[5]
+                width = peekw(&header+6)
+                height = peekw(&header+8)
+                palette_entries = header[10]
+                if palette_entries==0
+                    palette_entries=256
+                palette_start = header[11]
+                ; the data offset is not needed:  data_offset = peekw(&header+12)
+                compression = header[14]
+                border = header[15]
+                return true
+            } else
+                error_message = "unsupported bmx file version"
+        } else
+            error_message = "invalid bmx file"
+
+        return false
+    }
+
+    sub build_header() {
+        ; build the internal BMX header structure
+        ; normally you don't have to call this yourself
+        sys.memset(header, sizeof(header), 0)
+        uword data_offset = palette_entries*$0002
+        data_offset += sizeof(header)
+        header[0] = FILEID[0]
+        header[1] = FILEID[1]
+        header[2] = FILEID[2]
+        header[3] = 1        ; version 1
+        header[4] = bitsperpixel
+        header[5] = vera_colordepth
+        header[6] = lsb(width)
+        header[7] = msb(width)
+        header[8] = lsb(height)
+        header[9] = msb(height)
+        header[10] = lsb(palette_entries)
+        header[11] = palette_start
+        header[12] = lsb(data_offset)
+        header[13] = msb(data_offset)
+        header[14] = compression
+        header[15] = border
+    }
+}

compiler/res/prog8lib/cx16/diskio.p8

@@ -1,21 +1,28 @@
 ; Commander X16 disk drive I/O routines.
 ; Largely compatible with the default C64 ones, but adds more stuff specific to the X16 as well.
 
+; NOTE: If you experience weird behavior with these routines and you are using them
+;       in the X16 emulator using HostFs, please try again with an SD-card image instead first.
+;       It is possible that there are still small differences between HostFS and actual CBM DOS in the emulator.
+;
+; About the secondary addresses:
+; for writes (new files or overwrites), you can use 1 (without the mode string) or 2-14 (with the mode string)
+; for reads (existing files) you can use 0 or 2-14 (mode string is optional)
+; for modify mode (new files* or existing files), you must use 2-14, and the mode string ,s,m is required
+
+
 %import textio
+%import conv
 %import string
 %import syslib
 
 diskio {
-    %option no_symbol_prefixing
+    %option no_symbol_prefixing, ignore_unused
 
     const ubyte READ_IO_CHANNEL=12
     const ubyte WRITE_IO_CHANNEL=13
 
-    ubyte drivenumber = 8
-
-    sub set_drive(ubyte number) {
-        drivenumber = number
-    }
+    ubyte @shared drivenumber = 8           ; user programs can set this to the drive number they want to load/save to!
 
     sub reset_read_channel() {
         void cbm.CHKIN(READ_IO_CHANNEL)
@@ -35,8 +42,6 @@ diskio {
         if_cs
             goto io_error
         reset_read_channel()
-        if_cs
-            goto io_error
 
         repeat 4 {
             void cbm.CHRIN()     ; skip the 4 prologue bytes
@@ -87,25 +92,23 @@ io_error:
     sub diskname() -> uword {
         ; returns disk label name or 0 if error
         cbm.SETNAM(3, "$")
-        cbm.SETLFS(READ_IO_CHANNEL, diskio.drivenumber, 0)
+        cbm.SETLFS(READ_IO_CHANNEL, drivenumber, 0)
         ubyte status = 1
         void cbm.OPEN()          ; open 12,8,0,"$=c"
         if_cs
             goto io_error
         reset_read_channel()
-        if_cs
-            goto io_error
 
         while cbm.CHRIN()!='"' {
             ; skip up to entry name
         }
 
-        cx16.r0 = &diskio.list_filename
+        cx16.r0 = &list_filename
         repeat {
             @(cx16.r0) = cbm.CHRIN()
             if @(cx16.r0)=='"' {
                 @(cx16.r0) = ' '
-                while @(cx16.r0)==' ' and cx16.r0>=&diskio.list_filename {
+                while @(cx16.r0)==' ' and cx16.r0>=&list_filename {
                     @(cx16.r0) = 0
                     cx16.r0--
                 }
@@ -120,7 +123,7 @@ io_error:
         cbm.CLOSE(READ_IO_CHANNEL)
         if status and status & $40 == 0
             return 0
-        return diskio.list_filename
+        return list_filename
     }
 
     ; internal variables for the iterative file lister / loader
@@ -135,12 +138,13 @@ io_error:
 
     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.
+        ;    Files in the buffer are separated 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
+        filenames_buffer[0]=0
         if lf_start_list(pattern_ptr) {
             while lf_next_entry() {
                 if list_filetype!="dir" {
@@ -177,8 +181,6 @@ io_error:
         if_cs
             goto io_error
         reset_read_channel()
-        if_cs
-            goto io_error
 
         repeat 4 {
             void cbm.CHRIN()     ; skip the 4 prologue bytes
@@ -188,6 +190,7 @@ io_error:
             return true
 
 io_error:
+        cbm.CLOSE(READ_IO_CHANNEL)
         lf_end_list()
         return false
     }
@@ -270,30 +273,33 @@ close_end:
 
     ; ----- iterative file loader functions (uses the read io channel) -----
 
-    sub f_open(uword filenameptr) -> bool {
+    sub f_open(str filenameptr) -> bool {
         ; -- open a file for iterative reading with f_read
         ;    note: only a single iteration loop can be active at a time!
+        ;    Returns true if the file is successfully opened and readable.
+        ;    No need to check status(), unlike f_open_w() !
+        ;    NOTE: the default input isn't yet set to this logical file, you must use reset_read_channel() to do this,
+        ;          if you're going to read from it yourself instead of using f_read()!
         f_close()
 
         cbm.SETNAM(string.length(filenameptr), filenameptr)
         cbm.SETLFS(READ_IO_CHANNEL, drivenumber, READ_IO_CHANNEL)     ; note: has to be Channel,x,Channel because otherwise f_seek doesn't work
         void cbm.OPEN()          ; open 12,8,12,"filename"
         if_cc {
+            reset_read_channel()
             if cbm.READST()==0 {
                 iteration_in_progress = true
-                reset_read_channel()
-                if_cc {
-                    void cbm.CHRIN()        ; read first byte to test for file not found
-                    if not cbm.READST() {
-                        cbm.CLOSE(READ_IO_CHANNEL)           ; close file because we already consumed first byte
-                        void cbm.OPEN()         ; re-open the file
-                        cbm.CLRCHN()            ; reset default i/o channels
-                        return true
-                    }
+                void cbm.CHRIN()        ; read first byte to test for file not found
+                if cbm.READST()==0 {
+                    cbm.CLOSE(READ_IO_CHANNEL)    ; close file because we already consumed first byte
+                    void cbm.OPEN()         ; re-open the file
+                    cbm.CLRCHN()            ; reset default i/o channels
+                    return true
                 }
             }
         }
         f_close()
+        cbm.CLOSE(READ_IO_CHANNEL)
         return false
     }
 
@@ -308,13 +314,12 @@ close_end:
         reset_read_channel()
         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 readsize
         while num_bytes {
             readsize = 255
             if num_bytes<readsize
                 readsize = num_bytes
-            readsize = cx16.MACPTR(lsb(readsize), bufferpointer, false)
+            readsize = cx16.MACPTR(lsb(readsize), bufferpointer, false)     ; fast block reads
             if_cs
                 goto byte_read_loop     ; MACPTR block read not supported, do fallback loop
             list_blocks += readsize
@@ -412,15 +417,36 @@ _end        rts
 
     ; ----- iterative file writing functions (uses write io channel) -----
 
-    sub f_open_w(uword filenameptr) -> bool {
-        ; -- open a file for iterative writing with f_write
-        f_close_w()
+    sub f_open_w(str filename) -> bool {
+        ; -- Open a file for iterative writing with f_write
+        ;    WARNING: returns true if the open command was received by the device,
+        ;    but this can still mean the file wasn't successfully opened for writing!
+        ;    (for example, if it already exists). This is different than f_open()!
+        ;    To be 100% sure if this call was successful, you have to use status()
+        ;    and check the drive's status message!
+        return internal_f_open_w(filename, false)
+    }
 
-        cbm.SETNAM(string.length(filenameptr), filenameptr)
-        cbm.SETLFS(WRITE_IO_CHANNEL, drivenumber, 1)
-        void cbm.OPEN()             ; open 13,8,1,"filename"
-        if_cc
+    sub f_open_w_seek(str filename) -> bool {
+        ; -- Open an existing file for iterative writing with f_write, and seeking with f_seek_w.
+        return internal_f_open_w(filename, true)
+    }
+
+    sub internal_f_open_w(str filename, bool open_for_seeks) -> bool {
+        f_close_w()
+        list_filename = filename
+        str modifier = ",s,?"
+        modifier[3] = 'w'
+        if open_for_seeks
+            modifier[3] = 'm'
+        cx16.r0L = string.append(list_filename, modifier)   ; secondary 13 requires a mode suffix to signal we're writing/modifying
+        cbm.SETNAM(cx16.r0L, list_filename)
+        cbm.SETLFS(WRITE_IO_CHANNEL, drivenumber, WRITE_IO_CHANNEL)
+        void cbm.OPEN()             ; open 13,8,13,"filename"
+        if_cc {
             return not cbm.READST()
+        }
+        cbm.CLOSE(WRITE_IO_CHANNEL)
         f_close_w()
         return false
     }
@@ -429,6 +455,21 @@ _end        rts
         ; -- write the given number of bytes to the currently open file
         if num_bytes!=0 {
             reset_write_channel()
+            do {
+                cx16.r0 = cx16.MCIOUT(lsb(num_bytes), bufferpointer, false)     ; fast block writes
+                if_cs
+                    goto no_mciout
+                num_bytes -= cx16.r0
+                bufferpointer += cx16.r0
+                if msb(bufferpointer) == $c0
+                    bufferpointer = mkword($a0, lsb(bufferpointer))  ; wrap over bank boundary
+                if cbm.READST()!=0
+                    return false
+            } until num_bytes==0
+            return not cbm.READST()
+
+no_mciout:
+            ; the device doesn't support MCIOUT, use a normal per-byte write loop
             repeat num_bytes {
                 cbm.CHROUT(@(bufferpointer))
                 bufferpointer++
@@ -449,6 +490,14 @@ _end        rts
 
     sub status() -> uword {
         ; -- retrieve the disk drive's current status message
+        str device_not_present_error = "device not present #xx"
+        if cbm.READST()==128 {
+            device_not_present_error[len(device_not_present_error)-2] = 0
+            conv.str_ub(drivenumber)
+            void string.copy(conv.string_out, &device_not_present_error+len(device_not_present_error)-2)
+            return device_not_present_error
+        }
+
         uword messageptr = &list_filename
         cbm.SETNAM(0, list_filename)
         cbm.SETLFS(15, drivenumber, 15)
@@ -456,8 +505,6 @@ _end        rts
         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()
@@ -474,7 +521,8 @@ done:
         return list_filename
 
 io_error:
-        list_filename = "?disk error"
+        cbm.CLOSE(15)
+        list_filename = "io error"
         goto done
     }
 
@@ -516,9 +564,6 @@ io_error:
         if_cc
             cx16.r0L = cbm.READST()==0
 
-        cbm.CLRCHN()
-        cbm.CLOSE(1)
-
         return cx16.r0L
     }
 
@@ -567,8 +612,6 @@ io_error:
 +
         }}
 
-        cbm.CLRCHN()
-        cbm.CLOSE(1)
         return cx16.r1
     }
 
@@ -703,13 +746,11 @@ internal_vload:
         cx16.r12 = reversebuffer + MAX_PATH_LEN-1
         @(cx16.r12)=0
         cbm.SETNAM(3, "$=c")
-        cbm.SETLFS(READ_IO_CHANNEL, diskio.drivenumber, 0)
+        cbm.SETLFS(READ_IO_CHANNEL, drivenumber, 0)
         void cbm.OPEN()          ; open 12,8,0,"$=c"
         if_cs
             goto io_error
         reset_read_channel()
-        if_cs
-            goto io_error
 
         repeat 6 {
             void cbm.CHRIN()
@@ -777,7 +818,7 @@ io_error:
     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] = READ_IO_CHANNEL       ; f_open uses the read io channel
+        command[1] = READ_IO_CHANNEL       ; f_open uses this secondary address
         command[2] = lsb(pos_loword)
         command[3] = msb(pos_loword)
         command[4] = lsb(pos_hiword)
@@ -791,15 +832,18 @@ io_error:
     }
 
 
-    ; 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] = WRITE_IO_CHANNEL       ; f_open_w uses the write io channel
-;        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
-;    }
+    sub f_seek_w(uword pos_hiword, uword pos_loword) {
+        ; -- seek in the output file opened with f_open_w_seek, to the given 32-bits position
+        diskio.f_seek.command[1] = WRITE_IO_CHANNEL       ; f_open_w uses this secondary address
+        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)
+        cbm.SETNAM(sizeof(diskio.f_seek.command), &diskio.f_seek.command)
+        cbm.SETLFS(15, drivenumber, 15)
+        void cbm.OPEN()
+        cbm.CLOSE(15)
+        reset_write_channel()    ; back to the write io channel
+    }
 
 }

compiler/res/prog8lib/cx16/emudbg.p8

@@ -1,7 +1,8 @@
 ; Emulator debug interface.
-; Docs: https://github.com/X16Community/x16-emulator#debug-io-registers
+; Docs: https://github.com/X16Community/x16-emulator#emulator-io-registers
 
 emudbg {
+    %option ignore_unused
 
     const uword EMU_BASE = $9fb0
 
@@ -13,46 +14,52 @@ emudbg {
     &ubyte EMU_RECORD_GIF         = EMU_BASE + 5
     &ubyte EMU_RECORD_WAV         = EMU_BASE + 6
     &ubyte EMU_CMDKEYS_DISABLED   = EMU_BASE + 7
-    &ubyte EMU_CPUCLK_L           = EMU_BASE + 8     ; write: reset cpu clock to 0
-    &ubyte EMU_CPUCLK_M           = EMU_BASE + 9     ; write: outputs "User debug 1: $xx"
-    &ubyte EMU_CPUCLK_H           = EMU_BASE + 10    ; write: outputs "User debug 2: $xx"
-    &ubyte EMU_CPUCLK_U           = EMU_BASE + 11    ; write: outputs as character to console
+    &ubyte EMU_CPUCLK_L           = EMU_BASE + 8
+    &ubyte EMU_CPUCLK_M           = EMU_BASE + 9
+    &ubyte EMU_CPUCLK_H           = EMU_BASE + 10
+    &ubyte EMU_CPUCLK_U           = EMU_BASE + 11
+    &ubyte EMU_CPUCLK_RESET       = EMU_BASE + 8     ; write: reset cpu clock to 0
+    &ubyte EMU_DBGOUT1            = EMU_BASE + 9     ; write: outputs "User debug 1: $xx"
+    &ubyte EMU_DBGOUT2            = EMU_BASE + 10    ; write: outputs "User debug 2: $xx"
+    &ubyte EMU_CHROUT             = EMU_BASE + 11    ; write: outputs as character to console
     ; 12 is unused for now
     &ubyte EMU_KEYMAP             = EMU_BASE + 13
     &ubyte EMU_EMU_DETECT1        = EMU_BASE + 14
     &ubyte EMU_EMU_DETECT2        = EMU_BASE + 15
 
     sub is_emulator() -> bool {
+        ; Test for emulator presence.
+        ; It is recommended to only use the debug registers if this returns true.
         return EMU_EMU_DETECT1=='1' and EMU_EMU_DETECT2=='6'
     }
 
-    asmsub console_write(str isoString @R0) clobbers(Y) {
-        %asm {{
-            ldy  #0
--           lda  (cx16.r0),y
-            beq  +
-            sta  p8_EMU_CPUCLK_U
-            iny
-            bne  -
-+
-        }}
+    sub console_write(str isoString) {
+        ; note: make sure the text is in Iso encoding.
+        if is_emulator() {
+            ubyte chr
+            repeat {
+                chr = @(isoString)
+                if_z
+                    break
+                EMU_CPUCLK_U = chr
+                isoString++
+            }
+        }
     }
 
-    asmsub console_chrout(ubyte char @A) {
-        %asm {{
-            sta  p8_EMU_CPUCLK_U
-        }}
+    sub console_chrout(ubyte char) {
+        ; note: make sure the character is in Iso encoding.
+        if is_emulator()
+            EMU_CHROUT = char
     }
 
-    asmsub console_value1(ubyte value @A) {
-        %asm {{
-            sta  p8_EMU_CPUCLK_M
-        }}
+    sub console_value1(ubyte value) {
+        if is_emulator()
+            EMU_DBGOUT1 = value
     }
 
-    asmsub console_value2(ubyte value @A) {
-        %asm {{
-            sta  p8_EMU_CPUCLK_H
-        }}
+    sub console_value2(ubyte value) {
+        if is_emulator()
+            EMU_DBGOUT2 = value
     }
 }

compiler/res/prog8lib/cx16/floats.p8

@@ -1,15 +1,14 @@
 ; Prog8 definitions for floating point handling on the CommanderX16
 
-%option enable_floats
+%option enable_floats, no_symbol_prefixing, ignore_unused
 %import floats_functions
 
 floats {
 	; ---- this block contains C-64 compatible floating point related functions ----
 
-    %option no_symbol_prefixing
-
-        const float  PI     = 3.141592653589793
-        const float  TWOPI  = 6.283185307179586
+        const float  π      = 3.141592653589793
+        const float  PI     = π
+        const float  TWOPI  = 2*π
 
 
 ; ---- ROM float functions (same as on C128 except base page) ----
@@ -28,7 +27,7 @@ romsub $fe00 = AYINT() clobbers(A,X,Y)          ; fac1-> signed word in 100-101
 romsub $fe03 = GIVAYF(ubyte lo @ Y, ubyte hi @ A) clobbers(A,X,Y)
 
 romsub $fe06 = FOUT() clobbers(X) -> uword @ AY             ; fac1 -> string, address returned in AY
-; romsub $fe09 = VAL_1() clobbers(A,X,Y)                      ; convert ASCII string to floating point [not yet implemented!!!] see parse_f() instead
+romsub $fe09 = VAL_1(uword string @XY, ubyte length @A) clobbers(A,X,Y) -> float @FAC1      ; convert ASCII string in XY and length in A, to floating point in FAC1. WARNING: only implemented in ROM 47+. Safer to use floats.parse() instead.
 
 ; GETADR: 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)
@@ -71,10 +70,10 @@ romsub $fe6c = MOVAF() clobbers(A,X)                        ; copy fac1 to fac2
 ; X16 additions
 romsub $fe6f = FADDH() clobbers(A,X,Y)                      ; fac1 += 0.5, for rounding- call this before INT
 romsub $fe72 = ZEROFC() clobbers(A,X,Y)                     ; fac1 = 0
-romsub $fe75 = NORMAL() clobbers(A,X,Y)                     ; normalize fac1 (?)
-romsub $fe78 = NEGFAC() clobbers(A)                         ; switch the sign of fac1 (fac1 = -fac1) (juse use NEGOP() instead!)
+romsub $fe75 = NORMAL() clobbers(A,X,Y)                     ; normalize fac1
+romsub $fe78 = NEGFAC() clobbers(A)                         ; switch the sign of fac1 (fac1 = -fac1) (doesn't work, juse use NEGOP() instead!)
 romsub $fe7b = MUL10() clobbers(A,X,Y)                      ; fac1 *= 10
-romsub $fe7e = DIV10() clobbers(A,X,Y)                      ; fac1 /= 10 , CAUTION: result is always positive!
+romsub $fe7e = DIV10() clobbers(A,X,Y)                      ; fac1 /= 10 , CAUTION: result is always positive! Have to restore sign manually!
 romsub $fe81 = MOVEF() clobbers(A,X)                        ; copy fac1 to fac2
 romsub $fe84 = SGN() clobbers(A,X,Y)                        ; fac1 = SGN(fac1), result of SIGN (-1, 0 or 1)
 romsub $fe87 = FLOAT() clobbers(A,X,Y)                      ; FAC = (s8).A
@@ -136,29 +135,59 @@ asmsub  FREADUY (ubyte value @Y) {
     }}
 }
 
-asmsub parse_f(str value @AY) -> float @FAC1 {
+asmsub parse(str value @AY) -> float @FAC1 {
     ; -- parse a string value of a number to float in FAC1
-    ;    warning: uses an internal BASIC routine that may be rom version dependent
-    ;    ($ddf2 is inside the routine for VAL at $ddef)
+    ;    warning: on older <R47 kernals it uses an internal BASIC routine that is ROM version dependent,
+    ;    ($deb6 is inside the routine for VAL at $deb3)  See basic.sym from x16-rom
+    ;    TODO once ROM v47 is released, all the workarounds here can be removed. But probably keep the kernal VAL_1 existance check
     %asm {{
-        sta  $a9
-        sty  $aa
+        ldx  VAL_1
+        cpx  #$4c       ; is there an implementation in VAL_1? (test for JMP)
+        bne  +          ; no, do it ourselves
+        pha             ; yes, count the length and call rom VAL_1.
+        phy
         jsr  prog8_lib.strlen
         tya
-        jmp  $ddf2
+        ply
+        plx
+        jmp  VAL_1
++       sta  $a9    ; 'index' variable
+        sty  $aa
+        jsr  prog8_lib.strlen
+        lda  $deb6
+        cmp  #$d0   ; sanity check for kernal routine correct
+        bne  +
+        tya
+        jmp  $deb6   ; kernal version dependent...
++       ; print error message if routine is borked in kernal, and exit program
+        ldy  #0
+-       lda  _msg,y
+        beq  +
+        jsr  cbm.CHROUT
+        iny
+        bne  -
++       jmp  sys.exit
+
+_msg    .text 13,"?val kaputt",13,0
     }}
 }
 
 
 &uword AYINT_facmo = $c6      ; $c6/$c7 contain result of AYINT
 
-sub rndf() -> float {
+sub rnd() -> float {
     %asm {{
         lda  #1
         jmp  RND_0
     }}
 }
 
+asmsub normalize(float value @FAC1) -> float @ FAC1 {
+    %asm {{
+        jmp  floats.NORMAL
+    }}
+}
+
 %asminclude "library:c64/floats.asm"
 %asminclude "library:c64/floats_funcs.asm"
 

compiler/res/prog8lib/cx16/gfx2.p8

@@ -10,18 +10,21 @@
 ;       If you're memory constrained you should probably not use this built-in library,
 ;       but make a copy in your project only containing the code for the required resolution.
 ;
+; NOTE: For sake of speed, NO BOUNDS CHECKING is performed in most routines!
+;       You'll have to make sure yourself that you're not writing outside of bitmap boundaries!
+;
 ;
 ; SCREEN MODE LIST:
 ;   mode 0 = reset back to default text mode
 ;   mode 1 = bitmap 320 x 240 x 256c (8 bpp)
-;   mode 2 = bitmap 640 x 480 x 4c (2 bpp. there's not enough vram for more colors in hires mode.)
-;   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 4c (not yet implemented: just use 256c, there's enough vram for that)
-;   higher color dephts in highres are not supported due to lack of VRAM
+;   mode 2 = bitmap 640 x 480 x 4c (2 bpp)
+;   mode 3 = bitmap 320 x 240 x 4c (not yet implemented: just use 256c, there's enough vram for that)
+;   mode 4 = bitmap 320 x 240 x 16c (not yet implemented: just use 256c, there's enough vram for that)
+;   mode 5 = bitmap 640 x 400 x 16c (not yet implemented)
 
 gfx2 {
 
-    %option no_symbol_prefixing
+    %option ignore_unused
 
     ; read-only control variables:
     ubyte active_mode = 0
@@ -40,9 +43,6 @@ gfx2 {
                 cx16.VERA_L1_CONFIG = %00000111
                 cx16.VERA_L1_MAPBASE = 0
                 cx16.VERA_L1_TILEBASE = 0
-                width = 320
-                height = 240
-                bpp = 8
             }
             2 -> {
                 ; highres 4c
@@ -52,27 +52,44 @@ gfx2 {
                 cx16.VERA_L1_CONFIG = %00000101
                 cx16.VERA_L1_MAPBASE = 0
                 cx16.VERA_L1_TILEBASE = %00000001
-                width = 640
-                height = 480
-                bpp = 2
             }
             else -> {
                 ; back to default text mode
                 cx16.r15L = cx16.VERA_DC_VIDEO & %00000111 ; retain chroma + output mode
                 cbm.CINT()
                 cx16.VERA_DC_VIDEO = (cx16.VERA_DC_VIDEO & %11111000) | cx16.r15L
-                width = 0
-                height = 0
-                bpp = 0
-                mode = 0
             }
         }
 
-        active_mode = mode
-        if bpp
+        init_mode(mode)
+        if active_mode
             clear_screen(0)
     }
 
+    sub init_mode(ubyte mode) {
+        ; set the internal configuration variables corresponding to the given screenmode
+        ; doesn't manipulate Vera / the actual display mode
+        active_mode = mode
+        when mode {
+            1 -> {
+                width = 320
+                height = 240
+                bpp = 8
+            }
+            2 -> {
+                width = 640
+                height = 480
+                bpp = 2
+            }
+            else -> {
+                width = 0
+                height = 0
+                bpp = 0
+                active_mode = 0
+            }
+        }
+    }
+
     sub clear_screen(ubyte color) {
         position(0, 0)
         when active_mode {
@@ -124,8 +141,8 @@ gfx2 {
                 ; lores 256c
                 position(xx, yy)
                 %asm {{
-                    lda  color
-                    ldx  length+1
+                    lda  p8v_color
+                    ldx  p8v_length+1
                     beq  +
                     ldy  #0
 -                   sta  cx16.VERA_DATA0
@@ -133,7 +150,7 @@ gfx2 {
                     bne  -
                     dex
                     bne  -
-+                   ldy  length     ; remaining
++                   ldy  p8v_length     ; remaining
                     beq  +
 -                   sta  cx16.VERA_DATA0
                     dey
@@ -163,7 +180,7 @@ gfx2 {
                     sta  cx16.VERA_ADDR_L
                     lda  cx16.r0+1
                     sta  cx16.VERA_ADDR_M
-                    ldx  xx
+                    ldx  p8v_xx
                 }}
 
                 repeat length {
@@ -172,8 +189,8 @@ gfx2 {
                         and  #3
                         tay
                         lda  cx16.VERA_DATA0
-                        and  gfx2.plot.mask4c,y
-                        ora  colorbits,y
+                        and  p8b_gfx2.p8s_plot.p8v_mask4c,y
+                        ora  p8v_colorbits,y
                         sta  cx16.VERA_DATA0
                         cpy  #%00000011         ; next vera byte?
                         bne  ++
@@ -189,6 +206,24 @@ gfx2 {
         }
     }
 
+    sub safe_horizontal_line(uword xx, uword yy, uword length, ubyte color) {
+        ; does bounds checking and clipping
+        if msb(yy)&$80!=0 or yy>=height
+            return
+        if msb(xx)&$80!=0 {
+            length += xx
+            xx = 0
+        }
+        if xx>=width
+            return
+        if xx+length>width
+            length = width-xx
+        if length>width
+            return
+
+        horizontal_line(xx, yy, length, color)
+    }
+
     sub vertical_line(uword xx, uword yy, uword lheight, ubyte color) {
         when active_mode {
             1 -> {
@@ -197,9 +232,9 @@ gfx2 {
                 position(xx,yy)
                 cx16.VERA_ADDR_H = cx16.VERA_ADDR_H & %00000111 | (14<<4)
                 %asm {{
-                    ldy  lheight
+                    ldy  p8v_lheight
                     beq  +
-                    lda  color
+                    lda  p8v_color
 -                   sta  cx16.VERA_DATA0
                     dey
                     bne  -
@@ -225,8 +260,8 @@ gfx2 {
                 repeat lheight {
                     %asm {{
                         lda  cx16.VERA_DATA0
-                        and  mask
-                        ora  color
+                        and  p8v_mask
+                        ora  p8v_color
                         sta  cx16.VERA_DATA1
                     }}
                 }
@@ -336,6 +371,7 @@ gfx2 {
     }
 
     sub circle(uword @zp xcenter, uword @zp ycenter, ubyte radius, ubyte color) {
+        ; Warning: NO BOUNDS CHECKS. Make sure circle fits in the screen.
         ; Midpoint algorithm.
         if radius==0
             return
@@ -349,36 +385,91 @@ gfx2 {
         while xx>=yy {
             cx16.r14 = xcenter + xx
             cx16.r15 = ycenter + yy
-            plot(cx16.r14, cx16.r15, color)
+            plotq()
             cx16.r14 = xcenter - xx
-            plot(cx16.r14, cx16.r15, color)
+            plotq()
             cx16.r14 = xcenter + xx
             cx16.r15 = ycenter - yy
-            plot(cx16.r14, cx16.r15, color)
+            plotq()
             cx16.r14 = xcenter - xx
-            plot(cx16.r14, cx16.r15, color)
+            plotq()
             cx16.r14 = xcenter + yy
             cx16.r15 = ycenter + xx
-            plot(cx16.r14, cx16.r15, color)
+            plotq()
             cx16.r14 = xcenter - yy
-            plot(cx16.r14, cx16.r15, color)
+            plotq()
             cx16.r14 = xcenter + yy
             cx16.r15 = ycenter - xx
-            plot(cx16.r14, cx16.r15, color)
+            plotq()
             cx16.r14 = xcenter - yy
-            plot(cx16.r14, cx16.r15, color)
+            plotq()
 
             yy++
-            if decisionOver2<=0
-                decisionOver2 += (yy as word)*2+1
-            else {
+            if decisionOver2>=0 {
                 xx--
-                decisionOver2 += (yy as word -xx)*2+1
+                decisionOver2 -= xx*$0002
             }
+            decisionOver2 += yy*$0002
+            decisionOver2++
+        }
+
+        sub plotq() {
+            ; cx16.r14 = x, cx16.r15 = y, color=color.
+            plot(cx16.r14, cx16.r15, color)
+        }
+    }
+
+    sub safe_circle(uword @zp xcenter, uword @zp ycenter, ubyte radius, ubyte color) {
+        ; This version does bounds checks and clipping, but is a lot slower.
+        ; Midpoint algorithm.
+        if radius==0
+            return
+
+        ubyte @zp xx = radius
+        ubyte @zp yy = 0
+        word @zp decisionOver2 = (1 as word)-xx
+        ; R14 = internal plot X
+        ; R15 = internal plot Y
+
+        while xx>=yy {
+            cx16.r14 = xcenter + xx
+            cx16.r15 = ycenter + yy
+            plotq()
+            cx16.r14 = xcenter - xx
+            plotq()
+            cx16.r14 = xcenter + xx
+            cx16.r15 = ycenter - yy
+            plotq()
+            cx16.r14 = xcenter - xx
+            plotq()
+            cx16.r14 = xcenter + yy
+            cx16.r15 = ycenter + xx
+            plotq()
+            cx16.r14 = xcenter - yy
+            plotq()
+            cx16.r14 = xcenter + yy
+            cx16.r15 = ycenter - xx
+            plotq()
+            cx16.r14 = xcenter - yy
+            plotq()
+
+            yy++
+            if decisionOver2>=0 {
+                xx--
+                decisionOver2 -= xx*$0002
+            }
+            decisionOver2 += yy*$0002
+            decisionOver2++
+        }
+
+        sub plotq() {
+            ; cx16.r14 = x, cx16.r15 = y, color=color.
+            safe_plot(cx16.r14, cx16.r15, color)
         }
     }
 
     sub disc(uword @zp xcenter, uword @zp ycenter, ubyte @zp radius, ubyte color) {
+        ; Warning: NO BOUNDS CHECKS. Make sure circle fits in the screen.
         ; Midpoint algorithm, filled
         if radius==0
             return
@@ -391,12 +482,35 @@ gfx2 {
             horizontal_line(xcenter-yy, ycenter+radius, yy*$0002+1, color)
             horizontal_line(xcenter-yy, ycenter-radius, yy*$0002+1, color)
             yy++
-            if decisionOver2<=0
-                decisionOver2 += (yy as word)*2+1
-            else {
+            if decisionOver2>=0 {
                 radius--
-                decisionOver2 += (yy as word -radius)*2+1
+                decisionOver2 -= radius*$0002
             }
+            decisionOver2 += yy*$0002
+            decisionOver2++
+        }
+    }
+
+    sub safe_disc(uword @zp xcenter, uword @zp ycenter, ubyte @zp radius, ubyte color) {
+        ; This version does bounds checks and clipping, but is a lot slower.
+        ; Midpoint algorithm, filled
+        if radius==0
+            return
+        ubyte @zp yy = 0
+        word @zp decisionOver2 = (1 as word)-radius
+
+        while radius>=yy {
+            safe_horizontal_line(xcenter-radius, ycenter+yy, radius*$0002+1, color)
+            safe_horizontal_line(xcenter-radius, ycenter-yy, radius*$0002+1, color)
+            safe_horizontal_line(xcenter-yy, ycenter+radius, yy*$0002+1, color)
+            safe_horizontal_line(xcenter-yy, ycenter-radius, yy*$0002+1, color)
+            yy++
+            if decisionOver2>=0 {
+                radius--
+                decisionOver2 -= radius*$0002
+            }
+            decisionOver2 += yy*$0002
+            decisionOver2++
         }
     }
 
@@ -420,7 +534,7 @@ gfx2 {
                     sta  cx16.VERA_ADDR_M
                     lda  cx16.r0
                     sta  cx16.VERA_ADDR_L
-                    lda  color
+                    lda  p8v_color
                     sta  cx16.VERA_DATA0
                 }}
             }
@@ -444,15 +558,24 @@ gfx2 {
                     lda  cx16.r0L
                     sta  cx16.VERA_ADDR_L
                     ldy  cx16.r2L           ; xbits
-                    lda  mask4c,y
+                    lda  p8v_mask4c,y
                     and  cx16.VERA_DATA0
-                    ora  color
+                    ora  p8v_color
                     sta  cx16.VERA_DATA0
                 }}
             }
         }
     }
 
+    sub safe_plot(uword xx, uword yy, ubyte color) {
+        ; A plot that does bounds checks to see if the pixel is inside the screen.
+        if msb(xx)&$80!=0 or msb(yy)&$80!=0
+            return
+        if xx >= width or yy >= height
+            return
+        plot(xx, yy, color)
+    }
+
     sub pget(uword @zp xx, uword yy) -> ubyte {
         when active_mode {
             1 -> {
@@ -481,21 +604,32 @@ gfx2 {
                     lda  cx16.r0L
                     sta  cx16.VERA_ADDR_L
                     lda  cx16.VERA_DATA0
-                    sta  cx16.r0L
+                    pha
+                    lda  p8v_xx
+                    and  #3
+                    tay
+                    lda  p8b_gfx2.p8s_plot.p8v_shift4c,y
+                    tay
+                    pla
+                    cpy  #0
+                    beq  +
+-                   lsr  a
+                    dey
+                    bne  -
++                   and #3
                 }}
-                cx16.r1L = lsb(xx) & 3
-                cx16.r0L >>= gfx2.plot.shift4c[cx16.r1L]
-                return cx16.r0L & 3
             }
             else -> return 0
         }
     }
 
-    sub fill(word @zp xx, word @zp yy, ubyte new_color) {
+    sub fill(uword x, uword 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
+        const ubyte MAXDEPTH = 64
+        word @zp xx = x as word
+        word @zp yy = y as word
         word[MAXDEPTH] @split @shared stack_xl
         word[MAXDEPTH] @split @shared stack_xr
         word[MAXDEPTH] @split @shared stack_y
@@ -517,21 +651,21 @@ gfx2 {
 ;;                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
+                    lda  p8v_sxl
+                    sta  p8v_stack_xl_lsb,y
+                    lda  p8v_sxl+1
+                    sta  p8v_stack_xl_msb,y
+                    lda  p8v_sxr
+                    sta  p8v_stack_xr_lsb,y
+                    lda  p8v_sxr+1
+                    sta  p8v_stack_xr_msb,y
+                    lda  p8v_sy
+                    sta  p8v_stack_y_lsb,y
+                    lda  p8v_sy+1
+                    sta  p8v_stack_y_msb,y
                     ldy  cx16.r12L
-                    lda  sdy
-                    sta  stack_dy,y
+                    lda  p8v_sdy
+                    sta  p8v_stack_dy,y
                     inc  cx16.r12L
                 }}
             }
@@ -545,21 +679,21 @@ gfx2 {
             %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  yy
-                lda  stack_y_msb,y
-                sta  yy+1
+                lda  p8v_stack_xl_lsb,y
+                sta  p8v_x1
+                lda  p8v_stack_xl_msb,y
+                sta  p8v_x1+1
+                lda  p8v_stack_xr_lsb,y
+                sta  p8v_x2
+                lda  p8v_stack_xr_msb,y
+                sta  p8v_x2+1
+                lda  p8v_stack_y_lsb,y
+                sta  p8v_yy
+                lda  p8v_stack_y_msb,y
+                sta  p8v_yy+1
                 ldy  cx16.r12L
-                lda  stack_dy,y
-                sta  dy
+                lda  p8v_stack_dy,y
+                sta  p8v_dy
             }}
             yy+=dy
         }
@@ -574,11 +708,15 @@ gfx2 {
         while cx16.r12L {
             pop_stack()
             xx = x1
-            while xx >= 0 and pget(xx as uword, yy as uword) == cx16.r11L {
-                plot(xx as uword, yy as uword, cx16.r10L)
+            ; possible speed optimization: if mode==1 (256c) use vera autodecrement instead of pget(), but code bloat not worth it?
+            while xx >= 0 {
+                if pget(xx as uword, yy as uword) != cx16.r11L
+                    break
                 xx--
             }
-            if xx >= x1
+            if x1!=xx
+                horizontal_line(xx as uword+1, yy as uword, x1-xx as uword, cx16.r10L)
+            else
                 goto skip
 
             left = xx + 1
@@ -587,17 +725,26 @@ gfx2 {
             xx = x1 + 1
 
             do {
-                while xx <= width-1 and pget(xx as uword, yy as uword) == cx16.r11L {
-                    plot(xx as uword, yy as uword, cx16.r10L)
+                cx16.r9 = xx
+                ; possible speed optimization: if mode==1 (256c) use vera autoincrement instead of pget(), but code bloat not worth it?
+                while xx <= width-1 {
+                    if pget(xx as uword, yy as uword) != cx16.r11L
+                        break
                     xx++
                 }
+                if cx16.r9!=xx
+                    horizontal_line(cx16.r9, yy as uword, (xx as uword)-cx16.r9, cx16.r10L)
+
                 push_stack(left, xx - 1, yy, dy)
                 if xx > x2 + 1
                     push_stack(x2 + 1, xx - 1, yy, -dy)
 skip:
                 xx++
-                while xx <= x2 and pget(xx as uword, yy as uword) != cx16.r11L
+                while xx <= x2 {
+                    if pget(xx as uword, yy as uword) == cx16.r11L
+                        break
                     xx++
+                }
                 left = xx
             } until xx>x2
         }
@@ -705,7 +852,7 @@ skip:
                         position(xx,yy)
                         yy++
                         %asm {{
-                            ldx  color
+                            ldx  p8v_color
                             lda  cx16.VERA_DATA1
                             sta  P8ZP_SCRATCH_B1
                             ldy  #8
@@ -789,6 +936,8 @@ skip:
     }
 
     asmsub cs_innerloop640(ubyte color @A) clobbers(Y) {
+        ; using verafx 32 bits writes here would make this faster but it's safer to
+        ; use verafx only explicitly when you know what you're doing.
         %asm {{
             ldy  #80
 -           sta  cx16.VERA_DATA0

compiler/res/prog8lib/cx16/graphics.p8

@@ -7,10 +7,16 @@
 ; Unlike graphics module on the C64, you can use colors() to set new drawing colors for every draw operation.
 ; For other resolutions or other color modes, use the "gfx2" or "monogfx" module instead. (which is Cx16-specific)
 ; Note: there is no color palette manipulation here, you have to do that yourself or use the "palette" module.
+;
+; NOTE: For sake of speed, NO BOUNDS CHECKING is performed in most routines!
+;       You'll have to make sure yourself that you're not writing outside of bitmap boundaries!
+;
 
 
 graphics {
-    %option no_symbol_prefixing
+    %option ignore_unused
+
+    romsub $feff = FB_cursor_position2()  clobbers(A,X,Y)     ; alias for the normal FB_cursor_position() call but reuses existing r0 and r1
 
     const uword WIDTH = 320
     const ubyte HEIGHT = 240
@@ -69,6 +75,8 @@ graphics {
     }
 
     sub circle(uword xcenter, ubyte ycenter, ubyte radius) {
+        ; Warning: NO BOUNDS CHECKS. Make sure circle fits in the screen.
+
         ;cx16.r0 = xcenter - radius/2
         ;cx16.r1 = ycenter - radius/2
         ;cx16.r2 = radius*2
@@ -85,43 +93,47 @@ graphics {
         while xx>=yy {
             cx16.r0 = xcenter + xx
             cx16.r1 = ycenter + yy
-            cx16.FB_cursor_position2()
+            FB_cursor_position2()
             cx16.FB_set_pixel(stroke_color)
             cx16.r0 = xcenter - xx
-            cx16.FB_cursor_position2()
+            FB_cursor_position2()
             cx16.FB_set_pixel(stroke_color)
             cx16.r0 = xcenter + xx
             cx16.r1 = ycenter - yy
-            cx16.FB_cursor_position2()
+            FB_cursor_position2()
             cx16.FB_set_pixel(stroke_color)
             cx16.r0 = xcenter - xx
-            cx16.FB_cursor_position2()
+            FB_cursor_position2()
             cx16.FB_set_pixel(stroke_color)
             cx16.r0 = xcenter + yy
             cx16.r1 = ycenter + xx
-            cx16.FB_cursor_position2()
+            FB_cursor_position2()
             cx16.FB_set_pixel(stroke_color)
             cx16.r0 = xcenter - yy
-            cx16.FB_cursor_position2()
+            FB_cursor_position2()
             cx16.FB_set_pixel(stroke_color)
             cx16.r0 = xcenter + yy
             cx16.r1 = ycenter - xx
-            cx16.FB_cursor_position2()
+            FB_cursor_position2()
             cx16.FB_set_pixel(stroke_color)
             cx16.r0 = xcenter - yy
-            cx16.FB_cursor_position2()
+            FB_cursor_position2()
             cx16.FB_set_pixel(stroke_color)
+
             yy++
-            if decisionOver2<=0 {
-                decisionOver2 += (yy as word)*2+1
-            } else {
+            if decisionOver2>=0 {
                 xx--
-                decisionOver2 += (yy as word -xx)*2+1
+                decisionOver2 -= xx*$0002
             }
+            decisionOver2 += yy*$0002
+            decisionOver2++
         }
     }
 
     sub disc(uword xcenter, ubyte ycenter, ubyte radius) {
+        ; Warning: NO BOUNDS CHECKS. Make sure circle fits in the screen.
+        ; Midpoint algorithm, filled
+
         if radius==0
             return
         ubyte @zp yy = 0
@@ -132,20 +144,21 @@ graphics {
             horizontal_line(xcenter-radius, ycenter-yy, radius*2+1)
             horizontal_line(xcenter-yy, ycenter+radius, yy*2+1)
             horizontal_line(xcenter-yy, ycenter-radius, yy*2+1)
+
             yy++
-            if decisionOver2<=0
-                decisionOver2 += (yy as word)*2+1
-            else {
+            if decisionOver2>=0 {
                 radius--
-                decisionOver2 += (yy as word -radius)*2+1
+                decisionOver2 -= radius*$0002
             }
+            decisionOver2 += yy*$0002
+            decisionOver2++
         }
     }
 
     inline asmsub  plot(uword plotx @R0, uword ploty @R1) clobbers(A, X, Y) {
         %asm {{
             jsr  cx16.FB_cursor_position
-            lda  graphics.stroke_color
+            lda  p8b_graphics.p8v_stroke_color
             jsr  cx16.FB_set_pixel
         }}
     }

compiler/res/prog8lib/cx16/monogfx.p8

@@ -5,10 +5,13 @@
 ; For color bitmap graphics, see the gfx2 library.
 ;
 ; NOTE: a lot of the code here is similar or the same to that in gfx2
+; NOTE: For sake of speed, NO BOUNDS CHECKING is performed in most routines!
+;       You'll have to make sure yourself that you're not writing outside of bitmap boundaries!
+;
 
 monogfx {
 
-    %option no_symbol_prefixing
+    %option ignore_unused
 
     ; read-only control variables:
     uword width = 0
@@ -104,11 +107,11 @@ monogfx {
             ; just use 2 byte writes with shifted mask
             position2(xx,yy,true)
             %asm {{
-                ldy  length
-                lda  masked_ends,y
+                ldy  p8v_length
+                lda  p8v_masked_ends,y
                 sta  cx16.r0L           ; save left byte
                 stz  P8ZP_SCRATCH_B1
-                lda  xx
+                lda  p8v_xx
                 and  #7
                 beq  +
                 tay
@@ -119,10 +122,10 @@ monogfx {
                 bne  -
                 sta  cx16.r0L           ; new left byte
 +
-                lda  dont_stipple_flag
+                lda  p8v_dont_stipple_flag
                 bne  _dontstipple
                 ; determine stipple pattern
-                lda  yy
+                lda  p8v_yy
                 and  #1
                 beq  +
                 lda  #%10101010
@@ -136,7 +139,7 @@ monogfx {
                 and  P8ZP_SCRATCH_REG
                 sta  P8ZP_SCRATCH_B1
 _dontstipple
-                lda  draw
+                lda  p8v_draw
                 beq  _clear
                 lda  cx16.r0L           ; left byte
                 ora  cx16.VERA_DATA1
@@ -181,34 +184,34 @@ _clear
             separate_pixels = lsb(length) & 7
             xx += length & $fff8
             %asm {{
-                lsr  length+1
-                ror  length
-                lsr  length+1
-                ror  length
-                lsr  length+1
-                ror  length
-                lda  draw
+                lsr  p8v_length+1
+                ror  p8v_length
+                lsr  p8v_length+1
+                ror  p8v_length
+                lsr  p8v_length+1
+                ror  p8v_length
+                lda  p8v_draw
                 bne  +
                 ldy  #0     ; black
                 bra  _loop
-+               lda  dont_stipple_flag
++               lda  p8v_dont_stipple_flag
                 beq  _stipple
                 ldy  #255       ; don't stipple
                 bra  _loop
-_stipple        lda  yy
+_stipple        lda  p8v_yy
                 and  #1         ; determine stipple pattern to use
                 bne  +
                 ldy  #%01010101
                 bra  _loop
 +               ldy  #%10101010
-_loop           lda  length
-                ora  length+1
+_loop           lda  p8v_length
+                ora  p8v_length+1
                 beq  _done
                 sty  cx16.VERA_DATA0
-                lda  length
+                lda  p8v_length
                 bne  +
-                dec  length+1
-+               dec  length
+                dec  p8v_length+1
++               dec  p8v_length
                 bra  _loop
 _done
             }}
@@ -229,6 +232,24 @@ _done
         cx16.VERA_ADDR_H &= %00000111   ; vera auto-increment off again
     }
 
+    sub safe_horizontal_line(uword xx, uword yy, uword length, bool draw) {
+        ; does bounds checking and clipping
+        if msb(yy)&$80!=0 or yy>=height
+            return
+        if msb(xx)&$80!=0 {
+            length += xx
+            xx = 0
+        }
+        if xx>=width
+            return
+        if xx+length>width
+            length = width-xx
+        if length>width
+            return
+
+        horizontal_line(xx, yy, length, draw)
+    }
+
     sub vertical_line(uword xx, uword yy, uword lheight, bool draw) {
         cx16.r15L = monogfx.plot.maskbits[xx as ubyte & 7]           ; bitmask
         if draw {
@@ -385,6 +406,7 @@ _done
     }
 
     sub circle(uword @zp xcenter, uword @zp ycenter, ubyte radius, bool draw) {
+        ; Warning: NO BOUNDS CHECKS. Make sure circle fits in the screen.
         ; Midpoint algorithm.
         if radius==0
             return
@@ -398,36 +420,91 @@ _done
         while xx>=yy {
             cx16.r14 = xcenter + xx
             cx16.r15 = ycenter + yy
-            plot(cx16.r14, cx16.r15, draw)
+            plotq()
             cx16.r14 = xcenter - xx
-            plot(cx16.r14, cx16.r15, draw)
+            plotq()
             cx16.r14 = xcenter + xx
             cx16.r15 = ycenter - yy
-            plot(cx16.r14, cx16.r15, draw)
+            plotq()
             cx16.r14 = xcenter - xx
-            plot(cx16.r14, cx16.r15, draw)
+            plotq()
             cx16.r14 = xcenter + yy
             cx16.r15 = ycenter + xx
-            plot(cx16.r14, cx16.r15, draw)
+            plotq()
             cx16.r14 = xcenter - yy
-            plot(cx16.r14, cx16.r15, draw)
+            plotq()
             cx16.r14 = xcenter + yy
             cx16.r15 = ycenter - xx
-            plot(cx16.r14, cx16.r15, draw)
+            plotq()
             cx16.r14 = xcenter - yy
-            plot(cx16.r14, cx16.r15, draw)
+            plotq()
 
             yy++
-            if decisionOver2<=0
-                decisionOver2 += (yy as word)*2+1
-            else {
+            if decisionOver2>=0 {
                 xx--
-                decisionOver2 += (yy as word -xx)*2+1
+                decisionOver2 -= xx*$0002
             }
+            decisionOver2 += yy*$0002
+            decisionOver2++
+        }
+
+        sub plotq() {
+            ; cx16.r14 = x, cx16.r15 = y, draw=draw
+            plot(cx16.r14, cx16.r15, draw)
+        }
+    }
+
+    sub safe_circle(uword @zp xcenter, uword @zp ycenter, ubyte radius, bool draw) {
+        ; Does bounds checking and clipping.
+        ; Midpoint algorithm.
+        if radius==0
+            return
+
+        ubyte @zp xx = radius
+        ubyte @zp yy = 0
+        word @zp decisionOver2 = (1 as word)-xx
+        ; R14 = internal plot X
+        ; R15 = internal plot Y
+
+        while xx>=yy {
+            cx16.r14 = xcenter + xx
+            cx16.r15 = ycenter + yy
+            plotq()
+            cx16.r14 = xcenter - xx
+            plotq()
+            cx16.r14 = xcenter + xx
+            cx16.r15 = ycenter - yy
+            plotq()
+            cx16.r14 = xcenter - xx
+            plotq()
+            cx16.r14 = xcenter + yy
+            cx16.r15 = ycenter + xx
+            plotq()
+            cx16.r14 = xcenter - yy
+            plotq()
+            cx16.r14 = xcenter + yy
+            cx16.r15 = ycenter - xx
+            plotq()
+            cx16.r14 = xcenter - yy
+            plotq()
+
+            yy++
+            if decisionOver2>=0 {
+                xx--
+                decisionOver2 -= xx*$0002
+            }
+            decisionOver2 += yy*$0002
+            decisionOver2++
+        }
+
+        sub plotq() {
+            ; cx16.r14 = x, cx16.r15 = y, draw=draw
+            safe_plot(cx16.r14, cx16.r15, draw)
         }
     }
 
     sub disc(uword @zp xcenter, uword @zp ycenter, ubyte @zp radius, bool draw) {
+        ; Warning: NO BOUNDS CHECKS. Make sure circle fits in the screen.
         ; Midpoint algorithm, filled
         if radius==0
             return
@@ -440,12 +517,35 @@ _done
             horizontal_line(xcenter-yy, ycenter+radius, yy*$0002+1, draw)
             horizontal_line(xcenter-yy, ycenter-radius, yy*$0002+1, draw)
             yy++
-            if decisionOver2<=0
-                decisionOver2 += (yy as word)*2+1
-            else {
+            if decisionOver2>=0 {
                 radius--
-                decisionOver2 += (yy as word -radius)*2+1
+                decisionOver2 -= radius*$0002
             }
+            decisionOver2 += yy*$0002
+            decisionOver2++
+        }
+    }
+
+    sub safe_disc(uword @zp xcenter, uword @zp ycenter, ubyte @zp radius, bool draw) {
+        ; Does bounds checking and clipping.
+        ; Midpoint algorithm, filled
+        if radius==0
+            return
+        ubyte @zp yy = 0
+        word @zp decisionOver2 = (1 as word)-radius
+
+        while radius>=yy {
+            safe_horizontal_line(xcenter-radius, ycenter+yy, radius*$0002+1, draw)
+            safe_horizontal_line(xcenter-radius, ycenter-yy, radius*$0002+1, draw)
+            safe_horizontal_line(xcenter-yy, ycenter+radius, yy*$0002+1, draw)
+            safe_horizontal_line(xcenter-yy, ycenter-radius, yy*$0002+1, draw)
+            yy++
+            if decisionOver2>=0 {
+                radius--
+                decisionOver2 -= radius*$0002
+            }
+            decisionOver2 += yy*$0002
+            decisionOver2++
         }
     }
 
@@ -454,9 +554,9 @@ _done
         if draw {
             ; solid color or perhaps stipple
             %asm {{
-                lda  xx
-                eor  yy
-                ora  dont_stipple_flag
+                lda  p8v_xx
+                eor  p8v_yy
+                ora  p8v_dont_stipple_flag
                 and  #1
             }}
             if_nz {
@@ -475,7 +575,7 @@ _done
 
         sub prepare() {
             %asm {{
-                lda  xx
+                lda  p8v_xx
                 and  #7
                 pha     ; xbits
             }}
@@ -487,19 +587,28 @@ _done
             %asm {{
                 stz  cx16.VERA_CTRL
                 stz  cx16.VERA_ADDR_H
-                lda  xx+1
+                lda  p8v_xx+1
                 sta  cx16.VERA_ADDR_M
-                lda  xx
+                lda  p8v_xx
                 sta  cx16.VERA_ADDR_L
                 ply     ; xbits
-                lda  maskbits,y
+                lda  p8v_maskbits,y
             }}
         }
     }
 
+    sub safe_plot(uword xx, uword yy, bool draw) {
+        ; A plot that does bounds checks to see if the pixel is inside the screen.
+        if msb(xx)&$80!=0 or msb(yy)&$80!=0
+            return
+        if xx >= width or yy >= height
+            return
+        plot(xx, yy, draw)
+    }
+
     sub pget(uword @zp xx, uword yy) -> ubyte {
         %asm {{
-            lda  xx
+            lda  p8v_xx
             and  #7
             pha     ; xbits
         }}
@@ -512,12 +621,12 @@ _done
         %asm {{
             stz  cx16.VERA_CTRL
             stz  cx16.VERA_ADDR_H
-            lda  xx+1
+            lda  p8v_xx+1
             sta  cx16.VERA_ADDR_M
-            lda  xx
+            lda  p8v_xx
             sta  cx16.VERA_ADDR_L
             ply         ; xbits
-            lda  plot.maskbits,y
+            lda  p8s_plot.p8v_maskbits,y
             and  cx16.VERA_DATA0
             beq  +
             lda  #1
@@ -525,11 +634,13 @@ _done
         }}
     }
 
-    sub fill(word @zp xx, word @zp yy, bool draw) {
+    sub fill(uword x, uword y, bool draw) {
         ; 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
+        const ubyte MAXDEPTH = 64
+        word @zp xx = x as word
+        word @zp yy = y as word
         word[MAXDEPTH] @split @shared stack_xl
         word[MAXDEPTH] @split @shared stack_xr
         word[MAXDEPTH] @split @shared stack_y
@@ -551,21 +662,21 @@ _done
 ;;                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
+                    lda  p8v_sxl
+                    sta  p8v_stack_xl_lsb,y
+                    lda  p8v_sxl+1
+                    sta  p8v_stack_xl_msb,y
+                    lda  p8v_sxr
+                    sta  p8v_stack_xr_lsb,y
+                    lda  p8v_sxr+1
+                    sta  p8v_stack_xr_msb,y
+                    lda  p8v_sy
+                    sta  p8v_stack_y_lsb,y
+                    lda  p8v_sy+1
+                    sta  p8v_stack_y_msb,y
                     ldy  cx16.r12L
-                    lda  sdy
-                    sta  stack_dy,y
+                    lda  p8v_sdy
+                    sta  p8v_stack_dy,y
                     inc  cx16.r12L
                 }}
             }
@@ -579,21 +690,21 @@ _done
             %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  yy
-                lda  stack_y_msb,y
-                sta  yy+1
+                lda  p8v_stack_xl_lsb,y
+                sta  p8v_x1
+                lda  p8v_stack_xl_msb,y
+                sta  p8v_x1+1
+                lda  p8v_stack_xr_lsb,y
+                sta  p8v_x2
+                lda  p8v_stack_xr_msb,y
+                sta  p8v_x2+1
+                lda  p8v_stack_y_lsb,y
+                sta  p8v_yy
+                lda  p8v_stack_y_msb,y
+                sta  p8v_yy+1
                 ldy  cx16.r12L
-                lda  stack_dy,y
-                sta  dy
+                lda  p8v_stack_dy,y
+                sta  p8v_dy
             }}
             yy+=dy
         }
@@ -608,11 +719,14 @@ _done
         while cx16.r12L {
             pop_stack()
             xx = x1
-            while xx >= 0 and pget(xx as uword, yy as uword) == cx16.r11L {
-                plot(xx as uword, yy as uword, cx16.r10L)
+            while xx >= 0 {
+                if pget(xx as uword, yy as uword) != cx16.r11L
+                    break
                 xx--
             }
-            if xx >= x1
+            if x1!=xx
+                horizontal_line(xx as uword+1, yy as uword, x1-xx as uword, cx16.r10L)
+            else
                 goto skip
 
             left = xx + 1
@@ -621,17 +735,25 @@ _done
             xx = x1 + 1
 
             do {
-                while xx <= width-1 and pget(xx as uword, yy as uword) == cx16.r11L {
-                    plot(xx as uword, yy as uword, cx16.r10L)
+                cx16.r9 = xx
+                while xx <= width-1 {
+                    if pget(xx as uword, yy as uword) != cx16.r11L
+                        break
                     xx++
                 }
+                if cx16.r9!=xx
+                    horizontal_line(cx16.r9, yy as uword, (xx as uword)-cx16.r9, cx16.r10L)
+
                 push_stack(left, xx - 1, yy, dy)
                 if xx > x2 + 1
                     push_stack(x2 + 1, xx - 1, yy, -dy)
 skip:
                 xx++
-                while xx <= x2 and pget(xx as uword, yy as uword) != cx16.r11L
+                while xx <= x2 {
+                    if pget(xx as uword, yy as uword) == cx16.r11L
+                        break
                     xx++
+                }
                 left = xx
             } until xx>x2
         }
@@ -675,7 +797,7 @@ skip:
             cx16.vaddr_autoincr(charset_bank, chardataptr, 0, 1)
             %asm {{
                 ; pre-shift the bits
-                lda  text.xx
+                lda  p8s_text.p8v_xx
                 and  #7
                 sta  P8ZP_SCRATCH_B1
                 ldy  #0
@@ -688,9 +810,9 @@ skip:
                 ror  P8ZP_SCRATCH_REG
                 dex
                 bne  -
-+               sta  char_bitmap_bytes_left,y
++               sta  p8v_char_bitmap_bytes_left,y
                 lda  P8ZP_SCRATCH_REG
-                sta  char_bitmap_bytes_right,y
+                sta  p8v_char_bitmap_bytes_right,y
                 iny
                 cpy  #8
                 bne  --
@@ -701,7 +823,7 @@ skip:
             if draw {
                 %asm {{
                     ldy  #0
--                   lda  char_bitmap_bytes_left,y
+-                   lda  p8v_char_bitmap_bytes_left,y
                     ora  cx16.VERA_DATA1
                     sta  cx16.VERA_DATA0
                     iny
@@ -711,7 +833,7 @@ skip:
             } else {
                 %asm {{
                     ldy  #0
--                   lda  char_bitmap_bytes_left,y
+-                   lda  p8v_char_bitmap_bytes_left,y
                     eor  #255
                     and  cx16.VERA_DATA1
                     sta  cx16.VERA_DATA0
@@ -727,7 +849,7 @@ skip:
                 if draw {
                     %asm {{
                         ldy  #0
--                       lda  char_bitmap_bytes_right,y
+-                       lda  p8v_char_bitmap_bytes_right,y
                         ora  cx16.VERA_DATA1
                         sta  cx16.VERA_DATA0
                         iny
@@ -737,7 +859,7 @@ skip:
                 } else {
                     %asm {{
                         ldy  #0
--                       lda  char_bitmap_bytes_right,y
+-                       lda  p8v_char_bitmap_bytes_right,y
                         eor  #255
                         and  cx16.VERA_DATA1
                         sta  cx16.VERA_DATA0
@@ -768,6 +890,8 @@ skip:
     }
 
     asmsub cs_innerloop640(ubyte color @A) clobbers(Y) {
+        ; using verafx 32 bits writes here would make this faster but it's safer to
+        ; use verafx only explicitly when you know what you're doing.
         %asm {{
             cmp  #0
             beq  +

compiler/res/prog8lib/cx16/palette.p8

@@ -1,11 +1,11 @@
 ; Manipulate the Commander X16's display color palette.
-; Should you want to restore the default palette, you have to reinitialize the Vera yourself.
+; Should you want to restore the full default palette, you can call cbm.CINT()
+; The first 16 colors can be restored to their default with set_default16()
 
 palette {
-    %option no_symbol_prefixing
+    %option ignore_unused
 
     uword vera_palette_ptr
-    ubyte cc
 
     sub set_color(ubyte index, uword color) {
         vera_palette_ptr = $fa00+(index as uword * 2)
@@ -14,6 +14,11 @@ palette {
         cx16.vpoke(1, vera_palette_ptr, msb(color))
     }
 
+    sub get_color(ubyte index) -> uword {
+        vera_palette_ptr = $fa00+(index as uword * 2)
+        return mkword(cx16.vpeek(1, vera_palette_ptr+1), cx16.vpeek(1, vera_palette_ptr))
+    }
+
     sub set_rgb_be(uword palette_ptr, uword num_colors) {
         ; 1 word per color entry, $0rgb in big endian format
         vera_palette_ptr = $fa00
@@ -42,19 +47,30 @@ palette {
         ubyte red
         ubyte greenblue
         repeat num_colors {
-            red = @(palette_bytes_ptr) >> 4
-            palette_bytes_ptr++
-            greenblue = @(palette_bytes_ptr) & %11110000
-            palette_bytes_ptr++
-            greenblue |= @(palette_bytes_ptr) >> 4    ; add Blue
-            palette_bytes_ptr++
-            cx16.vpoke(1, vera_palette_ptr, greenblue)
+            cx16.r1 = color8to4(palette_bytes_ptr)
+            palette_bytes_ptr+=3
+            cx16.vpoke(1, vera_palette_ptr, cx16.r1H)       ; $GB
             vera_palette_ptr++
-            cx16.vpoke(1, vera_palette_ptr, red)
+            cx16.vpoke(1, vera_palette_ptr, cx16.r1L)       ; $0R
             vera_palette_ptr++
         }
     }
 
+    sub color8to4(uword colorpointer) -> uword {
+        ; accurately convert 24 bits (3 bytes) RGB color, in that order in memory, to 16 bits $GB;$0R colorvalue
+        cx16.r1 = colorpointer
+        cx16.r0 = channel8to4(@(cx16.r1))         ; (red)   -> $00:0R
+        cx16.r0H = channel8to4(@(cx16.r1+1))<<4   ; (green) -> $G0:0R
+        cx16.r0H |= channel8to4(@(cx16.r1+2))     ; (blue)  -> $GB:0R
+        return cx16.r0
+    }
+
+    sub channel8to4(ubyte channelvalue) -> ubyte {
+        ; accurately convert a single 8 bit color channel value to 4 bits,  see https://threadlocalmutex.com/?p=48
+        return msb(channelvalue * $000f + 135)
+    }
+
+
     sub set_monochrome(uword screencolorRGB, uword drawcolorRGB) {
         vera_palette_ptr = $fa00
         cx16.vpoke(1, vera_palette_ptr, lsb(screencolorRGB))   ; G,B
@@ -78,113 +94,84 @@ palette {
     }
 
     sub set_grayscale() {
+        ; set first 16 colors to a grayscale gradient from black to white
         vera_palette_ptr = $fa00
+        cx16.r2L=0
         repeat 16 {
-            cc=0
-            repeat 16 {
-                cx16.vpoke(1, vera_palette_ptr, cc)
-                vera_palette_ptr++
-                cx16.vpoke(1, vera_palette_ptr, cc)
-                vera_palette_ptr++
-                cc += $11
-            }
+            cx16.vpoke(1, vera_palette_ptr, cx16.r2L)
+            vera_palette_ptr++
+            cx16.vpoke(1, vera_palette_ptr, cx16.r2L)
+            vera_palette_ptr++
+            cx16.r2L += $11
         }
     }
 
-    uword[] C64_colorpalette_dark = [   ; this is a darker palette with more contrast
-        $000,  ; 0 = black
-        $FFF,  ; 1 = white
-        $632,  ; 2 = red
-        $7AB,  ; 3 = cyan
-        $638,  ; 4 = purple
-        $584,  ; 5 = green
-        $327,  ; 6 = blue
-        $BC6,  ; 7 = yellow
-        $642,  ; 8 = orange
-        $430,  ; 9 = brown
-        $965,  ; 10 = light red
-        $444,  ; 11 = dark grey
-        $666,  ; 12 = medium grey
-        $9D8,  ; 13 = light green
-        $65B,  ; 14 = light blue
-        $999   ; 15 = light grey
-    ]
-
-    uword[] C64_colorpalette_pepto = [  ; # this is Pepto's Commodore-64 palette  http://www.pepto.de/projects/colorvic/
-        $000,  ; 0 = black
-        $FFF,  ; 1 = white
-        $833,  ; 2 = red
-        $7cc,  ; 3 = cyan
-        $839,  ; 4 = purple
-        $5a4,  ; 5 = green
-        $229,  ; 6 = blue
-        $ef7,  ; 7 = yellow
-        $852,  ; 8 = orange
-        $530,  ; 9 = brown
-        $c67,  ; 10 = light red
-        $444,  ; 11 = dark grey
-        $777,  ; 12 = medium grey
-        $af9,  ; 13 = light green
-        $76e,  ; 14 = light blue
-        $bbb   ; 15 = light grey
-    ]
-
-    uword[] C64_colorpalette_light = [  ; this is a lighter palette
-        $000,  ; 0 = black
-        $FFF,  ; 1 = white
-        $944,  ; 2 = red
-        $7CC,  ; 3 = cyan
-        $95A,  ; 4 = purple
-        $6A5,  ; 5 = green
-        $549,  ; 6 = blue
-        $CD8,  ; 7 = yellow
-        $963,  ; 8 = orange
-        $650,  ; 9 = brown
-        $C77,  ; 10 = light red
-        $666,  ; 11 = dark grey
-        $888,  ; 12 = medium grey
-        $AE9,  ; 13 = light green
-        $87C,  ; 14 = light blue
-        $AAA   ; 15 = light grey
-    ]
-
     sub set_c64pepto() {
-        vera_palette_ptr = $fa00
-        repeat 16 {
-            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(ccp))     ; R
-                vera_palette_ptr++
-            }
-        }
+        ; set first 16 colors to the "Pepto" PAL commodore-64 palette  http://www.pepto.de/projects/colorvic/
+        uword[] colors = [
+            $000,  ; 0 = black
+            $FFF,  ; 1 = white
+            $833,  ; 2 = red
+            $7cc,  ; 3 = cyan
+            $839,  ; 4 = purple
+            $5a4,  ; 5 = green
+            $229,  ; 6 = blue
+            $ef7,  ; 7 = yellow
+            $852,  ; 8 = orange
+            $530,  ; 9 = brown
+            $c67,  ; 10 = light red
+            $444,  ; 11 = dark grey
+            $777,  ; 12 = medium grey
+            $af9,  ; 13 = light green
+            $76e,  ; 14 = light blue
+            $bbb   ; 15 = light grey
+        ]
+        set_rgb(colors, len(colors))
     }
 
-    sub set_c64light() {
-        vera_palette_ptr = $fa00
-        repeat 16 {
-            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(ccp))     ; R
-                vera_palette_ptr++
-            }
-        }
+    sub set_c64ntsc() {
+        ; set first 16 colors to a NTSC commodore-64 palette
+        uword[] colors = [
+            $000,   ; 0 = black
+            $FFF,   ; 1 = white
+            $934,   ; 2 = red
+            $9ff,   ; 3 = cyan
+            $73f,   ; 4 = purple
+            $4b1,   ; 5 = green
+            $20c,   ; 6 = blue
+            $ee6,   ; 7 = yellow
+            $b53,   ; 8 = orange
+            $830,   ; 9 = brown
+            $f8a,   ; 10 = light red
+            $444,   ; 11 = dark grey
+            $999,   ; 12 = medium grey
+            $9f9,   ; 13 = light green
+            $36f,   ; 14 = light blue
+            $ccc    ; 15 = light grey
+        ]
+        set_rgb(colors, len(colors))
     }
 
-    sub set_c64dark() {
-        vera_palette_ptr = $fa00
-        repeat 16 {
-            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(ccp))     ; R
-                vera_palette_ptr++
-            }
-        }
+    sub set_default16() {
+        ; set first 16 colors to the defaults on the X16
+        uword[] colors = [
+            $000,   ; 0 = black
+            $fff,   ; 1 = white
+            $800,   ; 2 = red
+            $afe,   ; 3 = cyan
+            $c4c,   ; 4 = purple
+            $0c5,   ; 5 = green
+            $00a,   ; 6 = blue
+            $ee7,   ; 7 = yellow
+            $d85,   ; 8 = orange
+            $640,   ; 9 = brown
+            $f77,   ; 10 = light red
+            $333,   ; 11 = dark grey
+            $777,   ; 12 = medium grey
+            $af6,   ; 13 = light green
+            $08f,   ; 14 = light blue
+            $bbb    ; 15 = light grey
+        ]
+        set_rgb(colors, len(colors))
     }
-
 }

compiler/res/prog8lib/cx16/psg.p8

@@ -1,7 +1,7 @@
 %import syslib
 
 psg {
-    %option no_symbol_prefixing
+    %option ignore_unused
 
     ; $1F9C0 - $1F9FF 	16 blocks of 4 PSG registers (16 voices)
     ; 00  frequency word LSB
@@ -104,20 +104,20 @@ psg {
         }
     }
 
-    sub envelopes_irq() {
+    sub envelopes_irq() -> bool {
         ; If you want to use real-time volume envelopes (Attack-Sustain-Release),
         ; 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)
+        ; Example: cx16.set_vsync_irq_handler(&psg.envelopes_irq)
         ; 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)
-        pushw(cx16.r9)
+        sys.pushw(cx16.r0)
+        sys.push(cx16.r1L)
+        sys.push(cx16.r2L)
+        sys.pushw(cx16.r9)
         ; calculate new volumes
         for cx16.r1L in 0 to 15 {
             when envelope_states[cx16.r1L] {
@@ -166,10 +166,11 @@ psg {
             cx16.VERA_DATA0 = cx16.VERA_DATA1 & %11000000 | msb(envelope_volumes[cx16.r1L])
         }
         cx16.restore_vera_context()
-        popw(cx16.r9)
-        pop(cx16.r2L)
-        pop(cx16.r1L)
-        popw(cx16.r0)
+        cx16.r9 = sys.popw()
+        cx16.r2L = sys.pop()
+        cx16.r1L = sys.pop()
+        cx16.r0 = sys.popw()
+        return true     ; run the system IRQ handler afterwards
     }
 
     ubyte[16] envelope_states

compiler/res/prog8lib/cx16/sprites.p8

@@ -7,6 +7,8 @@
 ; note: collision mask is not supported here yet.
 
 sprites {
+    %option ignore_unused
+
     const ubyte SIZE_8  = 0
     const ubyte SIZE_16 = 1
     const ubyte SIZE_32 = 2
@@ -39,6 +41,16 @@ sprites {
         cx16.vpoke_mask(1, sprite_reg+1, %11110000, msb(addr))    ; address 16:13
     }
 
+    sub get_data_ptr(ubyte spritenum) {
+        ; -- returns the VRAM address where the sprite's bitmap data is stored
+        ;    R1 (byte) = the vera bank (0 or 1), R0 (word) = the address.
+        sprite_reg = VERA_SPRITEREGS + spritenum*$0008
+        cx16.r0L = cx16.vpeek(1, sprite_reg)
+        cx16.r0H = cx16.vpeek(1, sprite_reg+1)
+        cx16.r1L = cx16.r0H & %00001000 !=0     ; bank
+        cx16.r0 <<= 5                           ; address
+    }
+
     sub pos(ubyte spritenum, word xpos, word ypos) {
         sprite_reg = VERA_SPRITEREGS + 2 + spritenum*$0008
         cx16.vpoke(1, sprite_reg, lsb(xpos))
@@ -81,6 +93,33 @@ sprites {
         cx16.vpoke(1, sprite_reg+1, msb(ypos))
     }
 
+    sub move(ubyte spritenum, word dx, word dy) {
+        ; move a sprite based on its current position
+        sprite_reg = VERA_SPRITEREGS + 2 + spritenum*$0008
+        cx16.r1 = mkword(cx16.vpeek(1, sprite_reg+1), cx16.vpeek(1, sprite_reg)) as word + dx
+        cx16.r2 = mkword(cx16.vpeek(1, sprite_reg+3), cx16.vpeek(1, sprite_reg+2)) as word + dy
+        cx16.vpoke(1, sprite_reg, cx16.r1L)
+        cx16.vpoke(1, sprite_reg+1, cx16.r1H)
+        cx16.vpoke(1, sprite_reg+2, cx16.r2L)
+        cx16.vpoke(1, sprite_reg+3, cx16.r2H)
+    }
+
+    sub movex(ubyte spritenum, word dx) {
+        ; move a sprite horizontally based on its current position
+        sprite_reg = VERA_SPRITEREGS + 2 + spritenum*$0008
+        cx16.r1 = mkword(cx16.vpeek(1, sprite_reg+1), cx16.vpeek(1, sprite_reg)) as word + dx
+        cx16.vpoke(1, sprite_reg, cx16.r1L)
+        cx16.vpoke(1, sprite_reg+1, cx16.r1H)
+    }
+
+    sub movey(ubyte spritenum, word dy) {
+        ; move a sprite vertically based on its current position
+        sprite_reg = VERA_SPRITEREGS + 4 + spritenum*$0008
+        cx16.r1 = mkword(cx16.vpeek(1, sprite_reg+1), cx16.vpeek(1, sprite_reg)) as word + dy
+        cx16.vpoke(1, sprite_reg, cx16.r1L)
+        cx16.vpoke(1, sprite_reg+1, cx16.r1H)
+    }
+
     sub getx(ubyte spritenum) -> word {
         sprite_reg = VERA_SPRITEREGS + 2 + spritenum*$0008
         return mkword(cx16.vpeek(1, sprite_reg+1), cx16.vpeek(1, sprite_reg)) as word
@@ -104,7 +143,49 @@ sprites {
     }
 
     sub set_palette_offset(ubyte spritenum, ubyte offset) {
-        sprite_reg = VERA_SPRITEREGS + 7 + spritenum*$0008
-        cx16.vpoke_mask(1, sprite_reg, %11110000, offset>>4)
+        cx16.vpoke_mask(1, VERA_SPRITEREGS + 7 + spritenum*$0008, %11110000, offset>>4)
+    }
+
+    sub set_mousepointer_hand() {
+        ; the array below is the compressed form of this sprite image:
+        ;    00, 16, 16, 00, 00, 00, 00, 00, 16, 16, 16, 00, 00, 00, 00, 00,
+        ;    16, 01, 01, 16, 00, 00, 16, 16, 15, 01, 01, 16, 00, 00, 00, 00,
+        ;    16, 01, 01, 01, 16, 16, 15, 01, 16, 15, 01, 01, 16, 00, 00, 00,
+        ;    16, 15, 01, 01, 01, 01, 16, 15, 01, 01, 01, 01, 16, 00, 00, 00,
+        ;    00, 16, 15, 01, 01, 01, 01, 01, 01, 01, 11, 01, 01, 16, 00, 00,
+        ;    00, 00, 16, 15, 01, 01, 01, 01, 11, 01, 01, 11, 01, 16, 00, 00,
+        ;    00, 00, 16, 16, 15, 01, 01, 01, 01, 11, 01, 01, 01, 16, 00, 00,
+        ;    00, 00, 16, 12, 12, 15, 01, 01, 01, 01, 01, 01, 01, 16, 16, 00,
+        ;    00, 00, 16, 12, 15, 15, 01, 01, 01, 01, 01, 01, 16, 01, 01, 16,
+        ;    00, 00, 00, 16, 15, 15, 01, 01, 01, 01, 01, 01, 01, 01, 01, 16,
+        ;    00, 00, 00, 00, 16, 15, 15, 15, 15, 15, 16, 01, 01, 15, 16, 00,
+        ;    00, 00, 00, 00, 00, 16, 16, 16, 16, 16, 15, 01, 15, 16, 00, 00,
+        ;    00, 00, 00, 00, 00, 00, 00, 00, 16, 12, 15, 15, 16, 00, 00, 00,
+        ;    00, 00, 00, 00, 00, 00, 00, 00, 16, 12, 12, 16, 00, 00, 00, 00,
+        ;    00, 00, 00, 00, 00, 00, 00, 00, 00, 16, 16, 00, 00, 00, 00, 00,
+        ;    00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00, 00
+        ubyte[] hand_image_lzsa = [
+            26, 31, 0, 16, 16, 0, 46, 16, 192, 58, 1, 1, 16, 156, 41, 15, 37,
+            137, 42, 1, 6, 115, 2, 43, 1, 152, 34, 4, 127, 4, 15, 11, 116, 233,
+            11, 37, 135, 6, 38, 7, 19, 12, 12, 67, 167, 35, 136, 237, 15, 77,
+            16, 157, 37, 71, 157, 47, 2, 33, 24, 9, 15, 69, 83, 66, 94, 6, 136,
+            77, 0, 186, 6, 69, 154, 6, 143, 70, 204, 15, 0, 191, 231, 232]
+
+        set_mousepointer_image(hand_image_lzsa, true)
+    }
+
+    sub set_mousepointer_image(uword data, bool compressed) {
+        get_data_ptr(0)  ; the mouse cursor is sprite 0
+        if cx16.r1L==0 and cx16.r0==0
+            return    ; mouse cursor not enabled
+        ubyte vbank = cx16.r1L
+        cx16.vaddr(vbank, cx16.r0, 0, 1)
+        if compressed {
+            void cx16.memory_decompress(data, &cx16.VERA_DATA0)      ; decompress directly into vram
+        } else {
+            for cx16.r0L in 0 to 255 {
+                cx16.VERA_DATA0 = data[cx16.r0L]
+            }
+        }
     }
 }

compiler/res/prog8lib/cx16/syslib.p8

@@ -1,19 +1,19 @@
 ; Prog8 definitions for the CommanderX16
 ; Including memory registers, I/O registers, Basic and Kernal subroutines.
 
-cbm {
+%option no_symbol_prefixing, ignore_unused
 
+cbm {
     ; Commodore (CBM) common variables, vectors and kernal routines
 
-    %option no_symbol_prefixing
 
 ; STROUT --> use txt.print
 ; CLEARSCR -> use txt.clear_screen
 ; HOMECRSR -> use txt.home or txt.plot
 
-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 $FF81 = CINT() clobbers(A,X,Y)                           ; (alias: SCINIT) initialize screen editor and video chip, including resetting to the default color palette. Note: also sets the video mode back to VGA
+romsub $FF84 = IOINIT() clobbers(A, X)                          ; initialize I/O devices (CIA, IRQ, ...)
+romsub $FF87 = RAMTAS() clobbers(A,X,Y)                         ; initialize RAM, 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
@@ -21,7 +21,7 @@ romsub $FF93 = SECOND(ubyte address @ A) clobbers(A)            ; (alias: LSTNSA
 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.   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 $FF9F = SCNKEY() clobbers(A,X,Y)                         ; scan the keyboard, also called  kbd_scan
 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
@@ -29,7 +29,7 @@ romsub $FFAB = UNTLK() clobbers(A)                              ; command serial
 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 $FFB7 = READST() -> ubyte @ A                            ; read I/O status word  (use CLEARST to reset it to 0)
 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
@@ -40,7 +40,7 @@ romsub $FFCC = CLRCHN() clobbers(A,X)                           ; (via 802 ($322
 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 character @ 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) clobbers (X, Y) -> bool @ Pc, ubyte @ A       ; (via 818 ($332)) save to a 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.  See also BSAVE
 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)
@@ -81,12 +81,19 @@ asmsub RDTIM16() clobbers(X) -> uword @AY {
     }}
 }
 
+sub CLEARST() {
+    ; -- Set the ST status variable back to 0. (there's no direct kernal call for this)
+    ;    Note: a drive error state (blinking led) isn't cleared! You can use diskio.status() to clear that.
+    SETNAM(0, $0000)
+    SETLFS(15, 3, 15)
+    void OPEN()
+    CLOSE(15)
+}
+
 }
 
 cx16 {
 
-    %option no_symbol_prefixing
-
 ; irq, system and hardware vectors:
     &uword  IERROR      = $0300
     &uword  IMAIN       = $0302
@@ -120,6 +127,9 @@ cx16 {
     &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  edkeyvec    = $ac03     ; (ram bank 0): for intercepting BASIN/CHRIN key strokes. See set_chrin_keyhandler()
+    &ubyte  edkeybk     = $ac05     ; ...the RAM bank of the handler routine, if not in low ram
+
 
 ; the sixteen virtual 16-bit registers in both normal unsigned mode and signed mode (s)
     &uword r0  = $0002
@@ -229,13 +239,15 @@ cx16 {
     const uword VERA_BASE       = $9F20
     &ubyte  VERA_ADDR_L         = VERA_BASE + $0000
     &ubyte  VERA_ADDR_M         = VERA_BASE + $0001
+    &uword  VERA_ADDR           = VERA_BASE + $0000 ; still need to do the _H separately
     &ubyte  VERA_ADDR_H         = VERA_BASE + $0002
     &ubyte  VERA_DATA0          = VERA_BASE + $0003
     &ubyte  VERA_DATA1          = VERA_BASE + $0004
     &ubyte  VERA_CTRL           = VERA_BASE + $0005
     &ubyte  VERA_IEN            = VERA_BASE + $0006
     &ubyte  VERA_ISR            = VERA_BASE + $0007
-    &ubyte  VERA_IRQ_LINE_L     = VERA_BASE + $0008
+    &ubyte  VERA_IRQLINE_L      = VERA_BASE + $0008 ; write only
+    &ubyte  VERA_SCANLINE_L     = VERA_BASE + $0008 ; read only
     &ubyte  VERA_DC_VIDEO       = VERA_BASE + $0009 ; DCSEL= 0
     &ubyte  VERA_DC_HSCALE      = VERA_BASE + $000A ; DCSEL= 0
     &ubyte  VERA_DC_VSCALE      = VERA_BASE + $000B ; DCSEL= 0
@@ -253,29 +265,53 @@ cx16 {
     &ubyte  VERA_L0_TILEBASE    = VERA_BASE + $000F
     &ubyte  VERA_L0_HSCROLL_L   = VERA_BASE + $0010
     &ubyte  VERA_L0_HSCROLL_H   = VERA_BASE + $0011
+    &uword  VERA_L0_HSCROLL     = VERA_BASE + $0010
     &ubyte  VERA_L0_VSCROLL_L   = VERA_BASE + $0012
     &ubyte  VERA_L0_VSCROLL_H   = VERA_BASE + $0013
+    &uword  VERA_L0_VSCROLL     = VERA_BASE + $0012
     &ubyte  VERA_L1_CONFIG      = VERA_BASE + $0014
     &ubyte  VERA_L1_MAPBASE     = VERA_BASE + $0015
     &ubyte  VERA_L1_TILEBASE    = VERA_BASE + $0016
     &ubyte  VERA_L1_HSCROLL_L   = VERA_BASE + $0017
     &ubyte  VERA_L1_HSCROLL_H   = VERA_BASE + $0018
+    &uword  VERA_L1_HSCROLL     = VERA_BASE + $0017
     &ubyte  VERA_L1_VSCROLL_L   = VERA_BASE + $0019
     &ubyte  VERA_L1_VSCROLL_H   = VERA_BASE + $001A
+    &uword  VERA_L1_VSCROLL     = VERA_BASE + $0019
     &ubyte  VERA_AUDIO_CTRL     = VERA_BASE + $001B
     &ubyte  VERA_AUDIO_RATE     = VERA_BASE + $001C
     &ubyte  VERA_AUDIO_DATA     = VERA_BASE + $001D
     &ubyte  VERA_SPI_DATA       = VERA_BASE + $001E
     &ubyte  VERA_SPI_CTRL       = VERA_BASE + $001F
 
-    ; experimental Vera FX registers: (depends on particular value set in VERA_CTRL!!!)
+    ; Vera FX registers: (accessing depends on particular DCSEL value set in VERA_CTRL!)
     &ubyte VERA_FX_CTRL         = VERA_BASE + $0009
-    &ubyte VERA_FX_MULT         = VERA_BASE + $000C
+    &ubyte VERA_FX_TILEBASE     = VERA_BASE + $000a
+    &ubyte VERA_FX_MAPBASE      = VERA_BASE + $000b
+    &ubyte VERA_FX_MULT         = VERA_BASE + $000c
+    &ubyte VERA_FX_X_INCR_L     = VERA_BASE + $0009
+    &ubyte VERA_FX_X_INCR_H     = VERA_BASE + $000a
+    &uword VERA_FX_X_INCR       = VERA_BASE + $0009
+    &ubyte VERA_FX_Y_INCR_L     = VERA_BASE + $000b
+    &ubyte VERA_FX_Y_INCR_H     = VERA_BASE + $000c
+    &uword VERA_FX_Y_INCR       = VERA_BASE + $000b
+    &ubyte VERA_FX_X_POS_L      = VERA_BASE + $0009
+    &ubyte VERA_FX_X_POS_H      = VERA_BASE + $000a
+    &uword VERA_FX_X_POS        = VERA_BASE + $0009
+    &ubyte VERA_FX_Y_POS_L      = VERA_BASE + $000b
+    &ubyte VERA_FX_Y_POS_H      = VERA_BASE + $000c
+    &uword VERA_FX_Y_POS        = VERA_BASE + $000b
+    &ubyte VERA_FX_X_POS_S      = VERA_BASE + $0009
+    &ubyte VERA_FX_Y_POS_S      = VERA_BASE + $000a
+    &ubyte VERA_FX_POLY_FILL_L  = VERA_BASE + $000b
+    &ubyte VERA_FX_POLY_FILL_H  = VERA_BASE + $000c
+    &uword VERA_FX_POLY_FILL    = VERA_BASE + $000b
     &ubyte VERA_FX_CACHE_L      = VERA_BASE + $0009
-    &ubyte VERA_FX_CACHE_M      = VERA_BASE + $000A
-    &ubyte VERA_FX_CACHE_H      = VERA_BASE + $000B
-    &ubyte VERA_FX_CACHE_U      = VERA_BASE + $000C
-    &ubyte VERA_FX_ACCUM_RESET  = VERA_BASE + $0009     ; (DCSEL=6)
+    &ubyte VERA_FX_CACHE_M      = VERA_BASE + $000a
+    &ubyte VERA_FX_CACHE_H      = VERA_BASE + $000b
+    &ubyte VERA_FX_CACHE_U      = VERA_BASE + $000c
+    &ubyte VERA_FX_ACCUM        = VERA_BASE + $000a
+    &ubyte VERA_FX_ACCUM_RESET  = VERA_BASE + $0009
 
 
 ; VERA_PSG_BASE     = $1F9C0
@@ -364,7 +400,6 @@ romsub $fef6 = FB_init()  clobbers(A,X,Y)
 romsub $fef9 = FB_get_info()  clobbers(X,Y) -> byte @A, uword @R0, uword @R1    ; width=r0, height=r1
 romsub $fefc = FB_set_palette(uword pointer @R0, ubyte index @A, ubyte colorcount @X)  clobbers(A,X,Y)
 romsub $feff = FB_cursor_position(uword x @R0, uword y @R1)  clobbers(A,X,Y)
-romsub $feff = FB_cursor_position2()  clobbers(A,X,Y)           ;  alias for the previous routine, but avoiding having to respecify both x and y every time
 romsub $ff02 = FB_cursor_next_line(uword x @R0)  clobbers(A,X,Y)
 romsub $ff05 = FB_get_pixel()  clobbers(X,Y) -> ubyte @A
 romsub $ff08 = FB_get_pixels(uword pointer @R0, uword count @R1)  clobbers(A,X,Y)
@@ -397,8 +432,8 @@ 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 $feb1 = MCIOUT(ubyte length @A, uword buffer @XY, bool dontAdvance @Pc)  clobbers(A) -> bool @Pc, uword @XY
 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()
+romsub $ff4d = clock_set_date_time(uword yearmonth @R0, uword dayhours @R1, uword minsecs @R2, uword jiffiesweekday @R3)  clobbers(A, X, Y)
+romsub $ff50 = clock_get_date_time()  clobbers(A, X, Y)  -> uword @R0, uword @R1, uword @R2, uword @R3   ; result registers see clock_set_date_time()
 
 ; keyboard, mouse, joystick
 ; note: also see the kbdbuf_clear() helper routine below!
@@ -407,19 +442,77 @@ romsub $febd = kbdbuf_peek2() -> uword @AX             ; alternative to above to
 romsub $fec0 = kbdbuf_get_modifiers() -> ubyte @A
 romsub $fec3 = kbdbuf_put(ubyte key @A) clobbers(X)
 romsub $fed2 = keymap(uword identifier @XY, bool read @Pc) -> bool @Pc
-romsub $ff68 = mouse_config(ubyte shape @A, ubyte resX @X, ubyte resY @Y)  clobbers (A, X, Y)
+romsub $ff68 = mouse_config(byte shape @A, ubyte resX @X, ubyte resY @Y)  clobbers (A, X, Y)
 romsub $ff6b = mouse_get(ubyte zpdataptr @X) -> ubyte @A
 romsub $ff71 = mouse_scan()  clobbers(A, X, Y)
 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 (rom bank 10)
-romsub $C04B = psg_init() clobbers(A,X,Y)                              ; (re)init Vera PSG
-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 both vera PSG and YM audio chips
-; TODO: add more of the audio routines?
+; X16Edit (rom bank 13/14 but you ideally should use the routine search_x16edit() to search for the correct bank)
+romsub $C000 = x16edit_default() clobbers(A,X,Y)
+romsub $C003 = x16edit_loadfile(ubyte firstbank @X, ubyte lastbank @Y, str filename @R0, ubyte filenameLength @R1) clobbers(A,X,Y)
+romsub $C006 = x16edit_loadfile_options(ubyte firstbank @X, ubyte lastbank @Y, str filename @R0,
+                uword filenameLengthAndOptions @R1, uword tabstopAndWordwrap @R2,
+                uword disknumberAndColors @R3, uword headerAndStatusColors @R4) clobbers(A,X,Y)
+
+; Audio (rom bank 10 - you have to activate this bank manually first! Or use callfar/JSRFAR.)
+romsub $C09F = audio_init() clobbers(A,X,Y) -> bool @Pc     ; (re)initialize both vera PSG and YM audio chips
+romsub $C000 = bas_fmfreq(ubyte channel @A, uword freq @XY, bool noretrigger @Pc) clobbers(A,X,Y) -> bool @Pc
+romsub $C003 = bas_fmnote(ubyte channel @A, ubyte note @X, ubyte fracsemitone @Y, bool noretrigger @Pc) clobbers(A,X,Y) -> bool @Pc
+romsub $C006 = bas_fmplaystring(ubyte length @A, str string @XY) clobbers(A,X,Y)
+romsub $C009 = bas_fmvib(ubyte speed @A, ubyte depth @X) clobbers(A,X,Y) -> bool @Pc
+romsub $C00C = bas_playstringvoice(ubyte channel @A) clobbers(Y)
+romsub $C00F = bas_psgfreq(ubyte voice @A, uword freq @XY) clobbers(A,X,Y) -> bool @Pc
+romsub $C012 = bas_psgnote(ubyte voice @A, ubyte note @X, ubyte fracsemitone @Y) clobbers(A,X,Y) -> bool @Pc
+romsub $C015 = bas_psgwav(ubyte voice @A, ubyte waveform @X) clobbers(A,X,Y) -> bool @Pc
+romsub $C018 = bas_psgplaystring(ubyte length @A, str string @XY) clobbers(A,X,Y)
+romsub $C08D = bas_fmchordstring(ubyte length @A, str string @XY) clobbers(A,X,Y)
+romsub $C090 = bas_psgchordstring(ubyte length @A, str string @XY) clobbers(A,X,Y)
+romsub $C01B = notecon_bas2fm(ubyte note @X) clobbers(A) -> ubyte @X, bool @Pc
+romsub $C01E = notecon_bas2midi(ubyte note @X) clobbers(A) -> ubyte @X, bool @Pc
+romsub $C021 = notecon_bas2psg(ubyte note @X, ubyte fracsemitone @Y) clobbers(A) -> uword @XY, bool @Pc
+romsub $C024 = notecon_fm2bas(ubyte note @X) clobbers(A) -> ubyte @X, bool @Pc
+romsub $C027 = notecon_fm2midi(ubyte note @X) clobbers(A) -> ubyte @X, bool @Pc
+romsub $C02A = notecon_fm2psg(ubyte note @X, ubyte fracsemitone @Y) clobbers(A) -> uword @XY, bool @Pc
+romsub $C02D = notecon_freq2bas(uword freqHz @XY) clobbers(A) -> ubyte @X, ubyte @Y, bool @Pc
+romsub $C030 = notecon_freq2fm(uword freqHz @XY) clobbers(A) -> ubyte @X, ubyte @Y, bool @Pc
+romsub $C033 = notecon_freq2midi(uword freqHz @XY) clobbers(A) -> ubyte @X, ubyte @Y, bool @Pc
+romsub $C036 = notecon_freq2psg(uword freqHz @XY) clobbers(A) -> uword @XY, bool @Pc
+romsub $C039 = notecon_midi2bas(ubyte note @X) clobbers(A) -> ubyte @X, bool @Pc
+romsub $C03C = notecon_midi2fm(ubyte note @X) clobbers(A) -> ubyte @X, bool @Pc
+romsub $C03F = notecon_midi2psg(ubyte note @X, ubyte fracsemitone @Y) clobbers(A) -> uword @XY, bool @Pc
+romsub $C042 = notecon_psg2bas(uword freq @XY) clobbers(A) -> ubyte @X, ubyte @Y, bool @Pc
+romsub $C045 = notecon_psg2fm(uword freq @XY) clobbers(A) -> ubyte @X, ubyte @Y, bool @Pc
+romsub $C048 = notecon_psg2midi(uword freq @XY) clobbers(A) -> ubyte @X, ubyte @Y, bool @Pc
+romsub $C04B = psg_init() clobbers(A,X,Y)               ; (re)init Vera PSG
+romsub $C04E = psg_playfreq(ubyte voice @A, uword freq @XY) clobbers(A,X,Y)
+romsub $C051 = psg_read(ubyte offset @X, bool cookedVol @Pc) clobbers(Y) -> ubyte @A
+romsub $C054 = psg_setatten(ubyte voice @A, ubyte attenuation @X) clobbers(A,X,Y)
+romsub $C057 = psg_setfreq(ubyte voice @A, uword freq @XY) clobbers(A,X,Y)
+romsub $C05A = psg_setpan(ubyte voice @A, ubyte panning @X) clobbers(A,X,Y)
+romsub $C05D = psg_setvol(ubyte voice @A, ubyte volume @X) clobbers(A,X,Y)
+romsub $C060 = psg_write(ubyte value @A, ubyte offset @X) clobbers(Y)
+romsub $C0A2 = psg_write_fast(ubyte value @A, ubyte offset @X) clobbers(Y)
+romsub $C093 = psg_getatten(ubyte voice @A) clobbers(Y) -> ubyte @X
+romsub $C096 = psg_getpan(ubyte voice @A) clobbers(Y) -> ubyte @X
+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 $C069 = ym_loadpatch(ubyte channel @A, uword patchOrAddress @XY, bool what @Pc) clobbers(A,X,Y)
+romsub $C06C = ym_loadpatchlfn(ubyte channel @A, ubyte lfn @X) clobbers(X,Y) -> ubyte @A, bool @Pc
+romsub $C06F = ym_playdrum(ubyte channel @A, ubyte note @X) clobbers(A,X,Y) -> bool @Pc
+romsub $C072 = ym_playnote(ubyte channel @A, ubyte kc @X, ubyte kf @Y, bool notrigger @Pc) clobbers(A,X,Y) -> bool @Pc
+romsub $C075 = ym_setatten(ubyte channel @A, ubyte attenuation @X) clobbers(Y) -> bool @Pc
+romsub $C078 = ym_setdrum(ubyte channel @A, ubyte note @X) clobbers(A,X,Y) -> bool @Pc
+romsub $C07B = ym_setnote(ubyte channel @A, ubyte kc @X, ubyte kf @Y) clobbers(A,X,Y) -> bool @Pc
+romsub $C07E = ym_setpan(ubyte channel @A, ubyte panning @X) clobbers(A,X,Y) -> bool @Pc
+romsub $C081 = ym_read(ubyte register @X, bool cooked @Pc) clobbers(Y) -> ubyte @A, bool @Pc
+romsub $C084 = ym_release(ubyte channel @A) clobbers(A,X,Y) -> bool @Pc
+romsub $C087 = ym_trigger(ubyte channel @A, bool noRelease @Pc) clobbers(A,X,Y) -> bool @Pc
+romsub $C08A = ym_write(ubyte value @A, ubyte register @X) clobbers(Y) -> bool @Pc
+romsub $C099 = ym_getatten(ubyte channel @A) clobbers(Y) -> ubyte @X
+romsub $C09C = ym_getpan(ubyte channel @A) clobbers(Y) -> ubyte @X
+romsub $C0A5 = ym_get_chip_type() clobbers(X) -> ubyte @A
 
 
 asmsub set_screen_mode(ubyte mode @A) clobbers(A,X,Y) -> bool @Pc {
@@ -450,7 +543,7 @@ asmsub kbdbuf_clear() {
     }}
 }
 
-asmsub mouse_config2(ubyte shape @A) clobbers (A, X, Y) {
+asmsub mouse_config2(byte shape @A) clobbers (A, X, Y) {
     ; -- convenience wrapper function that handles the screen resolution for mouse_config() for you
     %asm {{
         pha                         ; save shape
@@ -526,16 +619,11 @@ asmsub vpeek(ubyte bank @A, uword address @XY) -> ubyte @A {
         ; -- get a byte from VERA's video memory
         ;    note: inefficient when reading multiple sequential bytes!
         %asm {{
-                pha
-                lda  cx16.VERA_CTRL
-                ora  #1
-                sta  cx16.VERA_CTRL
-                pla
-                and  #1
+                stz  cx16.VERA_CTRL
                 sta  cx16.VERA_ADDR_H
                 sty  cx16.VERA_ADDR_M
                 stx  cx16.VERA_ADDR_L
-                lda  cx16.VERA_DATA1
+                lda  cx16.VERA_DATA0
                 rts
             }}
 }
@@ -545,7 +633,6 @@ asmsub vaddr(ubyte bank @A, uword address @R0, ubyte addrsel @R1, byte autoIncrO
         ;    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
             lda  cx16.r1
             and  #1
@@ -569,7 +656,6 @@ 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
@@ -582,6 +668,8 @@ asmsub vaddr_clone(ubyte port @A) clobbers (A,X,Y) {
         sty  VERA_ADDR_M
         ply
         sty  VERA_ADDR_H
+        eor  #1
+        sta  VERA_CTRL
         rts
     }}
 }
@@ -600,8 +688,7 @@ asmsub vaddr_autoincr(ubyte bank @A, uword address @R0, ubyte addrsel @R1, uword
             sta  cx16.VERA_ADDR_H
             rts
 
-_setup      and  #1
-            sta  P8ZP_SCRATCH_REG
+_setup      sta  P8ZP_SCRATCH_REG
             lda  cx16.r1
             and  #1
             sta  cx16.VERA_CTRL
@@ -671,7 +758,6 @@ asmsub vpoke(ubyte bank @A, uword address @R0, ubyte value @Y) clobbers(A) {
     ;    note: inefficient when writing multiple sequential bytes!
     %asm {{
         stz  cx16.VERA_CTRL
-        and  #1
         sta  cx16.VERA_ADDR_H
         lda  cx16.r0
         sta  cx16.VERA_ADDR_L
@@ -687,7 +773,6 @@ asmsub vpoke_or(ubyte bank @A, uword address @R0, ubyte value @Y) clobbers (A) {
     ;    note: inefficient when writing multiple sequential bytes!
     %asm {{
         stz  cx16.VERA_CTRL
-        and  #1
         sta  cx16.VERA_ADDR_H
         lda  cx16.r0
         sta  cx16.VERA_ADDR_L
@@ -705,7 +790,6 @@ asmsub vpoke_and(ubyte bank @A, uword address @R0, ubyte value @Y) clobbers(A) {
     ;    note: inefficient when writing multiple sequential bytes!
     %asm {{
         stz  cx16.VERA_CTRL
-        and  #1
         sta  cx16.VERA_ADDR_H
         lda  cx16.r0
         sta  cx16.VERA_ADDR_L
@@ -723,7 +807,6 @@ asmsub vpoke_xor(ubyte bank @A, uword address @R0, ubyte value @Y) clobbers (A)
     ;    note: inefficient when writing multiple sequential bytes!
     %asm {{
         stz  cx16.VERA_CTRL
-        and  #1
         sta  cx16.VERA_ADDR_H
         lda  cx16.r0
         sta  cx16.VERA_ADDR_L
@@ -742,7 +825,6 @@ asmsub vpoke_mask(ubyte bank @A, uword address @R0, ubyte mask @X, ubyte value @
     %asm {{
         sty  P8ZP_SCRATCH_B1
         stz  cx16.VERA_CTRL
-        and  #1
         sta  cx16.VERA_ADDR_H
         lda  cx16.r0
         sta  cx16.VERA_ADDR_L
@@ -832,13 +914,269 @@ asmsub restore_vera_context() clobbers(A) {
     }}
 }
 
+
+    asmsub set_chrin_keyhandler(ubyte handlerbank @A, uword handler @XY) clobbers(A) {
+        ; Install a custom CHRIN (BASIN) key handler in a safe manner. Call this before each line you want to read.
+        ; See https://github.com/X16Community/x16-docs/blob/master/X16%20Reference%20-%2002%20-%20Editor.md#custom-basin-petscii-code-override-handler
+        %asm {{
+            sei
+            sta  P8ZP_SCRATCH_REG
+            lda  $00
+            pha
+            stz  $00
+            lda  P8ZP_SCRATCH_REG
+            sta  cx16.edkeybk
+            stx  cx16.edkeyvec
+            sty  cx16.edkeyvec+1
+            pla
+            sta  $00
+            cli
+            rts
+        }}
+    }
+
+    asmsub get_chrin_keyhandler() -> ubyte @R0, uword @R1 {
+        ; --- retrieve the currently set CHRIN keyhandler in a safe manner, bank in r0L, handler address in R1.
+        %asm {{
+            sei
+            lda  $00
+            pha
+            stz  $00
+            lda  cx16.edkeybk
+            sta  cx16.r0L
+            lda  cx16.edkeyvec
+            ldy  cx16.edkeyvec+1
+            sta  cx16.r1
+            sty  cx16.r1+1
+            pla
+            sta  $00
+            cli
+            rts
+        }}
+    }
+
+
+    ; Commander X16 IRQ dispatcher routines
+
+inline asmsub  disable_irqs() clobbers(A) {
+    ; Disable all Vera IRQ sources. Note that it does NOT set the CPU IRQ disabled status bit!
+    %asm {{
+        lda  #%00001111
+        trb  cx16.VERA_IEN
+    }}
+}
+
+asmsub  enable_irq_handlers(bool disable_all_irq_sources @Pc) clobbers(A,Y)  {
+    ; Install the "master IRQ handler" that will dispatch IRQs
+    ; to the registered handler for each type.  (Only Vera IRQs supported for now).
+    ; The handlers don't need to clear its ISR bit, but have to return 0 or 1 in A,
+    ; where 1 means: continue with the system IRQ handler, 0 means: don't call that.
+	%asm {{
+        php
+        sei
+        bcc  +
+        lda  #%00001111
+        trb  cx16.VERA_IEN      ; disable all IRQ sources
++       lda  #<_irq_dispatcher
+        ldy  #>_irq_dispatcher
+        sta  cx16.CINV
+        sty  cx16.CINV+1
+        plp
+        rts
+
+_irq_dispatcher
+        ; order of handling: LINE, SPRCOL, AFLOW, VSYNC.
+        jsr  sys.save_prog8_internals
+        cld
+        lda  cx16.VERA_ISR
+        and  cx16.VERA_IEN          ; only consider the bits for sources that can actually raise the IRQ
+
+        bit  #2
+        beq  +
+_mod_line_jump
+        jsr  _default_line_handler      ; modified
+        ldy  #2
+        sty  cx16.VERA_ISR
+        bra  _dispatch_end
++
+        bit  #4
+        beq  +
+_mod_sprcol_jump
+        jsr  _default_sprcol_handler      ; modified
+        ldy  #4
+        sty  cx16.VERA_ISR
+        bra  _dispatch_end
++
+        bit  #8
+        beq  +
+_mod_aflow_jump
+        jsr  _default_aflow_handler      ; modified
+        ; note: AFLOW can only be cleared by filling the audio FIFO for at least 1/4. Not via the ISR bit.
+        bra  _dispatch_end
++
+        bit  #1
+        beq  +
+_mod_vsync_jump
+        jsr  _default_vsync_handler      ; modified
+        cmp  #0
+        bne  _dispatch_end
+        ldy  #1
+        sty  cx16.VERA_ISR
+        bra  _return_irq
++
+        lda  #0
+_dispatch_end
+        cmp  #0
+        beq  _return_irq
+        jsr  sys.restore_prog8_internals
+		jmp  (sys.restore_irq._orig_irqvec)   ; continue with normal kernal irq routine
+_return_irq
+        jsr  sys.restore_prog8_internals
+		ply
+		plx
+		pla
+		rti
+
+_default_vsync_handler
+        lda  #1
+        rts
+_default_line_handler
+        lda  #0
+        rts
+_default_sprcol_handler
+        lda  #0
+        rts
+_default_aflow_handler
+        lda  #0
+        rts
+    }}
+}
+
+asmsub set_vsync_irq_handler(uword address @AY) clobbers(A) {
+    ; Sets the VSYNC irq handler to use with enable_irq_handlers().  Also enables VSYNC irqs.
+    ; NOTE: unless a proper irq handler is already running, you should enclose this call in set_irqd() / clear_irqd() to avoid system crashes.
+    %asm {{
+        php
+        sei
+        sta  enable_irq_handlers._mod_vsync_jump+1
+        sty  enable_irq_handlers._mod_vsync_jump+2
+        lda  #1
+        tsb  cx16.VERA_IEN
+        plp
+        rts
+    }}
+}
+
+asmsub set_line_irq_handler(uword rasterline @R0, uword address @AY) clobbers(A,Y) {
+    ; Sets the LINE irq handler to use with enable_irq_handlers(), for the given rasterline.  Also enables LINE irqs.
+    ; You can use sys.set_rasterline() later to adjust the rasterline on which to trigger.
+    ; NOTE: unless a proper irq handler is already running, you should enclose this call in set_irqd() / clear_irqd() to avoid system crashes.
+    %asm {{
+        php
+        sei
+        sta  enable_irq_handlers._mod_line_jump+1
+        sty  enable_irq_handlers._mod_line_jump+2
+        lda  cx16.r0
+        ldy  cx16.r0+1
+        jsr  sys.set_rasterline
+        lda  #2
+        tsb  cx16.VERA_IEN
+        plp
+        rts
+    }}
+}
+
+asmsub set_sprcol_irq_handler(uword address @AY) clobbers(A) {
+    ; Sets the SPRCOL irq handler to use with enable_irq_handlers().  Also enables SPRCOL irqs.
+    ; NOTE: unless a proper irq handler is already running, you should enclose this call in set_irqd() / clear_irqd() to avoid system crashes.
+    %asm {{
+        php
+        sei
+        sta  enable_irq_handlers._mod_sprcol_jump+1
+        sty  enable_irq_handlers._mod_sprcol_jump+2
+        lda  #4
+        tsb  cx16.VERA_IEN
+        plp
+        rts
+    }}
+}
+
+asmsub set_aflow_irq_handler(uword address @AY) clobbers(A) {
+    ; Sets the AFLOW irq handler to use with enable_irq_handlers().  Also enables AFLOW irqs.
+    ; NOTE: unless a proper irq handler is already running, you should enclose this call in set_irqd() / clear_irqd() to avoid system crashes.
+    %asm {{
+        php
+        sei
+        sta  enable_irq_handlers._mod_aflow_jump+1
+        sty  enable_irq_handlers._mod_aflow_jump+2
+        lda  #8
+        tsb  cx16.VERA_IEN
+        plp
+        rts
+    }}
+}
+
+
+inline asmsub  disable_irq_handlers() {
+    ; back to the system default IRQ handler.
+    %asm {{
+        jsr  sys.restore_irq
+    }}
+}
+
+
+sub search_x16edit() -> ubyte {
+    ; -- Search the rom bank that contains x16edit. Returns bank number, or 255 if not found.
+    cx16.r0L = cx16.getrombank()
+    sys.set_irqd()
+    str @shared signature = petscii:"x16edit"
+    for cx16.r1L in 31 downto 0  {
+        cx16.rombank(cx16.r1L)
+        cx16.r2 = $fff0
+        %asm {{
+            ldy  #0
+-           lda  signature,y
+            cmp  (cx16.r2),y
+            bne  +
+            iny
+            cpy #7
+            bne  -
+            sec
+            bcs  ++
++           clc
++
+        }}
+        if_cs {
+            cx16.rombank(cx16.r0L)
+            sys.clear_irqd()
+            return cx16.r1L
+        }
+    }
+    sys.clear_irqd()
+    return 255
+}
+
+    asmsub cpu_is_65816() -> bool @A {
+        ; Returns true when you have a 65816 cpu, false when it's a 6502.
+        %asm {{
+			php
+			clv
+			.byte $e2, $ea  ; SEP #$ea, should be interpreted as 2 NOPs by 6502. 65c816 will set the Overflow flag.
+			bvc +
+			lda #1
+			plp
+			rts
++			lda #0
+			plp
+			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()  {
@@ -873,6 +1211,8 @@ asmsub  init_system()  {
         jsr  cbm.CHROUT
         lda  #147       ; clear screen
         jsr  cbm.CHROUT
+        lda  #8         ; disable charset case switch
+        jsr  cbm.CHROUT
         lda  #PROG8_VARSHIGH_RAMBANK
         sta  $00    ; select ram bank
         lda  #0
@@ -907,83 +1247,45 @@ asmsub  cleanup_at_exit() {
         lda  #4
         sta  $01        ; rom bank 4 (basic)
         stz  $2d        ; hack to reset machine code monitor bank to 0
+        jsr  cbm.CLRCHN		; reset i/o channels
+_exitcode = *+1
+        lda  #0        ; exit code possibly modified in exit()
         rts
     }}
 }
 
-asmsub  set_irq(uword handler @AY, bool useKernal @Pc) clobbers(A)  {
+asmsub  set_irq(uword handler @AY) clobbers(A)  {
+    ; Sets the handler for the VSYNC interrupt, and enable that interrupt.
 	%asm {{
-	        sta  _modified+1
-	        sty  _modified+2
-	        lda  #0
-	        rol  a
-	        sta  _use_kernal
-		sei
-		lda  #<_irq_handler
-		sta  cx16.CINV
-		lda  #>_irq_handler
-		sta  cx16.CINV+1
-                lda  cx16.VERA_IEN
-                ora  #%00000001     ; enable the vsync irq
-                sta  cx16.VERA_IEN
-		cli
-		rts
+        sei
+        sta  _modified+1
+        sty  _modified+2
+        lda  #<_irq_handler
+        sta  cx16.CINV
+        lda  #>_irq_handler
+        sta  cx16.CINV+1
+        lda  #1
+        tsb  cx16.VERA_IEN      ; enable the vsync irq
+        cli
+        rts
 
-_irq_handler    jsr  _irq_handler_init
-_modified	jsr  $ffff                      ; modified
-		jsr  _irq_handler_end
-		lda  _use_kernal
-		bne  +
-		; end irq processing - don't use kernal's irq handling
-		lda  #1
+_irq_handler
+        jsr  sys.save_prog8_internals
+        cld
+_modified
+        jsr  $ffff                      ; modified
+        pha
+		jsr  sys.restore_prog8_internals
+		pla
+		beq  +
+		jmp  (restore_irq._orig_irqvec)   ; continue with normal kernal irq routine
++		lda  #1
 		sta  cx16.VERA_ISR      ; clear Vera Vsync irq status
 		ply
 		plx
 		pla
 		rti
-+		jmp  (restore_irq._orig_irqvec)   ; continue with normal kernal irq routine
-
-_use_kernal     .byte  0
-
-_irq_handler_init
-		; save all zp scratch registers as these might be clobbered by the irq routine
-		lda  P8ZP_SCRATCH_B1
-		sta  IRQ_SCRATCH_ZPB1
-		lda  P8ZP_SCRATCH_REG
-		sta  IRQ_SCRATCH_ZPREG
-		lda  P8ZP_SCRATCH_W1
-		sta  IRQ_SCRATCH_ZPWORD1
-		lda  P8ZP_SCRATCH_W1+1
-		sta  IRQ_SCRATCH_ZPWORD1+1
-		lda  P8ZP_SCRATCH_W2
-		sta  IRQ_SCRATCH_ZPWORD2
-		lda  P8ZP_SCRATCH_W2+1
-		sta  IRQ_SCRATCH_ZPWORD2+1
-		cld
-		rts
-
-_irq_handler_end
-		; restore all zp scratch registers
-		lda  IRQ_SCRATCH_ZPB1
-		sta  P8ZP_SCRATCH_B1
-		lda  IRQ_SCRATCH_ZPREG
-		sta  P8ZP_SCRATCH_REG
-		lda  IRQ_SCRATCH_ZPWORD1
-		sta  P8ZP_SCRATCH_W1
-		lda  IRQ_SCRATCH_ZPWORD1+1
-		sta  P8ZP_SCRATCH_W1+1
-		lda  IRQ_SCRATCH_ZPWORD2
-		sta  P8ZP_SCRATCH_W2
-		lda  IRQ_SCRATCH_ZPWORD2+1
-		sta  P8ZP_SCRATCH_W2+1
-		rts
-
-IRQ_SCRATCH_ZPB1	.byte  0
-IRQ_SCRATCH_ZPREG	.byte  0
-IRQ_SCRATCH_ZPWORD1	.word  0
-IRQ_SCRATCH_ZPWORD2	.word  0
-
-		}}
+    }}
 }
 
 asmsub  restore_irq() clobbers(A) {
@@ -994,8 +1296,8 @@ asmsub  restore_irq() clobbers(A) {
 	    lda  _orig_irqvec+1
 	    sta  cx16.CINV+1
 	    lda  cx16.VERA_IEN
-	    and  #%11110000     ; disable all Vera IRQs
-	    ora  #%00000001     ; enable only the vsync Irq
+	    and  #%11110000     ; disable all Vera IRQs but the vsync
+	    ora  #%00000001
 	    sta  cx16.VERA_IEN
 	    cli
 	    rts
@@ -1004,15 +1306,16 @@ _orig_irqvec    .word  0
 }
 
 asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0) clobbers(A) {
+    ; Sets the handler for the LINE interrupt, and enable (only) that interrupt.
 	%asm {{
+            sei
             sta  _modified+1
             sty  _modified+2
             lda  cx16.r0
             ldy  cx16.r0+1
-            sei
             lda  cx16.VERA_IEN
-            and  #%11110000     ; clear other IRQs
-            ora  #%00000010     ; enable the line (raster) irq
+            and  #%11110000     ; disable all irqs but the line(raster) one
+            ora  #%00000010
             sta  cx16.VERA_IEN
             lda  cx16.r0
             ldy  cx16.r0+1
@@ -1025,13 +1328,13 @@ asmsub  set_rasterirq(uword handler @AY, uword rasterpos @R0) clobbers(A) {
             rts
 
 _raster_irq_handler
-            jsr  set_irq._irq_handler_init
-_modified   jsr  $ffff                      ; modified
-            jsr  set_irq._irq_handler_end
+            jsr  sys.save_prog8_internals
+            cld
+_modified   jsr  $ffff    ; modified
+            jsr  sys.restore_prog8_internals
             ; end irq processing - don't use kernal's irq handling
-            lda  cx16.VERA_ISR
-            ora  #%00000010
-            sta  cx16.VERA_ISR      ; clear Vera line irq status
+            lda  #2
+            tsb  cx16.VERA_ISR      ; clear Vera line irq status
             ply
             plx
             pla
@@ -1041,16 +1344,19 @@ _modified   jsr  $ffff                      ; modified
 
 asmsub  set_rasterline(uword line @AY) {
     %asm {{
-        sta  cx16.VERA_IRQ_LINE_L
-        lda  cx16.VERA_IEN
-        and  #%01111111
-        sta  cx16.VERA_IEN
+        php
+        sei
+        sta  cx16.VERA_IRQLINE_L
         tya
         lsr  a
-        ror  a
-        and  #%10000000
-        ora  cx16.VERA_IEN
-        sta  cx16.VERA_IEN
+        bcs  +
+        lda  #%10000000
+        trb  cx16.VERA_IEN
+        plp
+        rts
++       lda  #%10000000
+        tsb  cx16.VERA_IEN
+        plp
         rts
     }}
 }
@@ -1250,13 +1556,67 @@ _longcopy
         }}
     }
 
-    inline asmsub exit(ubyte returnvalue @A) {
+    inline asmsub disable_caseswitch() {
+        %asm {{
+            lda  #8
+            jsr  cbm.CHROUT
+        }}
+    }
+
+    inline asmsub enable_caseswitch() {
+        %asm {{
+            lda  #9
+            jsr  cbm.CHROUT
+        }}
+    }
+
+    asmsub save_prog8_internals() {
+        %asm {{
+            lda  P8ZP_SCRATCH_B1
+            sta  save_SCRATCH_ZPB1
+            lda  P8ZP_SCRATCH_REG
+            sta  save_SCRATCH_ZPREG
+            lda  P8ZP_SCRATCH_W1
+            sta  save_SCRATCH_ZPWORD1
+            lda  P8ZP_SCRATCH_W1+1
+            sta  save_SCRATCH_ZPWORD1+1
+            lda  P8ZP_SCRATCH_W2
+            sta  save_SCRATCH_ZPWORD2
+            lda  P8ZP_SCRATCH_W2+1
+            sta  save_SCRATCH_ZPWORD2+1
+            rts
+save_SCRATCH_ZPB1	.byte  0
+save_SCRATCH_ZPREG	.byte  0
+save_SCRATCH_ZPWORD1	.word  0
+save_SCRATCH_ZPWORD2	.word  0
+        }}
+    }
+
+    asmsub restore_prog8_internals() {
+        %asm {{
+            lda  save_prog8_internals.save_SCRATCH_ZPB1
+            sta  P8ZP_SCRATCH_B1
+            lda  save_prog8_internals.save_SCRATCH_ZPREG
+            sta  P8ZP_SCRATCH_REG
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD1
+            sta  P8ZP_SCRATCH_W1
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD1+1
+            sta  P8ZP_SCRATCH_W1+1
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD2
+            sta  P8ZP_SCRATCH_W2
+            lda  save_prog8_internals.save_SCRATCH_ZPWORD2+1
+            sta  P8ZP_SCRATCH_W2+1
+            rts
+        }}
+    }
+
+    asmsub exit(ubyte returnvalue @A) {
         ; -- immediately exit the program with a return code in the A register
         %asm {{
-            jsr  cbm.CLRCHN		; reset i/o channels
+            sta  cleanup_at_exit._exitcode
             ldx  prog8_lib.orig_stackpointer
             txs
-            rts		; return to original caller
+            jmp  cleanup_at_exit
         }}
     }
 
@@ -1266,4 +1626,31 @@ _longcopy
             ldy  #>prog8_program_end
         }}
     }
+
+    inline asmsub push(ubyte value @A) {
+        %asm {{
+            pha
+        }}
+    }
+
+    inline asmsub pushw(uword value @AY) {
+        %asm {{
+            pha
+            phy
+        }}
+    }
+
+    inline asmsub pop() -> ubyte @A {
+        %asm {{
+            pla
+        }}
+    }
+
+    inline asmsub popw() -> uword @AY {
+        %asm {{
+            ply
+            pla
+        }}
+    }
+
 }

compiler/res/prog8lib/cx16/textio.p8

@@ -1,4 +1,6 @@
 ; Prog8 definitions for the Text I/O and Screen routines for the CommanderX16
+; All routines work with Screencode character encoding, except `print`, `chrout` and `input_chars`,
+; these work with PETSCII encoding instead.
 
 %import syslib
 %import conv
@@ -6,7 +8,7 @@
 
 txt {
 
-    %option no_symbol_prefixing
+    %option no_symbol_prefixing, ignore_unused
 
 const ubyte DEFAULT_WIDTH = 80
 const ubyte DEFAULT_HEIGHT = 60
@@ -42,6 +44,43 @@ asmsub column(ubyte col @A) clobbers(A, X, Y) {
     }}
 }
 
+asmsub get_column() -> ubyte @Y {
+    %asm {{
+        sec
+        jmp cbm.PLOT
+    }}
+}
+
+asmsub row(ubyte rownum @A) clobbers(A, X, Y) {
+    ; ---- set the cursor on the given row (starting with 0) on the current line
+    %asm {{
+        sec
+        jsr  cbm.PLOT
+        tax
+        clc
+        jmp  cbm.PLOT
+    }}
+}
+
+asmsub get_row() -> ubyte @X {
+    %asm {{
+        sec
+        jmp cbm.PLOT
+    }}
+}
+
+sub get_cursor(uword colptr, uword rowptr) {
+    %asm {{
+        sec
+        jsr cbm.PLOT
+        tya
+        ldy #$00
+        sta (colptr),y
+        txa
+        sta (rowptr),y
+    }}
+}
+
 asmsub  fill_screen (ubyte character @ A, ubyte color @ Y) clobbers(A, X)  {
 	; ---- fill the character screen with the given fill character and character color.
 	%asm {{
@@ -384,10 +423,10 @@ _nextline
 	}}
 }
 
-romsub $FFD2 = chrout(ubyte character @ A)    ; for consistency. You can also use cbm.CHROUT directly ofcourse.
+romsub $FFD2 = chrout(ubyte character @ A)    ; for consistency. You can also use cbm.CHROUT directly ofcourse. Note: takes a PETSCII encoded character.
 
 asmsub  print (str text @ AY) clobbers(A,Y)  {
-	; ---- print null terminated string from A/Y
+	; ---- print null terminated string, in PETSCII encoding, 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 cbm.CHROUT of that single char.
@@ -572,7 +611,8 @@ asmsub  print_w  (word value @ AY) clobbers(A,X,Y)  {
 }
 
 asmsub  input_chars  (uword buffer @ AY) clobbers(A) -> ubyte @ Y  {
-	; ---- Input a string (max. 80 chars) from the keyboard. Returns length in Y. (string is terminated with a 0 byte as well)
+	; ---- Input a string (max. 80 chars) from the keyboard, in PETSCII encoding.
+	;      Returns length in Y. (string is terminated with a 0 byte as well)
 	;      It assumes the keyboard is selected as I/O channel!
 
 	%asm {{
@@ -756,4 +796,11 @@ asmsub height() clobbers(X, Y) -> ubyte @A {
     }}
 }
 
+asmsub waitkey() -> ubyte @A {
+    %asm {{
+-       jsr cbm.GETIN
+        beq -
+        rts
+    }}
+}
 }

compiler/res/prog8lib/cx16/verafx.p8

@@ -1,10 +1,10 @@
-; Experimental Vera FX support.
+; Somewhat experimental Vera FX support.
 ; Docs:
 ; https://github.com/X16Community/x16-docs/blob/master/VERA%20FX%20Reference.md
 ; https://docs.google.com/document/d/1q34uWOiM3Be2pnaHRVgSdHySI-qsiQWPTo_gfE54PTg/edit
 
 verafx {
-    %option no_symbol_prefixing
+    %option no_symbol_prefixing, ignore_unused
 
     sub available() -> bool {
         ; returns true if Vera FX is available (Vera V0.3.1 or later), false if not.
@@ -21,7 +21,7 @@ verafx {
         return cx16.r1L
     }
 
-    sub clear(ubyte vbank, uword vaddr, ubyte data, uword amountof32bits) {
+    sub clear(ubyte vbank, uword vaddr, ubyte data, uword num_longwords) {
         ; use cached 4-byte write to quickly clear a portion of the video memory to a given byte value
         ; this routine is around 3 times faster as gfx2.clear_screen()
         cx16.VERA_CTRL = 0
@@ -37,24 +37,24 @@ verafx {
         cx16.VERA_FX_MULT = 0
         cx16.VERA_FX_CTRL = %01000000    ; cache write enable
 
-        if (amountof32bits & %1111110000000011) == 0 {
-            repeat lsb(amountof32bits >> 2)
-                unroll 4 cx16.VERA_DATA0=0       ; write 4 bytes at a time, unrolled
+        if (num_longwords & %1111110000000011) == 0 {
+            repeat lsb(num_longwords >> 2)
+                unroll 4 cx16.VERA_DATA0=0       ; write 4*4 bytes at a time, unrolled
         }
-        else if (amountof32bits & %1111111000000001) == 0 {
-            repeat lsb(amountof32bits >> 1)
-                unroll 2 cx16.VERA_DATA0=0       ; write 4 bytes at a time, unrolled
+        else if (num_longwords & %1111111000000001) == 0 {
+            repeat lsb(num_longwords >> 1)
+                unroll 2 cx16.VERA_DATA0=0       ; write 2*4 bytes at a time, unrolled
         }
-        else if (lsb(amountof32bits) & 3) == 0 {
-            repeat amountof32bits >> 2
-                unroll 4 cx16.VERA_DATA0=0       ; write 4 bytes at a time, unrolled
+        else if (lsb(num_longwords) & 3) == 0 {
+            repeat num_longwords >> 2
+                unroll 4 cx16.VERA_DATA0=0       ; write 4*4 bytes at a time, unrolled
         }
-        else if (lsb(amountof32bits) & 1) == 0 {
-            repeat amountof32bits >> 1
-                unroll 2 cx16.VERA_DATA0=0       ; write 4 bytes at a time, unrolled
+        else if (lsb(num_longwords) & 1) == 0 {
+            repeat num_longwords >> 1
+                unroll 2 cx16.VERA_DATA0=0       ; write 2*4 bytes at a time, unrolled
         }
         else {
-            repeat amountof32bits
+            repeat num_longwords
                 cx16.VERA_DATA0=0       ; write 4 bytes at a time
         }
 
@@ -62,13 +62,66 @@ verafx {
         cx16.VERA_CTRL = 0
     }
 
+    sub copy(ubyte srcbank, uword srcaddr, ubyte tgtbank, uword tgtaddr, uword num_longwords) {
+        ; use cached 4-byte writes to quickly copy a portion of the video memory to somewhere else
+        ; this routine is about 50% faster as a plain byte-by-byte copy
+        cx16.VERA_CTRL = 1
+        cx16.VERA_ADDR_H = srcbank | %00010000       ; source: 1-byte increment
+        cx16.VERA_ADDR_M = msb(srcaddr)
+        cx16.VERA_ADDR_L = lsb(srcaddr)
+        cx16.VERA_CTRL = 0
+        cx16.VERA_ADDR_H = tgtbank | %00110000       ; target: 4-byte increment
+        cx16.VERA_ADDR_M = msb(tgtaddr)
+        cx16.VERA_ADDR_L = lsb(tgtaddr)
+        cx16.VERA_CTRL = 2<<1       ; dcsel = 2
+        cx16.VERA_FX_MULT = 0
+        cx16.VERA_FX_CTRL = %01100000    ; cache write enable + cache fill enable
+        cx16.r0 = num_longwords
+
+        if (cx16.r0L & 1) == 0 {
+            repeat cx16.r0>>1 {
+                %asm {{
+                    lda  cx16.VERA_DATA1    ; fill cache with 4 source bytes...
+                    lda  cx16.VERA_DATA1
+                    lda  cx16.VERA_DATA1
+                    lda  cx16.VERA_DATA1
+                    stz  cx16.VERA_DATA0    ; write 4 bytes at once.
+                    lda  cx16.VERA_DATA1    ; fill cache with 4 source bytes...
+                    lda  cx16.VERA_DATA1
+                    lda  cx16.VERA_DATA1
+                    lda  cx16.VERA_DATA1
+                    stz  cx16.VERA_DATA0    ; write 4 bytes at once.
+                }}
+            }
+        } else {
+            repeat cx16.r0 {
+                %asm {{
+                    lda  cx16.VERA_DATA1    ; fill cache with 4 source bytes...
+                    lda  cx16.VERA_DATA1
+                    lda  cx16.VERA_DATA1
+                    lda  cx16.VERA_DATA1
+                    stz  cx16.VERA_DATA0    ; write 4 bytes at once.
+                }}
+            }
+        }
+
+        cx16.VERA_FX_CTRL = 0    ; cache write disable
+        cx16.VERA_CTRL = 0
+    }
+
     ; unsigned multiplication just passes the values as signed to muls
     ; if you do this yourself in your call to muls, it will save a few instructions.
     sub mult(uword value1, uword value2) -> uword {
+        ; Returns the lower 16 bits of the 32 bits result,
+        ; the upper 16 bits are stored in cx16.r0 so you can access those separately.
+        ; It's not part of the subroutine's signature to avoid awkward use of multiple returnvalues.
         return muls(value1 as word, value2 as word) as uword
     }
 
-    asmsub muls(word value1 @R0, word value2 @R1) -> word @AY {
+    asmsub muls(word value1 @R0, word value2 @R1) clobbers(X) -> word @AY {
+        ; Returns the lower 16 bits of the 32 bits result in AY,
+        ; the upper 16 bits are stored in cx16.r0 so you can access those separately.
+        ; It's not part of the subroutine's signature to avoid awkward use of multiple returnvalues.
         %asm {{
             lda  #(2 << 1)
             sta  cx16.VERA_CTRL        ; $9F25
@@ -104,14 +157,13 @@ verafx {
             sta  cx16.VERA_ADDR_H     ; so we can read out the result
             lda  cx16.VERA_DATA0
             ldy  cx16.VERA_DATA0
+            ldx  cx16.VERA_DATA0      ; store the upper 16 bits of the result in r0
+            stx  cx16.r0
+            ldx  cx16.VERA_DATA0
+            stx  cx16.r0+1
             stz  cx16.VERA_FX_CTRL    ; Cache write disable
             stz  cx16.VERA_CTRL       ; reset DCSEL
             rts
-; we skip the upper 16 bits of the result:
-;            lda  cx16.VERA_DATA0
-;            sta  $0402
-;            lda  cx16.VERA_DATA0
-;            sta  $0403
         }}
     }