Character literals

CHANGELOG.md

@@ -6,6 +6,8 @@
 
 * Added multiple new text codecs.
 
+* Added character literals.
+
 * Fixed several bugs, most importantly invalid offsets for branching instructions.
 
 * Other improvements.

docs/lang/literals.md

@@ -36,7 +36,7 @@ Currently available encodings:
 
 * `bbc` – BBC Micro and ZX Spectrum character set
 
-* `jis` – JIS X 0201
+* `jis` or `jisx` – JIS X 0201
 
 * `iso_de`, `iso_no`, `iso_se`, `iso_yu` – various variants of ISO/IEC-646
  
@@ -46,6 +46,13 @@ When programming for Commodore,
 use `pet` for strings you're printing using standard I/O routines
 and `scr` for strings you're copying to screen memory directly.
 
+## Character literals
+
+Character literals are surrounded by single quotes and followed by the name of the encoding: 
+
+    'x' ascii
+
+From the type system point of view, they are constants of type byte.
 
 ## Array initialisers 
 

src/main/scala/millfork/parser/MfParser.scala

@@ -53,6 +53,28 @@ case class MfParser(filename: String, input: String, currentDirectory: String, o
 
   //  def operator: P[String] = P(CharsWhileIn("!-+*/><=~|&^", min=1).!) // TODO: only valid operators
 
+  private val invalidCharLiteralTypes = Set[Int](
+    Character.LINE_SEPARATOR,
+    Character.PARAGRAPH_SEPARATOR,
+    Character.CONTROL,
+    Character.PRIVATE_USE,
+    Character.SURROGATE,
+    Character.UNASSIGNED)
+
+  def charAtom: P[LiteralExpression] = for {
+    p <- position()
+    c <- "'" ~/ CharPred(c => c >= ' ' && !invalidCharLiteralTypes(Character.getType(c))).! ~/ "'"
+    co <- HWS ~ codec
+  } yield {
+    co.encode(Some(p), c.charAt(0)) match {
+      case List(value) =>
+        LiteralExpression(value, 1)
+      case _ =>
+        ErrorReporting.error(s"Character `$c` cannot be encoded as one byte", Some(p))
+        LiteralExpression(0, 1)
+    }
+  }
+
   // TODO: 3-byte types
   def size(value: Int, wordLiteral: Boolean, longLiteral: Boolean): Int =
     if (value > 255 || value < -128 || wordLiteral)
@@ -120,7 +142,7 @@ case class MfParser(filename: String, input: String, currentDirectory: String, o
       LiteralExpression(value, size(value, s.length > 4, s.length > 8)).pos(p)
     }
 
-  val literalAtom: P[LiteralExpression] = binaryAtom | hexAtom | octalAtom | quaternaryAtom | decimalAtom
+  val literalAtom: P[LiteralExpression] = charAtom | binaryAtom | hexAtom | octalAtom | quaternaryAtom | decimalAtom
 
   val atom: P[Expression] = P(literalAtom | (position() ~ identifier).map { case (p, i) => VariableExpression(i).pos(p) })
 
@@ -200,7 +222,7 @@ case class MfParser(filename: String, input: String, currentDirectory: String, o
 
   val doubleQuotedString: P[List[Char]] = P("\"" ~/ CharsWhile(c => c != '\"' && c != '\n' && c != '\r').! ~ "\"").map(_.toList)
 
-  val codec: P[TextCodec] = P(position() ~ identifier).map {
+  def codec: P[TextCodec] = P(position() ~ identifier).map {
     case (_, "ascii") => TextCodec.Ascii
     case (_, "petscii") => TextCodec.Petscii
     case (_, "pet") => TextCodec.Petscii

src/test/scala/millfork/test/TextCodecSuite.scala

@@ -0,0 +1,25 @@
+package millfork.test
+
+import millfork.test.emu.{EmuBenchmarkRun, EmuUnoptimizedRun}
+import org.scalatest.{FunSuite, Matchers}
+
+/**
+  * @author Karol Stasiak
+  */
+class TextCodecSuite extends FunSuite with Matchers {
+
+  test("Char literals") {
+    val m = EmuUnoptimizedRun(
+      """
+        | void main() {
+        |   if 'a'ascii != 'a' ascii  { poke($bfff, 0) }
+        |   if '¥'jis   != '\' ascii  { poke($bffe, 0) }
+        |   if '£'pet   != '\' ascii  { poke($bffd, 0) }
+        |   if '£'bbc   != 'é' iso_se { poke($bffc, 0) }
+        | }
+        | macro asm void poke(word const addr, byte a) {
+        |   STA addr
+        | }
+      """.stripMargin)
+  }
+}