Correctly handle delimiters in character operands

We weren't checking to see if character operands matched their
delimiters, so bad code like "LDA #'''" was being generated.

There wasn't a test for this in 2006-operand-formats, so the test
has been updated with single and double quotes in low and high ASCII.

Asm65/Formatter.cs

@@ -119,10 +119,9 @@ namespace Asm65 {
             public char OpenDelim { get; private set; }
             public char CloseDelim { get; private set; }
             public string Suffix { get; private set; }
+            public string FormatStr { get; private set; }
 
-            public DelimiterDef(char delim) {
+            public DelimiterDef(char delim) : this(string.Empty, delim, delim, string.Empty) {
-                OpenDelim = CloseDelim = delim;
-                Prefix = Suffix = string.Empty;
             }
             public DelimiterDef(string prefix, char openDelim, char closeDelim, string suffix) {
                 Debug.Assert(prefix != null);
@@ -131,6 +130,15 @@ namespace Asm65 {
                 OpenDelim = openDelim;
                 CloseDelim = closeDelim;
                 Suffix = suffix;
+
+                // Generate format string.
+                StringBuilder sb = new StringBuilder();
+                sb.Append(Prefix);
+                sb.Append(OpenDelim);
+                sb.Append("{0}");
+                sb.Append(CloseDelim);
+                sb.Append(Suffix);
+                FormatStr = sb.ToString();
             }
             public override string ToString() {
                 return Prefix + OpenDelim + '#' + CloseDelim + Suffix;
@@ -291,13 +299,6 @@ namespace Asm65 {
         private string mAddrFormatNoBank;
         private string mAddrFormatWithBank;
 
-        // Character data delimiter format strings, processed from the delimiter patterns.
-        private string mCharDelimAsciiFmt;
-        private string mCharDelimHighAsciiFmt;
-        private string mCharDelimC64PetsciiFmt;
-        private string mCharDelimC64ScreenCodeFmt;
-
-
         // Generated opcode strings.  The index is the bitwise OR of the opcode value and
         // the disambiguation value.  In most cases this just helps us avoid calling
         // ToUpper incessantly.
@@ -431,24 +432,6 @@ namespace Asm65 {
                 Debug.WriteLine("NOTE: char delimiters not set");
                 chrDelim = DelimiterSet.GetDefaultCharDelimiters();
             }
-            mCharDelimAsciiFmt =
-                DelimiterDefToFormat(chrDelim.Get(CharEncoding.Encoding.Ascii));
-            mCharDelimHighAsciiFmt =
-                DelimiterDefToFormat(chrDelim.Get(CharEncoding.Encoding.HighAscii));
-            mCharDelimC64PetsciiFmt =
-                DelimiterDefToFormat(chrDelim.Get(CharEncoding.Encoding.C64Petscii));
-            mCharDelimC64ScreenCodeFmt =
-                DelimiterDefToFormat(chrDelim.Get(CharEncoding.Encoding.C64ScreenCode));
-        }
-
-        private string DelimiterDefToFormat(DelimiterDef def) {
-            StringBuilder sb = new StringBuilder();
-            sb.Append(def.Prefix);
-            sb.Append(def.OpenDelim);
-            sb.Append("{0}");
-            sb.Append(def.CloseDelim);
-            sb.Append(def.Suffix);
-            return sb.ToString();
         }
 
         /// <summary>
@@ -544,37 +527,36 @@ namespace Asm65 {
                 return FormatHexValue(value, 2);
             }
 
-            string fmt;
+            DelimiterDef delimDef = mFormatConfig.mCharDelimiters.Get(enc);
+            string fmt = delimDef.FormatStr;
+            Debug.Assert(fmt != null);
+
             CharEncoding.Convert conv;
             switch (enc) {
                 case CharEncoding.Encoding.Ascii:
-                    fmt = mCharDelimAsciiFmt;
                     conv = CharEncoding.ConvertAscii;
                     break;
                 case CharEncoding.Encoding.HighAscii:
-                    fmt = mCharDelimHighAsciiFmt;
                     conv = CharEncoding.ConvertHighAscii;
                     break;
                 case CharEncoding.Encoding.C64Petscii:
-                    fmt = mCharDelimC64PetsciiFmt;
                     conv = CharEncoding.ConvertC64Petscii;
                     break;
                 case CharEncoding.Encoding.C64ScreenCode:
-                    fmt = mCharDelimC64ScreenCodeFmt;
                     conv = CharEncoding.ConvertC64ScreenCode;
                     break;
                 default:
                     return FormatHexValue(value, 2);
             }
-            if (string.IsNullOrEmpty(fmt)) {
-                return FormatHexValue(value, 2);
-            }
 
             char ch = conv((byte)value);
-            if (ch == CharEncoding.UNPRINTABLE_CHAR) {
+            if (ch == CharEncoding.UNPRINTABLE_CHAR || ch == delimDef.OpenDelim ||
+                    ch == delimDef.CloseDelim) {
+                // We might be able to do better with delimiter clashes, e.g. '\'', but
+                // that's assembler-specific.
                 return FormatHexValue(value, 2);
             } else {
-                // possible optimization: replace fmt with a prefix/suffix pair, and just concat
+                // Possible optimization: replace fmt with a prefix/suffix pair, and just concat
                 return string.Format(fmt, ch);
             }
         }

SourceGen/SGTestData/2006-operand-formats.dis65

@@ -1,6 +1,6 @@
 ### 6502bench SourceGen dis65 v1.0 ###
 {
-"_ContentVersion":1,"FileDataLength":125,"FileDataCrc32":611887040,"ProjectProps":{
+"_ContentVersion":1,"FileDataLength":133,"FileDataCrc32":-811370049,"ProjectProps":{
 "CpuName":"65816","IncludeUndocumentedInstr":false,"EntryFlags":33489103,"AnalysisParams":{
 "AnalyzeUncategorizedData":true,"MinCharsForString":4,"SeekNearbyTargets":true},
 "PlatformSymbolFileIdentifiers":[],"ExtensionScriptFileIdentifiers":[],"ProjectSyms":{
@@ -19,8 +19,8 @@
 "UserLabels":{
 "63":{
 "Label":"skipdata","Value":4159,"Source":"User","Type":"LocalOrGlobalAddr"},
-"104":{
+"112":{
-"Label":"more_ascii","Value":4200,"Source":"User","Type":"LocalOrGlobalAddr"}},
+"Label":"more_ascii","Value":4208,"Source":"User","Type":"LocalOrGlobalAddr"}},
 "OperandFormats":{
 "4":{
 "Length":2,"Format":"NumericLE","SubFormat":"Hex","SymbolRef":null},
@@ -93,33 +93,43 @@
 "92":{
 "Length":2,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
 "94":{
-"Length":3,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
-"97":{
-"Length":3,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
-"104":{
-"Length":1,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
-"105":{
-"Length":1,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
-"106":{
 "Length":2,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
-"109":{
+"96":{
-"Length":2,"Format":"NumericLE","SubFormat":"Address","SymbolRef":null},
+"Length":2,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
-"111":{
+"98":{
-"Length":3,"Format":"NumericLE","SubFormat":"Address","SymbolRef":null},
+"Length":2,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
+"100":{
+"Length":2,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
+"102":{
+"Length":3,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
+"105":{
+"Length":3,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
+"108":{
+"Length":3,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
+"112":{
+"Length":1,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
+"113":{
+"Length":1,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
 "114":{
+"Length":2,"Format":"NumericLE","SubFormat":"Ascii","SymbolRef":null},
+"117":{
+"Length":2,"Format":"NumericLE","SubFormat":"Address","SymbolRef":null},
+"119":{
+"Length":3,"Format":"NumericLE","SubFormat":"Address","SymbolRef":null},
+"122":{
 "Length":2,"Format":"NumericBE","SubFormat":"Address","SymbolRef":null},
-"116":{
+"124":{
 "Length":1,"Format":"NumericLE","SubFormat":"Symbol","SymbolRef":{
 "Label":"more_ascii","Part":"Low"}},
-"117":{
+"125":{
 "Length":1,"Format":"NumericLE","SubFormat":"Symbol","SymbolRef":{
 "Label":"more_ascii","Part":"High"}},
-"118":{
+"126":{
 "Length":2,"Format":"NumericLE","SubFormat":"Symbol","SymbolRef":{
 "Label":"more_ascii","Part":"Low"}},
-"120":{
+"128":{
 "Length":3,"Format":"NumericLE","SubFormat":"Symbol","SymbolRef":{
 "Label":"more_ascii","Part":"Low"}},
-"123":{
+"131":{
 "Length":2,"Format":"NumericBE","SubFormat":"Symbol","SymbolRef":{
 "Label":"more_ascii","Part":"Low"}}}}

SourceGen/SGTestData/Expected/2006-operand-formats_64tass.S

@@ -36,11 +36,15 @@ skipdata lda    #'h'
         lda     @l'h'
         lda     #$1f
         lda     #' '
+        lda     #'"'
+        lda     #$27
         lda     #'~'
         lda     #$7f
         lda     #$80
         lda     #$9f
         lda     #' ' | $80
+        lda     #'"' | $80
+        lda     #$a7
         lda     #'~' | $80
         lda     #$ff
         rep     #'0'
@@ -62,4 +66,4 @@ more_ascii .byte 'h'
         .byte   >more_ascii
         .word   more_ascii
         .long   more_ascii
-        .byte   $10,$68
+        .byte   $10,$70

SourceGen/SGTestData/Expected/2006-operand-formats_Merlin32.S

@@ -33,11 +33,15 @@ skipdata lda   #'h'
          ldal  'h'
          lda   #$1f
          lda   #' '
+         lda   #'"'
+         lda   #$27
          lda   #'~'
          lda   #$7f
          lda   #$80
          lda   #$9f
          lda   #" "
+         lda   #$a2
+         lda   #"'"
          lda   #"~"
          lda   #$ff
          rep   #'0'

SourceGen/SGTestData/Expected/2006-operand-formats_acme.S

@@ -36,11 +36,15 @@ skipdata lda    #'h'
         lda+3   'h'
         lda     #$1f
         lda     #' '
+        lda     #'"'
+        lda     #$27
         lda     #'~'
         lda     #$7f
         lda     #$80
         lda     #$9f
         lda     #' ' | $80
+        lda     #'"' | $80
+        lda     #$a7
         lda     #'~' | $80
         lda     #$ff
         rep     #'0'
@@ -62,4 +66,4 @@ more_ascii !byte 'h'
         !byte   >more_ascii
         !word   more_ascii
         !24     more_ascii
-        !byte   $10,$68
+        !byte   $10,$70

SourceGen/SGTestData/Expected/2006-operand-formats_cc65.S

@@ -37,11 +37,15 @@ skipdata: lda    #'h'
          lda     f:'h'
          lda     #$1f
          lda     #' '
+         lda     #'"'
+         lda     #$27
          lda     #'~'
          lda     #$7f
          lda     #$80
          lda     #$9f
          lda     #' ' | $80
+         lda     #'"' | $80
+         lda     #$a7
          lda     #'~' | $80
          lda     #$ff
          rep     #'0'

SourceGen/SGTestData/Expected/2006-operand-formats_cc65.cfg

@@ -1,7 +1,7 @@
 # 6502bench SourceGen generated linker script for 2006-operand-formats
 MEMORY {
     MAIN: file=%O, start=%S, size=65536;
-#    MEM000: file=%O, start=$1000, size=125;
+#    MEM000: file=%O, start=$1000, size=133;
 }
 SEGMENTS {
     CODE: load=MAIN, type=rw;

SourceGen/SGTestData/Source/2006-operand-formats.S

@@ -43,11 +43,15 @@
 
         lda     #$1f
         lda     #$20
+        lda     #$22
+        lda     #$27
         lda     #$7e
         lda     #$7f
         lda     #$80
         lda     #$9f
         lda     #$a0
+        lda     #$a2
+        lda     #$a7
         lda     #$fe
         lda     #$ff