Finish the underscore handling in the label localizer

Correctly handle pre-existing underscores and avoidance of
"reserved" labels.

Also, add more underscores to 2012-label-localizer to exercise
the code.

(issue #16)

SourceGen/AsmGen/GenAndAsm.cs

@@ -211,7 +211,7 @@ namespace SourceGen.AsmGen {
         /// </summary>
         private void UpdateAssemblerControls() {
             bool asmConf = IsAssemblerConfigured();
-            Debug.WriteLine("ID=" + mSelectedAssemblerId + " asmConf=" + asmConf);
+            //Debug.WriteLine("ID=" + mSelectedAssemblerId + " asmConf=" + asmConf);
             configureAsmLinkLabel.Visible = !asmConf;
             if (mGenerationResults == null || !asmConf) {
                 runAssemblerButton.Enabled = false;

SourceGen/AsmGen/LabelLocalizer.cs

@@ -167,6 +167,8 @@ namespace SourceGen.AsmGen {
         /// Analyzes labels to identify which ones may be treated as non-global.
         /// </summary>
         public void Analyze() {
+            Debug.Assert(LocalPrefix.Length > 0);
+
             mGlobalFlags.SetAll(false);
 
             // Currently we only support the "local labels have scope that ends at a global
@@ -317,6 +319,8 @@ namespace SourceGen.AsmGen {
         /// 
         /// This may be called even if label localization is disabled.  In that case we just
         /// create an empty label map and populate as needed.
+        /// 
+        /// Only call this if underscores are used to indicate local labels.
         /// </summary>
         public void MaskLeadingUnderscores() {
             bool allGlobal = false;
@@ -325,6 +329,14 @@ namespace SourceGen.AsmGen {
                 LabelMap = new Dictionary<string, string>();
             }
 
+            // Throw out the original local label generation.
+            LabelMap.Clear();
+
+            // Use this to test for uniqueness.  We add all labels here as we go, not just the
+            // ones being remapped.  For each label we either add the original or the localized
+            // form.
+            SortedList<string, string> allLabels = new SortedList<string, string>();
+
             for (int i = 0; i < mProject.FileDataLength; i++) {
                 Symbol sym = mProject.GetAnattrib(i).Symbol;
                 if (sym == null) {
@@ -332,12 +344,43 @@ namespace SourceGen.AsmGen {
                     continue;
                 }
 
-                if (sym.Label.StartsWith("_") && (allGlobal || mGlobalFlags[i])) {
+                string newLabel;
-                    Debug.WriteLine("Adjusting " + sym.Label);
+                if (allGlobal || mGlobalFlags[i]) {
-                    // TODO: uniquify
+                    // Global symbol.  Don't let it start with '_'.
-                    LabelMap[sym.Label] = "X" + sym.Label;
+                    if (sym.Label.StartsWith("_")) {
+                        // There's an underscore here that was added by the user.  Stick some
+                        // other character in front.
+                        newLabel = "X" + sym.Label;
+                    } else {
+                        // No change needed.
+                        newLabel = sym.Label;
+                    }
+                } else {
+                    // Local symbol.
+                    if (sym.Label.StartsWith("_")) {
+                        // The original starts with one or more underscores.  Adding another
+                        // will create a "__" label, which is reserved in 64tass.
+                        newLabel = "_X" + sym.Label;
+                    } else {
+                        newLabel = "_" + sym.Label;
+                    }
+                }
+
+                // Make sure it's unique.
+                string uniqueLabel = newLabel;
+                int uval = 1;
+                while (allLabels.ContainsKey(uniqueLabel)) {
+                    uniqueLabel = newLabel + uval.ToString();
+                }
+                allLabels.Add(uniqueLabel, uniqueLabel);
+
+                // If it's different, add it to the label map.
+                if (sym.Label != uniqueLabel) {
+                    LabelMap.Add(sym.Label, uniqueLabel);
                 }
             }
+
+            Debug.WriteLine("UMAP: allcount=" + allLabels.Count + " mapcount=" + LabelMap.Count);
         }
     }
 }

SourceGen/RuntimeData/Help/codegen.html

@@ -139,6 +139,8 @@ code, but also needs to know how to handle the corner cases.</p>
     label is local.  If you create labels with leading underscores that
     are not local, the labels must be altered to start with some other
     character, and made unique.</li>
+  <li>Labels starting with two underscores are "reserved".  Trying to
+    use them causes an error.</li>
   <li>By default, 64tass sets the first two bytes of the output file to
     the load address.  The <code>--nostart</code> flag is used to
     suppress this.</li>

SourceGen/SGTestData/2012-label-localizer.dis65

@@ -30,7 +30,7 @@
 "16":{
 "Label":"pastglob","Value":4112,"Source":"User","Type":"LocalOrGlobalAddr"},
 "15":{
-"Label":"nextglob","Value":4111,"Source":"User","Type":"GlobalAddr"},
+"Label":"X_start","Value":4111,"Source":"User","Type":"GlobalAddr"},
 "20":{
 "Label":"nlocal","Value":4116,"Source":"User","Type":"GlobalAddr"},
 "22":{
@@ -40,9 +40,9 @@
 "30":{
 "Label":"reach3G","Value":4126,"Source":"User","Type":"GlobalAddr"},
 "31":{
-"Label":"reach4","Value":4127,"Source":"User","Type":"LocalOrGlobalAddr"},
+"Label":"_reach4","Value":4127,"Source":"User","Type":"LocalOrGlobalAddr"},
 "1":{
-"Label":"start","Value":4097,"Source":"User","Type":"LocalOrGlobalAddr"},
+"Label":"_start","Value":4097,"Source":"User","Type":"LocalOrGlobalAddr"},
 "50":{
 "Label":"gtest1","Value":4146,"Source":"User","Type":"GlobalAddr"},
 "54":{
@@ -52,7 +52,7 @@
 "61":{
 "Label":"topglob","Value":4157,"Source":"User","Type":"GlobalAddr"},
 "70":{
-"Label":"globalnm","Value":4166,"Source":"User","Type":"LocalOrGlobalAddr"},
+"Label":"globalnm","Value":4166,"Source":"User","Type":"GlobalAddr"},
 "76":{
 "Label":"nglobal","Value":4172,"Source":"User","Type":"LocalOrGlobalAddr"},
 "82":{
@@ -60,7 +60,11 @@
 "77":{
 "Label":"globlat","Value":4173,"Source":"User","Type":"GlobalAddr"},
 "83":{
-"Label":"EXCESSIVELY_LONG_LABEL","Value":4179,"Source":"User","Type":"GlobalAddr"}},
+"Label":"EXCESSIVELY_LONG_LABEL","Value":4179,"Source":"User","Type":"GlobalAddr"},
+"14":{
+"Label":"__nopped","Value":4110,"Source":"User","Type":"LocalOrGlobalAddr"},
+"13":{
+"Label":"start","Value":4109,"Source":"User","Type":"LocalOrGlobalAddr"}},
 "OperandFormats":{
 "23":{
 "Length":3,"Format":"NumericLE","SubFormat":"Symbol","SymbolRef":{
@@ -70,6 +74,6 @@
 "Label":"reach2","Part":"Low"}},
 "32":{
 "Length":3,"Format":"NumericLE","SubFormat":"Symbol","SymbolRef":{
-"Label":"reach4","Part":"Low"}},
+"Label":"_reach4","Part":"Low"}},
 "58":{
 "Length":3,"Format":"NumericLE","SubFormat":"Hex","SymbolRef":null}}}

SourceGen/SGTestData/Expected/2012-label-localizer_64tass.S

@@ -5,13 +5,14 @@ REALLYLONGLABELNAME = $8888 ;that's a long name
         .as
         .xs
         nop
-start   lda     _L100D
+X_start lda     _start
-        lda     nextglob
+        lda     X_start1
         lda     pastglob
-        lda     _L100E
+        lda     _X__nopped
-_L100D  nop
+_start  nop
-_L100E  nop
+_X__nopped
-nextglob
+        nop
+X_start1
         nop
 pastglob
         nop
@@ -22,12 +23,13 @@ reach1G nop
         lda     _reach2+2
 _reach2 nop
 reach3G nop
-_reach4 nop
+_X_reach4
-        lda     _reach4-2
+        nop
+        lda     _X_reach4-2
         lda     $00
         beq     _L102D
-        jsr     _reach4
+        jsr     _X_reach4
-        jsr     start
+        jsr     X_start
 _L102D  lda     #$22
         lda     gtest2
 gtest1  nop
@@ -42,7 +44,7 @@ topglob nop
         nop
 _L1043  nop
         lda     #$44
-_globalnm
+globalnm
         jsr     _L104A
         nop
 _L104A  nop

SourceGen/SGTestData/Expected/2012-label-localizer_Merlin32.S

@@ -2,13 +2,14 @@ REALLYLONGLABELNAME equ $8888 ;that's a long name
 
          org   $1000
          nop
-start    lda   :L100D
+_start   lda   :start
-         lda   nextglob
+         lda   X_start
          lda   pastglob
-         lda   :L100E
+         lda   :__nopped
-:L100D   nop
+:start   nop
-:L100E   nop
+:__nopped
-nextglob nop
+         nop
+X_start  nop
 pastglob nop
          lda   nlocal
 nlocal   lda   #$11
@@ -17,12 +18,12 @@ reach1G  nop
          lda   :reach2+2
 :reach2  nop
 reach3G  nop
-:reach4  nop
+:_reach4 nop
-         lda   :reach4-2
+         lda   :_reach4-2
          lda   $00
          beq   :L102D
-         jsr   :reach4
+         jsr   :_reach4
-         jsr   start
+         jsr   _start
 :L102D   lda   #$22
          lda   gtest2
 gtest1   nop
@@ -37,8 +38,7 @@ topglob  nop
          nop
 :L1043   nop
          lda   #$44
-:globalnm
+globalnm jsr   :L104A
-         jsr   :L104A
          nop
 :L104A   nop
          nop

SourceGen/SGTestData/Expected/2012-label-localizer_cc65.S

@@ -5,14 +5,14 @@ REALLYLONGLABELNAME = $8888 ;that's a long name
          .a8
          .i8
          nop
-start:   lda     @L100D
+_start:  lda     @start
-         lda     nextglob
+         lda     X_start
          lda     pastglob
-         lda     @L100E
+         lda     @__nopped
-@L100D:  nop
+@start:  nop
-@L100E:  nop
+@__nopped:
-nextglob:
          nop
+X_start: nop
 pastglob:
          nop
          lda     nlocal
@@ -22,12 +22,13 @@ reach1G: nop
          lda     @reach2+2
 @reach2: nop
 reach3G: nop
-@reach4: nop
+@_reach4:
-         lda     @reach4-2
+         nop
+         lda     @_reach4-2
          lda     $00
          beq     @L102D
-         jsr     @reach4
+         jsr     @_reach4
-         jsr     start
+         jsr     _start
 @L102D:  lda     #$22
          lda     gtest2
 gtest1:  nop
@@ -42,7 +43,7 @@ topglob: nop
          nop
 @L1043:  nop
          lda     #$44
-@globalnm:
+globalnm:
          jsr     @L104A
          nop
 @L104A:  nop