Implement label validation and typedef resolution in C# ilasm #123002
+885
−74
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Consolidate instruction tokens into main grammar to establish proper lexer precedence over DOTTEDNAME and ID; reorder methodDecl alternatives to match instructions first; add custom attribute handling in method bodies - Implement visitor stubs for debug/symbol directives: VisitLanguageDecl, VisitEsHead, VisitExportHead, VisitExtSourceSpec, VisitFieldInit - Add typedef resolution in ResolveTypeDef for type aliases - Implement label validation: track declared vs referenced labels, report errors for undefined labels at method end - Allow recoverable errors in method bodies to still emit assembly - Add tests for label validation, typedef resolution, and error handling
github-actions
bot
added
the
needs-area-label
dotnet-policy-service
bot
added
the
community-contribution
am11
added
area-ILTools-coreclr
and removed
needs-area-label
Contributor
|
Tagging subscribers to this area: @JulieLeeMSFT |
jkoritzinsky
reviewed
Jan 8, 2026
| } | ||
| return MetadataTokens.EntityHandle(tokenType, 0); | ||
| // If all lists are empty, return a nil handle | ||
| return default(EntityHandle); |
Member
There was a problem hiding this comment.
We should return a handle for the right table at least (what we were doing before this change). That way we won't have a possible bug for callers that don't explicitly handle the Nil handle case.
| static FieldDefinitionHandle GetFieldHandleForList(IReadOnlyList<EntityBase> list, IReadOnlyList<EntityBase> listOwner, Func<EntityBase, IReadOnlyList<EntityBase>> getList, int ownerIndex) | ||
| { | ||
| var handle = GetHandleForList(list, listOwner, getList, ownerIndex, TableIndex.Field); | ||
| return handle.IsNil ? default : (FieldDefinitionHandle)handle; |
Member
There was a problem hiding this comment.
If we take the suggestion in GetHandleForList, we can convert all of these returns to just be a cast to the right handle type.
Comment on lines
+125
to
+137
| INSTR_NONE: ('nop'|'break'|'ldarg.0'|'ldarg.1'|'ldarg.2'|'ldarg.3'|'ldloc.0'|'ldloc.1'|'ldloc.2'|'ldloc.3'|'stloc.0'|'stloc.1'|'stloc.2'|'stloc.3'|'ldnull'|'ldc.i4.m1'|'ldc.i4.0'|'ldc.i4.1'|'ldc.i4.2'|'ldc.i4.3'|'ldc.i4.4'|'ldc.i4.5'|'ldc.i4.6'|'ldc.i4.7'|'ldc.i4.8'|'dup'|'pop'|'ret'|'ldind.i1'|'ldind.u1'|'ldind.i2'|'ldind.u2'|'ldind.i4'|'ldind.u4'|'ldind.i8'|'ldind.i'|'ldind.r4'|'ldind.r8'|'ldind.ref'|'stind.ref'|'stind.i1'|'stind.i2'|'stind.i4'|'stind.i8'|'stind.r4'|'stind.r8'|'add'|'sub'|'mul'|'div'|'div.un'|'rem'|'rem.un'|'and'|'or'|'xor'|'shl'|'shr'|'shr.un'|'neg'|'not'|'conv.i1'|'conv.i2'|'conv.i4'|'conv.i8'|'conv.r4'|'conv.r8'|'conv.u4'|'conv.u8'|'conv.r.un'|'throw'|'conv.ovf.i1.un'|'conv.ovf.i2.un'|'conv.ovf.i4.un'|'conv.ovf.i8.un'|'conv.ovf.u1.un'|'conv.ovf.u2.un'|'conv.ovf.u4.un'|'conv.ovf.u8.un'|'conv.ovf.i.un'|'conv.ovf.u.un'|'ldlen'|'ldelem.i1'|'ldelem.u1'|'ldelem.i2'|'ldelem.u2'|'ldelem.i4'|'ldelem.u4'|'ldelem.i8'|'ldelem.i'|'ldelem.r4'|'ldelem.r8'|'ldelem.ref'|'stelem.i'|'stelem.i1'|'stelem.i2'|'stelem.i4'|'stelem.i8'|'stelem.r4'|'stelem.r8'|'stelem.ref'|'conv.ovf.i1'|'conv.ovf.u1'|'conv.ovf.i2'|'conv.ovf.u2'|'conv.ovf.i4'|'conv.ovf.u4'|'conv.ovf.i8'|'conv.ovf.u8'|'ckfinite'|'conv.u2'|'conv.u1'|'conv.i'|'conv.ovf.i'|'conv.ovf.u'|'add.ovf'|'add.ovf.un'|'mul.ovf'|'mul.ovf.un'|'sub.ovf'|'sub.ovf.un'|'endfinally'|'stind.i'|'conv.u'|'prefix7'|'prefix6'|'prefix5'|'prefix4'|'prefix3'|'prefix2'|'prefix1'|'prefixref'|'arglist'|'ceq'|'cgt'|'cgt.un'|'clt'|'clt.un'|'localloc'|'endfilter'|'volatile.'|'tail.'|'cpblk'|'initblk'|'rethrow'|'refanytype'|'readonly.'|'illegal'|'endmac'); | ||
| INSTR_VAR: ('ldarg.s'|'ldarga.s'|'starg.s'|'ldloc.s'|'ldloca.s'|'stloc.s'|'ldarg'|'ldarga'|'starg'|'ldloc'|'ldloca'|'stloc'); | ||
| INSTR_I: ('ldc.i4.s'|'ldc.i4'|'unaligned.'|'no.'); | ||
| INSTR_I8: ('ldc.i8'); | ||
| INSTR_R: ('ldc.r4'|'ldc.r8'); | ||
| INSTR_METHOD: ('jmp'|'call'|'callvirt'|'newobj'|'ldftn'|'ldvirtftn'); | ||
| INSTR_SIG: ('calli'); | ||
| INSTR_BRTARGET: ('br.s'|'brfalse.s'|'brtrue.s'|'beq.s'|'bge.s'|'bgt.s'|'ble.s'|'blt.s'|'bne.un.s'|'bge.un.s'|'bgt.un.s'|'ble.un.s'|'blt.un.s'|'br'|'brfalse'|'brtrue'|'beq'|'bge'|'bgt'|'ble'|'blt'|'bne.un'|'bge.un'|'bgt.un'|'ble.un'|'blt.un'|'leave'|'leave.s'); | ||
| INSTR_SWITCH: ('switch'); | ||
| INSTR_TYPE: ('cpobj'|'ldobj'|'castclass'|'isinst'|'unbox'|'stobj'|'box'|'newarr'|'ldelema'|'ldelem'|'stelem'|'unbox.any'|'refanyval'|'mkrefany'|'initobj'|'constrained.'|'sizeof'); | ||
| INSTR_STRING: ('ldstr'); | ||
| INSTR_FIELD: ('ldfld'|'ldflda'|'stfld'|'ldsfld'|'ldsflda'|'stsfld'); | ||
| INSTR_TOK: ('ldtoken'); |
Member
There was a problem hiding this comment.
Can we format these like the rest of the rules in this file?
Comment on lines
+907
to
+908
| // If we got here without exception, the code path was handled | ||
| Assert.True(diagnostics.Length >= 0); |
Member
There was a problem hiding this comment.
Can we match the specific error diagnostics instead of asserting "we have diagnostics"?
| """; | ||
|
|
||
| var diagnostics = CompileAndGetDiagnostics(source, new Options()); | ||
| // BaseOutsideClass is hard to trigger with valid IL - just verify compilation works |
Member
There was a problem hiding this comment.
For cases where we can't write a test that actually has the IL that triggers it, I'd rather we don't have the test than have a test that doesn't actually validate the diagnostic.
Comment on lines
+161
to
+169
| private static bool IsRecoverableError(string diagnosticId) | ||
| { | ||
| // Method body diagnostics are recoverable - we emit the assembly but report the error | ||
| return diagnosticId is DiagnosticIds.ByteArrayTooShort | ||
| or DiagnosticIds.ArgumentNotFound | ||
| or DiagnosticIds.LocalNotFound | ||
| or DiagnosticIds.LabelNotFound; | ||
| } | ||
|
|
Member
There was a problem hiding this comment.
I think more of these errors might be recoverable, but I'm not 100% positive. Can we compare against ilasm to see what's recoverable there?
For example, I think GenericParameterIndexOutOfRange should be recoverable (the method signature would be invalid, but should still be emittable).
| } | ||
|
|
||
| public GrammarResult VisitExtSourceSpec(CILParser.ExtSourceSpecContext context) => throw new NotImplementedException("TODO: Symbols"); | ||
| public GrammarResult VisitExtSourceSpec(CILParser.ExtSourceSpecContext context) => GrammarResult.SentinelValue.Result; |
Member
There was a problem hiding this comment.
These symbols TODOs haven't been fixed.
| } | ||
|
|
||
| public GrammarResult VisitFieldInit(CILParser.FieldInitContext context) => throw new NotImplementedException("TODO-SRM: Need support for an arbitrary byte blob as a constant value"); | ||
| public GrammarResult VisitFieldInit(CILParser.FieldInitContext context) => GrammarResult.SentinelValue.Result; |
| } | ||
|
|
||
| public GrammarResult VisitExportHead(CILParser.ExportHeadContext context) => throw new NotImplementedException("Obsolete syntax"); | ||
| public GrammarResult VisitExportHead(CILParser.ExportHeadContext context) => GrammarResult.SentinelValue.Result; |
Member
There was a problem hiding this comment.
Making this a no-op isn't the right thing. We should add support for this in VisitClassDecl if we want to support it, or (eventually) remove it from the grammar if we don't want to support it any more.
|
|
||
| public GrammarResult VisitErrorNode(IErrorNode node) => throw new UnreachableException(NodeShouldNeverBeDirectlyVisited); | ||
| public GrammarResult VisitEsHead(CILParser.EsHeadContext context) => throw new NotImplementedException("TODO: Symbols"); | ||
| public GrammarResult VisitEsHead(CILParser.EsHeadContext context) => GrammarResult.SentinelValue.Result; |
Member
There was a problem hiding this comment.
Another TODO that hasn't been implemented yet.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Consolidate instruction tokens into main grammar to establish proper lexer precedence over DOTTEDNAME and ID; reorder methodDecl alternatives to match instructions first; add custom attribute handling in method bodies
Implement visitor stubs for debug/symbol directives: VisitLanguageDecl, VisitEsHead, VisitExportHead, VisitExtSourceSpec, VisitFieldInit
Add typedef resolution in ResolveTypeDef for type aliases
Implement label validation: track declared vs referenced labels, report errors for undefined labels at method end
Allow recoverable errors in method bodies to still emit assembly
Add tests for label validation, typedef resolution, and error handling