================
@@ -18544,8 +18544,14 @@ Sema::ActOnTag(Scope *S, unsigned TagSpec, TagUseKind
TUK, SourceLocation KWLoc,
if (PrevDecl)
CheckRedeclarationInModule(New, PrevDecl);
- if (TUK == TagUseKind::Definition && (!SkipBody || !SkipBody->ShouldSkip))
- New->startDefinition();
+ if (TUK == TagUseKind::Definition) {
+ if (!SkipBody || !SkipBody->ShouldSkip) {
+ New->startDefinition();
+ } else {
+ New->setCompleteDefinition();
+ New->demoteThisDefinitionToDeclaration();
----------------
mizvekov wrote:
The case we were hitting on this regression was from a recent improvement
implemented here:
https://github.com/llvm/llvm-project/commit/2db239acd1e4cb61e8c5c55dcc4b08c7d64919b6
But this skipping of parsing of bodies is built upon much older patch:
https://github.com/llvm/llvm-project/commit/d9ba224f66af76e66352dcabd52a030e382ee235
But yes, we should come here for other tag kinds as well, based on the tests
added in the first commit.
https://github.com/llvm/llvm-project/pull/155900
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
