cor3ntin added inline comments.
================ Comment at: clang/lib/Parse/ParseExprCXX.cpp:1397 + for (LateParsedAttribute *Attr : LateParsedAttrs) { + ParseLambdaLexedAttribute(*Attr, Attributes, D); + delete Attr; ---------------- aaron.ballman wrote: > cor3ntin wrote: > > aaron.ballman wrote: > > > I think you should add an assert before this call to ensure that all > > > attributes are GNU ones. > > I'm not sure that gains much? > > In the future if late parsing is implemented for `__declspec` this would > > just work , unless we add an assert. I'd be fine adding it though. > Because `ParseLambdaLexedAttribute()` specifically parses GNU attribute args > and not declspec ones, I think that function should be renamed. Once that > function is renamed to say GNU explicitly, then there's a question of what > happens when we hand it late parsed declspec attribute. Which is why I was > suggesting a comment above to mention that declspec attributes don't late > parse anyway. > > Another way to handle this that's more general would be to change > `ParseLambdaLexedAttribute()` to parse different args depending on the syntax > of the attribute (then it can keep its generic name). Okay, after actually reading the code, we can't actually assert on that. `LateParsedAttribute` is really `LateParsedGNUAttribute`. And maybe we want to rename that after this patch. It can never contain anything else, and because `__attribute__` was eaten, we don't really have anything to check for. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D119136/new/ https://reviews.llvm.org/D119136 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits