sammccall added a comment. Having done some out-of-tree testing, it seems this silently breaks enough ASTMatchers that it'll be hard to get it to stick.
(We don't strictly need to block on out-of-tree failures, but I'm pretty sure a lot of stuff in-tree is broken too, and the out-of-tree users just have better test coverage). It would be great if things like `expr(hasType(hasDeclaration(cxxRecordDecl(...))))`, `loc(qualType(hasDeclaration(cxxRecordDecl(...))))` still matched the cases they used to. Even with this change at front of mind, it's really surprising and inconvenient to have to explicitly unwrap this level of sugar. ElaboratedType gets implicitly unwrapped by hasDeclaration, I suspect UsingType is more usefully treated like this than like TypedefType. Then some other special-purpose matcher would be used to traverse the UsingType->UsingShadowDecl edge instead of hasDeclaration. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D114251/new/ https://reviews.llvm.org/D114251 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
