ymand wrote: > > Can you expand on your concern about the clash? > > The clash is happening in `NodesMap`. See, > > 1. the matcher which is associated with `rewriteDescendants` gets "Tag0", > 2. tags for matchers from `applyfirst` are shifted, they get "Tag1", ... > 3. that is why `transformer::detail::findSelectedCase` selects incorrect > `Tag` from `NodesMap` (this also leads to error "ID not bound", for expected > bound ID from other matcher) > > Have I made anything clearer? Do you want for me to update the description of > the commit?
Thanks for the explanation (and for your patience). Yes, this would be good to specify in the description. That said, I'm still a little unclear: 1. Will any of the new tests trigger the bug? 2. Can you clarify where the matchers associated with the nested rule will be shifted? As far as I can tell, the original match uses Tag0 but happens in a separate context (that is, MatchFinder) than the nested match so I don't see how they are interacting. https://github.com/llvm/llvm-project/pull/117658 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
