localspook wrote: I wouldn't be against banning Transformer* checks. It's yet another layer of abstraction (which inevitably adds some overhead) and yet another API to learn, and I just don't see what it's buying us. It doesn't provide any significant new _functionality,_ there's no part of the API that can't be directly translated into regular AST matcher-based code (I admit I'm not intimately familiar with Transformer, so if anyone believes I'm being ignorant here, do correct me). The only benefit I see is slightly more concise code, and I don't think that's enough of a justification.
A good experiment would be to rewrite a Transformer check to use AST matchers while (unlike this PR) preserving the general logic, and compare things like performance and LOC before and after. \* That's the name of the RewriteRule API I believe https://github.com/llvm/llvm-project/pull/192889 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
