fhahn wrote: > I think this is a strict improvement over the general 'fstrict-aliasing' mode > we already have, and I see little value in maintaining separate levels of > strict aliasing conformance long term. > > I think it would not be a good situation for users to rely on some level of > strict aliasing violations, by clang accepting some bugs, rejecting others, > without a clear line separating them. > > I'd favor not having the flag, with 'fno-strict-aliasing' still being > enforced. I don't have strong objections against the new flag, if it's only > temporary, with a clear goalpost for removing it.
I agree that long-term we should remove the separate flag, maybe after one or two releases. Initially I expect it will help users during the transition phase (and also us during bug triage to quickly identify if an issue is related to the new behavior) https://github.com/llvm/llvm-project/pull/117244 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
