AaronBallman wrote: > This I think qualifies as a serious problem as it introduces a crash which > may occur in _any_ existent codebase. Anyone using the compiler at trunk and > encountering the crash has no way to know how to deal with it.
I think it's worth remembering that downstreams carry additional changes that the community has no visibility into, so the mere presence of a crash does not actually identify this change as the culprit without more investigation. The community cannot perform that investigation ourselves, we need Google to do that from their downstream or we need a reproducer that happens with vanilla Clang. Do you see this crash with community clang and no downstream changes? In the short term (next 24hrs), I think the existing flag to disable color diagnostics should get you a path forward; but if you disagree, your downstream can certainly revert this patch locally until you're able to provide us with a way to reproduce the issue or have at least verified that the issue is not caused by changes in your downstream. But if we get any evidence that other downstreams or users are hitting this or the issue is happening for Google's test case with a vanilla Clang, then I agree that a revert is appropriate until we have a solution. https://github.com/llvm/llvm-project/pull/66514 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
