thieta added a comment. In D130689#3706424 <https://reviews.llvm.org/D130689#3706424>, @aaron.ballman wrote:
> +1, thank you for thinking about how we can improve this process in the > future! Given that C++17 adoption across compilers has been far better than > C++20, I suspect the next time we bump the language version will be even more > of a challenge with these sort of weird issues. I'll happily ignore that for another 3 years 😅 > Yeah, this is rather deeply concerning to be honest. I triple-checked and the > only changes in my tree were yours to compiler-rt and when I looked at the > configure logs, they clearly showed `compiler-rt project is disabled` in the > logs. It is really weird that a change to compiler-rt's cmake would have any > impact on Clang's ability to build when compiler-rt is disabled. I think > that's worth some investigative work. Hmmm - something is really funny, with my cache nuking change it seems like this bot is back to working again: https://lab.llvm.org/buildbot/#/builders/123/builds/12164 > It seems like we might need that, but it also seems like it would be good to > understand why compiler-rt is impacting the rest of Clang when it's disabled. > That said, the goal is to get the build lab back to green as quickly as > possible, so I think it may make sense to land sooner rather than later. I'll hold off a bit on this - it seems like my commit above have fixed at least some of the issues. There are still some msvc bots failing - but they seem to be failing because they have a too old version of MSVC. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D130689/new/ https://reviews.llvm.org/D130689 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
