thurstond wrote: > @thurstond Well can you provide a more recent buildbot failure link where > #130182 is not on the list of possible culpirits? Because the link @fmayer > provided for sure included it.
@mizvekov I localized the MSan error to this patch by re-running the buildbot script with this patch and immediately prior, in both cases also reverting #130182 to remove noise: Repro: ``` cd ~/llvm-project # This patch git checkout f302f35526553abcb46dab278c4494c3d01deb45 # Revert #130182 git revert 76fa9530c9ac7f81a49b840556f51f4838efbfe1 # Buildbot mkdir ~/buildbot_repro cd ~/buildbot_repro git clone https://github.com/llvm/llvm-zorg.git BUILDBOT_MONO_REPO_PATH=~/llvm-project bash llvm-zorg/zorg/buildbot/builders/sanitizers/buildbot_fast.sh # For speed, I commented out build_stage2_asan_ubsan and check_stage2_asan_ubsan in the script above ``` This repro'ed the MSan failures: ``` Clang :: ASTMerge/class-template-partial-spec/test.cpp Clang :: Analysis/ctu-inherited-default-ctor.cpp ``` The MSan failures are not present when using the immediately prior patch: ``` # Immediately prior to this patch git checkout 990a086d9da0bc2fd53a6a4c95ecbbe23a297a83 git revert 76fa9530c9ac7f81a49b840556f51f4838efbfe1 # Same buildbot steps as before ``` Please let me know if you encounter any issues reproducing the error. https://github.com/llvm/llvm-project/pull/132748 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
