sammccall marked an inline comment as done. sammccall added a comment. In D73873#1857042 <https://reviews.llvm.org/D73873#1857042>, @thakis wrote:
> I agree it's a bit strange. The linux bot didn't recover after the revert > either, so something's amiss. > > My Win bot has been hanging since this landed and I just CRD'd in and it's > hanging in check-clangd and there are tens of ClangdTests processes hanging > around and doing nothing. > > On the more official win bot, instead the link of ClangdTests has been > failing since this landed, which on Windows is likely caused by the processes > still hanging around > (http://lab.llvm.org:8011/builders/clang-x64-windows-msvc/builds/14120/steps/stage%201%20check/logs/stdio). > This is the first bad build: > http://lab.llvm.org:8011/builders/clang-x64-windows-msvc/builds/14113/ Windows: the stacktrace in the first bad build is almost certainly the bug above. The processes hanging around after a crash (on the main thread) is surprising. The crash is on the main thread, though there are others. Something I should look into. Linux: (some of) the stacktraces seem very unrelated to this code, and a revert didn't fix them, but a clean rebuild did (thanks @thakis!). Next step: wait for the gn linux bot to be back in a good state, reland with the double-unlock fixed. ================ Comment at: clang-tools-extra/clangd/TUScheduler.cpp:528 + RebuildTimes.push_back(RebuildDuration); + Mutex.unlock(); + } ---------------- Here's a double-unlock (thanks @kadircet) Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D73873/new/ https://reviews.llvm.org/D73873 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
