tom-anders added a comment. In D134137#3877726 <https://reviews.llvm.org/D134137#3877726>, @sammccall wrote:
> It's not surprising that we often don't crash here, the "obvious" lowering of > `S.front()` is `*S.data()` which is perfectly valid (will be 0 for an empty > string). > > One way to crash on this is to build with libstdc++'s debug iterators, which > assert on this. Fedora seems to be building (some of) their packages in this > mode, such that they crash but (sometimes?) do not print the assertion > message. This was the cause in > https://github.com/clangd/vscode-clangd/issues/400. Interesting, FYI I was able to reproduce this on Arch Linux. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D134137/new/ https://reviews.llvm.org/D134137 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
