sammccall added a comment. In D109506#3087055 <https://reviews.llvm.org/D109506#3087055>, @0x1eaf wrote:
> I think this should work-around end-to-end test failure on Windows: > > diff --git a/clang-tools-extra/clangd/test/crash.test > b/clang-tools-extra/clangd/test/crash.test > index 1197e1ab07c3..68ef54808f09 100644 > --- a/clang-tools-extra/clangd/test/crash.test > +++ b/clang-tools-extra/clangd/test/crash.test > @@ -1,3 +1,4 @@ > +# REQUIRES: shell > # Overflow the recursive json parser, prevent `yes` error due to broken > pipe and `clangd` SIGSEGV from being treated as a failure. > # RUN: (yes '[' || :) | head -n 50000 | (clangd --input-style=delimited > 2>&1 || :) | FileCheck %s > # CHECK: Signalled while processing message: > > Alternatively, the test can be completely disabled for the time being. > > I don't have commit access, should I create a new revision for review with > the test change? I just removed this test for now, we're also seeing problems on mac (I believe because stack overflow can do something other than signal), and flakiness on linux+tsan. This functionality is covered by unittests that are not failing. We should probably work on an integration test that is a bit more principled, probably using `#pragma clang __debug llvm_fatal_error` which calls abort(). It's annoying because we disable those pragmas for obvious reasons... @smeenai thanks! Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D109506/new/ https://reviews.llvm.org/D109506 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
