DeinAlptraum wrote: > Given the fact that OSError already contains the message that indicates the > problem, I think we should explore why it's not properly emitted in the first > place, before introducing bespoke ways of delivering error messages. > @DeinAlptraum any thoughts here?
This has been bothering me for a while but I didn't think too much about it. Took this opportunity to finally take a look and opened a fix PR #142371 . TLDR; one of our tests captures `stderr` to keep the test output clean, but if the test crashes before it is released (usually when loading libclang.so raises as is the case here), the remaining tests in the run are effectively silenced. So this is an issue that only affected the tests. https://github.com/llvm/llvm-project/pull/142353 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
