[PATCH] D94324: [InitLLVM] Ensure SIGPIPE handler installed before sigaction()

Fri Jan 8 13:03:21 PST 2021

vsk added a comment.

In D94324#2487572 <https://reviews.llvm.org/D94324#2487572>, @aganea wrote:

> What are the exact conditions to repro Nick's failure? On Fedora 32, I ran `py llvm-lit sigpipe-handling.c` before this patch and before rGbf401256edd00e921a5d3a0bf4cf6ee66ae51cd6 <https://reviews.llvm.org/rGbf401256edd00e921a5d3a0bf4cf6ee66ae51cd6> , the test suceeds, it should not, should it?

In order for the test to fail pre-patch, the OS would need to send a SIGPIPE to clang when it writes to stdout after `head` exits. It's possible that certain OSes don't do this (perhaps they are patched to SIGPIPE less aggressively? it's the sort of patch an OS maintainer might plausibly ship as a stopgap to appease users).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D94324/new/

https://reviews.llvm.org/D94324