[PATCH] D35085: Respect exitcode sanitizer option in UBSan

[Francis Ricci via Phabricator via llvm-commits]

fjricci added inline comments.


================
Comment at: lib/ubsan/ubsan_init.cc:44
   InitializeSuppressions();
+  Atexit(ReportFailure);
 }
----------------
rsmith wrote:
> Is this the right place to do this? If UBSan is hosted inside another sanitizer, does that sanitizer already set up such a callback?
That's a good question. I know that all of the tests pass for me on Linux and Darwin (both for UBSan hosted inside other sanitizers and for the other sanitizers themselves), but that's been the extent of my testing. Should that be sufficient, or should I investigate further?


https://reviews.llvm.org/D35085