In Bug 200543 (and other updates) we changed our sandbox code to use vend extensions to processes by audit token. During this change, we used the WTFLogAlways() logging function when a sandbox extension could not be created. We now see that any output to stderr causes the test runner used by the micro-benchmark system in PerformanceTests to trigger a failure. Now that we have debugged this code, we should switch back to our standard logging behavior.
<rdar://problem/58714766>
Created attachment 388328 [details] Patch
Comment on attachment 388328 [details] Patch Great! R=me.
Comment on attachment 388328 [details] Patch Clearing flags on attachment: 388328 Committed r254872: <https://trac.webkit.org/changeset/254872>
All reviewed patches have been landed. Closing bug.