Bug 140218

Summary: When WebProcess is slow to respond to IPC, that's mistakenly reported as crash
Product: WebKit Reporter: Alexey Proskuryakov <ap>
Component: Tools / TestsAssignee: Alexey Proskuryakov <ap>
Status: RESOLVED FIXED    
Severity: Normal CC: commit-queue, dbates, glenn, ossy, simon.fraser
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
proposed fix darin: review+

Description Alexey Proskuryakov 2015-01-07 15:25:56 PST
When WebKitTestRunner doesn't get a response to IPC, it dumps "#PROCESS UNRESPONSIVE" to stderr. When run-webkit-tests sees that, it for some reason reports the test as crashing.

It seems like there may have been ports that couldn't detect crashes (chromium perhaps?). However, detecting crashes is a core WebKit2 behavior, and if that doesn't work, we shouldn't be lying about that in run-webkit-tests.

There are practical cases where IPC takes a long time and triggers this - e.g. with ASan builds, despite them now having a higher IPC timeout. Generating accurate output will help reduce confusion.
Comment 1 Alexey Proskuryakov 2015-01-07 15:33:15 PST
Created attachment 244211 [details]
proposed fix
Comment 2 Darin Adler 2015-01-07 20:47:22 PST
Comment on attachment 244211 [details]
proposed fix

View in context: https://bugs.webkit.org/attachment.cgi?id=244211&action=review

> Tools/WebKitTestRunner/TestController.cpp:1392
> +        fputs("#CRASHED - %s\n", webProcessName(), stderr);

This needs to be changed to fprintf.
Comment 3 Alexey Proskuryakov 2015-01-08 09:10:29 PST
Committed <http://trac.webkit.org/r178116>.
Comment 4 Alexey Proskuryakov 2015-01-08 09:48:08 PST
Looking into webkitpy failures - I forgot to run these tests. We really need to have them on EWS.
Comment 5 Alexey Proskuryakov 2015-01-08 10:02:23 PST
Fixed the failures in r178118.