Bug 63837 – We shouldn't wait for ReportCrash while the timeout is running

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Adam Barth]

We shouldn't wait for ReportCrash while the timeout is running

[Adam Barth]

Created attachment 99513 [details]
Patch

[Eric Seidel (no email)]

Comment on attachment 99513 [details]
Patch

Makes sense.  How do we test this?

[Dirk Pranke]

Comment on attachment 99513 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=99513&action=review

> Tools/Scripts/webkitpy/layout_tests/layout_package/worker.py:133
> +        self._port.executive().wait_newest(self._port.is_crash_reporter)

Does it make sense to only do this if the previous test crashed? 

Or was it your intent to block *every* worker if a crash reporter is running? It will be interesting to see what kind of impact that has on test run time on the many-core machines like the big mac pros and xserves.

[Adam Barth]

> Or was it your intent to block *every* worker if a crash reporter is running?

Yes.

> It will be interesting to see what kind of impact that has on test run time on the many-core machines like the big mac pros and xserves.

In my machine (a big mac pro), this appears to be necessary for test stability.  When tests crash (rare), it adds ~minutes of waiting.

[WebKit Review Bot]

Comment on attachment 99513 [details]
Patch

Clearing flags on attachment: 99513

Committed r90277: <http://trac.webkit.org/changeset/90277>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.