Run-webkit-tests is falsely reporting crashes in some tests due to a bug in webkitpy. From radar 27786762.
Created attachment 285847 [details] Patch
Comment on attachment 285847 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=285847&action=review Good catch! > Tools/Scripts/webkitpy/port/driver.py:447 > + self._server_process.write('#SAMPLE FINISHED\n', True) # Must be able to ignore a broken pipe here. This comment does not explain why we need to ignore a failure to write. How can it be improved to explain it?
Created attachment 285852 [details] Patch
Comment on attachment 285852 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=285852&action=review > Tools/Scripts/webkitpy/port/driver.py:447 > + self._server_process.write('#SAMPLE FINISHED\n', True) # Must be able to ignore a broken pipe here, target process may already be closed. A slightly shorter way to say the same is "Ignoring write errors, as target process may already be terminated."
Comment on attachment 285852 [details] Patch Clearing flags on attachment: 285852 Committed r204514: <http://trac.webkit.org/changeset/204514>
All reviewed patches have been landed. Closing bug.
I feel like this could have had a better changelog entry. I can't easily tell what the bug was and how you fixed it.