Bug 161257 – Database process crash was copied, but didn't get into Other Crashes

• Component: Tools / Tests
• Status: RESOLVED WONTFIX

[Alexey Proskuryakov]

I see a Database process crash here <https://build.webkit.org/results/Apple%20Yosemite%20Debug%20WK2%20(Tests)/r205023%20(14606)/>, but it's not in results.html: <https://build.webkit.org/results/Apple%20Yosemite%20Debug%20WK2%20(Tests)/r205023%20(14606)/results.html>.

This seems strange, clearly webkitpy knew about this crash if it copied it.

[Jonathan Bedard]

Here is the standard out for this test: https://build.webkit.org/builders/Apple%20Yosemite%20Debug%20WK2%20%28Tests%29/builds/14606/steps/layout-test/logs/stdio

What is very strange about this particular error is that the crash file has a timestamp of "2016-08-25 15:00:37.971," meaning this particular crash file started generating the previous afternoon.  There is a possibility that https://bugs.webkit.org/show_bug.cgi?id=161218 would fix this bug.

There is also a strong likely-hood that this bug has to do with bot184 and bot185 being restarted this morning.

[Jonathan Bedard]

Layout-test-results is deleted after it is archived during archive-test-results.  We should delete it when we remove stale test results as well.

[Jonathan Bedard]

Created attachment 287154 [details]
Patch

[Jonathan Bedard]

This has not been an issue since.

It seems this bug was caused by the bots being interrupted in the middle of their test run.