Bug 124845

Summary: Tester bots triggered incorrectly
Product: WebKit Reporter: Csaba Osztrogonác <ossy>
Component: New BugsAssignee: Nobody <webkit-unassigned>
Status: NEW    
Severity: Normal CC: ap, beidson, dburkart, ddkilzer, jonlee, lforschler, matthew_hanson, mrowe, ossy, rniwa
Priority: P1    
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   

Csaba Osztrogonác
Reported 2013-11-25 09:33:10 PST
It seems something bad happens during triggering the Mac Debug tester bots. - http://build.webkit.org/builders/Apple%20Mavericks%20Debug%20%28Build%29/builds/1279 built r159749 - http://build.webkit.org/builders/Apple%20Mavericks%20Debug%20%28Build%29/builds/1280 built r159750 Both of them triggered the tester bot during it was running tests for an earlier revision. And here is the result: http://build.webkit.org/builders/Apple%20Mavericks%20Debug%20WK1%20%28Tests%29/builds/748 Revision 159749 Got Revision 159749 All Changes: - Change #39546 - r159750 - Change #39545 - r159749 It seems the tester ran the oldest one, not the newest one. Similar results on ML debug bots.
Attachments
Add attachment proposed patch, testcase, etc.
Csaba Osztrogonác
Comment 1 2013-11-25 09:42:13 PST
I have a feeling that the bug is around http://trac.webkit.org/changeset/117508 .
Csaba Osztrogonác
Comment 2 2014-09-30 06:25:22 PDT
The bug is still valid on Apple Windows bots too: https://build.webkit.org/builders/Apple%20Win%207%20Release%20%28Tests%29/builds/47633 Revision 174101 Got Revision 174101 All Changes: - Change #52898 Revision 174102 - Change #52897 Revision 174101 The tester bot should have picked 174102 instead of 174101.
Csaba Osztrogonác
Comment 3 2015-03-06 09:00:33 PST
this bug is still valid: Let's see two builders: building r181145 finished before r181143 https://build.webkit.org/builders/Apple%20Mavericks%20Release%20%28Build%29/builds/13446 r181143 Start Fri Mar 6 01:23:52 2015 End Fri Mar 6 01:57:23 2015 Elapsed 33 mins, 31 secs https://build.webkit.org/builders/Apple%20Mavericks%20Release%20%28Build%29/builds/13447 r181145 Start Fri Mar 6 01:24:24 2015 End Fri Mar 6 01:40:29 2015 Elapsed 16 mins, 5 secs https://build.webkit.org/builders/Apple%20Mavericks%20Release%20%28Build%29/builds/13448 r181147 Start Fri Mar 6 01:40:29 2015 End Fri Mar 6 01:49:11 2015 Elapsed 8 mins, 41 secs And then the tester bot ran the tests in the same order as the builder finished: https://build.webkit.org/builders/Apple%20Mavericks%20Release%20WK1%20%28Tests%29/builds/12735 r181145 Start Fri Mar 6 01:40:30 2015 End Fri Mar 6 01:53:19 2015 Elapsed 12 mins, 49 secs https://build.webkit.org/builders/Apple%20Mavericks%20Release%20WK1%20%28Tests%29/builds/12737 r181143 Start Fri Mar 6 01:57:23 2015 End Fri Mar 6 02:10:39 2015 Elapsed 13 mins, 15 secs https://build.webkit.org/builders/Apple%20Mavericks%20Release%20WK1%20%28Tests%29/builds/12738 181149 Start Fri Mar 6 02:47:22 2015 End Fri Mar 6 03:00:06 2015 Elapsed 12 mins, 43 secs
Csaba Osztrogonác
Comment 4 2015-03-06 09:04:23 PST
(In reply to comment #3) > this bug is still valid: > > Let's see two builders: building r181145 finished before r181143 ... No, it is a different bug. I don't know how can be it be fixed. The tester don't have any chance to check if a builder will finish with an older revision than a newer revision.
Alexey Proskuryakov
Comment 5 2015-03-09 12:21:39 PDT
The tester can't know that, but it seems like buildbot master could be smarter about such things. I'm not quite sure what this bug is tracking at this point though, could you rephrase it?
Note You need to log in before you can comment on or make changes to this bug.