972814 - Timeout: bridge.set when loading mozmill.js and frame.js

Status: RESOLVED WONTFIX

(Testing Graveyard :: Mozmill, defect)

Description

[Andrei Eftimie]

Attachment: jenkins_log_timeout.txt

This failed during an update test:
> 08:07:49 Timeout: bridge.set("8fe80d2e-94c8-11e3-a2fe-a8206612836f", Components.utils.import("resource://mozmill/driver/mozmill.js"))
> 08:07:49
> 08:07:49 Timeout: bridge.set("8fe815ee-94c8-11e3-9dd2-a8206612836f", Components.utils.import("resource://mozmill/modules/frame.js"))

Which led the test failing with:
> 08:08:07 ERROR | Test Failure | {
> 08:08:07   "exception": {
> 08:08:07     "message": "persisted.updateIndex is undefined",
> 08:08:07     "lineNumber": 19,
> 08:08:07     "name": "TypeError",
> 08:08:07     "fileName": "resource://mozmill/modules/frame.js -> file:///Users/mozauto/jenkins/workspace/ondemand_update/data/mozmill-tests/firefox/tests/update/testFallbackUpdate/test2.js"
> 08:08:07   }

I'm not sure what to investigate here.

Comment 1

[Henrik Skupin [:whimboo][⌚️UTC+2]]

We missed to run test1.js because of that failure, so any remaining test for this restart test will fail. Andrei, this also happened on 10.8-1 were we had the Python crash, right? Was it around the same time?

Comment 2

[Andrei Eftimie]

That's how I found the python crash, I was trying to investigate this issue.

There's about a 12 hour gap between this failure and when I've noticed the Python crash.
The python crash has a time signature of:
> Date/Time:       2014-02-13 15:09:08.227 -0800

This testrun ran between 7:55-8:08 (this is the time reported by Jenkins).
Jenkins reports the same timezone:
> user.timezone	America/Los_Angeles

Time-wise it doesn't look likely.

We've also had a few passing testruns in-between, so it doesn't look like these are related.

Comment 3

[Andreea Matei [:AndreeaMatei]]

I have retriggered something similar (although no test actually ran), and it failed again. Checked the machine but it was fine.

05:21:17 *** Creating profile: /Users/Mozauto/jenkins/workspace/mozilla-central_addons/data/profile
05:21:17 2014-02-14 05:40:46.738 firefox[53585:5607] invalid pixel format
05:21:17 2014-02-14 05:40:46.740 firefox[53585:5607] invalid context
05:21:17 2014-02-14 05:40:46.741 firefox[53585:5607] invalid pixel format
05:21:17 2014-02-14 05:40:46.741 firefox[53585:5607] invalid context
05:21:17 Timeout: bridge.set("9bc235f5-957d-11e3-ad0a-c42c031efd37", Components.utils.import("resource://mozmill/driver/mozmill.js"))
05:21:17
05:21:17 Timeout: bridge.set("bf6f165e-957d-11e3-9928-c42c031efd37", Components.utils.import("resource://mozmill/modules/frame.js"))
05:21:17
05:21:17 Timeout: bridge.set("c007c1cf-957d-11e3-ae46-c42c031efd37", Components.utils.import("resource://mozmill/modules/frame.js"))
05:21:17
05:21:17 RESULTS | Passed: 0
05:21:17 RESULTS | Failed: 2
05:21:17 RESULTS | Skipped: 0

Comment 4

[Henrik Skupin [:whimboo][⌚️UTC+2]]

I can see exactly this problem with frame.js while working on bug 794020. Not exactly sure what it is but I will try to find out.

Comment 5

[Andreea Matei [:AndreeaMatei]]

Sorry, forgot the machine from comment 3, it was mm-osx-106-4.

Comment 6

[Andrei Eftimie]

We've had another instance of these failures. 3 testruns failed on mm-osx-109-3, all in the exactly the same place: 
> 04:09:50 TEST-START | testPreferences/testPasswordSavedAndDeleted.js | testSaveAndDeletePassword
> 04:11:53 Timeout: bridge.execFunction("58f6b130-bd7c-11e3-885c-3c07540e96bd", bridge.registry["{d2297029-0a7a-494e-bdd0-1cdd576cd339}"]["runTestFile"], ["/Users/mozauto/jenkins/workspace/mozilla-central_functional/data/mozmill-tests/firefox/tests/functional/testPreferences/testPasswordSavedAndDeleted.js", null])
> 04:11:53 
> 04:11:53 Timeout: bridge.set("7dfc07f0-bd7c-11e3-9ed2-3c07540e96bd", Components.utils.import("resource://mozmill/modules/frame.js"))

Which is actually a disconnect, and all 3 instances missed to run /restartTests/testPreferences_masterPassword/test3.js

This looks to be fallout from bug 972912.

Comment 7

[Henrik Skupin [:whimboo][⌚️UTC+2]]

We are not going back to use Mozmill for our UI testing of Firefox and this is just an enhancement and not a real blocker for a possible release.