Closed Bug 1288619 Opened 4 years ago Closed 4 years ago
.test page load | Test timed out (after: START)
would it be possible that this is caused by bug 1288480 ?
(In reply to Carsten Book [:Tomcat] from comment #1) > would it be possible that this is caused by bug 1288480 ? It is very unlikely, the code path changed in bug 1288480 is never invoked in this test. Luca, are you familiar with this test? (I'm not even sure how to run the jetpack tests locally)
Flags: needinfo?(aswan) → needinfo?(lgreco)
The failure doesn't seem to be related to Bug 1288480. I tried to reproduce it locally (even by trying to run the test on a win32 pgo build, which is where it seems that it fails), but without any luck.
Looking back at inbound's history, the first instance came shortly after a set of worker-related patches of baku's landed - bug 1288033, bug 1288031. Any thoughts, baku?
is it possible that this is related to bug 1288708? that summary is " Intermittent page-worker/main.test postMessage | Test timed out (after: START)"
Yes, they generally both happen as a pair.
Attachment #8809109 - Flags: review?(dtownsend) → review?(lgreco)
Comment on attachment 8809109 [details] Bug 1288619 - disable page-worker test due to high intermittent rate. Sounds good, lets disable it for now. (nit: "tst" -> "test" in the patch summary message) I'm fighting with this test from a while now, I'm still not able to reproduce it locally and so I tried to get some additional details about what is happening when it fails by pushing to try with "--rebuild" (which is able to reproduce it consistently enough by re-running the mochitest-jetpack test suite at least 10 times per build). When it fails, it seems that the SDK "sdk/content/page-worker/" module has never reached this line for the created page-worker: https://dxr.mozilla.org/mozilla-central/source/addon-sdk/source/lib/sdk/content/page-worker.js#130 The possible reasons seems to be: - it has not received a "document-element-inserted" observer notification for the expected document - none of the existents page seems to be related to the "document-element-inserted" document (e.g. the observer has received a document but it doesn't pass the check "contentWindow.document == target") By looking at the logs that I collected when the test passes and when it fails, it looks like the intermittents can be related to some weird behavior of the windowless browser created and initialized as an about:blank in the PageWorker constructor (it wouldn't be the first time that the windowlessBrowser and/or "about:blank" show some weird behavior): - https://dxr.mozilla.org/mozilla-central/source/addon-sdk/source/lib/sdk/content/page-worker.js#38 - https://dxr.mozilla.org/mozilla-central/source/addon-sdk/source/lib/sdk/content/page-worker.js#41-43 I'll keep fighting on it, I have the feeling that the actual issue is not far from what described above.
Attachment #8809109 - Flags: review?(lgreco) → review+
Comment on attachment 8809109 [details] Bug 1288619 - disable page-worker test due to high intermittent rate. https://reviewboard.mozilla.org/r/91756/#review92100 carry forward r+
Attachment #8809109 - Flags: review?(lgreco) → review?(jmaher)
Comment on attachment 8809109 [details] Bug 1288619 - disable page-worker test due to high intermittent rate. https://reviewboard.mozilla.org/r/91756/#review92104 goofy mozreview, I guess I need to assign this to myself to carry over the r+ as I just changed a typo in the commit message!!!
Attachment #8809109 - Attachment is obsolete: true
updated patch with proper commit message.
Attachment #8809578 - Flags: review+
Pushed by email@example.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/28dc1d71a791 disable page-worker test due to high intermittent rate. r=rpl
This appears to have been caused by 1285373.
As opposed to turning off the test, was any attempt made to backtrack and see what patch started causing the intermittent failures?
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
You did see that this got fixed and re-enabled in bug 1316875, right?
> You did see that this got fixed and re-enabled in bug 1316875, right? No I did not. Thanks for pointing that out. That might help get to the bottom of why some other things were broken by 1285373. I should have taken a look at the current state of the test (and noticed the see also bugs).
Status: REOPENED → RESOLVED
Closed: 4 years ago → 4 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.