Closed Bug 1155863 Opened 10 years ago Closed 10 years ago

Intermittent ua-style-sheet-margin-8.html,ua-style-sheet-margin-10.html | image comparison (==), max difference: 255, number of differing pixels: 523

Categories

(Core :: Layout, defect)

ARM
Gonk (Firefox OS)
defect
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla40
Tracking Status
firefox38 --- unaffected
firefox39 --- unaffected
firefox40 --- fixed
firefox-esr31 --- unaffected

People

(Reporter: RyanVM, Assigned: jfkthame)

Details

(Keywords: intermittent-failure)

Attachments

(3 files)

Attached image test screenshot
10:20:05 INFO - REFTEST TEST-START | http://10.0.2.2:8888/tests/layout/reftests/writing-mode/ua-style-sheet-margin-10.html 10:20:05 INFO - REFTEST TEST-LOAD | http://10.0.2.2:8888/tests/layout/reftests/writing-mode/ua-style-sheet-margin-10.html | 235 / 677 (34%) 10:20:05 INFO - REFTEST TEST-LOAD | http://10.0.2.2:8888/tests/layout/reftests/writing-mode/ua-style-sheet-margin-10-ref.html | 235 / 677 (34%) 10:20:05 INFO - REFTEST TEST-UNEXPECTED-FAIL | http://10.0.2.2:8888/tests/layout/reftests/writing-mode/ua-style-sheet-margin-10.html | image comparison (==), max difference: 255, number of differing pixels: 523 10:20:06 INFO - REFTEST IMAGE 1 (TEST): 10:20:06 INFO - REFTEST IMAGE 2 (REFERENCE): 10:20:06 INFO - REFTEST INFO | Saved log: START http://10.0.2.2:8888/tests/layout/reftests/writing-mode/ua-style-sheet-margin-10.html 10:20:06 INFO - REFTEST INFO | Saved log: [CONTENT] OnDocumentLoad triggering AfterOnLoadScripts 10:20:06 INFO - REFTEST INFO | Saved log: Initializing canvas snapshot 10:20:06 INFO - REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000 10:20:06 INFO - REFTEST INFO | Saved log: [CONTENT] RecordResult fired 10:20:06 INFO - REFTEST INFO | Saved log: RecordResult fired 10:20:06 INFO - REFTEST INFO | Saved log: START http://10.0.2.2:8888/tests/layout/reftests/writing-mode/ua-style-sheet-margin-10-ref.html 10:21:05 INFO - REFTEST INFO | Saved log: [CONTENT] OnDocumentLoad triggering AfterOnLoadScripts 10:21:05 INFO - REFTEST INFO | Saved log: Initializing canvas snapshot 10:21:05 INFO - REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000 10:21:05 INFO - REFTEST INFO | Saved log: [CONTENT] RecordResult fired 10:21:05 INFO - REFTEST INFO | Saved log: RecordResult fired 10:21:05 INFO - REFTEST INFO | Loading a blank page 10:21:05 INFO - REFTEST TEST-END | http://10.0.2.2:8888/tests/layout/reftests/writing-mode/ua-style-sheet-margin-10.html
Flags: needinfo?(jfkthame)
Attached image reference screenshot
Summary: Intermittent ua-style-sheet-margin-10.html | image comparison (==), max difference: 255, number of differing pixels: 523 → Intermittent ua-style-sheet-margin-8.html,ua-style-sheet-margin-10.html | image comparison (==), max difference: 255, number of differing pixels: 523
AFAICS, the first occurrence of this was on gwright's push 6dc252c85f6c (Bug 1155359 - Set NS_FRAME_IS_BIDI on nsTextBoxFrame if text direction is set to RTL). Maybe that has upset something, though it's not immediately obvious to me why it should affect these tests. However, I also notice that (a) the test run seems to be ending with a crash, and (b) there was a mozharness update just before this (cset d0a6539a1ca8, Bug 1155679 - "mozharness update to ref 4f1cf3369955"). Would a "mozharness update" affect the b2g reftest runs? If so, I'd be inclined to suspect that of having triggered the failures here. I've retriggered a bunch of jobs to try and confirm whether the failures ever occur prior to that change; let's see how things look in a couple of hours (or however long they take...). Leaving n-i? in place for now as a reminder.
OK, I got several failures on 6f89775dedd6, which predates the mozharness bump, so I guess that's not the culprit. Note that these tests have been running since bug 1134598, which landed more than a month ago (2015-03-12), and AFAIK this failure has only just arisen ... so apparently *something* has changed recently to make them unreliable. Looking further.....
This seems to begin with bug 1144501 or bug 1151993, as the first occurrences I've managed to find via retriggers are on https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=27220470506a. I'm not sure why the issue here is either B2G-specific or intermittent, but I propose to disable (comment out) the vertical-mode fragment of these testcases for the time being. This was already done for the corresponding fragment of ua-style-sheet-margin-9.html, as these fragments can't test what they're intended to until it's implemented anyway. When bug 1077521 is done and we re-enable the vertical parts of these tests, we'll see whether any real issue remains to be addressed here.
Flags: needinfo?(jfkthame)
Given how frequent this failure has become, I think we should disable the (currently irrelevant) portions of these tests; we'll need to re-enable them once tables are implemented, and address any remaining issues of unreliability at that point.
Attachment #8594505 - Flags: review?(cam)
Assignee: nobody → jfkthame
Status: NEW → ASSIGNED
Attachment #8594505 - Flags: review?(cam) → review+
Status: ASSIGNED → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla40
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: