Closed Bug 782196 Opened 12 years ago Closed 11 months ago

Intermittent REFTEST TEST-UNEXPECTED-FAIL | reftest/tests/layout/reftests/bugs/598726-1.html | image comparison (==)

Categories

(Core :: CSS Parsing and Computation, defect, P3)

defect

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: philor, Unassigned)

References

()

Details

(Keywords: intermittent-failure, Whiteboard: [test disabled on Windows])

Attachments

(2 files, 1 obsolete file)

Attached file reftest log
+++ This bug was initially created as a clone of Bug #702184 +++

See also bug 702184 comment 628, and who knows how many of the other comments there.

https://tbpl.mozilla.org/php/getParsedLog.php?id=14333419&tree=Mozilla-Inbound
Rev3 WINNT 6.1 mozilla-inbound debug test reftest on 2012-08-12 23:12:18 PDT for push 9e0799cf1096
slave: talos-r3-w7-045

REFTEST TEST-START | file:///c:/talos-slave/test/build/reftest/tests/layout/reftests/bugs/598726-1-ref.html | 2815 / 8122 (34%)
++DOMWINDOW == 45 (0AB6A5A0) [serial = 7726] [outer = 0A22AEF0]
REFTEST TEST-UNEXPECTED-FAIL | file:///c:/talos-slave/test/build/reftest/tests/layout/reftests/bugs/598726-1.html | image comparison (==)
...
REFTEST max difference: 9 number of differing pixels: 44

Along both ends of the input.
And suspicion falls on something in roc's https://tbpl.mozilla.org/?tree=Mozilla-Inbound&rev=94e4dbce3b94 - that's once on the push, and once on the next push.
(In reply to Ed Morley [:edmorley] from comment #107)
> Is virtually permaorange as of:
> https://hg.mozilla.org/integration/mozilla-inbound/
> pushloghtml?fromchange=2bc119b67709&tochange=e2e8d066279b
> 
> ...and for m-c:
> https://hg.mozilla.org/mozilla-central/
> pushloghtml?fromchange=f89feda9d997&tochange=75cdb3f932c6
> 
> Have requested some rebuilds in the coalesced ranges to help narrow.

Some of these got coalesced, so triggered more.

Links to save me digging them out again:
https://tbpl.mozilla.org/?rev=75cdb3f932c6&jobname=WINNT
https://tbpl.mozilla.org/?tree=Mozilla-Inbound&rev=e2e8d066279b&jobname=WINNT
Tried a backout of that push but had several conflicts, so just going to disable the test for now, to get us away from near-permaorange, whilst roc takes a look.
Attached patch Disable on Win7 (obsolete) — Splinter Review
Temporarily disable on Win7
Attachment #652479 - Flags: review?(philringnalda)
Attachment #652479 - Flags: review?(philringnalda) → review+
Whiteboard: [orange] → [orange][test disabled on Win7]
Too many failures
Attachment #664557 - Flags: review?(philringnalda)
Comment on attachment 664557 [details] [diff] [review]
Disable on Windows entirely

WHAT WAS THAT? I CAN'T HEAR YOU OVER THIS BOT YAMMERING ON AND ON!
Attachment #664557 - Flags: review?(philringnalda) → review+
Whiteboard: [orange][test disabled on Win7] → [orange][test disabled on Windows]
Whiteboard: [orange][test disabled on Windows] → [test disabled on Windows]
Attachment #652479 - Attachment is obsolete: true
Attachment #664557 - Flags: checkin+
Does anybody know what triggered the flood of these yesterday and the day before?
That little spurt was brought to you by bug 849399.
Bulk assigning P3 to all open intermittent bugs without a priority set in Firefox components per bug 1298978.
Priority: -- → P3
Severity: normal → S3

The failure in comment 652 (the most recent failure) was actually load failed: timed out waiting for reftest-wait to be removed for this test (598726-1.html) and one other test that gets run three times (613433-1.html vs. its 3 reference cases). All four of those reftest.list lines are flagged as needs-focus, so I suspect really they were blocked on the reftest window failing to gain focus for some reason, and the test timed out for that reason.

That failure was on aarch64. The previous comment 2 years earlier was also on aarch64, so it may have been the same issue -- can't know for sure since the logs have expired.

But in any case, the comment 652 [and maybe 651] "timed out" issue (possibly focus-related, possibly aarch64-specific) seems unrelated to what this bug was originally tracking, which was a reftest failure with "image comparison (==)". So let's close the original issue as WFM. And hopefully any remaining aarch64/focus-related timeouts that trip over this test will spawn a new bug report which can be investigated with more recent data rather than 650 comments' worth of unrelated decades-old data. :)

Status: NEW → RESOLVED
Closed: 11 months ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: