Closed Bug 626103 Opened 13 years ago Closed 12 years ago

Intermittent select-required-multiple-invalid-changed.html select-required-multiple-valid.html select-required-multiple-invalid.html select-required-multiple-valid-changed.html select-focus.html select-dyn-not-disabled.html select-fieldset-legend.html

Categories

(Core :: Layout: Form Controls, defect)

Product:
Core
Component:
Layout: Form Controls
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED WORKSFORME
Tracking Flags:
Tracking Status
firefox12 --- affected

People

(Reporter: dbaron, Unassigned)

References

(Depends on 1 open bug)

Details

(Keywords: intermittent-failure)

Attachments

(2 files)

log snippet
16.57 KB, text/plain
Details
log snippet
49.46 KB, text/plain
Details 
We've seen two reftest runs this morning in which some of the css-ui-valid or css-ui-invalid reftests involving select elements failed, on Mac OS X 10.5.8, due to focus differences (first option focused vs. not focused in test vs. reference).
Attached file log snippet — 
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1295109788.1295110619.27842.gz
Rev3 MacOSX Leopard 10.5.8 mozilla-central opt test reftest

REFTEST TEST-UNEXPECTED-FAIL | file:///Users/cltbld/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-invalid/select/select-required-multiple-invalid-changed.html | image comparison (==)

In this case the first option in the test has focus, but the first option in the reference does not.
Attached file log snippet — 
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1295112941.1295113764.8453.gz
Rev3 MacOSX Leopard 10.5.8 mozilla-central opt test reftest

s: talos-r3-leopard-049
REFTEST TEST-UNEXPECTED-FAIL | file:///Users/cltbld/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-valid/select/select-required-multiple-invalid.html | image comparison (==)
REFTEST TEST-UNEXPECTED-FAIL | file:///Users/cltbld/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-valid/select/select-required-multiple-valid.html | image comparison (==)
REFTEST TEST-UNEXPECTED-FAIL | file:///Users/cltbld/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-valid/select/select-required-multiple-valid-changed.html | image comparison (==)

In this case, for three tests in a row, the reference had focus and the test did not.
Attachment #504151 - Attachment mime type: application/octet-stream → text/plain
That is obviously something I should try to fix.
Assignee: nobody → mounir.lamouri
This mochitest-plain-1 failure may also be related:
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1295118426.1295119122.30606.gz


This is a recent regression; I think I'm most suspicious of the last (top) three changesets in:
http://hg.mozilla.org/mozilla-central/pushloghtml?changeset=4df430b64d1b
(In reply to comment #4)
> This mochitest-plain-1 failure may also be related:
> http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1295118426.1295119122.30606.gz

FYI, the images here basically make the error message:
expected [image of button in normal state] but got [blank image]
I would think only changeset 43a6717878fa could be relevant.
Four additional reftest runs against the build on
http://hg.mozilla.org/mozilla-central/rev/4df430b64d1b
did not show this failure.  I suspect it may actually be lower frequency than it appears.
Is it possible the backout done for bug 626100 also fixed this?  Might we have problems with native theme drawing when we're leaking memory?
I wouldn't have thought so on Mac, but maybe it's possible.
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1295399909.1295400785.15957.gz
Rev3 MacOSX Leopard 10.5.8 mozilla-central opt test reftest on 2011/01/18 17:18:29
s: talos-r3-leopard-037

REFTEST TEST-UNEXPECTED-FAIL | file:///Users/cltbld/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-invalid/select/select-required-multiple-valid.html | image comparison (==)

(And we should probably fix up the summary to let tbpl catch these, for all I know someone has filed dupes of it by now.)
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1297863911.1297864542.4627.gz
s: talos-r3-leopard-051
REFTEST TEST-UNEXPECTED-FAIL | file:///Users/cltbld/talos-slave/test/build/reftest/tests/layout/reftests/css-valid/select/select-required-multiple-invalid.html | image comparison (==)
REFTEST TEST-UNEXPECTED-FAIL | file:///Users/cltbld/talos-slave/test/build/reftest/tests/layout/reftests/css-valid/select/select-required-multiple-valid.html | image comparison (==)
Summary: various validity reftests involving selects intermittently failing due to focus differences → intermittent failures in reftests involving selects, due to focus differences (select-required-multiple-invalid-changed.html, select-required-multiple-valid.html, select-required-multiple-invalid.html, select-required-multiple-valid-changed.h)
Summary: intermittent failures in reftests involving selects, due to focus differences (select-required-multiple-invalid-changed.html, select-required-multiple-valid.html, select-required-multiple-invalid.html, select-required-multiple-valid-changed.h) → intermittent failures in reftests involving selects, due to focus differences (select-required-multiple-invalid-changed.html, select-required-multiple-valid.html, select-required-multiple-invalid.html, select-required-multiple-valid-changed.html)
OS: Mac OS X → All
Hardware: x86 → All
Summary: intermittent failures in reftests involving selects, due to focus differences (select-required-multiple-invalid-changed.html, select-required-multiple-valid.html, select-required-multiple-invalid.html, select-required-multiple-valid-changed.html) → intermittent failures in reftests with selects due to focus diff (select-required-multiple-invalid-changed.html, select-required-multiple-valid.html, select-required-multiple-invalid.html, select-required-multiple-valid-changed.html, select-focus.html)
http://tinderbox.mozilla.org/showlog.cgi?log=Birch/1300429652.1300431107.30927.gz
Rev3 Fedora 12x64 birch debug test reftest

s: talos-r3-fed64-055
REFTEST TEST-UNEXPECTED-FAIL | file:///home/cltbld/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-valid/select/select-dyn-not-disabled.html | image comparison (==)
Happened in textarea-focus.html too:

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1300911918.1300914234.16543.gz
Summary: intermittent failures in reftests with selects due to focus diff (select-required-multiple-invalid-changed.html, select-required-multiple-valid.html, select-required-multiple-invalid.html, select-required-multiple-valid-changed.html, select-focus.html) → failures in reftests with selects due to focus diff (select-required-multiple-invalid-changed, select-required-multiple-valid, select-required-multiple-invalid, select-required-multiple-valid-changed, select-focus, textarea-focus)
Summary: failures in reftests with selects due to focus diff (select-required-multiple-invalid-changed, select-required-multiple-valid, select-required-multiple-invalid, select-required-multiple-valid-changed, select-focus, textarea-focus) → failures in reftests with :-moz-ui-invalid due to focus diff (select-required-multiple-invalid-changed, select-required-multiple-valid, select-required-multiple-invalid, select-required-multiple-valid-changed, select-focus, textarea-focus)
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1300992474.1300993953.27512.gz

better fix this before we run out of maxlength for the summary field.  ;-)
Summary: failures in reftests with :-moz-ui-invalid due to focus diff (select-required-multiple-invalid-changed, select-required-multiple-valid, select-required-multiple-invalid, select-required-multiple-valid-changed, select-focus, textarea-focus) → failures in reftests with :-moz-ui-invalid due to focus diff (select-{required-multiple-invalid-changed,required-multiple-valid,required-multiple-invalid,required-multiple-valid-changed,focus,dyn-not-disabled}, textarea-focus)
http://tinderbox.mozilla.org/showlog.cgi?log=TraceMonkey/1301005522.1301008332.1454.gz
Rev3 WINNT 5.1 tracemonkey debug test reftest

s: talos-r3-xp-034
REFTEST TEST-UNEXPECTED-FAIL | file:///C:/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-valid/select/select-required-multiple-invalid.html | image comparison (==)
REFTEST TEST-UNEXPECTED-FAIL | file:///C:/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-valid/select/select-required-multiple-valid.html | image comparison (==)
REFTEST TEST-UNEXPECTED-FAIL | file:///C:/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-valid/select/select-required-multiple-valid-changed.html | image comparison (==)
PROCESS-CRASH | Main app process exited normally | application crashed (minidump found)
Thread 0 (crashed)
PROCESS-CRASH | Main app process exited normally | application crashed (minidump found)
Thread 0 (crashed)
PROCESS-CRASH | Main app process exited normally | application crashed (minidump found)
Thread 0 (crashed)
PROCESS-CRASH | Main app process exited normally | application crashed (minidump found)
Thread 0 (crashed)
(In reply to comment #23)
> http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1300992474.1300993953.27512.gz
> 
> better fix this before we run out of maxlength for the summary field.  ;-)

Yep :)
I thought it was coming from my tests at the first glance but obviously not. I guess it's showing an error that might be deeper so I'm removing myself from assignee.
Assignee: mounir.lamouri → nobody
Well, it should get better on its own now that the summary is broken for tbpl, so instead of having one busy bug and the dupe bug 636278, we'll have seven dupes.
http://tinderbox.mozilla.org/showlog.cgi?log=TraceMonkey/1301055805.1301056499.23123.gz
Rev3 Fedora 12x64 tracemonkey opt test reftest on 2011/03/25 05:23:25
s: talos-r3-fed64-030

REFTEST TEST-UNEXPECTED-FAIL | file:///home/cltbld/talos-slave/test/build/reftest/tests/layout/reftests/css-ui-invalid/select/select-required-multiple-invalid.html | image comparison (==)
(In reply to comment #26)
> Well, it should get better on its own now that the summary is broken for tbpl,
> so instead of having one busy bug and the dupe bug 636278, we'll have seven
> dupes.

Do you have a better suggestion about what the summary should look like, given that the length is limited?
Maybe.
Summary: failures in reftests with :-moz-ui-invalid due to focus diff (select-{required-multiple-invalid-changed,required-multiple-valid,required-multiple-invalid,required-multiple-valid-changed,focus,dyn-not-disabled}, textarea-focus) → select-required-multiple-invalid-changed.html select-required-multiple-valid.html select-required-multiple-invalid.html select-required-multiple-valid-changed.html select-focus.html select-dyn-not-disabled.html textarea-focus.html input-focus.html
Do we need to wait for the window to get focused here or some such?
(In reply to comment #93)
> Do we need to wait for the window to get focused here or some such?

Yes.  We may just need to annotate the reftest.list files with appropriate needs-focus tags...
(In reply to comment #97)
> (In reply to comment #93)
> > Do we need to wait for the window to get focused here or some such?
> 
> Yes.  We may just need to annotate the reftest.list files with appropriate
> needs-focus tags...

Mounir, is this something that you can do, please?
Failing for select-fieldset-legend.html too but the title is already too long...
Blocks: 659708
(In reply to comment #97)
> (In reply to comment #93)
> > Do we need to wait for the window to get focused here or some such?
> 
> Yes.  We may just need to annotate the reftest.list files with appropriate
> needs-focus tags...

They already do have needs-focus.
Blocks: 660224
No longer blocks: 660224
Summary: select-required-multiple-invalid-changed.html select-required-multiple-valid.html select-required-multiple-invalid.html select-required-multiple-valid-changed.html select-focus.html select-dyn-not-disabled.html textarea-focus.html input-focus.html → Intermittent select-required-multiple-invalid-changed.html select-required-multiple-valid.html select-required-multiple-invalid.html select-required-multiple-valid-changed.html select-focus.html select-dyn-not-disabled.html select-fieldset-legend.html
Depends on: 660226
I've open bug 660224 for the input-focus.html and textarea-focus.html given that the issue is different (even if the origin might be the same).
No longer blocks: 659708
(In reply to comment #122)
> This might help:
> https://hg.mozilla.org/mozilla-central/rev/bf282a9b7ae4

Actually, it won't.  needs-focus should've really been called ignore-if-fails-on-Android or something...
(In reply to comment #122)
> This might help:
> https://hg.mozilla.org/mozilla-central/rev/bf282a9b7ae4

How could that help given that no focus is needed in those tests?
(In reply to comment #124)
> (In reply to comment #122)
> > This might help:
> > https://hg.mozilla.org/mozilla-central/rev/bf282a9b7ae4
> 
> How could that help given that no focus is needed in those tests?

The point is that because of bug 623625, needs-focus doesn't really mean anything useful for desktop Firefox.
(In reply to comment #125)
> The point is that because of bug 623625, needs-focus doesn't really mean
> anything useful for desktop Firefox.

It would help to fix bug 660224 but I still don't understand how is that linked with this one...
Ehsan, could you explain how bug 623625 is going to fix these reftests failure?
And David, why https://hg.mozilla.org/mozilla-central/rev/bf282a9b7ae4 might help?
(In reply to comment #128)
> Ehsan, could you explain how bug 623625 is going to fix these reftests failure?

If we have a reftest which depends on the Window being focused in order to render correctly, and the application window running the reftests is not focused when we're running them, then the reftest fails.  The needs-focus attribute pretends that it specifies such tests, but it just lies, and bug 623625 will fix it.

> And David, why https://hg.mozilla.org/mozilla-central/rev/bf282a9b7ae4 might
> help?

Once that bug is fixed, what David landed makes sure that these tests are only loaded when the window is focused.
(In reply to comment #130)
> (In reply to comment #128)
> > Ehsan, could you explain how bug 623625 is going to fix these reftests failure?
> 
> If we have a reftest which depends on the Window being focused in order to
> render correctly, and the application window running the reftests is not
> focused when we're running them, then the reftest fails.  The needs-focus
> attribute pretends that it specifies such tests, but it just lies, and bug
> 623625 will fix it.

Most of the failing tests listed in this bug do not need focus, see bug 	660226. I do not see how making sure the reftests are focused is going to fix this intermittent issue when most of them shouldn't care about focus...
(In reply to comment #131)
> (In reply to comment #130)
> > (In reply to comment #128)
> > > Ehsan, could you explain how bug 623625 is going to fix these reftests failure?
> > 
> > If we have a reftest which depends on the Window being focused in order to
> > render correctly, and the application window running the reftests is not
> > focused when we're running them, then the reftest fails.  The needs-focus
> > attribute pretends that it specifies such tests, but it just lies, and bug
> > 623625 will fix it.
> 
> Most of the failing tests listed in this bug do not need focus, see bug    
> 660226. I do not see how making sure the reftests are focused is going to fix
> this intermittent issue when most of them shouldn't care about focus...

Oh, OK, I stand corrected then :)
FWIW, bug 716054 is very similar to this one.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → WORKSFORME
Whiteboard: [orange]
Blocks: 819452
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: