Closed Bug 525175 Opened 15 years ago Closed 12 years ago

Fix timing issue with test_name_nsRootAcc.xul on Linux.

Categories

(Core :: Disability Access APIs, defect)

Product:
Core
Component:
Disability Access APIs
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla13

People

(Reporter: philor, Assigned: surkov)

References

(Blocks 1 open bug)

Details

(Whiteboard: [fixed by bug 719754] [test disabled on Linux on m-1.9.3+])

Attachments

(5 files, 1 obsolete file)

patch [Checkin: Comment 4]
1.01 KB, patch
MarcoZ
: review+
Details | Diff | Splinter Review
a long shot [Checkin: Comment 18]
1.24 KB, patch
MarcoZ
: review+
surkov
: review+
Details | Diff | Splinter Review
fix added windows failures [Checkin: Comment 28]
884 bytes, patch
surkov
: review+
Details | Diff | Splinter Review
investigative [Checkin: Comment 63]
1.66 KB, patch
MarcoZ
: review+
Details | Diff | Splinter Review
bail on linux2 [Checkin: Comment 78]
1.53 KB, patch
davidb
: review+
Details | Diff | Splinter Review 
At least three times today (that I spotted looking back through the ignored errors), a11y tests on Linux debug have blown up, reporting tens or hundreds of thousands of errors, all starting with

1889 ERROR TEST-UNEXPECTED-FAIL | chrome://mochikit/content/a11y/accessible/test_name_nsRootAcc.xul | Test timed out

followed by thousands of

[SimpleTest/SimpleTest.js, window.onerror] An error occurred - nsIAccessibleEvent is not defined at chrome://mochikit/content/a11y/accessible/events.js:766

Logs I saw:
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1256785900.1256791549.21285.gz
Linux mozilla-central test debug everythingelse on 2009/10/28 20:11:40  

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1256763311.1256769275.28481.gz
Linux mozilla-central test debug everythingelse on 2009/10/28 13:55:11  

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1256751737.1256758120.31701.gz
Linux mozilla-central test debug everythingelse on 2009/10/28 10:42:17  

I haven't seen it before, and the first failure was just a couple of runs after bug 524665 landed, so I suppose (without understanding the code in the least) it could be a result of that, rather than just coincidence.
Blocks: 438871
Whiteboard: [orange]
(In reply to comment #0)
> followed by thousands of

That behavior is tracked in bug 492956.
Depends on: 492956
this might help
Attachment #409031 - Flags: review?(marco.zehe)
Attachment #409031 - Flags: review?(marco.zehe) → review+
Comment on attachment 409031 [details] [diff] [review]
patch
[Checkin: Comment 4]

r=me. Thanks!
landed on 1.9.3 - http://hg.mozilla.org/mozilla-central/rev/aaaa9779a9a2

hope this will fix it
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1257791471.1257797839.4994.gz
Linux mozilla-central debug test everythingelse on 2009/11/09 10:31:11
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1257876663.1257883552.16238.gz
Linux mozilla-central debug test everythingelse on 2009/11/10 10:11:03
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1257886520.1257893272.26917.gz
Linux mozilla-central debug test everythingelse on 2009/11/10 12:55:20

Many many a11y failures as described, but unlike comment 0 the first failure is:

TEST-UNEXPECTED-FAIL | chrome://mochikit/content/browser/browser/components/places/tests/browser/browser_library_views_liveupdate.js | Timed out
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1257884834.1257891765.10041.gz#err1
Linux mozilla-central debug test everythingelse on 2009/11/10 12:27:14
Well this is annoying.I'm curious to see if this happens on the 192 branch as well.

Ginn this seems to hit linux only, have you seen it on open solaris?
worth a try?
Attachment #411593 - Flags: review?(surkov.alexander)
Attachment #411593 - Flags: review?(marco.zehe)
(In reply to comment #11)
> Created an attachment (id=411593) [details]
> a long shot
> 
> worth a try?

Can you describe how this is going to help?
(In reply to comment #10)
> Ginn this seems to hit linux only, have you seen it on open solaris?

I don't have a box running mochitest-plain or mochitest-a11y all the time.
I haven't seen this timeout for my manually runs on Solaris.
Comment on attachment 411593 [details] [diff] [review]
a long shot
[Checkin: Comment 18]

This is worth a shot. I was also thinking whether we could check if the window actually opened, and bail if it didn't within a certain amount of time. But I don't know if toolkit allows us to do that.
Attachment #411593 - Flags: review?(marco.zehe) → review+
(In reply to comment #12)
> (In reply to comment #11)
> > Created an attachment (id=411593) [details] [details]
> > a long shot
> > 
> > worth a try?
> 
> Can you describe how this is going to help?

Not sure but I think the focus refactor might have changed eventing for non foreground windows; or a least there might be some strange interaction between gecko eventing and the Linux window manager. Making sure the opened window has is front-most and has focus might help.

Like I said, it is a long shot :)
Neil your thoughts on focus events and comment #15?
Comment on attachment 411593 [details] [diff] [review]
a long shot
[Checkin: Comment 18]

ok, at least I don't see anything bad here
Attachment #411593 - Flags: review?(surkov.alexander) → review+
Landed on central:
http://hg.mozilla.org/mozilla-central/rev/e12ee7ca0834

Note: Adding bug 521233 as a (likely) dependency -- I just discovered this bug now.
Depends on: 521233
(In reply to comment #18)
> Landed on central:
> http://hg.mozilla.org/mozilla-central/rev/e12ee7ca0834
> 

This didn't work either.
Not sure if my push caused this but so far no Linux fails, but new to me, Windows fails.

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1257981877.1257984401.28000.gz
WINNT 5.2 mozilla-central test everythingelse on 2009/11/11 15:24:37
OS: Linux → All
Hardware: x86 → All
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1257986719.1257992406.20502.gz#err0
WINNT 5.2 mozilla-central debug test everythingelse on 2009/11/11 16:45:19
The silly part of me want to try only doing the focus only if (LINUX). Let's see if the instances have truly flipped only to Windows.
If so, then it will tell use more about where to look (OS specific)
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1257989754.1257993019.27386.gz
WINNT 5.2 mozilla-central opt test everythingelse on 2009/11/11 17:35:54
Attachment #411871 - Flags: review?(surkov.alexander) → review+
Comment on attachment 411871 [details] [diff] [review]
fix added windows failures
[Checkin: Comment 28]

let's try
Comment on attachment 411871 [details] [diff] [review]
fix added windows failures
[Checkin: Comment 28]

My thoughts exactly. :) Let's land this and see what happens.
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1258038507.1258044884.29070.gz
Linux mozilla-central debug test everythingelse on 2009/11/12 07:08:27
s: moz2-linux-slave01

Looks like comment 28 didn't fix it. :(
Daniel is it less noisy at least?
From a quick glance at the log in comment 29 versus the one in comment 24, it appears to be just as noisy as before.
However, it looks like the changeset in comment 28 was targeted at fixing Windows failures -- and the recent orange (comment 29) was on Linux.  (I don't really know the backstory here, so I don't know if that's significant.)
(In reply to comment #31)
> From a quick glance at the log in comment 29 versus the one in comment 24, it
> appears to be just as noisy as before.

Just to be clear, I meant noisy == the frequency of this bug :)
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1258581730.1258593696.15607.gz
Linux mozilla-central debug test everythingelse on 2009/11/18 14:02:10
s: moz2-linux-slave13
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1259675396.1259682883.24231.gz
Linux mozilla-central debug test everythingelse on 2009/12/01 05:49:56
s: moz2-linux-slave02
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1259716936.1259724448.17969.gz
Linux mozilla-central debug test everythingelse on 2009/12/01 17:22:16
s: moz2-linux-slave10
bug 513605 looks like the same underlying issue (or at least same symptom), with a different test timing out.
See Also: → 513605
See Also: 513605 → 
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1261102092.1261111207.23385.gz
Linux mozilla-central debug test everythingelse on 2009/12/17 18:08:12
s: moz2-linux-slave01
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1262034124.1262041846.31771.gz
Linux mozilla-central debug test everythingelse on 2009/12/28 13:02:04
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1264184354.1264187294.31064.gz
Linux mozilla-central debug test mochitest-other on 2010/01/22 10:19:14
s: moz2-linux-slave15
Blocks: a11yrandomorange
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1264635959.1264638529.24737.gz
Linux mozilla-central debug test mochitest-other on 2010/01/27 15:45:59
s: moz2-linux-slave41
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265747526.1265751128.17627.gz
Linux mozilla-central debug test mochitest-other on 2010/02/09 12:32:06
s: moz2-linux-slave06
Summary: Intermittent timeout in test_name_nsRootAcc.xul followed by thousands of nsIAccessibleEvent is not defined errors → Intermittent timeout in test_name_nsRootAcc.xul
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265759399.1265762175.11113.gz
Linux mozilla-central debug test mochitest-other on 2010/02/09 15:49:59
s: moz2-linux-slave06
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265767077.1265771803.23716.gz
Linux mozilla-central debug test mochitest-other on 2010/02/09 17:57:57
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265828593.1265831653.510.gz
Linux mozilla-central debug test mochitest-other on 2010/02/10 11:03:13
s: moz2-linux-slave13
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265835889.1265838318.12383.gz
Linux mozilla-central debug test mochitest-other on 2010/02/10 13:04:49
s: moz2-linux-slave08
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265839144.1265841969.21024.gz
Linux mozilla-central debug test mochitest-other on 2010/02/10 13:59:04
s: moz2-linux-slave09
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265884122.1265886905.6646.gz
Linux mozilla-central debug test mochitest-other on 2010/02/11 02:28:42
s: moz2-linux-slave06
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265876440.1265879188.14732.gz
Linux mozilla-central debug test mochitest-other on 2010/02/11 00:20:40
s: moz2-linux-slave41
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265902178.1265905849.7173.gz
Linux mozilla-central debug test mochitest-other on 2010/02/11 07:29:38
s: moz2-linux-slave11
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265899813.1265903479.12893.gz
Linux mozilla-central debug test mochitest-other on 2010/02/11 06:50:13
s: moz2-linux-slave01
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265928360.1265933013.26893.gz
Linux mozilla-central debug test mochitest-other on 2010/02/11 14:46:00
s: moz2-linux-slave07
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265933791.1265937772.16930.gz
Linux mozilla-central debug test mochitest-other on 2010/02/11 16:16:31
s: moz2-linux-slave09
Comment on attachment 426560 [details] [diff] [review]
investigative
[Checkin: Comment 63]

In absence of a better idea to try, let's try this. I really am not sure why this is failing so often.
Attachment #426560 - Flags: review?(marco.zehe) → review+
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265957634.1265960179.5947.gz#err0
Linux mozilla-central debug test mochitest-other on 2010/02/11 22:53:54  
s: moz2-linux-slave12
(In reply to comment #60)
> investigative: http://hg.mozilla.org/mozilla-central/rev/e24c49718527

I can't see any changes.
(In reply to comment #61)
> (In reply to comment #60)
> > investigative: http://hg.mozilla.org/mozilla-central/rev/e24c49718527
> 
> I can't see any changes.

Wild. Let's try again.

investigative: http://hg.mozilla.org/mozilla-central/rev/145556db94fb
Sigh.
Linux mozilla-central debug test mochitest-other
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265990792.1265993637.31660.gz

(Maybe it will get more frequent now that we don't especially focus the dialog on linux)
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1265999005.1266003076.14356.gz
Linux mozilla-central debug test mochitest-other on 2010/02/12 10:23:25
s: moz2-linux-slave09
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266002363.1266005337.8409.gz
Linux mozilla-central debug test mochitest-other on 2010/02/12 11:19:23
s: moz2-linux-slave08
Attached patch bail on linux (obsolete) — — Splinter Review 
We should probably stop the noise and figure this out offline; or take stabs at it during off peak hours.
Attachment #426727 - Flags: review?(surkov.alexander)
Comment on attachment 426727 [details] [diff] [review]
bail on linux

Cancelling review; I have to step away - Alexander can tidy this up and land it.
Attachment #426727 - Flags: review?(surkov.alexander)

        Attachment #426727 -
        Attachment is obsolete: true

    
    

    
  
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266446100.1266448570.901.gz&fulltext=1
Linux mozilla-central debug test mochitest-other [testfailed] Started 14:35, finished 15:17

    
    

    
  
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266483229.1266485625.24659.gz
Linux mozilla-central debug test mochitest-other on 2010/02/18 00:53:49
s: moz2-linux-slave08

    
    

    
  
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266528454.1266532186.23863.gz
Linux mozilla-central debug test mochitest-other on 2010/02/18 13:27:34
s: moz2-linux-slave02

    
    

    
  
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266543457.1266546436.13845.gz
s: moz2-linux-slave102966 ERROR TEST-UNEXPECTED-FAIL | chrome://mochikit/content/a11y/accessible/test_name_nsRootAcc.xul | Test timed out.
2969 ERROR TEST-UNEXPECTED-FAIL | chrome://mochikit/content/a11y/accessible/test_nsIAccessNode_utils.html | [SimpleTest/SimpleTest.js, window.onerror] An error occurred - uncaught exception: [accessible/events.js, gA11yEventObserver.observe] This is expected if a previous test has been aborted... Initial exception was: [ ReferenceError: nsIAccessibleEvent is not defined ] at :0
PROCESS-CRASH | automation.py | application crashed (minidump found)
Thread 1 (crashed)
PROCESS-CRASH | automation.py | application crashed (minidump found)
Thread 1 (crashed)
PROCESS-CRASH | automation.py | application crashed (minidump found)
Thread 1 (crashed)

    
    

    
  
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266576297.1266578783.30714.gz
Linux mozilla-central debug test mochitest-other on 2010/02/19 02:44:57  
s: moz2-linux-slave15

    
    

    
  
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266601430.1266604090.16626.gz
Linux mozilla-central debug test mochitest-other on 2010/02/19 09:43:50  
s: moz2-linux-slave02

    
    

    
  
Comment on attachment 426825 [details] [diff] [review]
bail on linux2
[Checkin: Comment 78]

Yep, thanks.
Pushed as http://hg.mozilla.org/mozilla-central/rev/899cf3b6f797

Sorry for the oranges folks.

        Attachment #426825 -
        Flags: review+

    
    

    
  
Should this still be a random orange bug?

    
    

    
  
Removed orange keyword and morphed bug summary.
No longer blocks: 438871
Summary: Intermittent timeout in test_name_nsRootAcc.xul → Fix timing issue with test_name_nsRootAcc.xul on Linux.
Whiteboard: [orange]

    
    

    
  
Linux mozilla-central debug test mochitest-other [testfailed] Started 12:02, finished 12:52
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1266609738.1266612717.13303.gz

        Attachment #426825 -
        Attachment description: bail on linux2 → bail on linux2
[Checkin: Comment 78]

    
    

    
  
(In reply to comment #80)
> Removed orange keyword and morphed bug summary.

I would not have changed them, as this is worked around, not fixed:
yet less need to find this bug if the failure doesn't occur anymore,
it may just miss some attention.
Whiteboard: [test disabled on Linux on m-1.9.3+]

        Attachment #426560 -
        Attachment description: investigative → investigative
[Checkin: Comment 63]

        Attachment #411871 -
        Attachment description: fix added windows failures → fix added windows failures
[Checkin: Comment 28]

        Attachment #411593 -
        Attachment description: a long shot → a long shot
[Checkin: Comment 18]

        Attachment #409031 -
        Attachment description: patch → patch
[Checkin: Comment 4]

    
    

    
  
this does not look like a browser chrome test, how can it depend on bug 521233?

    
    

    
  
(In reply to comment #83)
> this does not look like a browser chrome test, how can it depend on bug 521233?

a11y tests are chrome tests.

    
    

    
  
(In reply to comment #84)
> (In reply to comment #83)
> > this does not look like a browser chrome test, how can it depend on bug 521233?
> 
> a11y tests are chrome tests.

and a11y tests use waitForFocus function so it might be related, theoretically I think

    
    

    
  
Right, I set the dependency so I would remember retry this test when 532233 is fixed.

    
    

    
  
(In reply to comment #84)
> (In reply to comment #83)
> > this does not look like a browser chrome test, how can it depend on bug 521233?
> 
> a11y tests are chrome tests.

that bug is only about BROWSER chrome tests, nobody ever reported issues with waitForFocus in Chrome tests. if you can tell the issue is waitForFocus, than we should change that bug to be generic.

    
  
Blocks: 586818
Depends on: 633725

    
    

    
  
fixed by bug 719754
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla13
No longer blocks: 513605
No longer blocks: 586818
No longer blocks: 542726
Assignee: nobody → surkov.alexander
Flags: in-testsuite-
Whiteboard: [test disabled on Linux on m-1.9.3+] → [fixed by bug 719754] [test disabled on Linux on m-1.9.3+]
No longer depends on: 521233

          You need to log in
          before you can comment on or make changes to this bug.
        






  

    
  







  

    

      
Attachment

      

      
      
      
      
    

    

      

        

          

            
General

            

              Creator: 



            

            
Created: 

            
Updated: 

            
Size: