510128 - mochitest-a11y: intermittent "test_events_caretmove.html | Wrong caret offset for textbox - got 5, expected -1"

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Description

[Jeff Muizelaar [:jrmuizel]]

Intermittent failure on os x

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1250115564.1250121515.890.gz

Comment 1

[David Bolter [:davidb] (NeedInfo me for attention)]

Marco, Alex, any ideas what might have caused this? Any recent work in text box, and text area caret computation?  (Assuming this is a recent failure)

Comment 2

[alexander :surkov (:asurkov)]

I don't think so. It sounds like timing issue. Rarely I get this on windows platform locally.

Comment 3

[David Bolter [:davidb] (NeedInfo me for attention)]

Just thinking aloud... I sometimes get errors when the test harness window isn't in the foreground. Could this happen on a tinderbox?

Comment 4

[David Bolter [:davidb] (NeedInfo me for attention)]

When running locally I get this bug in suite but not when run on its own.

Comment 5

[David Bolter [:davidb] (NeedInfo me for attention)]

Ditto for my fresh windows 7 build too.

Comment 6

[Daniel Holbert [:dholbert]]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1252419853.1252427717.2284.gz
OS X 10.5.2 mozilla-central unit test on 2009/09/08 07:24:13

> 1232 ERROR TEST-UNEXPECTED-FAIL | chrome://mochikit/content/a11y/accessible/test_events_caretmove.html | Wrong caret offset for textbox - got 5, expected -1
> 1233 ERROR TEST-UNEXPECTED-FAIL | chrome://mochikit/content/a11y/accessible/test_events_caretmove.html | Wrong caret offset for textarea - got 12, expected -1

Note that this failure happened after an "Error: Unable to restore focus, expect failures and timeouts." in test_unsafeBidiChars.xhtml, so it's possible that this failure was due to whatever went wrong there.  The error messages in test_events_caretmove.html exactly match those in comment 0's log, though, so I think this is the same problem.

Comment 7

[David Bolter [:davidb] (NeedInfo me for attention)]

Attachment: patch (disable fragile tests for mac)

Marco, I regularly get this bug locally on Mac. Until we can put time into our Mac a11y implementation I think we should mitigate the tbox tree noise. Thoughts?

Comment 8

[Marco Zehe (:MarcoZ)]

Comment on attachment 402166 [details] [diff] [review]
patch (disable fragile tests for mac)

Yep makes sense! r=me

Comment 9

[David Bolter [:davidb] (NeedInfo me for attention)]

Pushed to trunk: http://hg.mozilla.org/mozilla-central/rev/a1283b105a34

Leaving, bug open. Please report instances on other platforms.

Comment 10

[David Bolter [:davidb] (NeedInfo me for attention)]

Landed on 1.9.2 to mitigate noise there too http://hg.mozilla.org/releases/mozilla-1.9.2/rev/b60603331352

Comment 11

[David Bolter [:davidb] (NeedInfo me for attention)]

Just happened on windows again. I think I'll disable these tests for all platforms.

Comment 12

[David Bolter [:davidb] (NeedInfo me for attention)]

Attachment: cover all platforms

Comment 13

[Serge Gautherie (:sgautherie)]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1253718770.1253721521.16968.gz
Linux mozilla-central test everythingelse on 2009/09/23 08:12:50

Comment 14

[Marco Zehe (:MarcoZ)]

Comment on attachment 402372 [details] [diff] [review]
cover all platforms

OK!

Comment 15

[David Bolter [:davidb] (NeedInfo me for attention)]

Thanks.

Pushed to mozilla-central: http://hg.mozilla.org/mozilla-central/rev/ebd1844746e0

Comment 16

[David Bolter [:davidb] (NeedInfo me for attention)]

Attachment: less random timing

This should remove one aspect of timing randomness.
(Related to bug 474952 comment 3)
Thoughts?

Comment 17

[alexander :surkov (:asurkov)]

Comment on attachment 414772 [details] [diff] [review]
less random timing

r=me, might make a sense, let's check

Comment 18

[Marco Zehe (:MarcoZ)]

Comment on attachment 414772 [details] [diff] [review]
less random timing

Yes, let's try this! Thanks!

Comment 19

[David Bolter [:davidb] (NeedInfo me for attention)]

Tryserver ran with failures on Win and Linux but not Mac.

Comment 20

[David Bolter [:davidb] (NeedInfo me for attention)]

Attachment: removed non deterministic tests.

Comment 21

[alexander :surkov (:asurkov)]

(In reply to comment #20)
> Created an attachment (id=415127) [details]
> removed non deterministic tests.

Why are they non deterministic and why should they be removed?

Comment 22

[David Bolter [:davidb] (NeedInfo me for attention)]

(In reply to comment #21)
> (In reply to comment #20)
> > Created an attachment (id=415127) [details] [details]
> > removed non deterministic tests.
> 
> Why are they non deterministic

Probably not truly non deterministic (what is?).

> and why should they be removed?

I'm not sure what we gain by testing for -1 offsets.

Comment 23

[alexander :surkov (:asurkov)]

(In reply to comment #22)

> Probably not truly non deterministic (what is?).

Still not sure I fully understand deterministic epithet.

> I'm not sure what we gain by testing for -1 offsets.

bug 448744

Comment 24

[David Bolter [:davidb] (NeedInfo me for attention)]

I see, then it would appear we have bug (sometimes).

Comment 25

[David Bolter [:davidb] (NeedInfo me for attention)]

http://hg.mozilla.org/mozilla-central/rev/8fc7bc73fd89
(Not a complete fix for this bug)

Comment 26

[Marco Zehe (:MarcoZ)]

Hasn't popped up in over two years. Closing WFM.

Comment 27

[alexander :surkov (:asurkov)]

(In reply to Marco Zehe (:MarcoZ) from comment #26)
> Hasn't popped up in over two years. Closing WFM.

absolutely and it wont' pop up until tests are enabled

Comment 28

[alexander :surkov (:asurkov)]

fixed by bug 725581