After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 577899 - Strange cursor movements in Firefox
Strange cursor movements in Firefox
Status: RESOLVED FIXED
Product: orca
Classification: Applications
Component: general
2.26.x
Other All
: Normal normal
: 2.28.0
Assigned To: Joanmarie Diggs (IRC: joanie)
Orca Maintainers
Depends on:
Blocks: 404403
 
 
Reported: 2009-04-04 03:34 UTC by Nolan Darilek
Modified: 2009-11-09 21:35 UTC
See Also:
GNOME target: ---
GNOME version: 2.25/2.26


Attachments
Debug output (242.00 KB, text/plain)
2009-04-04 03:37 UTC, Nolan Darilek
  Details
possible fix?? (1.27 KB, patch)
2009-08-14 21:08 UTC, Joanmarie Diggs (IRC: joanie)
committed Details | Review

Description Nolan Darilek 2009-04-04 03:34:51 UTC
I've seen this bug for quite some time, but finally got and edited a log of the behavior.

Sometimes, if I load a page in Firefox and immediately begin arrowing down, the cursor position seems to move higher on the page randomly. Also, the title of the HTML content area is spoken several times for no apparent reason. This behavior seems to happen almost immediately after the page loads. If I wait some unspecified time but usually too long for my patience, I can usually avoid it. For kicks, I wondered if it would stop after a time or continue, and have successfully navigated down a page for several minutes while this behavior continues, finally giving up in impatience. The solution either seems to be waiting an unspecified amount of time for some condition to be resolved, or pressing left or right arrows. It also seems to happen more when under heavier loads, and some behaviors (E.g. speaking "<Page title> HTML content") will appear in continuous read. I've seen this in multiple versions of Firefox. This is 3.0.8 under Ubuntu 9.04.

The following log was obtained from the URL:

http://www.virtualbox.org/wiki/Documentation

Notice the following. This text is spoken several times, seemingly at random:

SPEECH OUTPUT: ''
SPEECH OUTPUT: 'Documentation - VirtualBox html content'

This text is also spoken several times. If I am arrowing down through the page, prresumably advancing, I'd only expect it to be spoken once:

SPEECH OUTPUT: 'Login link Settings link Register link Help/Guide link'
Comment 1 Nolan Darilek 2009-04-04 03:37:26 UTC
Created attachment 132059 [details]
Debug output
Comment 2 Willie Walker 2009-06-17 14:31:54 UTC
I believe the same cursor movement problem (or at least something very similar) will happen regardless of whether Orca is running or not.  :-(
Comment 3 Joanmarie Diggs (IRC: joanie) 2009-08-14 21:08:43 UTC
Created attachment 140803 [details] [review]
possible fix??

This is hard to reproduce. However, based on the description, along with some experiments on planet.gnome.org (which can take a while to fully load), I hope that this patch may address the problem -- or at least make things better.

Not yet regression tested, but logically I would think it's safe. Will, thoughts please.
Comment 4 Nolan Darilek 2009-08-14 21:15:56 UTC
Sorry, meant to get to this earlier in the week.

I think this is also related to WebbVisum. One big issue I noticed when using it, for instance, was that sometimes the cursor movements yanked focus out of the location, search or in-page find areas, to the point where I'd conditioned myself to wait a second or two before typing, lest my URL be interpreted as structural nav commands.

The issue has mostly gone away with the disabling of WV, though it still happens very occasionally. If the patch adds value then cool, but I'm fairly sure this isn't Orca's issue and wouldn't object to it being closed (unless WV is just another piece of the puzzle.)
Comment 5 Willie Walker 2009-08-15 01:04:31 UTC
(In reply to comment #3)
> Created an attachment (id=140803) [edit]
> possible fix??
> 
> This is hard to reproduce. However, based on the description, along with some
> experiments on planet.gnome.org (which can take a while to fully load), I hope
> that this patch may address the problem -- or at least make things better.
> 
> Not yet regression tested, but logically I would think it's safe. Will,
> thoughts please.

This seems fair to me, and I say commit. :-)
Comment 6 Joanmarie Diggs (IRC: joanie) 2009-08-15 08:59:42 UTC
Thanks Will.

I ran all the regression tests to be safe. I haven't yet worked out how to create a reliable regression test to verify this bug/fix, but I'm still pondering it. Maybe something clever involving writing the page content via javascript with some sort of delay.... Anyhoo, in the meantime this patch has been committed to master. Closing as FIXED.