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 153920 - Item values repeated twice by screen reader
Item values repeated twice by screen reader
Status: RESOLVED FIXED
Product: gnopernicus
Classification: Deprecated
Component: speech
unspecified
Other Solaris
: Normal normal
: ---
Assigned To: remus draica
remus draica
AP2
Depends on:
Blocks:
 
 
Reported: 2004-09-28 05:45 UTC by Partha Pratim Dey
Modified: 2004-12-22 21:47 UTC
See Also:
GNOME target: ---
GNOME version: 2.5/2.6


Attachments
proposed patch (1.80 KB, patch)
2004-09-29 12:11 UTC, remus draica
none Details | Review
a different proposed patch (1.69 KB, patch)
2004-09-30 10:19 UTC, remus draica
accepted-commit_now Details | Review

Description Partha Pratim Dey 2004-09-28 05:45:15 UTC
Installed JDS Release 3 (GNOME 2.6)on HP intel P4 machin.

1. Start Gnopernicus,enable acessibility support,logout and login again.
2. Start Gnopernicus again.
3. Open Gedit.
4. Click on 'Preferences' from Edit menu.
5. In the Preferences window click on 'Syntax Highlighting' Tab. 
6. In the resulting window click check the option 'Foreground' and click on the 
color button  next to it. 
7. In the 'Pick a color' window navigate through the items

Observation: Screen reader reads the text and the value of the focusssed item 
but while moving to the next item it first reads the value of the previous item 
and then it reads the  next items text and value.This is very much confusing 
for user.
Comment 1 remus draica 2004-09-29 12:11:50 UTC
Created attachment 32067 [details] [review]
proposed patch
Comment 2 bill.haneman 2004-09-29 13:21:10 UTC
remus, can you explain what is going on in the 'else' statement which this patch
modifies?  I don't understand why you would present any information at all if
the object is not watched, and not focussed... 
Comment 3 remus draica 2004-09-30 08:14:43 UTC
In the case of else, if event is for watched object, then the speech output has
a higher priority (the events for watched object should not be interrupted by
focus-tracking events). Otherwise the event is for focused object.

NOTE: When gnopernicus starts to present an object, the object may be focused
(exception for watched objects). But, while gnopernicus is building XML
presentation string, the object may change its state. This situation is
happening here. If an event for an object which doesn't have FOCUSED state is
proposed for presentation then it is for an watched object (in current
implementation). 

In this bug, the state was changing during XML construction, and, the logic was
supposing that the event is for an watched object, and it was sent to speech
with a higher priority. So, a better detection of events for watched object was
needed here.

A better solution is to not present events like this (for an unfocused and
unwatched object) at all. I will check the behaviour for this solution.

Event from at-spi, when (only) tabbing from one spin-button to another:
  selection-changed from current _focused_ object
  focus  for the _new_ _focused_ object.
These 2 events are send by application one after other, with no delay between.
The situation described above (at NOTE) occurs in this case.
Comment 4 remus draica 2004-09-30 10:19:35 UTC
Created attachment 32102 [details] [review]
a different proposed patch
Comment 5 bill.haneman 2004-10-01 10:51:15 UTC
Either patch looks OK to me; please apply whichever one you think is working
better in tests.
Comment 6 remus draica 2004-10-05 13:49:01 UTC
Comment on attachment 32102 [details] [review]
a different proposed patch


This patch work better in my oppinion. It was applied to cvs head and gnome-2-8
branch.