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 455337 - [lynn][mike] flat-review issues with an app containing horizontal and vertical sliders
[lynn][mike] flat-review issues with an app containing horizontal and vertica...
Status: RESOLVED NOTABUG
Product: orca
Classification: Applications
Component: general
unspecified
Other Linux
: Normal normal
: ---
Assigned To: Lynn Monsanto
Orca Maintainers
Depends on:
Blocks:
 
 
Reported: 2007-07-09 22:50 UTC by Lynn Monsanto
Modified: 2007-07-19 01:00 UTC
See Also:
GNOME target: ---
GNOME version: 2.17/2.18


Attachments
JSlider Demo (8.66 KB, text/plain)
2007-07-17 23:10 UTC, Lynn Monsanto
Details

Description Lynn Monsanto 2007-07-09 22:50:29 UTC
It's possible to flat-review the vertical sliders (right half) of the SwingSet2 JSlider Demo, but not the horizontal sliders (left half).

It is possible to tab navigate all the sliders in the demo, and Orca does speak them correctly.
Comment 1 Lynn Monsanto 2007-07-17 23:09:40 UTC
I've attached a test program that just shows the JSlider demo part of SwingSet2. When the application starts up, the horizontal slider in the upper-right-hand corner gets keyboard focus. When flat-review is started, pressing KP-9 or KP-6 moves flat-review to the next horizontal slider. This continues until the next horizontal slider is reached. The next KP-9 or KP-6 moves flat-review to a filler in the right-hand panel containing the vertical sliders. Repeated KP-9 or KP-6 key presses just speak filler components in the right-hand panel.

This is different behavior than tab navigation where repeated tab key presses move  down through the horizontal sliders and then right to the vertical sliders. 

It turns out that all the vertical slider zones are in first line, while the first horizontal slider is in the second line. This is because the y-coordinate of the top of the vertical sliders is less than the y-coordinate of the top of the first horizontal slider. So, when the test application starts of with the first horizontal slider having focus, pressing KP-7 or KP-4 moves flat-review to the first line containing the top of the vertical sliders. 

I'm thinking that the flat-review behavior is not a bug, but a little surprising from a user-interaction perspective.
Comment 2 Lynn Monsanto 2007-07-17 23:10:45 UTC
Created attachment 91909 [details]
JSlider Demo

This is the demo application referenced in the previous comment.
Comment 3 Willie Walker 2007-07-18 12:23:07 UTC
I believe this behavior is the same as what we see with all vertical sliders and scrollbars everywhere.  For example, I think it is also what we see with gnome-terminal.  Mike, can you confirm this is the expected behavior?

In addition, ideally we could also present the orientation of the slider/scrollbar, but I'm not sure we can get this information (it would be STATE_VERTICAL and STATE_HORIZONTAL).  This might be something to log as an RFE -- Mike, can you confirm this is desirable state information that should be added to flat review?

In addition, when running the SwingSet demo, however, I did notice something horribly wrong with flat review, which is that we do not review panel labels (e.g., "Horizontal" and "Vertical", "Plain", "Major Ticks", ...) or the status bar.  This seems horribly bad.  Lynn, can you please verify this is also broken for you and please log a bug if it is indeed broken?
Comment 4 Mike Pedersen 2007-07-18 22:00:38 UTC
> I believe this behavior is the same as what we see with all vertical sliders
> and scrollbars everywhere.  For example, I think it is also what we see with
> gnome-terminal.  Mike, can you confirm this is the expected behavior?
Yes this is the expected behavior.
> 
> In addition, ideally we could also present the orientation of the
> slider/scrollbar, but I'm not sure we can get this information (it would be
> STATE_VERTICAL and STATE_HORIZONTAL).  This might be something to log as an RFE
> -- Mike, can you confirm this is desirable state information that should be
Yes this would be quite nice to have.  Lynn could you file a separate bug on this?  
Comment 5 Lynn Monsanto 2007-07-19 01:00:30 UTC
Closing as NOTABUG since the behavior is expected.