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 519587 - Navigating nested lists with Collection not functioning properly
Navigating nested lists with Collection not functioning properly
Status: RESOLVED FIXED
Product: orca
Classification: Applications
Component: general
unspecified
Other Linux
: Normal normal
: 2.24.0
Assigned To: Scott Haeger
Orca Maintainers
Depends on:
Blocks: 490565
 
 
Reported: 2008-02-29 18:23 UTC by Scott Haeger
Modified: 2008-03-19 17:00 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
first version of Navigating nested lists with Collection not functioning properly (12.99 KB, patch)
2008-02-29 22:40 UTC, Scott Haeger
none Details | Review
second version of Navigating nested lists with Collection not functioning properly (12.58 KB, patch)
2008-03-01 21:08 UTC, Scott Haeger
none Details | Review

Description Scott Haeger 2008-02-29 18:23:08 UTC
Navigating nested lists with Collections is not functioning properly.  This problem was seen in the original "'go to' with Collections" bug.  The original comment is posted below:

Okay, cycle 3 was fine with one exception:  Going to the previous list with
Shift+L in a page with nested lists.  The test is html_struct_nav_lists.py 
Other structural navigation (at least in terms of the regression tests) worked
as expected.
Comment 1 Scott Haeger 2008-02-29 22:40:53 UTC
Created attachment 106304 [details] [review]
first version of Navigating nested lists with Collection not functioning properly

The patch fixes the nested list navigation problem but I occasionally see a very bizarre problem.  On test/html/lists.html I occassionally see the following:

Actual: "0. remember what the heck we are doing each day"
Expected: "1. remember what the heck we are doing each day"

and

Actual: "0. And use roman numerals,"
Expected: "VI. And use roman numerals,"

The problem occurs for Braille and Speech (sorry, I couldn't reproduce the actual problem for this posting).  The bogus '0' comes directly from the text interface and I have only seen it during the tests, never through Accerciser or manual testing.  I am sure it is the same version of Firefox so I am not sure what the problem is.

The fix also includes a similar fix for listitems even though they passed on the regression tests.  I wonder if we should beef up the listitem regression tests to include listitems that are focusable.  I believe the test would have failed if this type of listitem were included.

pylinted to 10.00
Comment 2 Joanmarie Diggs (IRC: joanie) 2008-03-01 20:02:05 UTC
The bogus '0' problem used to occur for me but disappeared 2 or 3 weeks ago (FF fix I think rather than anything we're doing differently, but I'm not sure).  What FF build are you using?

As for beefed up tests, we need a lot more tests and some beefed up versions of our existing tests.  If you already have such a test for list items, that would be awesome.  Otherwise, I'll add a beefier list item test to my list (so to speak). :-)  Thanks for pointing it out.
Comment 3 Scott Haeger 2008-03-01 21:08:21 UTC
Created attachment 106351 [details] [review]
second version of Navigating nested lists with Collection not functioning properly

same as first patch except for removal of test/harness/runone.sh debug change.

I used a fully updated version of Firefox and I am certain the regression test version and manual testing version were the same.  In addition, I also used the regression test Firefox profile during manual testing.  I never once saw the '0.' issue during manual testing, but consistently got it during regression testing.  The issue went away only after I fully solved the Collection navigation issue.  I then ran the regression test two times successfully.  Strange!
Comment 4 Mike Pedersen 2008-03-03 17:43:40 UTC
I've tested this on several pages with lists.  It seems to be working OK.  I'd say after it passes the regression tests it should be ready for trunk.  
Comment 5 Scott Haeger 2008-03-03 18:37:31 UTC
committed to trunk.  Bug report to remain open to track gnome2-22 back porting.
Comment 6 Willie Walker 2008-03-11 14:06:36 UTC
First coarse pass at GNOME 2.24 planning.
Comment 7 Willie Walker 2008-03-19 15:42:06 UTC
Just to verify - we were going to backport this into Orca v2.22.1, correct?  My understanding is that the reason we didn't do it for v2.22.0 was just that we were so close to the deadline.  Scott, can you verify this is the right thinking please?
Comment 8 Scott Haeger 2008-03-19 15:59:55 UTC
That is correct Will.  Let me know if you need me to do the work.
Comment 9 Willie Walker 2008-03-19 16:20:42 UTC
(In reply to comment #8)
> That is correct Will.  Let me know if you need me to do the work.
> 

Hey - if you're into it and have the time, go for it!  :-)  Much appreciated.
Comment 10 Scott Haeger 2008-03-19 16:49:43 UTC
Committed to .../branches/gnome-2-22.
pylinted to 10.00
Passes list/list item regression tests.
Marked as FIXED.