GNOME Bugzilla – Bug 538058
The role for accessibles of ROLE_LIST_ITEM should not be spoken when the accessible is given focus.
Last modified: 2008-06-23 16:03:25 UTC
Steps to reproduce: 1. Launch Impress and create a new presentation. 2. Click on the layout panel to give it focus. 3. Arrow amongst the different layouts Expected results: Orca would not say "list item" after each item Actual results: Orca says "list item" after each item Note, this is not limited to Impress. I've seen it here and there (OOo wizards and/or dialogs, Firefox's bookmarks dialog) and it's always driven me batty. Turns out Mike doesn't like it either. :-)
Created attachment 112706 [details] Orca debug log generated whilst testing this problem. What do you mean by "too many times". As far as I can see (or rather hear), it only says it once for each layout type (for example, see line 872). Does that mean you don't want it to say: SPEECH OUTPUT: 'Title, Text, Chart list item' but just: SPEECH OUTPUT: 'Title, Text, Chart' ?
> What do you mean by "too many times". :-) "too many times" == several times per item, whereas "too darned much" == "if I hear Orca say 'list item' one more time my head will explode". ;-) Admittedly a fine line. We're suffering from the latter. I'll re-title the bug. Was planning on doing so anyway (if you didn't beat me to it). The opening comment is the (serious) description of the problem and the expected vs. actual results. > Does that mean you don't want it to say: > > SPEECH OUTPUT: 'Title, Text, Chart list item' > > but just: > > SPEECH OUTPUT: 'Title, Text, Chart' > > ? Yes please. Note that the speaking of the role should not be changed in the whereAmI information. Just when the item is given focus (tabbed to, arrowed to, etc.).
Okay. Still needs further clarification though. Do you want this change generically for ALL list items in ALL applications, whereever they may appear? Or just to be applied to the specific example you gave for OOo Impress? If it's for ALL list items, a more general description might be less confusing.
(In reply to comment #3) > Do you want this change generically for ALL > list items in ALL applications, Yes, please. :-) But that really is a question for Mike so I'll put his name on the bug. FWIW the way I see it is this: 1. If focus wasn't in the list before and now it is in a list (i.e. the user tabbed to a list and is now on a list item), the speech context should inform the user that he/she is now in a list. That implies that he/she is on a list item because lists contain list items. I believe most users therefore don't need to hear "list item" in this case. 2. If focus is already in the list and the user is arrowing around, the only thing that user can be landing on are other list items. Therefore speaking "list item" in this instance is redundant. 3. If the user has no idea where he/she is and uses whereAmI to work it out, knowing that he/she is on a list item is valuable. So we DO want to speak "list item" for whereAmI. > If it's for ALL list items, a more general > description might be less confusing. Do you mean the summary or the opening comment? It's obviously too late for the opening comment, but I certainly wouldn't object to your re-re-titling the bug to say what you think it should. I was merely attempting to be precise. Sorry! :-(
I agree 100% with Joanie's comments above.
Created attachment 112998 [details] [review] Revision #1. Tested with OOo 3.0beta_20080429 on Ubuntu Hardy. This just adjusts speech. Mike, what adjustments (if any) do you want for the braille line?
Rich, you have just made my day. Tested in OOo and Firefox. Seems to work quite nicely -- and boy what a huge difference! Because there are various focusable lists in Firefox for which we should be defering to the default speechgenerator, I do want to run the FF regression tests using this patch as an added check. I'll do that tonight. Thanks for this fix!
I'm also quite happy with this change. I don't think any braille changes are required.
Rich, I just finished running all of the FF regression tests. I have a few tests to update to remove the role, but the changes are all appropriate (and improvements). It all looks good from my perspective and I vote for a commit.
Thanks Joanie. Patch committed to SVN trunk. Moving to "[pending]".
I'm still getting these on some web sites but I believe that is a separate issue that Joanie is looking at.