GNOME Bugzilla – Bug 434780
Orca does not provide access to text attributes in Firefox
Last modified: 2009-03-10 14:25:23 UTC
This is a tracking bug for what is currently* Mozilla issue 340809: ATK: Implement all AtkObject and AtkText attributes (https://bugzilla.mozilla.org/show_bug.cgi?id=340809). Until Firefox exposes text attribute information to us, we cannot provide access to it. Because the bug has been around a bit, I pinged Aaron to find out what the status is. He replied: > I'd like to have this implemented by mid-June for sure. We want to stabilize > what we have first, and fix some other really bad bugs, before taking on > something new. *In an earlier comment he indicated that text attributes would be filed as a separate bug. When that happens, I'll update our bug.
Removing target milestone from [blocked] bugs. We have little control over them, so we're better off letting priority and severity be our guide for poking the related components.
As of today's build of 3.1a we have attributes being exposed to us. Yea! Orca+F ain't JustWork(tm)'ing however, so I need to figure out why. Having access to attributes in FF (and I assume Thunderbird, but I've not yet checked that) would be a great addition for the 2.24 release I think. Besides, if there's anything wrong with the Mozilla implementation, it would be best to find out sooner rather than later. Therefore, I'm targeting this for 2.23.90.
Created attachment 121167 [details] [review] work in progress only Putting this here for safe keeping. Not ready for testing, but I'm getting there.
Created attachment 121174 [details] [review] revision 1 *NOT* regression tested. But it seems to work. Will, when you're back and caught up, mind giving this a look? It's not elegant; the goal was just to get all the Gecko names and values mapped to and from Atk equivalents for consistency and localization. The one "known issue" is this: settings.py has a non-setting called allTextAttributes. Many of the attributes in allTextAttributes are not in Gecko. Some attributes in Gecko (e.g. invalid) are not in everything else. Therefore it seems (to me) that allTextAttributes should be a property of the script rather than settings. I've left it where it is and applied what was done in the soffice script.py to allTextAttributes in the Gecko script.py. It seems to work with one exception: If a Gecko app has focus and you get into the generic, Orca-wide preferences dialog, the text attributes listed (i.e. all of the possible attributes) are those associated with Gecko and not the generic ones from settings. Seems that at the moment we're building that treestore the active script is still the one associated with the app. I'm not yet sure what the right thing to do here is. Part of me wants to call it a separate bug and worry about it later, but that may just be my fatigue talking. ;-) Thanks! P.S. (If you enable "invalid" you can now identify misspelled words - those with red squigglies. Woo hoo!)
Will and I went over this over chai today. :-) Thanks Will! Pending the running of the regression tests, which I'll do this evening, the plan is to commit this patch to trunk. We'll also request a string freeze break for the new strings (3 as I recall) introduced by this patch so that we can get this in for 2.24.2.
Patch committed to trunk for the 2.25.1 release. Will volunteered to submit the freeze break request for the (four) new strings for 2.24.2. Thanks!!
Forgot to mention something: The 'known issue' identified in comment #4 is not a bug resulting from this patch, but rather a bug exposed by it. Will and I also discussed this today and agreed that a new bug should be opened to address that issue. I've just opened bug 558497.
Because this is seen as a new feature rather than a bug fix, we're only able to commit it trunk and not the 2.24 branch. That was done back on 30 Oct. Therefore updating the target and closing as FIXED.