GNOME Bugzilla – Bug 555466
Orca should attempt to correct the Firefox breakage introduced by bug 535827 via script mapping
Last modified: 2008-10-14 20:25:41 UTC
Steps to reproduce: 1. Get the very latest Ubuntu Intrepid 2. Launch Orca 3. Launch the version of Firefox which ships with Intrepid Expected results: All of the Firefox commands, navigation, speech/braille, etc. would work Actual results: None of the Firefox commands, navigation, speech/braille, etc. work. This seems to be the result of bug 535827. Minefield seems to continue to function, however. The difference between the two applications is that we map the name Minefield to the Mozilla script, but don't do the same for name Firefox/firefox.
Created attachment 120167 [details] [review] Seems to work.... I re-broke a11y on my laptop. This patch causes us to treat FF 3 and Minefield the same (i.e. we don't seem to detect the toolkit, but we use Mozilla.py which imports Gecko). Quick tests of structural navigation and line navigation in FF 3.0 seem promising. I have not yet tried this for localized Firefox. Will, do you think that will be an issue? I also need to try this in OpenSolaris.... The regression tests are based on 3.1 and an app called Minefield, but I'll see about modifying my tests to do a quick (relatively speaking) run using 3.0.
(In reply to comment #1) > Created an attachment (id=120167) [edit] > Seems to work.... Works for me, too. THANKS! With the gtk-modules schemas in place, this seems to give me Firefox support back. This patch is benign enough that I think it should be committed. I'd like to make sure more testing is done with this to see if this completely bypasses the Orca problems seen as a result of the new FF side effect introduced by bug #535827. That is, does it help stop an accessibility disaster completely for FF/Orca?
On the sanity (I don't want to make anything worse) front, I have done cursory testing with FF 3.0 in OpenSolaris as well as in Intrepid using the following locales: Spanish, German, and Arabic. In all cases, we are loading the correct script now and basic line nav and structural nav are working. Based on this and your suggestion, I have just committed the patch to trunk. I'll send out a note to the list. If there are some users who solved the problem by getting Minefield (as opposed to removing the distro at-spi and reinstalling from svn trunk), then they should have a broken Firefox with which they could provide feedback. Moving to pending because "testing required" refers to patches which have not yet been committed.