[orca-list] GtkTreeView problems and fixes
- From: Joanmarie Diggs <jdiggs igalia com>
- To: orca-list <orca-list gnome org>
- Subject: [orca-list] GtkTreeView problems and fixes
- Date: Sat, 11 Feb 2012 10:07:41 -0500
Hey guys.
For those of you on the bleeding edge of your distros (including precise
and rawhide), there is a rather unpleasant bug or two in Gtk+ 3 which
causes Orca to behave very badly in GtkTreeView's which have multiple
columns. The best case is that Orca just presents the wrong thing. The
almost worst case is that Orca crashes (what I was seeing). The really
ugly case is that Orca hangs (what people were seeing in Precise). And I
don't see a way to hack around this in Orca.
The good news is that Benjamin Otte has already fixed these things in
Gtk+ 3. In addition, Mike Gorse added a check in AT-SPI2 which doesn't
solve the wrong-cell presentation, but should stop the crashes and
hangs. But this all occurred after GNOME 3.3.5 (unstable release) went
out the door. So....
If you are on the bleeding edge and haven't updated in a while, my
advice to you would be "don't update." If you have updated recently, my
advice would be either get the latest AT-SPI2 or avoid GtkTreeView's
with multiple columns for now. If you are on the bleeding edge and want
to risk trashing your system, you can also build the latest Gtk+ 3 (and
a whole slew of dependencies) from source. I did this to test and verify
everything is fixed, but it is not (not not) something I would recommend
as you really might bork your system. And if you do completely bork your
system, I cannot help you unbork it.
Finally, for package maintainers who are not just grabbing upstream
wholesale, I have included the AT-SPI2 and Gtk+ commits you'll need on
the following bug: https://bugzilla.gnome.org/show_bug.cgi?id=669794
Take care.
--joanie
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]