Re: [orca-list] A question with checking items containing table related

[Joanmarie Diggs <jdiggs igalia com>]

Hey Attila.

> What can possible doing in Orca side to Orca present the first column 
> state changes? Or this is again a Gtk3 bug?

I just opened up a Gtk+ bug for that:
https://bugzilla.gnome.org/show_bug.cgi?id=672367

We are seeing the state changes. The problem is that the 'toggle' action
went away. And it's the toggle action that we are looking to in order to
determine of a table cell can be checked/unchecked.

So.... Orca could potentially hack around the bug you reported, however
doing so would not fix all the problems; just the state-changed problem.
In other words, every time you arrowed to an unchecked cell, Orca would
still not have a state changed event to tell it unchecked, nor would it
have a state to look at to tell it unchecked, and thus Orca will never
be able to tell you that the cell you just moved to is unchecked.

When this bug is properly fixed in Gtk+ all the problems -- including
the state-changed problem -- will magically be fixed. In addition,
adding such a hack to Orca within 24 hours of hard code freeze seems
like a great way to introduce a regression in Orca which I then cannot
undo without getting blessing from the Release Team. And based on past
experience, I am not always successful in getting that blessing. So even
if we are forced to hack around this issue, and can at best only fix
part of it, I will not do it before code freeze. I will instead talk to
the Gtk+ developer and see what can be done.

Thanks for the report!
--joanie