Re: [orca-list] bug: smplayer applications flood orca when they exit

[Joanmarie Diggs <jdiggs igalia com>]

Thanks Jerem.

If you pressed and released Down, heard "1300" and then immediately
pressed and released Down but had to wait two seconds to hear "1400,"
then I'm afraid my guess is correct. Because after you hear "1300",
Orca is told absolutely nothing: there are no accessible object events
(e.g. focus change, caret-moved), and there are no accessible input
events (e.g. Down was pressed). After those two seconds pass, Orca is
given object events from objects which are now dead. Orca ignores
nearly all of these immediately. Then Orca gets the input event from
the Down arrow and the caret-moved event and says "1400."

So the problem/bug is that "Orca is told nothing for two seconds." I
don't know what I can do in Orca to make other apps tell it stuff
sooner. The only thing I can try to do is reproduce this locally, then
demonstrate it happens without Orca (e.g. via a pyatspi event
listener), and file a bug against AT-SPI so Mike can determine if the
problem is in AT-SPI itself, or if the Qt app is the problem. So....
Are you just listening to something using smplayer in the background
and when it finishes playing this lag happens?

--joanie

On Sat, 2022-04-23 at 09:27 +0200, Jérémy Prego via orca-list wrote:
> hi joanie,
>
> I just carried out a debug by reading lines in pluma, and I think we
> can 
> see the problem better.
>
> between SPEECH OUTPUT: '1300' and SPEECH OUTPUT: '1400'
>
> https://www.dropbox.com/s/u7vfbjglqt1oqkd/debug-2022-04-23-09%3A05%3A38.out?dl=1
>
> thanks again for the help, and a correction if possible :)
>
> Jerem
> Le 21/04/2022 à 14:39, Joanmarie Diggs a écrit :
> > Hey Jerem.
> >
> > I'd be happy to look at a new one, but my guess is that I'll see
> > the
> > same thing.  Let's see....
> >
> > --joanie
> >
> > On Thu, 2022-04-21 at 14:25 +0200, Jérémy Prego via orca-list
> > wrote:
> > > hello joanie?
> > >
> > > I didn't want to do anything else while smplayer was closing so
> > > as
> > > not
> > > to overload the debug.
> > >
> > > on the other hand I can make a debug which shows the problem by
> > > reading
> > > a text in pluma for example
> > >
> > > could this help to understand if there is a problem?
> > >
> > > Jerem
> > > Le 21/04/2022 à 13:51, Joanmarie Diggs a écrit :
> > > > Hey Jerem.
> > > >
> > > > Looking at your debug.out, the very slowest thing I see in
> > > > terms of
> > > > the
> > > > work Orca is doing is when you quit Orca. That took 0.15
> > > > seconds.
> > > >
> > > > I see a mini flood which I assume is coming from the app
> > > > destroying
> > > > itself, but Orca is quickly ignoring all of those.
> > > >
> > > > The other, more relevant, thing I'm seeing however is that
> > > > around
> > > > two
> > > > seconds are passing between the time I assume smplayer closed
> > > > and
> > > > the
> > > > time Orca receives the aforementioned mini flood. During that
> > > > time
> > > > Orca
> > > > is logging nothing. That means there were no new accessible
> > > > events
> > > > from
> > > > AT-SPI2 and Orca had already processed all the events it had
> > > > received.
> > > > So it looks like the freeze in question is outside of Orca and
> > > > beyond
> > > > Orca's control. If this is a normal Qt app behavior, I'm
> > > > guessing
> > > > it's
> > > > either in the Qt AT-SPI2 implementation (they bypassed ATK) or
> > > > in
> > > > AT-
> > > > SPI2 itself.
> > > >
> > > > Sorry!
> > > > --joanie
> > > >
> > > > On Sun, 2022-04-17 at 02:39 +0200, Jérémy Prego via orca-list
> > > > wrote:
> > > > > hello,
> > > > >
> > > > > when a qt application (smplayer in my example) closes, orca
> > > > > freezes
> > > > > for
> > > > > 2-3 seconds
> > > > >
> > > > > I don't know if this behavior is recent and applies to other
> > > > > applications than smplayer, but I think so.
> > > > >
> > > > > I took the time to do an orca debug showing the problem
> > > > >
> > > > > https://www.dropbox.com/s/g8rqqwo93aoaxir/debug-2022-04-16-21%3A45%3A14.out?dl=1
> > > > >
> > > > > ideally, I know that the problem should be corrected at the
> > > > > source,
> > > > > but
> > > > > I hope that we can do something on the orca side to correct
> > > > > this
> > > > > little
> > > > > inconvenience :)
> > > > >
> > > > > thanks for the help,
> > > > >
> > > > > Jerem
> > > > > _______________________________________________
> > > > > orca-list mailing list
> > > > > orca-list gnome org
> > > > > https://mail.gnome.org/mailman/listinfo/orca-list
> > > > > Orca wiki: https://wiki.gnome.org/Projects/Orca
> > > > > Orca documentation: https://help.gnome.org/users/orca/stable/
> > > > > GNOME Universal Access guide:
> > > > > https://help.gnome.org/users/gnome-help/stable/a11y.html
> > > _______________________________________________
> > > orca-list mailing list
> > > orca-list gnome org
> > > https://mail.gnome.org/mailman/listinfo/orca-list
> > > Orca wiki: https://wiki.gnome.org/Projects/Orca
> > > Orca documentation: https://help.gnome.org/users/orca/stable/
> > > GNOME Universal Access guide:
> > > https://help.gnome.org/users/gnome-help/stable/a11y.html
>
> _______________________________________________
> orca-list mailing list
> orca-list gnome org
> https://mail.gnome.org/mailman/listinfo/orca-list
> Orca wiki: https://wiki.gnome.org/Projects/Orca
> Orca documentation: https://help.gnome.org/users/orca/stable/
> GNOME Universal Access guide:
> https://help.gnome.org/users/gnome-help/stable/a11y.html