Re: [orca-list] slow tree views

[Mike Gorse <mgorse alum wpi edu>]

Hi Peter,

I just looked into this briefly--it seems that orca is being flooded with 
object:property-change:accessible-parent events. It's a separate issue 
from my last commit and from bgo#650090 (now fixed). I'm not sure who is 
to blame--need to investigate more.

On Fri, 28 Oct 2016, Peter Vágner wrote:

> Hello,
>
> I don't understand this matter at all I think, however with Thunderbird and 
> at-spi2-atk master installed some two days ago I can see no improvements or 
> they might be barelly noticeable in the day to day usage.
> However I'm not yet running at-spi2-core from git, so perhavs I do also have 
> to update it as well.
>
> Greetings
>
> Peter
>
>
>
> D?a 28. októbra 2016 11:58:01 dopoludnia používate? Krishnakant 
> <krmane openmailbox org> napísal:
>
> > On Wednesday 26 October 2016 10:32 PM, Peter Vágner wrote:
> > > Hello,
> > >
> > >
> > > Thunderbird message list is originally an XUL control at the mozilla
> > > side of things. However at the orca side of things it's also a table.
> > >
> > > Might your fix help with this issue no one can shed more light on?
> > >
> > > https://bugzilla.mozilla.org/show_bug.cgi?id=924915#c24
> > >
> > >
> > > Greetings
> > >
> > >
> > > Peter
> >
> > I have been looking closely at Orca source code off late and trying to
> > study it.
> > As Peter rightly said, that's exactly the case and for me, the Orca side
> > is pretty much neutral on this.
> > So my guess is that this would help a lot in speeding up things, unless
> > TB sends some thing very specific for screen readers to make an exception.
> > Happy hacking.
> > Krishnakant.
> >
> > > On 26.10.2016 at 17:54 Mike Gorse wrote:
> > > > Hi all,
> > > >
> > > > While trying to fix evolution slowness with very large address books
> > > > for my day job, I noticed that orca was also getting bogged down. I'm
> > > > wondering if this has affected anyone else, with orca appearing to
> > > > hang when focusing on a large tree view.
> > > >
> > > > There was a bug with AT-SPI's caching of roles, and orca thought that
> > > > the tree view was just a panel and proceeded to poke through what
> > > > might be thousands of cells. On the gtk side, the atk role was
> > > > originally set to ATK_ROLE_PANEL and later set to ATK_ROLE_TABLE. An
> > > > event was fired, but the event was not being sent over dbus so that
> > > > orca would see it.
> > > >
> > > > I've fixed this in the master branch of at-spi2-atk, for anyone
> > > > pulling from git. I'm debating whether to push this for 3.22.2
> > > > (although the code now works as intended, I'm slightly worried that
> > > > it might cause slowness elsewhere through event floods).
> > > >
> > > > If any of you test this, then I'd appreciate knowing whether it makes
> > > > a difference either way.
> > > >
> > > > Thanks,
> > > > -Mike
> > > > _______________________________________________
> > > > 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
> > > > Log bugs and feature requests at http://bugzilla.gnome.org
> > >
> > > _______________________________________________
> > > 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
> > > Log bugs and feature requests at http://bugzilla.gnome.org