Re: [orca-list] orca completely stopps functioning when i launch thunderbird

[Joanmarie Diggs <jdiggs igalia com>]

Hey Majid.

Taking a closer look, there seems to actually be two different problems.

Problem 1: There are periods in your debug.out where nothing happens 
(four-second gaps at the beginning, later a series of two-second gaps). 
During this time, Orca is being told exactly nothing. So I don't know if 
you were just sitting there not using your computer, or if your computer 
was busy because Thunderbird was doing something and not telling AT-SPI2 
anything, or if AT-SPI2 was busy and not telling Orca anything. There is 
nothing I can do about this in Orca and, in answer to your question, I 
wouldn't know where to file it or what to complain about. "Some app or 
library is doing something bad" is not a good bug report. ;) Because I'm 
quite busy, I don't have time to try to figure out what the app is or 
what it's doing. If someone else with debugging chops (Thunderbird, 
AT-SPI2, DBus) has time to look into that, it would be awesome.

Problem 2: I am seeing a ton of spam from what I assume is the following 
scenario: You don't automatically filter your messages into folders and 
have a lot of messages coming into your inbox each time you launch 
Thunderbird. As a result, Orca is getting two events each time your 
folder goes from being, say, Inbox (4000) to Inbox (4001) to Inbox 
(4002) to Inbox (4003) to .... So if you have 100 new messages, Orca may 
wind up getting 200 new events.

Those events are not wrong. And if they went away entirely, Orca would 
have stale names and potentially tell you the wrong information. Thus 
stale names are bad. So if I were to file a bug against Thunderbird, it 
would be something to the effect of: Hold off on the name changes until 
we have the final new message in that particular folder. The thing is, 
knowing when they have the final new message *might* be tricky. Or not. 
I dunno. Regardless, if you want to file that bug against Thunderbird, 
that would be great. What's super important, however, is that they don't 
stop emitting name changes entirely. If they do, it will potentially be 
a major regression. :(

In the meantime, I just made a change in Orca master to ignore those 
events. Orca will still be spammed and a delay will thus still be 
present, but it should immediately discard them rather than discarding 
them after a tenth of a second. Given the number, maybe it will make the 
situation suck less until when (and if) the problem can be addressed in 
Thunderbird itself. Please test and let me know.

Thanks!
--joanie

On 9/6/19 11:50 AM, Majid Hussain wrote:
> hi there,
> my bad,
> was in a hurry sorry about that! :)
> are thunderbird aware of the bugg? e.g is the bug present or will one
> require fileing?
>
> Majid :)
>
>
> On 06/09/2019, Joanmarie Diggs <jdiggs igalia com> wrote:
> > Hi Majid.
> >
> > So I think it's the same issue we're aware of and which probably needs
> > to be fixed in Thunderbird so that it's not spamming AT-SPI2 and/or Orca.
> >
> > On a related note, you've been CCing orcalist rather than orca-list. As
> > a result, those messages are bouncing.
> >
> > --joanie
> >
> > On 9/6/19 11:42 AM, Majid Hussain wrote:
> > > hi there,
> > > i'm now running orca v 3.32.0, and this issue is present  so it
> > > happens in master and on this latest stable release.
> > > and happens exactly the same as on master :(
> > > Majid
> > >
> > > On 05/09/2019, Joanmarie Diggs <jdiggs igalia com> wrote:
> > > > Hi Majid.
> > > >
> > > > When there are lots of messages in a folder in Thunderbird and you arrow
> > > > to that folder, Thunderbird creates lots of accessible objects and
> > > > notifies AT-SPI2 about them. Orca tries to filter out the event spam as
> > > > best it can.....
> > > >
> > > > What's important to me, due to the impending stable release is to know
> > > > whether or not this is a regression in Orca (i.e. the problem only
> > > > exists in master, but not in Orca 3.32.). Is that the case?
> > > >
> > > > --joanie
> > > >
> > > > On 9/5/19 10:28 PM, Majid Hussain wrote:
> > > > > hi there,
> > > > > I doo have alot of messages in my inbox could that be the issue?
> > > > > is there no way of working around this issue?
> > > > > is this a orca issue?
> > > > > thanks,
> > > > > Majid Hussain
> > > > >
> > > > > On 05/09/2019, Joanmarie Diggs <jdiggs igalia com> wrote:
> > > > > > Do you have a huge number of messages in your inbox? Is this a
> > > > > > regression in Orca? I don't see anything fatal jumping out at me in
> > > > > > your
> > > > > > debug.out, nor am I able to reproduce the problem using Orca (master)
> > > > > > and Thunderbird (nightly).
> > > > > >
> > > > > > --joanie
> > > > > >
> > > > > > On 9/5/19 9:22 PM, Majid Hussain wrote:
> > > > > > > hi again,
> > > > > > > attached is the debug file,
> > > > > > > left it for a few minutes,
> > > > > > > file is quite small,
> > > > > > > 1 mb decompressed.
> > > > > > > to be presise, this issue happens when i arrow to my inbox complete
> > > > > > > freeze,
> > > > > > > hope this helps
> > > > > > > Majid
> > > > > > >
> > > > > > > On 05/09/2019, Majid Hussain <mhussaincov93 gmail com> wrote:
> > > > > > > > hi there,
> > > > > > > > when I launch thunderbird, orca completely goes caputt.
> > > > > > > > my set-up for thunderbird is pop3 non standerd i'm aware,
> > > > > > > > everything is normal apart from pop3 being used.
> > > > > > > > not sure what  the issue could be, i'll send a d-bug .out when i've
> > > > > > > > sent this message, clozed all other windows and relaunched orca with
> > > > > > > > the d-bug switch enabled.
> > > > > > > > i will have to reboot my machine after, because it completely stopps
> > > > > > > > working i'll keep the orca and thunderbird going for a while so you
> > > > > > > > are able to get what you can.
> > > > > > > > thanks,
> > > > > > > > sorry for the long email!
> > > > > > > > p.s,
> > > > > > > > running the very,very latest orca from master.
> > > > > > > > Majid
> > > > > > > >
> > > > > > > > --
> > > > > > > > kind regards,
> > > > > > > > Majid Hussain