Re: [Pan-users] Latest pan 100% CPU usage

[Gerald Livingston]

On Fri, 26 Apr 2013 16:52:31 +0000 (UTC)
Bob <> wrote:

> On Fri, 26 Apr 2013 08:09:16 +0000, Duncan wrote:
> 
> > The last time people were reporting 100% cpu, it was traced to some
> > sort of conflict with the gnome accessibility stuff.  As I don't
> > run gnome I've never seen the problem, so this is all second hand
> > to me, but the solution back then was to turn off whatever the
> > gnome accessibility daemon was called (a11y ring a bell?), before
> > trying to run pan.
> > 
> > I had thought someone finally traced down that problem using gdb
> > and a bugfix had been committed, but it can't hurt to see if it's
> > running and try turning it off if so, and see if that fixes pan.

> Bob wrote:
> 
> Hi Duncan...
> 
> Thinking about what you mentioned, I checked, and I don't seem to have
> a11y anywhere on the system.  

Old message from the list. Still no name for the daemon:

--------------------------------------------------------------------
From: Heinrich Mueller <>
To: address@hidden
Subject: Re: [Pan-users] Gnome and Ubuntu foks, some good pan news =:^)
Date: Thu, 01 Dec 2011 08:50:16 +0100
Reply-To: address@hidden, address@hidden
Sender: address@hidden
User-Agent: Mozilla/5.0 (X11; Linux x86_64;
 rv:8.0) Gecko/20111108 Thunderbird/8.0

Am 01.12.2011 08:10, schrieb Duncan:
> For some years now, pan has been known not to get along well with the
> gnome accessibility daemon (IDR what it's called, exactly), and ever
> since we found out that was the problem, the recommendation has been
> to turn it off, at least when running pan.  But while that's the gnome
> accessibility app, it seems nearly all reported issues with it have
> been on ubuntu.  I could speculate that ubuntu activates
> accessibility by default, thus all the problem reports from there,
> but it's pure speculation.
>
> Anyway, it seems someone finally traced down the root problem -- it
> has been a known issue with one of the gtk widgets (treeview),
> apparently, there's a non-pan (rythmbox) bug on it too -- and has
> devised a workaround patch for the problem.
>
> The judgefudge/hmueller repo has the patch in it now, after pkovar ask
> him to test it, but I'm not sure whether it's in khaley or the
> official gnome repo (pkovar's bailiwick) yet, but as I said, pkovar's
> aware of it, so it should hopefully hit with the next official
> release. =:^)
>
> Here's the judgefudge git commit header; the comment being just a
> link to the bug with the discussion and patch, which basically sets
> the treeview widget that's triggering the issue to null while
> updating, thus avoiding all that traffic over corba/dbus to the
> accessibility daemon (a11y, based on the bugs):
>
> commit 8e092ff555a6a156534d151d3fc1fccd05fe66b1
> Author: Heinrich Müller<address@hidden>
> Date:   Mon Nov 28 12:43:05 2011 +0100
>
>      * https://bugzilla.gnome.org/show_bug.cgi?id=660030 (Jindrich
> Makovicka)
>
>  
Thanks for pointing that out.
I had independently implemented this rather common fix for other
treeviews in pan already but didn't know about this one, so it really
helped a lot. It isn't in any other repo yet.
------------------------------------------------------------------

-- 
I am on the email list. Unless you are sending information which should
not be shared with everyone else please reply to the list, not directly
to me.