Re: [Evolution] serial 138770 error_code 8 request_code 59 minor_code 0

[Adam Tauno Williams <awilliam whitemice org>]

On Thu, 2012-12-27 at 00:46 +0000, Pete Biggs wrote:
> > > > The program 'evolution' received an X Window System error.
> > > > This probably reflects a bug in the program.
> > > > The error was 'BadMatch (invalid parameter attributes)'.
> > > > (Details: serial 138770 error_code 8 request_code 59 minor_code 0)
> > > >
> > > > What can/must I do ?
> > >
> > > What version of Evolution?  What desktop are you using? (i.e. Gnome,
> > > KDE, whatever)
> >
> > Evolution 2.32.2 on GNOME-Version 2.32.1 from Ubuntu 
> That's a very old version of Evolution. 

*VERY* old [where the * !&(*@(*@ do all these ancient users come from?
Really, it baffles me.  No updates for four+ years?  Really?]  I
understand lagging for stability and getting-work-done, but... multiple
years?

>  It is unlikely that the
> Evolution devs will fix a bug (if there is one) in that version.  If the
> problem corresponds to a bug fixed in a later version, then it would be
> up to the Ubuntu maintainers to back port the patch to the version they
> distribute.

If it worked previously and now doesn't - it is *not* an Evolution bug
[because your Evolution hasn't changed]. Not creating an X error.  This
is a bug in X, your video driver, Gtk, etc... especially if it
corresponds to scrolling; it is further down the stack than Evolution.
I'd wager some library got updated [gasp!] to a slightly incompatible
version, your video hardware is dying, or the user [admin?] mucked up
the libraries by forcing something or making a 'manual' change.

Do not discount your video hardware dying [or somehow becoming
unstable].  I've twice been able to crash a workstation by doing things
on the screen with particular applications - once because the video card
installed was "not supported" and once because a RAM stick was dying.
Replacing the *physical* components resolved the 'bug'.  Why those
particular actions invoked the doom?  Way above my pay grade,  but I
went to college for Electronics Manufacturing and know enough about the
insanity of bus timing on that scale to have no problem believing it.

> > > Does Evolution crash completely? Is the crash reproducible?
> > <<<<<<<<<<<<<<<<^^^^^^^^^^^^^^^^ YES ! Haven't tried it yet, 'coz I'm
> > happy with a running system. But I remember what I was doing before the
> > crash was only rapidly scrolling down in a large mail ?:-|
> A one off crash on an X call from an unknown part of the code is not
> going to be easy to get to the bottom of. The only way to deal with it
> is if the crash is reproducible and you can get a debugger back trace of
> the crashed code.

-- 
Adam Tauno Williams  GPG D95ED383
Systems Administrator, Python Developer, LPI / NCLA