Re: [Evolution] bus error causing crashes

[Adam Tauno Williams <awilliam whitemice org>]

On Sat, 2012-12-15 at 07:11 +0000, Pete Biggs wrote:
> > > using Debian Sid/unstable with Evolution 3.4.4 (but happened since
> > > 3.4.x), wanting to reply to some message, Evolution often crashes due to
> > > a Âbus errorÂ. Does somebody experience something similar?
> > I think that is a euphamism for a segmentation fault.
> No, they are different things - a seg fault is when a program attempts
> to access a memory segment it doesn't own and a bus error is when a
> program attempts to access memory beyond the capabilities of the
> machine.  Both are caused by similar things, such as dereferencing
> invalid pointers, but bus errors sometimes point to hardware problems.

Does 3.4.x do something different regarding caches / databases than 3.6?
I don't show folders.db, etc... as memory mapped for my running 3.6
[fuser will tell you if an open file is memory mapped or not].

The only memory mapped 'data' files I see are:
  /home/awilliam/.local/share/gvfs-metadata/home-897372f8.log (deleted)
  /home/awilliam/.local/share/gvfs-metadata/home (deleted)
  /home/awilliam/.local/share/mime/mime.cache
  /home/awilliam/.config/dconf/user (deleted)
  /usr/share/icons/gnome/icon-theme.cache
  /var/cache/fontconfig/7ef2298fde41cc6eeb7af42e48b7d293-x86_64.cache-3
  /usr/share/mime/mime.cache
  /usr/lib64/gconv/gconv-modules.cache
  ... and a few font files?

See /proc/{evolutionPID}/maps

Will a backtrace record a SIG-BUS?  I don't know that I've ever had to
deal with a memory-map issue outside of a straight-up bug.

<http://www.whitemiceconsulting.com/2012/10/d-is-for-debugging.html>

-- 
Adam Tauno Williams <awilliam whitemice org>