Re: [Evolution] Slow closing Evolution...?

[Milan Crha <mcrha redhat com>]

On Fri, 2012-05-11 at 16:16 +0100, Steve T wrote:
> Sorry about the delay. I'm not sure either whether the backtrace is
> correct as I had to download the debug-info stuff today, and I think
> Evolution was already running when I installed them. So if the
> attached isn't detailed enough, then let me know and I'll redo it
> again tomorrow.

        Hi,
thanks, the backtrace looks fine. One thread is synchronizing all your
Search Folders, while another thread is storing summary information for
one of those search folders. I guess the CPU usage is high during this
time, maybe for whole 20 minutes. This all depends on the number of
search folders you've defined. The other threads are basically idle,
thus the long time is only about search folders and their
synchronization (the current "camel_folder_refresh_info_sync"
implementation in search folders seems to rebuild itself, which means to
clear itself and refill from scratch - it seems to me that this refresh
call is not needed on search folders when going to quit, on the first
look).

I'm not aware of any workaround for this. It would be good if you could
file this in bugzilla [1], thus it'll not be forgotten. If you could
attach there two backtraces, one after those 30 seconds, another say
after 5 minutes of "inactivity", and confirm or falsify the high CPU
usage, with a note of how many search folders with approximately how
many folders in them you have, and what those real folders are from (On
This Computer, IMAP, ...) and an approximate message count in folders.
(The backtrace also shows that it's trying to update Unmatched search
folder).

Please do not forget to post here the bug link, for a reference.
        Thanks and bye,
        Milan

[1] https://bugzilla.gnome.org/enter_bug.cgi?product=Evolution