crashes when accepting meeting invitation

Thank you for taking the time to report this bug and helping to make Ubuntu better. Please try to obtain a backtrace following the instructions at http://wiki.ubuntu.com/DebuggingProgramCrash and upload the backtrace (as an attachment) to the bug report. This will greatly help us in tracking down your problem.

Sorry, I think I've made an error where I say it crashes. It doesn't crash, it hangs and has to be killed.

I'll try to have some time to analyze it with valgrind's tool callgrind to see where it hangs.

Hi,

I cannot take anything from valgrind :(

This is really a showstopper for me as I cannot approve meetings nor can I manage my calendar week.

why can't you get a valgrind log?

I can get a valgrind cachegrind report, what I can't is understand it from Kcachegrind. I cannot say with certainty where evolution is blocking.

Although it now blocs on almost all meeting invitations I receive :(

Running it in debug mode, shows no debug information at all as no error message is written to the console. Can it be that the problem isn't in evolution itself but in the evolution-data-server not returning the calendar information to the front end? I say this because all front ends (if more than one is open) I have to "killall evolution" to bring it back.

Ok, I seem to have something using GDB to backtrace where it hanged, or at least I'm hopping so.

This is the backtrace for:

> opening evolution
> selecting the email message from the meeting
> clicking "Accept" the meeting invitation

The problem blocked and I Ctrl+C and bt this from GDB

Hope this helps

does the program crash or hang? this bug starts being really confusing, those are different things ...

I know, sorry for that, I've changed the title to a proper (and correct) one.

It hangs evolution until I run "killall evolution" or "evolution --force-shutdown".

the stacktrace would suggest that evolution-data-server is crashing, could you see if that's the issue?

So it seems.

In annex is the GDB backtrace for evolution-data-server-2.6.

ok, that's useful thanks, that shows a libical assertion, do you get that one any meeting invitation? do you think you can add an email example there? do you know how to rebuild a debug package, having libical0 in debug version would be useful there

Hi Sebastien.

You have the email example in the original bug report attachment.

I have multiple situations resulting in the client hang:
1) Opening a recursive meeting invitation will hang evolution immediately after opening the mail;
2) Accepting a single meeting invitation;

I can try to download the code for libical0 compile it with -g but I cannot affect the system integrity. If you could point some more info on how, I'll can help.

I'm still finding my way using .deb packaging (former archlinux user).

Hi again,

I've managed to download from packages.ubuntu.com the source code for the current jaunty version of libical. Compiled with CFLAGS=-g -O0 moved the generated libical, libicalss and libicalvcal to /usr/lib replaced each symbolic link to use the new build, and it worked! :)

Can it be some issue with the library compilation?

I doubt of that, could be an optimization issue, the usual way to do a debug build is:
- add a deb-src apt source
- apt-source libical
- sudo apt-get build-dep libical
- cd libical-version
- DEB_BUILD_OPTIONS="nostrip noopt" debuild

you might not want "noopt" if that workaround the issue

The bug has been sent on http://bugzilla.gnome.org/show_bug.cgi?id=579176

Do you still need me to recompile using apt-source or the current information is enough to solve the issue?

upstream doesn't get the issue using the calendar event you attached there, do you get the bug using that directly and how? or do you get it only on an email which has the calendar entry? having a debug stacktrace would still be useful

upstream has a patch candidate, do you think you could rebuild an evolution-data-server package using the change to verify if it fixes the issue?

Sure, I'll try. Although it may take a bit, depend on workload...

Sorry if this is stupid, but: where do I find that patch candidate and how do I apply it?

the GNOME bug http://bugzilla.gnome.org/attachment.cgi?id=132771&action=view has the change

Alright, is there a standard procedure for building a package from source and applying that patch?

(Sorry, but my Ubuntu-Development-Fu isn't that strong.)

Sorry for my blindness. I found the note above and was able to interpret it. I'm building right now.

The patch worked and calendar has been stable for me since then. I tested it by moving my archived copy of ~/.evolution/calendars/local back into the evolution directory.

Great to hear.

I haven't got the time right now to also test it... big release coming up :(

I'm still hanging with the workaround in libical for now.

On Wed, 2009-04-22 at 09:59 +0000, gothicknight wrote:
> Great to hear.
>
> I haven't got the time right now to also test it... big release coming
> up :(
>
> I'm still hanging with the workaround in libical for now.

The patch was simple to install given the instructions above. Took me
only about 2 minutes. :)

--
Christian Theune · <email address hidden>
gocept gmbh & co. kg · forsterstraße 29 · 06112 halle (saale) · germany
http://gocept.com · tel +49 345 1229889 7 · fax +49 345 1229889 1
Zope and Plone consulting and development

Accepted evolution-data-server into jaunty-proposed-proposed; please test and give feedback here. Please see https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

Hi Martin

I cannot find the new .deb package in repositories. Can you post the URL for it?

the upload has just been accepted, it usually takes some hour to get it build and available to download

Hi all,

I've upgraded all packages to jaunty-proposed, including e-d-s, and unfortunately the issue is still there :(

did you restart your session? are you sure you got the new version? could you get an update stacktrace?

Yes I did update to jaunty-proposed and force shutdown all evolution's front and back-end, even rebooted to grant no preaload was being done.

Will the GDB backtrace be enough?

the new stacktrace is a different crash, I've commented on the GNOME bug

Ok, the libical recompilation still works, I'll stick with that workaround for now.

could somebody confirm that the original crash has been fixed so the update can be copied to jaunty updates?

I had a different case of evolution hanging which was also fixed by upgrading to this jaunty-proposed version.

In my case, clicking on Calendar view would hang. It started two days ago, and continued through yesterday and today. Quitting and restarting evolution didn't help avoid the hang. Rebooting didn't avoid the hang. This new version seems to have fixed it. Basically this meant I had no access to my calendar at all, which was quite frustrating.

It is possible that the event that triggered the bug on my system was receiving a meeting request that rescheduled an existing one-time meeting. Not sure, though.

Anyway, thanks for making the patched version available and providing easy instructions (and mechanisms) for me to try it without having to wait for the official upgrade.

Upgrading the the proposed version fixed the issue of Evolution hanging when I click on Calendar too

the change fixing the issue would be rather bug #368508 could people confirm that?

Upgrading fixed issue for me as well

This bug was fixed in the package evolution-data-server - 2.26.1-0ubuntu2

---------------
evolution-data-server (2.26.1-0ubuntu2) jaunty-proposed; urgency=low

  * debian/patches/90_upstream_change_fix_calendar_crasher.patch:
    - upstream change to fix a calendar crasher leading to evolution hanging
      (lp: #357636)

 -- Sebastien Bacher <email address hidden> Thu, 23 Apr 2009 11:57:36 +0200

Copied to karmic