Bug 256969 - Adding a second Imap ressource fails
Summary: Adding a second Imap ressource fails
Status: VERIFIED FIXED
Alias: None
Product: KDE PIM Mobile
Classification: Miscellaneous
Component: general (show other bugs)
Version: unspecified
Platform: Windows CE Microsoft Windows CE
: HI major
Target Milestone: ---
Assignee: kdepim bugs
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-11-15 12:55 UTC by Andre Heinecke
Modified: 2011-02-15 16:33 UTC (History)
5 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Andre Heinecke 2010-11-15 12:55:19 UTC
I've tried this serveral times and was never able to archive two imap accounts on Windows CE.
When adding a second Kolab account the Accountwizard fails with the message:

Creating Ressource instance for IMAP-E-Mail-Server ...
Agent instance creation timed out
Comment 1 Till Adam 2011-01-11 16:00:45 UTC
For the record, this works fine on Maemo.
Comment 2 Bernhard E. Reiter 2011-01-17 17:07:31 UTC
We should get up to five accounts (identifies, transports) working, because
often group accounts will be fully entered.
Comment 3 Andre Heinecke 2011-01-18 17:35:56 UTC
Added accounts are shown as offline in akonadiconsole but with the right id (akonadi_imap_resourceXY)
Comment 4 Andre Heinecke 2011-01-19 17:03:44 UTC
What does not work:
The akonadi_imap_resource can be started several times if in the main it does only sleep, after the sleep when the eventloop is hit all but one terminate with return code 0.

Loading the application with the wincetools loader does not make a difference in the behavior.

If a second imap resource is started directly without the correct command line parameters it terminates even before the output that the command line parameters are missing is shown.

If a second imap resource is started with the accountwizard or by akonadiconsole it get's loaded the process is created but no output is shown and as soon as all modules are loaded they are unloaded again.

What does work:
Applications that don't use qt. ( At least tested with very simple ones but it does not appear to be an Operating System limitation )

Test case (akonadi stack is running in the background):
Two simple Qt Hello World widgets http://paste.kde.org/2836

-> Start two times
- two widgets are shown
-> wait 0-60 seconds 
- The application that was already running terminates with return code 0

Test case (akonadi stack is not running in the background)
Same Application
-> Start two times
- two widgets are shown
-> wait >3 minutes 
- both widgets are still shown
-> wildly click on the device start new processes
- both processes are still running
-> Start kleopatra (as a big qt application)
- both processes are still running
-> start akonadictl
- after a minute or so one process terminates
Comment 5 Andre Heinecke 2011-01-20 13:26:17 UTC
Additional Info:
The difference between both processes open and one closing seems not to be akonadi, i've had two processes running while akonadiserver was running but when i started the 28th process the problems begun.

QEventLoop debug output when a test process is terminating:
http://paste.kde.org/2891/

It also seems to have nothing to do with the application name as soon as i open any 28th process one of my test applications terminates.
So this test case leads to unrelated information.
Comment 6 Andreas Holzammer 2011-01-20 17:07:58 UTC
should be fixed with svn-1215972