[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-pim
Subject:    Re: [Kde-pim] Review Request: Fix for Control::start() not
From:       "Volker Krause" <vkrause () kde ! org>
Date:       2010-06-07 7:38:37
Message-ID: 20100607073837.11822.69540 () localhost
[Download RAW message or body]


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviewboard.kde.org/r/4250/#review6004
-----------------------------------------------------------

Ship it!


Looks good, trunk already has the reloading on service owner changed logic but lacks \
the re-creation of the interface. Should be ported there as well I guess.

- Volker


On 2010-06-06 23:29:39, Kevin Krammer wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviewboard.kde.org/r/4250/
> -----------------------------------------------------------
> 
> (Updated 2010-06-06 23:29:39)
> 
> 
> Review request for KDE PIM.
> 
> 
> Summary
> -------
> 
> The observed problem is that when KMail starts, it attempts to start Akonadi, but \
> after a timeout tells the user that it failed because Akonadi couldn't find any \
> resources. 
> Some distributions have increased the timeout to give Akonadi more time to complete \
> startup, but the problem is that ServerManager is never signalling server startup \
> at all. 
> ServerManager does check the server's status on several occasions, however, one of \
> the checks is asking AgentManager for the agent types and verifying that at least \
> one of them is a resource. Unfortunately AgentManager is always empty, because its \
> D-Bus interface gets created right in AgentManager's contructor, which is invoked \
> from ServerManager's constructor, which means there hasn't even been an attempt of \
> starting yet. 
> So the D-Bus interface is invalid and QtDBus doesn't make it valid at a later time. \
> None of the slots connected to the interface's signals get ever called, the \
> AgentManager instance remains empty. Which mean ServerManager can never satisfy its \
> "operational" check and Control always runs into its timeout. 
> The patch makes AgentManager check the validity of the interface and if it is not, \
> waits for the control D-Bus name to be registered before recreating th e interface \
> and rechecking for registered agent types. 
> 
> This addresses bug 236538.
> https://bugs.kde.org/show_bug.cgi?id=236538
> 
> 
> Diffs
> -----
> 
> /branches/KDE/4.4/kdepimlibs/akonadi/agentmanager.h 1135240 
> /branches/KDE/4.4/kdepimlibs/akonadi/agentmanager.cpp 1135240 
> /branches/KDE/4.4/kdepimlibs/akonadi/agentmanager_p.h 1135240 
> 
> Diff: http://reviewboard.kde.org/r/4250/diff
> 
> 
> Testing
> -------
> 
> Applied the changed to the Debian source package for kdepimlibs and installed the \
> modified libakonadi-kde package to test with the otherwise unchanged KDE 4.4 setup \
> of Debian unstable. 
> 
> Thanks,
> 
> Kevin
> 
> 

_______________________________________________
KDE PIM mailing list kde-pim@kde.org
https://mail.kde.org/mailman/listinfo/kde-pim
KDE PIM home page at http://pim.kde.org/


[prev in list] [next in list] [prev in thread] [next in thread]