Review Request 122065: Fix activity start and stop.

Xuetian Weng wengxt at gmail.com
Sat Jan 24 06:01:22 UTC 2015



> On Jan. 23, 2015, 7:40 a.m., Martin Gräßlin wrote:
> > This looks still wrong to me. The service is still called "org.kde.kwin" while it should be "org.kde.KWin". Ideally this gets changed to a generated adaptor from the DBus interface KWin installs.
> > 
> > To make things worse: the service name is not guaranteed to be "org.kde.KWin" - it can be mangled on multi-head systems. I had considered eporting it to a root window property, maybe I should implement that.
> 
> Martin Gräßlin wrote:
>     And just implemented the announcment of the DBus service: https://git.reviewboard.kde.org/r/122215/
> 
> Ivan Čukić wrote:
>     How will this work?
>     
>     KAMD is a unique instance per dbus session, kwin can have more instances. Which instance handles the session stuff? etc.
>     
>     I don't like the idea of introducing another connection from kamd to X.
> 
> Martin Gräßlin wrote:
>     If KAMD doesn't want to support multi-head (which is fine) it should pick the kwin instance responsible for the X head it's on.
> 
> Ivan Čukić wrote:
>     Sorry, I need more details here.
>     
>     What confuses me is the 'X head it's on' - it is a service, it has no UI to be on a specific head. Or am I misunderstanding what a 'head' is? :)
>     
>     Can you describe (nothing detailed - just a few words) the situation when kwin has multiple instances / multiple dbus services. Does it imply separate X sessions?
> 
> Martin Gräßlin wrote:
>     yes in a multi-head environment one has kind of separate X sessions. E.g. you cannot move windows from one head to another and the windows cannot interact with each other. It's a rare situation and most of our daemons do not support it properly.
> 
> Xuetian Weng wrote:
>     I must misread the qdbus autocompletion, so org.kde.kwin is fine. And it seems to be guaranteed that default screen of kwin uses org.kde.kwin. (others will use org.kde.kwin-screen-XX)

Ah, sorry. The patch is to fix the wrong interface name, not the dbus service name.

I think using org.kde.kwin is fine unless this service name will be removed from kwin. Because it's the dbus name for most common default screen if KAMD don't care about multihead.

And If two different kwin are in same dbus-session, they will talk to same ksmserver and close session twice?.. It seems there are still many missing pieces in support activities with multihead.

Most people are moving to xinerema so it's really a rare case IMHO...


- Xuetian


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/122065/#review74574
-----------------------------------------------------------


On Jan. 22, 2015, 11:42 p.m., Xuetian Weng wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://git.reviewboard.kde.org/r/122065/
> -----------------------------------------------------------
> 
> (Updated Jan. 22, 2015, 11:42 p.m.)
> 
> 
> Review request for Plasma, Martin Gräßlin and Ivan Čukić.
> 
> 
> Repository: kactivities
> 
> 
> Description
> -------
> 
> 1. KWin Interface name is wrong
> 2. kactivitymanagerd doesn't listen on ksmserver anymore, thus subSession{Opened,Closed,CloseCanceled} are not called anymore. It need to send Started/Stopped event by itself when kwin dbus call returns.
> 
> 
> Diffs
> -----
> 
>   src/service/ksmserver/KSMServer_p.h c0f5598 
>   src/service/ksmserver/KSMServer.cpp b5e1467 
> 
> Diff: https://git.reviewboard.kde.org/r/122065/diff/
> 
> 
> Testing
> -------
> 
> Now starting and stopping activity works properly.
> 
> 
> Thanks,
> 
> Xuetian Weng
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20150124/b9a6902c/attachment.html>


More information about the Plasma-devel mailing list