> On Nov. 19, 2011, 3:50 p.m., Christian Mollekopf wrote:
> > You might want to filter the collections out using: 
> > 
> > > Akonadi::EntityMimeTypeFilterModel *collectionsModel = new 
> > > konadi::EntityMimeTypeFilterModel(this);
> > > collectionsModel->addMimeTypeExclusionFilter( 
> > > Akonadi::Collection::mimeType() );
> > 
> > Otherwise you have the model flattened, with the collections still in.
> > No idea if it is relevant for this case, but with the 
> > EntityMimeTypeFilterModel you would get exactly the same result as an 
> > invisible collection fetch.
> 
> David Narváez wrote:
>     I thought of adding the filter, but then I noticed that in this case the 
> KDescendantsProxyFilter is flattening a CalendarModel which is, in turn, 
> constructed from a Monitor which is a mime-filtered model itself (see the 
> first context lines of the relevant hunk, which are dealing with the monitor 
> before using it in the calendar). Is that enough for a mime filter as you 
> describe it?
> 
> Christian Mollekopf wrote:
>     I assume your referring to the setMimeTypeMonitored calls on the monitor. 
> This means the monitor is configured to only include items of the configured 
> mimetypes, however they're still in the collection hierarchy. So collections 
> are not mimetype checked (since they're only the containers). Therefore you 
> should still have the collections in the flattened model.
>     So filtering the collections afterwards seems cleaner to me, although the 
> Calendar might work just fine with empty collections in the list.

Yes, I was referring to that. Let me update the patch to include MIME 
filtering. Thanks!


- David


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/103088/#review8314
-----------------------------------------------------------


On Nov. 10, 2011, 6:12 a.m., David Narváez wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/103088/
> -----------------------------------------------------------
> 
> (Updated Nov. 10, 2011, 6:12 a.m.)
> 
> 
> Review request for Plasma and Christian Mollekopf.
> 
> 
> Description
> -------
> 
> As the InvisibleCollectionFetch fetch strategy is currently broken, I 
> implemented a suggestion by Christian Mollekopf of working around the issues 
> by flatting the model through a KDescendantsProxyModel to fix crashes in 
> Plasma related to calendars.
> 
> 
> This addresses bug 279322.
>     http://bugs.kde.org/show_bug.cgi?id=279322
> 
> 
> Diffs
> -----
> 
>   plasma/generic/dataengines/calendar/calendarengine.cpp 53f98c0 
> 
> Diff: http://git.reviewboard.kde.org/r/103088/diff/diff
> 
> 
> Testing
> -------
> 
> 1. Add an iCal resource from the internet
> 2. Refresh the iCal resource
> 3. Refresh the iCal resource again if Plasma didn't crash on step 2
> 
> Withouth this patch, Plasma crashes when trying to remove an item that has no 
> collection in the Calendar Model embedded in the Dataengine.
> 
> You should also be able to replicate bug 279322 without this patch and this 
> patch should fix the situation described there.
> 
> 
> Thanks,
> 
> David Narváez
> 
>

_______________________________________________
Plasma-devel mailing list
Plasma-devel@kde.org
https://mail.kde.org/mailman/listinfo/plasma-devel

Reply via email to