playground/base/plasma/dataengines/akonadi

Sebastian Kügler sebas at kde.org
Wed Feb 11 10:55:35 CET 2009


On Tuesday 10 February 2009 07:55:57 David Baron wrote:
> > The loading of data in the dataengine is async. Don't block on it but
> > react to signals announcing new datasources instead. Look at lionmail.cpp
> > for an example.
>
> What if the initial get sources always returned something, in this case
> EmailCollections, ContactCollections which is a handful of items? The
> caller will always have something to test and can choose which if any to
> request.

I want to do this as async and on-demand as possible (as database access can 
potentially be slow, doing i/o and stuff when data is requested) and I don't 
like this at startup. (This could potentially make Plasma and thus KDE startup 
slower -- I hate that.) Other dataengines behave similar, btw.

Why would you do that other than debugging?

The issue really is documentation for dataengine. Maybe we should a source 
"Documentation" to all dataengines? It would explain how to use it, i.e. a bit 
of text that explains how to go from an empty plasmaengineexplorer...
-- 
sebas

 http://www.kde.org | http://vizZzion.org |  GPG Key ID: 9119 0EF9 



More information about the Plasma-devel mailing list