Separate Plasma::DataSource and data engines
Marco Martin
notmart at gmail.com
Mon Jan 19 10:28:41 UTC 2015
On Monday 19 January 2015, Pier Luigi Fiorini wrote:
> Hi,
>
> I would like to use Plasma::DataSource and the data engine plugins in
> my project because it suits my needs very well.
>
> However it introduce large dependencies (plasma-framework and
> plasma-workspace).
>
> Would you consider moving the code into a framework pretty much like
> what you did with KPackage?
It's something that I would kinda like, however there are some problems with
it.
Splitting KPackage (thing that is still actually far from done, there is still
a lot of porting to do) has been a thing that taken a really long time.
Literally few days to do the actual split, months and months to make sure
there aren't too many behavior chages, that everything is perfectly
retrocompatible, and porting every user.
I think it was worth it since it's of global interest and relatively small.
On dataengines, personally I would love that as well, but on one hand, they
are more complicated than kpackage, therefore it would take even more time
(that I'm not sure i have in the next months among the other stuff to do..)
On the other hand I'm not sure it would be really worth it, because I see in
general people don't like them, I don't see people wanting to write new ones
(as opposed to a qml import) even for problems where it would match.
Personally I still think they are the way to go for certain kind of problems
(my deep dislike for qml import proliferation I think it's well known ;) since
they offer out of the box all the asyncness that like it or not, is simply
necessary in a gui project, that in a qml import would have be to be either
reimplemented from scratch or (usually) just ignored producing fun freezes.
On an ideal world, I would make them in a separate framework, redesigning some
parts that i think can be improved and all..
But anyways, at the moment (I'm always open to change idea) I don't have much
motivation to undergo a massive porting effort on something that very few
would ever want to use.
--
Marco Martin
More information about the Plasma-devel
mailing list