KDE Frameworks 5 Switch, When, how, what?

Milian Wolff mail at milianw.de
Mon Nov 11 09:45:11 UTC 2013

On Monday 11 November 2013 01:13:27 Aleix Pol wrote:
> On Sat, Nov 9, 2013 at 1:04 PM, Christoph Cullmann 
<cullmann at absint.com>wrote:
> > Hi,
> > 
> > KDE Frameworks 5 gets more traction in the last months and perhaps we
> > should plan ahead how
> > we want to go the 5.x road.
> > 
> > We have a lot of "cleanups" and "fixups" listed for the KTextEditor
> > interfaces we want to do for
> > the 5.x release and until now I think we have no clear plan when to go for
> > this.
> > 
> > I think Kate/KatePart are at the moment in a relative good state for the
> > 4.x line and not that much stuff is missing
> > that we actually need to implement in 4.x (beside perhaps some plugin
> > enhancements or bugfixes).
> > 
> > Perhaps we should draft some small timeline for the future when we want to
> > go to switch the main focus
> > on porting + cleaning for 5.0.
> > 
> > We should coordinate a bit with the KDevelop guys, not sure how their
> > plans for 5.x are, either.
> > 
> > I would propose for Kate, to create some "frameworks" branch in December,
> > in which the idea is to
> > 
> > a) first get it compiling with KDE 5 Frameworks + Qt 5.2
> > b) then cleanup and fix the stuff in the interfaces
> > 
> > This "frameworks" branch would have no BC/SC requirements until we arrive
> > in the 5.0 Beta phase I think.
> > 
> > "master" should stay KDE 4.x until the last 4.xx release is done and then
> > we can switch that over.
> > 
> > All stuff that goes into master after the frameworks branch is there must
> > be synced to frameworks by
> > the people doing the change, there won't be an auto-merge by me or someone
> > else I think.
> > 
> > This is just a bit brainstorming to kick this topic off, comments /
> > changes are welcome :P
> Hi,
> Good thing to see that you're interested in the porting. I would suggest to
> create a frameworks branch in kdelibs (just like kde-workspaces and kdelibs
> did) and just start working on it.

Probably Aleix meant to create a frameworks branch in kate ;-) I agree with 
that. Starting early wouldn't hurt.

> As for KDevelop, we're quite depending on Kate, so I guess more than
> coordinating, we'll have to work on the Kate side if you guys don't start
> before.

Yes. I hope to help a bit here and there, but can't give any promises as I 
need to continue hacking on KDevelop :) I hope to have added KDE5 notes where 
ever I spotted any issues in the past though. Also there is some 
readme/porting text file in the sources afaik. Can't find it now though.

> Regarding dependencies, I'd just go forward and whenever we see the result
> we end up, we can decide to cut them up. Either way, Just depending on KIO
> and XmlGui makes us depend on many many many things.

Yes that is true. But personally I wouldn't say its a big loss. Porting Kate 
stuff to KF5 more or less as-is first would give us the nice text editor, 
deeply integrated into KDE. Once we have that, we can think about cutting down 
on dependencies.

The idea of having a "second" interface sounds nice to me. I.e. some more low-
level libraries providing a "text editor widget" library and maybe separately 
a "text highlighting" library. 

To achieve this though, one would need to refactor the KTextEditor interfaces 
to not depend on KParts but be "pure" interfaces. Not sure how nice this will 
work out code-wise. KIO could then be used optionally if available. Actions 
could be provided by another interface (similar to actionCollection()).

Anyhow, just my $0.02. Cheers!
Milian Wolff
mail at milianw.de

More information about the KDevelop-devel mailing list