getting rid of qt3support in kde-workspace

[Boudewijn Rempt]

On Wednesday 25 May 2011 May, Albert Astals Cid wrote:
> A Wednesday, May 25, 2011, Boudewijn Rempt va escriure:
> > On Wed, 25 May 2011, Albert Astals Cid wrote:
> > > A Wednesday, May 25, 2011, Aaron J. Seigo va escriure:
> > >> hi :)
> > >> 
> > >> i just finished pushing a set of changes to the (ironically named?)
> > >> qt3support branch in kde-workspace which leaves no traces of qt3support
> > >> or kde3support left in that module. huzzah!
> > >> 
> > >> i'm ok with holding on to the branch until 4.8, but if there is interest
> > >> in getting this into 4.7 it should be a fairly safe merge and would
> > >> allow distributors to ship kde-workspace without a dependency on
> > >> qt3/kde3support.
> > > 
> > > Is there a real demand for this?
> > 
> > Yes, I'd say there's a real demand. qt3/kde3 support takes space and
> > qt3-support doesn't exist on OSX anymore, and is going to disappear
> > everywhere else. 
> 
> That's just facts, not saying who is demanding KDE 4.7 to have no qt3support.

Yes, it's a fact that on OS X there's no qt3support.

> 
> > After Qt 4.8, no Qt3 support anymore. 
> 
> Yes, but KDE 4.8 will come before Qt 5 does so that's a non issue too.

No, it is not a non-issue. It took many years to port from Qt3 to Qt4. It took me about a year to get rid of most qt3/kde3-support use in calligra. It's much better to start now and make code that uses kdelibs, like kde workspace, qt3/kde3-support free.

> > Best get it removed everywhere as soon as possible. 
> 
> Disagree, removing it for the sake of removing it, is just opening the can to 
> bugs for no real reason.

Well, and I obviously disagree with that. Get your stuff cleaned up as soon as possible makes a lot of sense to me.

> 
> > (As well as QWidget/QPainter code, too, of course.)
> 
> Hope you are really joking with this one.

No, I'm not joking. If QWidget/QPainter are declared "done" now, then with Qt6, they will be gone, just like qt3support will be gone in Qt5. And that will be in a relatively short period of time: five years or so. Which might sound like a long time, but isn't when porting complex, real world applications.

-- 
Boudewijn Rempt | http://www.valdyas.org, http://www.krita.org