Current status of kdebindings

[Stefano Crocco]

On lunedì 25 marzo 2024 04:11:14 CET Mathieu Jobin wrote:
> Reviving this old thread.
> 
> Now that Plasma 6 is out. Would there be enough interests to see QtRuby
> being brought back to life for Qt6 ?
> 
> I did not get a chance to test the port to Qt5, anybody can comment on how
> well it worked?
> 
> If we could make it easier on people to install it and start off. I'm
> willing to put some of my time, if there is a team or interested people to
> work with.
> 
> cheers,


Unfortunately, my attempt to document and improve the old bindings was 
unsuccessful: I realized that I was able to understand what a single function 
did, but I couldn't see how everything fit together. In the end, I gave up. The 
bindings themselves worked well enough, at least until I switched to ruby 3.1. 
From that version of ruby, they practically stopped working because creating a 
second object of any class produced an error complaining about a missing 
allocator (or something like that). Just a few days ago, I had to update all 
my system and realized that they don't even build correctly with the current 
version of Qt5, for reasons I can't understand.

I'd love for QtRuby to be brought back, but I'm not sure whether starting from 
the old bindings is the best way to proceed. I was wandering whether using 
Swig [1] wouldn't be a better idea. I tried creating wrappers for a few 
methods of simple classes like QPoint, QSize and QRect and they were very easy 
to do. Things become more difficult for QObject derived classes, as you have to 
take care of memory management and so on, not to mention signal-slot 
connections, which I didn't try to do.

In case there's a group working on this, I can lend a hand, too. 
Unfortunately, until the end of June I won't have much free time, so my help 
will be limited. Maybe, it would be a good idea to look for other interested 
people on the ruby mailing list.

Stefano

[1] https://www.swig.org/