[Kde-bindings] Re: py(Qt|KDE) and KDE CVS

[Marcus]

On Monday 22 March 2004 9:09 pm, Simon Edwards wrote:

> I'll try to explain the reasoning for wanting to have the python bindings
> distributed separately from a monolithic kde-bindings tar ball.

The Qt# project distributes a copy of QtC from outside of kdebindings is that 
the configure script for kdebindings seems to insist upon a working KDE 
installation, and many people who wanted to try Qt#, which has QtC as a 
dependency, were not KDE users and did not want to install KDE just to try 
Qt#.


> Distro support for kdebindings appears to be spotty. Mandrake don't package
> it. FreeBSD don't either. By wanting the python bindings tarballs to be
> distributed as one self contained 'blob' with dependancies only on python
> and C++, it would be much much easier to for people and packagers to build
> and package.

One problem with kdebindings is that the various components are in different 
states of readiness. This is not meant as a slight toward anyone, of course. 


> Conectiva 9 have kdebindings-c-devel, kdebindings-csharp-devel,
> kdebindings-ruby-devel etc packages. And I think it makes sense for
> bindings to be distributed not as one large tarball, but on a language by
> language basis.

As a practical matter, some distributors might not want to inflate the number 
of packages from 1 (kdebindings) to 8 (one packager per binding/component). 
On the other hand, if I just want to play with COBOLQt, why should I have to 
install AdaKDE? (It's too bad that the packaging schemes don't permit 
selectively installing subpackages...)