Re: [Bug-gnubg] Did I just mess up?

[Joern Thyssen]

On Mon, Jul 28, 2003 at 05:16:01PM +0200, Holger wrote
> At 14:53 28.07.2003 +0000, Joern Thyssen wrote:
> >On Mon, Jul 28, 2003 at 04:10:42PM +0200, Holger wrote
> >> But on the other hand: in <glib/gutils.h> both
> >> g_path_get_dirname/g_path_get_basename are defined. Why not use these? Or
> >> are there really systems left without glib?
> >
> >I don't know, but the answer is probably yes. On one hand is nice to
> 
> OK, more specifically: ... any platform that we want gnubg to compile on.
> The answer may still be yes. But as long as I don't see a good 
> counter-example I suppose this must be rather exotic.

The target platforms for GNU Backgammon is GNU/Hurd and GNU/Linux, and I
assume that glib is present on those (since the very first glib).

Candidates for a missing glib is AIX, Sun, IRIX, and other commercial unix
flavours, but like windows they are not among the target platforms.

I admit it would be nice to be able to use glib...

> >ensure that gnubg compiles with any plain vanilla ANSI C compiler on a
> >minimal system, but on the other hand, it would be nice to use some of
> >the glib stuff, for example:
> 
> So the question arises: Do we want to support the very last system (with 
> only the no-gui version), or  bear the chance that gnubg doesn't compile on 
> everything that consists of silicon.
> At least for the GUI version there's no need to hesitate - GTK+ relies on 
> glib anyway.

Yes, I use lots of glib functions in the GTK+ code I write.

Jørn