[glade--]Re: [gtkmm] Problem with gtkmm handling comboboxes from glade

[Christof Petig <christof petig-baender de>]

Murray Cumming wrote:

> On Fri, 2001-11-30 at 21:38, Daniel Elstner wrote:
> > Am Fre, 2001-11-30 um 18.25 schrieb John Bartelt:
> > > On Friday 30 November 2001 07:51 am, you wrote:
> > > > Ahh, now I remember that one.
> > > > Try if it works after removing the parentheses around items[].
> > >
> > > No, it doesn't help. I don't think it has to do with scope, but with
> > > constness.
> > >
> > > I had a reply from vadim vk crocodile org saying that he had the same problem
> > > and fixed it by running a perl script on the generated code to change " const
> > > char * const " to " const char * ". I've confirmed that this works.
> > >
> > > Now the question: which should be fixed, glademm or gtkmm?
> >
> > Certainly gtkmm.  I'll look at it.
> >
> > I think we can make this change without breaking the ABI since SArray
> > seems to be all inline anyway.  Murray?
>
> I'd rather not risk breaking the ABI, or other code that uses SArray. If
> it can be fixed with a change to glademm then that would be better.

I will not remove a sensible const (and ruin the generated code forever) because RH
shipped gcc-2.96.

I would tolerate a --for-gcc-2-96 command line switch (which might try to
auto-detect the compiler used). Are there any RH fans around which step ahead to
implement this?

Perhaps a configure error is also a good idea, like "compiler does not accept valid
const declarations".

I would even accept fixing gtkmm/gcc by generating another template/inline function,
can you tell me how the fix looks like?

Furiously cursing the public existance of this compiler version
    Christof