Re: gmmproc parsing problem in cluttermm

[Kjell Ahlstedt <kjell ahlstedt bredband net>]

Den 2016-01-11 kl. 12:16, skrev Kjell Ahlstedt:
> Den 2016-01-08 kl. 00:06, skrev Ian Martin:
> > Hi,
> > I've finally got some time to relook at the cluttermm bindings 
> > again.  I have a problem with regenerating the defs files that I 
> > think is probably due to some funky code in clutter: the clutter API 
> > uses inheritance in the C code, but in a very C sort of way. For 
> > example, this creates a new ClutterBindConstraint, but returns a 
> > ClutterConstraint * :
> >
> > > ClutterConstraint * clutter_bind_constraint_new            
> > > (ClutterActor *source,
> > > ClutterBindCoordinate  coordinate,
> > > gfloat                 offset);
> >
> > which I think is why when h2def.py parses the code all the 
> > ClutterBindConstraint methods are parsed as ( global ) functions, not 
> > methods.
> >
> > I'm OK with having to manually go through the defs file and correct 
> > it all, but would like to know first that I'm not missing something?
> >
> > Ian
> > _______________________________________________
> I tested to generate cluttermm/clutter/src/clutter_methods.defs. 
> Something in clutter-bind-constraint.h confuses h2def.py. Most of 
> what's generated from that file is wrong. Hopefully h2def.py can be 
> fixed, so it can handle all clutter files. I'll see what I can do. 
> Manually fixing the .defs file should be a last resort.
>
> Kjell
Fixed with this glibmm patch:
https://git.gnome.org/browse/glibmm/commit/?id=379d93938569d8a63dfe7c6c7304a7738eb10aeb

Kjell