Re: GObject-Introspection 0.5.0

[Murray Cumming <murrayc murrayc com>]

On Mon, 2008-09-08 at 10:09 -0500, Mike Kestner wrote:
> On Sun, 2008-09-07 at 17:39 +0200, Johan Dahlin wrote:
> > Paul Pogonyshev wrote:
> > > Johan Dahlin wrote:
> > >> [...]
> > >>
> > >  > I'm leaning towards using the "ownership" terminology instead of "transfer". 
> > >> typedef enum {
> > >>    GI_OWNERSHIP_CALLER, /* caller owns it, caller should free it after use */
> > >>    GI_OWNERSHIP_CALLEE  /* callee owns it, caller should leave it as it is */
> > >> } GITypeOwnership;
> > > 
> > > Just as a nitpick, these two names look very similar and quite confusing
> > > for non-native English speakers.  Maybe you could come up with something
> > > different, especially in place of 'callee'?
> > 
> > Agreed, I'm open to suggestions.
> > caller/subroutine ?
> 
> In GAPI, we have two ownership attributes on list elements: owned and
> elements_owned.

or nothing is owned. Yes, we have these 3 in gtkmm too. Johan's system
offers a fourth combination, list-not-owned, element-owned, so-far not
seen in the wild), but it seems fine.

>   It's either true or false, and it is from the
> perspective of the recipient of the object.  Thus, an owned list in a
> return value is owned by the caller.  An owned list in a method "in"
> parameter is owned by the function.
> 
-- 
murrayc murrayc com
www.murrayc.com
www.openismus.com