Re: [PATCH 04/15] core: introducing GrlOperationOptions

[Guillaume Emont <guijemont igalia com>]

Excerpts from Simon Pena's message of Wed Dec 07 19:42:11 +0100 2011:
> Hi,
> 
> Please find below some comments:
> 
> On 02/12/11 19:36, gemont igalia com wrote:
> [...]
> > +/* ========== API ========== */
> > +
> > +/**
> > + * grl_operation_options_new:
> > + * @caps: (allow-none): caps that options will have to match. If %NULL, all
> > + * options will be accepted.
> > + *
> > + * Creates a new GrlOperationOptions object.
> > + *
> > + * Returns: a new GrlOperationOptions instance.
> > + */
> 
> Are you missing here a "(transfer full):" annotation?
> 

Well, yes and no... That stuff is not much documented, but the current
behaviour of the parser is that the default transfer mode for the return values
of constructors is full. It's not guaranteed to remain like that as I think
there is no documentation on the subject, but I really see no sane reason to
change that behaviour. I am encouraged in that thought by the fact that the
scanner does not raise any warning for that.

I can add the annotation for clarity if you want, but I don't think it is
needed.