Re: [PATCH 2/3] tests: added a small test program for local-metadata

[Iago Toral <itoral igalia com>]

On Thu, 27 Jan 2011 12:51:09 +0100, Guillaume Emont <gemont igalia com> 
wrote:
> On 27/01/2011 09:17, Iago Toral wrote:
> > On Wed, 26 Jan 2011 19:32:27 +0100, Guillaume Emont
> > <gemont igalia com> wrote:
> > (...)
> > > +static GrlMedia *
> > > +make_grl_image (const gchar *path)
> > > +{
> > > +  GFile *file;
> > > +  GrlMedia *media;
> > > +  gchar *full_path, *uri;
> > > +
> > > +
> > > +  file = g_file_new_for_path (path);
> > > +
> > > +  media = grl_media_image_new ();
> > > +
> > > +  full_path = g_file_get_path (file);
> > > +  grl_media_set_id (media, full_path);
> > > +  g_free (full_path);
> > > +
> > > +  uri = g_file_get_uri (file);
> > > +  grl_media_set_url (media, uri);
> > > +  g_free (uri);
> > > +
> > > +  return media;
> > > +}
> >
> > Not that it is very important, but just in case you missed it, you
> > could've used this instead for creating the media object:
> >
> > GrlMedia *
> > grl_media_source_get_media_from_uri_sync (GrlMediaSource *source,
> >                                           const gchar *uri,
> >                                           const GList *keys,
> >                                           GrlMetadataResolutionFlags
> > flags,
> >                                           GError **error)
> >
> > basically, it does the same thing without you having to set the 
> > media
> > id directly in the GrlMedia in the client code.
> Hmm, that call needs a GrlMediaSource. I guess I could easily get a 
> hold
> of the filesystem source for that, but I'd rather have the test of a
> given plugin not depend on another plugin, or even on its 
> existence...

Sure, I think that's reasonable.

Iago