Re: GEP 7 - icons and thumbnails - gnome_icon_loader
Alexander Larsson <[email protected]> Mon, 7 Oct 2002 04:06:26 -0400 (EDT)
| Newsgroups | gmane.comp.gnome.lib.devel |
|---|---|
| Message-ID | <Pine.LNX.4.44.0210070404170.31799-100000@devserv.devel.redhat.com> |
On 3 Oct 2002, Michael Meeks wrote:
> Hi Alex,
>
> On Tue, 2002-10-01 at 09:59, Alexander Larsson wrote:
> > > Will you take a patch to fully abstract this ? I can't think of any
> > > good performance reason to have it like this.
> >
> > Gladly.
>
> So; I ran into a sticky API mess at the nautilus scaling; since the
> attach points rather rely on scaling; would you be adverse to having
> something like:
>
> GnomeIconDataPoint *
> gnome_icon_data_get_attach_points (const GnomeIconData *icon_data,
> int *n_attach_points,
> double scale_x,
> double scale_y);
Mmm. sounds fine.
> Which created a copy that we could keep around tied to the cached icon
> data ?
>
> Failing that I'd prefer to return the attach points as an array of
> floating point proportional offsets into the icon, since that would
> avoid copying and scaling them around the place, leaving a unified
> source of point data.
>
> Oh, and since I can't commit without this; I also binned the
> x0,x1,y0,y1 thing in favour of a GnomeIconRectangle typedef - which
> saves typing for the accessors / headers, and improves the contract
> clarity [ to my mind ], is that ok ?
Yupp. I'd like a way to know whether there was a embedding rect specified
at all too.
> And finally, it looks like create_embedded_text_cache_icon should be
> doing a gnome_icon_data_copy instead of the struct copy it currently
> does.
Very much so.
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Alexander Larsson Red Hat, Inc
[email protected] [email protected]
He's an uncontrollable crooked vagrant who must take medication to keep him
sane. She's a plucky extravagent lawyer descended from a line of powerful
witches. They fight crime!