[Bf-committers] Cursor proposal and OSX Cursor size
LarstiQ
bf-committers@blender.org
Mon, 1 Dec 2003 17:06:21 +0100
On Mon, Dec 01, 2003 at 09:37:50AM -0500, Robert Wenzlaff wrote:
<snip>
> If only I had a mac to test it on...
Do you personally need a mac, or someone who can test things for you?
> > Robert will come with a further proposal to mailing list later.
>
> ************* Proposal ******************
>
> To keep it simple for now, we'll keep the small cursor rule at 16x16, and only
> raise that limit when/if we get OSX working at 32x32. After that, the 16x16
> rule gets degraded to a "moderately strict guideline".
>
> That said, the rest of the proposal is:
>
> 1) All cursors will become custom cursors so they match across platforms.
> 2) Cursors will have a large and small version selectable by a userpref.
> This userpref will be disabled on platforms that cannot support large
> cursors. The data will consist of a bitmap and transparency mask, as it
> does now. (My MakeCursor.py is available to paint Cursors. Can this be
> put on Blender.org?).
We have a source/tools directory in cvs (and intern/tools), is that an
ok place? Or perhaps a more visible 'tools' section on the webpages
(along with a python script repository!, www.pygame.org/pcr/repository.php for reference)
> 3) Cursors will have definable fg and bg colors.
> 4) Cursors will have definable hot-spots.
Definable per cursor, or a global setting the user can switch? Hmm,
perhaps not that useful.
> 5) Each cursor will be a member in a global array of structs defined in
> cursors.c
> 6) Each cursor will have a define in cursor.h that is its index number in
> the array.
> 7) Cursor setting API code will go in cursor.c And will be implimented with
> the existing GHOST custom cursor calls (modified to handle the size
> change).
> 8) API will be:
> GetBlenderCursor() - returns defined number of current cursor.
> SetBlenderCursor(int i) - sets cursor to the ith cursor.
> Defines will be set so all cursors can be refered to by name.
> 9) Current cursor API will be phased out, but will remain active during
> changeover so things don't get broken.
Looks good to me.
LarstiQ