Re: simple rotation support for gdk-pixbuf

[Matthias Clasen <mclasen redhat com>]

Am So, den 20.06.2004 schrieb muppet um 13:42:
> On Jun 20, 2004, at 7:53 AM, John Cupitt wrote:
> 
> > On 20 Jun 2004 10:27:48 +0200, Sven Neumann <sven gimp org> wrote:
> >> or similar to avoid the problem of how to interpret the angles?
> >
> > Even that confuses me :-( does it mean the top of the image moving to
> > the right, or the bottom? Image processing textbooks always have
> > increasing angles meaning anticlockwise rotation (hope I've got that
> > the right way around), I'd say that was the safest model.
> 
> given your own uncertainty, i think you proved the "don't use the 
> numbers" argument.
> 
> what about
> 
>      GDK_PIXBUF_ROTATE_NONE
>      GDK_PIXBUF_ROTATE_CLOCKWISE
>      GDK_PIXBUF_ROTATE_UPSIDEDOWN
>      GDK_PIXBUF_ROTATE_COUNTERCLOCKWISE
> 

Yes, thats the alternative I would have proposed myself, although I it
has to be 

GDK_PIXBUF_ROTATE_NONE
GDK_PIXBUF_ROTATE_COUNTERCLOCKWISE
GDK_PIXBUF_ROTATE_UPSIDEDOWN
GDK_PIXBUF_ROTATE_CLOCKWISE

if you want to go with increasing numbers...

I also noted that both the api and the implementation of my _flip()
proposal become simpler if it is changed to 

typedef enum {
  GDK_PIXBUF_FLIP_HORIZONTAL,
  GDK_PIXBUF_FLIP_VERTICAL
} GdkPixbufFlip;

GdkPixbuf *gdk_pixbuf_flip (const GdkPixbuf *src,
		            GdkPixbufFlip   *flip)

One could consider adding 
GDK_PIXBUF_FLIP_ASCENDING_DIAGONAL
GDK_PIXBUF_FLIP_DESCENDING_DIAGONAL

Matthias