[cairo] PDF Backend for cairo

[Kristian Høgsberg]

Carl Worth wrote:
> On Mon, 29 Nov 2004 14:59:09 -0500, =?ISO-8859-1?Q?Kristian_H=F8gsberg?= wrote:
> 
>>I've been working on a PDF backend for cairo over the weekend, and at 
>>this point I think it's appropriate post a snapshot of the current code. 
> 
> Fantastic! It's great to see more progress continuing on this front.

Well, here's a little more progress:

	- I found the problem with the "leak" in the polygons, I misunderstood 
the trapezoid format.
	- As described in my other mail, I changed the generated PDF to draw 
all trapezoids as subpaths of one big path and the fill that for each 
call to composite_trapezoids() which solves the problem with cracks in 
the polygons.
	- I added support for constant alpha values
	- I added support for clipping paths

I've uploaded a new patch with these changes:

   http://people.redhat.com/krh/cairo-pdf/cairo-pdf-2.patch

and there's an updated snippets.pdf available:

   http://people.redhat.com/krh/cairo-pdf/snippets.pdf

The clipping support requires a change to the backend API: I added a 
set_clip_trapezoids() call to cairo_surface_backend_t which works much 
like set_clip_region(), except it takes a list of trapezoids for the 
clipping area.  If a backend doesn't support this it must return 
CAIRO_INT_STATUS_UNSUPPORTED as usual.  This change isn't fully done 
yet, there's a few issues to work out still.

By the way, in _cairo_gstate_clip() line 1761, the rectangle case ends with:

         if (status != CAIRO_INT_STATUS_UNSUPPORTED) {
             _cairo_traps_fini (&traps);
         }
         return status;

Shouldn't this be

         if (status != CAIRO_INT_STATUS_UNSUPPORTED) {
             _cairo_traps_fini (&traps);
             return status;
         }

instead?

Oh, and another thing: it looks like the show_glyphs function pointer in 
the cairo_surface_backend_t struct can be NULL, in which case it is 
treated as returning CAIRO_INT_STATUS_NOT_SUPPORTED.  Is there a reason 
show_glyphs doesn't follow the convention that all function pointers 
must be non-NULL?

cheers,
Kristian