[cairo] Problem with stretched images

Bill Spitzak spitzak at d2.com
Tue Sep 5 15:06:57 PDT 2006


Thanks, I should have checked that.

What I would propose is that NONE be replaced with Vlad's proposal (do 
PAD but also intersect the mask with the transformed rectangle of the 
source image).

The current behavior of NONE can be replicated by making a source image 
with an extra row of black pixels on all sides, so I think it can safely 
be removed. Even if it is not removed, the new behavior should be the 
default, not the NONE behavior.

Carl Worth wrote:
> On Tue, 05 Sep 2006 13:22:08 -0700, Bill Spitzak wrote:
>> People are much more interested in tiling, and reflected tiling, than in
>>   the EXTEND_PAD behavior. If EXTEND_PAD really is needed it should be
>> merged with whatever tiling api there is.
> 
> It is already the case that this API is merged. The available extend
> modes are NONE, REPEAT, REFLECT, and PAD. Those apply to all
> patterns[*] so each needs semantics for gradients as well as for
> surface tiling.
> 
> -Carl
> 
> [*] Ignoring for the moment the bug in that PAD does not yet work with
> surface patterns.
> 
> --
> cworth at redhat.com


More information about the cairo mailing list