[cairo] Problem with stretched images

[Ed Voas]

One note about this as I was just thinking about it...

In my findings, it was only affecting the bilinear stretch, not the 
nearest neighbor one. If what is mentioned below is true, then it should 
have happened to both equally, no?

Ed Voas wrote:

> On Sep 5, 2006, at 11:54 AM, Carl Worth wrote:
>
>> On Tue, 5 Sep 2006 11:12:02 -0700, Ed Voas wrote:
>>
>>> I briefly searched the archives, but I'm seeing an issue where images
>>> that are stretched don't handle the boundary cases properly. As a
>>> result, the edges of the image are effectively blended with a zero
>>> pixel, causing the edges to fade. I found the issue inside pixman and
>>> fixed it for my case, but I'm wondering if this is already known and
>>> fixed (I am using 1.10 at present, I believe), or if not, how can I
>>> get the fix back into cairo/pixman itself.
>>
>>
>> This is a known, open issue. The latest conversation on the topic can
>> be seen in the following messages:
>>
>> http://lists.freedesktop.org/archives/cairo/2006-August/007672.html
>> http://lists.freedesktop.org/archives/cairo/2006-August/007673.html
>>
>> Under the current model, the behavior you are seeing is "correct" for
>> the default CAIRO_EXTEND_NONE semantics for surface patterns, (where
>> all pixels "outside" the surface are transparent). With this model the
>> approach to get the behavior you want is to use CAIRO_EXTEND_PAD
>> instead, (but there's a bug in that this doesn't actually work
>> yet---oops!).
>>
>> So we need to get the surface PAD stuff fixed.
>>
>> Meanwhile, in the above message, Vlad proposes changing the model to
>> get this behavior by default in order to avoid confusion such as what
>> you have hit here. And I think the principle of least surprise makes a
>> good argument that direction.
>
>
> I would agree with that. And good to finally understand why it's  
> happening (it all finally makes sense). Thanks.
>
> -- Ed
>
> _______________________________________________
> cairo mailing list
> cairo at cairographics.org
> http://cairographics.org/cgi-bin/mailman/listinfo/cairo