[cairo] Alpha masks

[Owen Taylor]

On Thu, 2005-01-06 at 11:51 -0800, Keith Packard wrote:
> Around 13 o'clock on Jan 6, Carl Worth wrote:
> 
> > Right. But I'm thinking about "display image at 50% opacity". Having
> > to create a pattern and a mask for this operation would make it just
> > as hard as the original complaint about drawing a solid color through
> > a mask.
> 
> As cairo_set_rgb_color creates a source pattern, should cairo_set_alpha 
> create a mask pattern?  Should we expose this 'mask pattern' as a part of 
> the drawing pipeline?
> 
> 	dst = (src IN (path IN mask IN CLIP)) OP dst
> 
> This has the minor advantage of clarifying the relationship between 
> cairo_set_rgb_color and cairo_set_alpha as to whether the rgb values 
> should be premultiplied.

I think in terms of the original requested operation (basically, use
an A8 surface in place of a polygon or set of glyphs), this doesn't
fit very will.

In order to allow the alpha value to affect images/patterns with
alpha, it needs to replace 'mask' in the above, rather than being
the 4th component of 'src'. That's why show_surface_mask() should
replace 'path' in the above.

(Replacing 'path' also makes it very natural to do path arithmetic
on a temporary surface and then use the result as a path-like object)

Given the ability to replace 'path' in the above with an image, I
don't think it hurts to have a generalized mask rather than an
alpha value as I had it. But I don't think it's that useful either.

Regards,
					Owen

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.freedesktop.org/archives/cairo/attachments/20050106/3b2ff2cf/attachment.pgp