[cairo] CAIRO_FORMAT_A8 / stride

[Nikolaj Thygesen]

    All the processors I know access single bytes at any address, so perhaps 
the 32-bit rule could be enforced only when needed?? I haven't thoroughly 
looked at the code, but the stride thing just makes 
cairo_image_surface_create_for_data() a bit too implementation dependent; at 
least if the idea was for the program to utilize any private buffer for 
rendering. As a minimum this limitation should be mentioned in the docs, but 
"fixing" it would be way better :o)

    br - Nikolaj Thygesen

>From: Keith Packard <keithp at keithp.com>
>To: Owen Taylor <otaylor at redhat.com>
>CC: cairo at cairographics.org, keithp at keithp.com
>Subject: Re: [cairo] CAIRO_FORMAT_A8 / stride
>Date: Sun, 02 Jul 2006 14:48:11 -0700
>
>On Sun, 2006-07-02 at 13:26 -0400, Owen Taylor wrote:
>
> > I don't quite understand what your problem is ... as far as I know,
> > Cairo will work fine with non-word-aligned rowstrides, though it
> > may be less efficient.
>
>no, pixman only handles 32-bit aligned strides for all rendering.
>
>Fixing that would be hard in general as the pixman code often accesses
>pixel data 32 bits at a time, and many processors don't like unaligned
>data accesses.
>
>keith.packard at intel.com


><< signature.asc >>




>_______________________________________________
>cairo mailing list
>cairo at cairographics.org
>http://cairographics.org/cgi-bin/mailman/listinfo/cairo

_________________________________________________________________
Opret en personlig blog og del dine billeder på MSN Spaces:  
http://spaces.msn.com/