- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Thu, 18 Aug 2011 14:58:58 -0700
- To: Cameron McCormack <cam@mcc.id.au>
- Cc: www-svg <www-svg@w3.org>
Argh, I keep missing these calls! They're on my calendar now, so I should get next week. On Thu, Aug 18, 2011 at 2:34 PM, Cameron McCormack <cam@mcc.id.au> wrote: > radial gradients My approach for degenerate cases has been to specify the limit image, so the behavior is as continuous as possible. I consider any behavior that is sensitive to rounding errors a bug. For example, if a repeating gradient approaches 0 distance between first and last stops, it will assume the average color on an infinite-resolution device. Thus, on our finite resolution devices, when the distance is small enough that we can't faithfully render it any longer, we should just render the average color. I have similar cases for repeating radial gradients that approach zero width or zero height. In this case, when the focus approaches the edge of the shape, the image converges to a half-plane of circles, and a half-plane of average color. ~TJ
Received on Thursday, 18 August 2011 21:59:45 UTC