- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Thu, 12 Feb 2009 16:47:09 -0600
- To: robert@ocallahan.org
- Cc: Brad Kemper <brad.kemper@gmail.com>, David Hyatt <hyatt@apple.com>, fantasai <fantasai.lists@inkedblade.net>, "www-style@w3.org" <www-style@w3.org>
On Thu, Feb 12, 2009 at 4:44 PM, Robert O'Callahan <robert@ocallahan.org> wrote: > On Fri, Feb 13, 2009 at 11:24 AM, Tab Atkins Jr. <jackalmage@gmail.com> > wrote: >> >> On Thu, Feb 12, 2009 at 4:15 PM, Robert O'Callahan <robert@ocallahan.org> >> wrote: >> > So one question just came to mind: how do we deal with box-shadow >> > "spread" >> > if we're using the border-images to generate a non-rectangular mask? I >> > don't >> > see any reasonable option. >> >> I can't think of anything that wouldn't be computationally ridiculous. >> Would it be horrible to just ignore it? > > > You mean ignore "spread" if border-images are present? Slightly unclean, but > perhaps the best option. Yeah. However, could we achieve what we want by simply having each pixel inherit the highest alpha value of any pixel within X of it, with X depending on the spread? Negative spreads would inherit the lowest alpha instead. Obviously more expensive than ignoring it, but not as bad as, say, edge-detecting and trying to intelligently grow it. ~TJ
Received on Thursday, 12 February 2009 22:47:49 UTC