- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Thu, 2 Aug 2012 10:00:20 -0700
- To: robert@ocallahan.org
- Cc: Simon Fraser <smfr@me.com>, Rik Cabanier <cabanier@gmail.com>, www-style list <www-style@w3.org>
On Thu, Aug 2, 2012 at 2:42 AM, Robert O'Callahan <robert@ocallahan.org> wrote: > On Thu, Aug 2, 2012 at 5:49 AM, Simon Fraser <smfr@me.com> wrote: >> However, I wonder if this doesn't conflict with our earlier decision that >> when you render the element(), it's as if you're starting painting from the >> root, but only painting the target element and its descendants. Things like >> filters etc. will create stacking context. So if the element() target has an >> ancestor with a filter, we paint its positioned descendants in their normal >> painting order (affected by the fact that the filter creates stacking >> context), but we don't actually apply the effect of the filter. > > Good point. > > What we actually do in Gecko is treat the element target as a stacking > context when rendering its contents for element(). Ah, that's an important detail. In discussion with James Robinson, we were wondering about the element being forced into a stacking context. We were also wondering if we should take this farther, and actually require the element to be a stacking context in reality, as it might let us reuse existing rendering more often. ~TJ
Received on Thursday, 2 August 2012 17:01:11 UTC