- From: Ian Hickson <ian@hixie.ch>
- Date: Tue, 2 Dec 2008 02:36:28 +0000 (UTC)
- To: Simon Pieters <simonp@opera.com>
- Cc: public-html <public-html@w3.org>
On Mon, 13 Oct 2008, Simon Pieters wrote: > > http://html5.org/tools/web-apps-tracker?from=925&to=926 > > In order to get the layout right, what is actually needed is a unit that > acts like 'em' but ignores ancestor h1-h6 elements when calculating the > computed style of font-size. Using 'rem' or similar isn't good enough > because pages can set the font-size on root to something that isn't used > for the actual text on the page, or some pages have a sidebar in a > smaller font, etc. > > So we've decided to do what webkit does. (Any element in between whose > end tag isn't implied by <h1> makes them nest.) Oops, the spec was intended to match what WebKit did here, but I missed the case of direct nesting. Fixed. On Mon, 13 Oct 2008, Simon Pieters wrote: > > What about <h1><p><font><h2>x > > I'm tempted to say that the <h2> should cause the imply-end-tag-for-p > rules to be followed, so that the h1 is at the top of the stack (or is > it bottom?), and so the tree would be > > <h1> > <p> > <font> > <h2> > <font> > "x" > > But I'm not sure. Yeah, that's what the spec says now (because it checks for h1-h6 after closing <p>s). > PS. It seems that in webkit, <p><font><h2>x is parsed into > > <p> > <font> > <h2> > "x" > > and <p><font><p>x is parsed into > > <p> > <font> > <font> > <p> > "x" > > which causes less elements in the tree than the spec for > <p><u><u><u><p>x<p>x<p>x... Yeah, but compare: <h1><p><font color=navy><h2>x</font>y <h1><span><p><font color=navy><h2>x</font>y There are always edge cases where one algorithm beats another; the spec beats webkit's (less cloning) in the case of inlines being opened outside blocks and closed inside blocks, as in: <p><i>...<p>...</i> -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 2 December 2008 02:37:06 UTC