- From: Andrew Fedoniouk <news@terrainformatica.com>
- Date: Mon, 26 Nov 2012 10:51:43 -0800
- To: "L. David Baron" <dbaron@dbaron.org>
- Cc: www-style@w3.org
On Mon, Nov 26, 2012 at 12:14 AM, L. David Baron <dbaron@dbaron.org> wrote: > On Sunday 2012-11-25 23:21 -0800, Andrew Fedoniouk wrote: >> This sample: >> >> https://dl.dropbox.com/u/14981836/line-height-test.htm >> >> contains two identical paragraphs, the only difference is that >> first one has line-height:normal and second one - line-height:1em; >> >> According to the spec [1]: >> 'line-height' specifies the *minimal* height of line boxes within >> the element. >> ... >> >> I suspect that I miss something between lines of the spec. >> My understanding of the spec wording is this: >> >> used-line-height = max( {normal-line-height}, {defined-line-height} ); >> > > (1) when you're testing line-height, you should really test > standards mode rather than quirks mode; quirks mode behavior is > substantially different (and not fully explained by the first two > items in http://quirks.spec.whatwg.org/#css ). > > (2) 'normal' and '1em' are *very* different in terms of how they > inherit when the font size changes (since '1em' inherits as the > computed value, which is no longer relative to the changed > descendent font size). 'normal' and '1' are much more similar. > > -David > Thanks, David. I've updated the sample with HTML5 doctype. That actually makes no difference (only IE started to match all other UAs here). In fact I've missed that part in the spec: "On a non-replaced inline element, 'line-height' specifies the height that is used in the calculation of the line box height." So for such elements (like <big> in my sample) line-height defines not "min-line-height" but just 'line-height' allowing to reduce it. Together with the fact that line-height is an inheritable property that creates such strange effect. Actually it is a bug in spec design: if property is inheritable then it shall not change meaning on child (min-line-height to line-height transformation). -- Andrew Fedoniouk. http://terrainformatica.com
Received on Monday, 26 November 2012 18:52:11 UTC