W3C home > Mailing lists > Public > www-style@w3.org > October 2012

Re: [css3-break] Fragment content displayed outside fragmentainer boundary

From: Alan Stearns <stearns@adobe.com>
Date: Mon, 15 Oct 2012 13:30:38 -0700
To: Alan Stearns <stearns@adobe.com>, "www-style@w3.org" <www-style@w3.org>
CC: "hyatt@apple.com" <hyatt@apple.com>
Message-ID: <CCA1BE6C.17D40%stearns@adobe.com>
On 9/13/12 3:58 PM, "Alan Stearns" <stearns@adobe.com> wrote:

>I think we should have a section in css3-break (perhaps 5.4?) that defines
>what happens when content in a particular fragment is displayed outside
>the fragmentainer boundary. Two instances of this are relatively
>positioned elements and line boxes at the top of the fragmentainer with a
>line-height insufficient to enclose the content. The test file in a recent
>css3-regions thread [1] shows non-interoperable behavior with multicol
>elements across all browsers.
>I think some of the differences are just bugs. But what happens with
>relatively-positioned elements appears to be a result of two solutions
>arrived at separately because there is no spec that says what should
>happen. Since this is shared behavior between columns and regions, I think
>it belongs in css3-break.
>The question on relative positioning is whether the offsets apply relative
>to the fragmentainer or the fragmented flow. In the former solution, a
>line box shifted down at the bottom of a fragmentainer would appear below
>the fragmentainer box. In the latter solution the line box would appear at
>the top of the next fragmentainer in the fragmentation context. I think
>that the former solution is preferable.
>Here's my initial stab at some text to add for relative positioning:
>  Once the fragment contents for a particular fragmentainer are
>  any offsets from relative positioning are applied relative to the
>  of the content in the fragmentainer, not the position of the content in
>  the fragmented flow.
>[1] http://lists.w3.org/Archives/Public/www-style/2012Sep/0268.html

Here's another (perhaps more relevant) example of the problem. Say you are
using relative positioning to get a particular superscript effect and the
line with this effect happens to land at the top of a fragmentainer.
Currently there is no specification that says what happens to the
'sup-sup' glyphs in the example below. This has led to two separate
interpretations. I'd like to have a single correct result defined in

<!DOCTYPE html>
  <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  <title>Relpos Superscript in Columns</title>
    div {
      -moz-column-count: 2;
      -webkit-column-count: 2;
    .sup-sup {
      top: -10px;
    <p>relative pos superscript</p>
    <p>10<sup>10</sup><sup class="sup-sup">10</sup></p>


Received on Monday, 15 October 2012 20:31:05 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:22 UTC