Re: [css-tables][css2.1] What should the measurement of the table be?

Le 2015-03-23 15:42, Robert Hogan a écrit :
> OK, so I'd like to propose the following as the behaviour to converge 
> on.
> 
> Case One:
> 
> There is nothing in CSS 2.1 that forces the cell to be anything greater
> than 1px in this test. While the size of the row and the table may be 
> up
> for grabs (as behaviour with percentage height is not defined) the cell 
> has
> a specified height and has no reason to be higher than 1 px. It is the
> containing block for its content, so the content has no reason to 
> exceed
> 1px either.

The height of table element, table-row and table-cell do not work like 
the height of block containers. The table box (not the anonymous table 
wrapper box) and table-row objects are not block containers. Overflow do 
not apply to table box and table-row.


> Gerard, in connection with this test you mentioned: “In your test, 
> because
> the table only has 1 row and 1 single cell, the single cell is 
> constrained
> to honor the 100% height declaration on the table element, the cell 
> height
> has to increase dramatically.” Why do you believe the cell is 
> constrained
> to ignore its 1px height in this case?
> 
> http://jsbin.com/wabileduma/1/edit?html,output

I'm sorry. This was more the result of a visual interpretation of what 
was rendered in a few browsers than the result of a solid spec reading 
and interpretation.


> <!DOCTYPE html>
> 
> <html>
> 
> <style>
> 
>    .content {
> 
>        background-color: green;
> 
>    }
> 
>    .table {
> 
>        display: table;
> 
>    }
> 
>    .row {
> 
>        display: table-row;
> 
>    }
> 
>    .cell {
> 
>        display: table-cell;
> 
>        height: 1px;
> 
>    }
> 
>    div, html, body {
> 
>        height: 100%;
> 
>        width: 100%;
> 
>    }
> 
> </style>
> 
> <body> <!-- Should be viewport height -->
> 
>    <div class="table"> <!-- Should be viewport height -->
> 
>        <div class="row"> <!-- Should be viewport height -->
> 
>            <div class="cell"> <!-- Should be 1px height -->
> 
>                <div class="content"> <!-- Should be 1px height -->
> 
>                </div>
> 
>            </div>
> 
>        </div>
> 
>    </div>
> 
> </body>
> 
> </html>
> 
> 
> Case Two:
> 
> As with Case One the container in the cell respects the specific height 
> of
> the cell.
> 
> <!DOCTYPE html>
> 
> <style>
> 
> table {
> 
>    width: 250px;
> 
> }
> 
> td {
> 
>    width: 242px;
> 
> }
> 
> .cell-specified-height {
> 
>    height: 202px;
> 
> }
> 
> .container {
> 
>    height: 100%;
> 
>    overflow-y: hidden;
> 
> }
> 
> .content {
> 
>    background-color: black;
> 
>    height: 2000px;
> 
>    width: 225px;
> 
> }
> 
> </style>
> 
> <table>
> 
>    <td class="cell-specified-height" data-expected-height=206>

Did you mean that the expected offsetHeight of cell is 206 ?

202px plus default 1px padding-left plus default 1px padding-right 
should give 204px, no?

> 
>        <div class="container" data-expected-height=204>

Why expected height is not 202 ?

> 
>            <div class="content" data-expected-height=2000></div>
> 
>        </div>
> 
>    </td>
> 
> </table>
> 
> <div id="output"></div>


In your case 2, the table's height is 'auto'; the table-row's height is 
'auto'. So, the minimum required height by content in cell or the 
specified content 'height' applying to cell should be honored without a 
problem.

Your test (with several small modifications):

http://www.gtalbot.org/BrowserBugsSection/css21testsuite/Case2-container-respects-specified-height-cell.xht

I think this is a overflow-y: hidden bug in Chrome 41.0.2272.101. The 
specified 100% height applying in div#container is honored. Overflow-y 
declaration is not honored. What it seems I get with Chrome 41 is 
'overflow-y: visible' instead. Chrome 41 behaves as if overflow-y: 
visible was effective.

If I set overflow-y to auto, then Chrome 41 shrinks the div#container 
width from 100px to 85px without rendering a vertical scrollbar. Firefox 
36.0.4 keeps the width of div#container at 100px and does not render a 
vertical scrollbar.

Gérard

> I'm happy to go with whatever makes sense for everybody so please 
> regard
> this as an attempt to reduce the issues discussed in this thread down 
> to
> two digestible test cases that we can attempt to agree a rendering for.
> 
> Thanks,
> Robert

Received on Monday, 23 March 2015 23:21:35 UTC