W3C home > Mailing lists > Public > www-style@w3.org > May 2009

RE: "Re: [CSS21] rules="all" attribute specification and border-collapse property"

From: Gérard Talbot <info@gtalbot.org>
Date: Mon, 4 May 2009 20:21:48 -0400 (EDT)
Message-ID: <3216.70.83.193.175.1241482908.squirrel@69.59.28.71>
To: "www-style@w3.org" <www-style@w3.org>
Cc: "Sylvain Galineau" <sylvaing@microsoft.com>, "fantasai" <fantasai.lists@inkedblade.net>, "Patrick Garies" <pgaries@fastmail.us>

> Load the
> following in a couple of browsers:
>
> <!doctype html>
> <html>
> <body>
> <table rules="all">
> <tr><td>A</td><td>B</td></tr>
> <tr><td>C</td><td>D</td></tr>
> </table>
> </body>
> </html>
>
> The result is identical in Firefox 3.5b4, Opera 9.64 and IE8. WebKit and
> IE7 seem to honor the author's explicit rules="all" in this case by
> setting border to 1; IE8, Firefox and Opera do something else that I don't
> find quite logical (yet) but they all agree.


I still think that the table rules should not be collapsed in your first
minimalist example. If the initial, default value is border-collapse:
separate (and it is so in HTML 4 [3] and in HTML 5 [4]), then the table
rules should not collapse. And they collapse in IE 8 RTW, Firefox 3.0.10,
Opera 9.64, Safari 4.0 beta build 528.16.


> I'm trying to figure out
> why.
>
> Now try this in the same browsers:
>
> <!doctype html>
> <html>
> <body>
> <table rules="all" style="border-collapse:separate">
> <tr><td>A</td><td>B</td></tr>
> <tr><td>C</td><td>D</td></tr>
> </table>
> </body>
> </html>
>
> The visible cell borders go away in IE8 and Firefox;


Then that's a bug too. Maybe a distinct bug besides bug 155507 [2] should
be filed, I suppose, to cover this case at bugzilla.mozilla.org

> Opera shows the left
> border of each cell.

Then that's a bug too.

Opera ignores rules="none" in some border-collapse: separate cases:

http://www.gtalbot.org/BrowserBugsSection/Opera9Bugs/Opera7RulesNoneIgnored.html
which is bug 214944 in Opera BTS

> WebKit renders this differently than the first case
> even though border-collapse's initial value is separate so one would
> expect the same result.
>
> So while it's useful to agree on what should happen when border is
> explicitly defined as non-zero and rules="all" is set, other cases may
> need consideration for this HTML/CSS area to interoperate properly.

Sylvain, I fully agree with you on the interoperability matter.
The thing is that Firefox has bugs too, well identified and filed bugs,
namely bug 43178 [1] and bug 155507 [2].

regards, Gérard

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=43178
[2] https://bugzilla.mozilla.org/show_bug.cgi?id=155507
[3] http://www.w3.org/TR/CSS21/changes.html#q127
[4]
http://www.w3.org/TR/2009/WD-html5-20090423/the-xhtml-syntax.html#margins-and-padding
[5]
https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=409470
[6] http://www.w3.org/TR/html4/sgml/dtd.html#TFrame
Received on Tuesday, 5 May 2009 00:22:39 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:18 GMT