W3C home > Mailing lists > Public > www-style@w3.org > May 2008

Re: [CSS 21] overflow:hidden adjacent to a float

From: Alan Gresley <alan@css-class.com>
Date: Tue, 20 May 2008 19:06:18 +1000
Message-ID: <4832948A.6090508@css-class.com>
To: Ingo Chao <i4chao@googlemail.com>
CC: Boris Zbarsky <bzbarsky@mit.edu>, "www-style@w3.org" <www-style@w3.org>

Ingo Chao wrote:
> My concern is:
> The variance has grown. What happened after the change in CSS2.1:9.5
> is that while Fx changed from margin-box to border-box, Opera and IE
> changed from border-box to margin-box. And it makes me nervous to see
> an additional /right/ margin, where no one is declared, in Safari, but
> in Opera 9.24 too. Different implementations conclude to the same bug?
> Finally, I think it cannot be expected that designers test in latest
> betas if (and when, to what extend) the box is narrowed.

I agree.

> There is nothing invalid here, and my concerns are not based on a
> behavior of a particular browser. Actually the screenshot shows that
> there isn't even the tendency for interoperability among the browsers.
> That makes overflow next to a float currently unusable (and it is
> irrelevant that I believe that overflow is greatly misused for
> containing floats). My suggestion was to consider defining the
> expected rendering. How can I place content into this overflow box
> without knowing if the next browser clips it or not while narrowing
> the box, when the narrowing itself is undefined?

This is a very valid point.

> I'd prefer a box that establishes a new block formatting context to be
> not narrowed at all. This would not even have to be defined, just
> cancel the last two sentences in paragraph 5 of CSS2.1:9.5 and let the
> paragraph end with:
> "If necessary, implementations should clear the said element by
> placing it below any preceding floats."
> Thanks

I agree. That would mean that in your test case Ingo the desired 
rendering of the overflow:hidden box is like IE7 in the screenshot you 
provided. Not clipped as in FF3 RC or showing whitespace to the right as 
in Safari or a margin-left of 20px with Opera 9.5. No author can really 
use these properties with such inconstancy between implementations due 
to an undefined spec.

Received on Tuesday, 20 May 2008 09:07:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:55:06 GMT