Re: where is overflow:none ?

> [Original Message]
> From: Boris Zbarsky <bzbarsky@MIT.EDU>
> Ernest Cline wrote:
> > The default value of overflow is "auto"
> Not according to
> Also not according to CSS3 Box 
> <>.

For whatever reason, I seem to be doing a lousy job of reading the
standards of late, for which I apologize.  I will admit to a built in bias
towards assuming that a value of "auto" when it is an allowed choice
will be the default choice, but I certainly should have checked and
caught this before stating it.

> > For a UA that implements 'max-width' (which has a default of "auto")
> You mean "none"?  See 
> <>.  Or are
> you reading the CSS3 Box Working Draft?  I wonder why that changes
> the default value.... 
> (and removes the existing "none" value).

Perhaps the author has the same built-in bias towards "auto" that I do. :)

> > then if I am interpreting the Box module draft correctly, in a case
> > 'max-width' is "auto" then the intrinsic width of an item would become
> > in effect a minimum width
> Why?  What part of the draft leads you to believe that?  That would be
> pretty inconsistent with CSS2.1...

But {max-width:auto} isn't in CSS 2.1, so whatever behavior I assumed
would be inconsistent with it.  :)

> > Gecko has a 'max-width' property, but it does not appear to follow
> > the current CSS 3 Box working draft as it appears to use a default
> > value of "none" which isn't in either official Box module working draft.
> Which documents are you considering official here?

Even tho I looked for it in CSS2.1, somehow I failed to find 'max-width'
in there.  I did find it  in both of the publicly available CSS 3 Box

The details of how to implement a 'max-width' of "auto" are sketchy
in the draft and could be interpreted in several ways.  I attributed to
"auto" that it would behave differently from the behavior I observed
with "none", since otherwise why have both values?

Taking a closer look at this, it would appear that Gecko does conform
to what the CSS 2.1 standard calls for here, altho I must say the defaults
it calls for do not appear to me to be the optimal choice, but as long as
it is implementable, and the truly desirable choices are choosable
and implementable, that will do.

Received on Monday, 17 May 2004 01:06:58 UTC