On Tue, Mar 23, 2010 at 4:02 PM, Philippe Wittenbergh [EMAIL-REMOVED]> wrote:
> On Mar 23, 2010, at 11:46 PM, Bruno Fassino wrote:
>> Yes, with positive margins there is more consistency amongst modern browsers.
>> The only anomaly that I see, with positive margin on the same side of
>> the float, is that Safari 4 makes the b.f.c. box narrower than
>> necessary, so there is a gap on the other side of the float, see
>> http://brunildo.org/test/FloatMarginOverflow.html . Of course this is
>> "allowed" by the spec.
> The phantom margin/gap on the opposite side of the float (on the right of the BFC with your
left-margin + float:left) ? I don't think that is allowed by the spec. And it is considered a bug by
at least some people of the WebKit community.
Yes, exactly that (and your bug report is almost two year old :-)
It is surely undesirable. The reason I said it is "allowed" is that
the spec at <http://www.w3.org/TR/CSS21/visuren.html#bfc-next-to-float>,
says "CSS2 does not define ... by how much said element may become
narrower". So making a b.f.c. arbitrarily narrower than defined by
section 10.3.3 is not "prohibited" (and when the box is made
"narrower" than strictly required by the presence of float and
margins, we have those phantom margins/gaps on one side or another, ).
Of course this is an attempt of a "formal" reading of the spec. I
consider that a bug :-)
Bruno Fassino http://www.brunildo.org/test
List wiki/FAQ -- http://css-discuss.incutio.com/
List policies -- http://css-discuss.org/policies.html
Supported by evolt.org -- http://www.evolt.org/help_support_evolt/