134645 - Floats taken into account calculating the height of a float

Status: RESOLVED INVALID

(Core :: Layout: Floats, defect, P2)

Description

[Eric A. Meyer (dead account)]

Tests show that Mozilla is considering floated elements when calculating the
height of floated non-replaced elements, in contradiction to CSS2:10.6.3
(http://www.w3.org/TR/REC-CSS2/visudet.html#q17), which states that only
children in the normal flow are taken into account when calculating the height
of floated non-replaced elements.  This was apparently not a problem in the
0.9.4 era.  Testcase to follow.

Comment 1

[Eric A. Meyer (dead account)]

Attachment: Floated div and normal-flow div showing different behaviors

This testcase, derived from a case created by Brian Costner, shows that a
right-floated 'div' is differently drawn than a normal-flow 'div' where both
have exactly the came content.	CSS2:10.6.3 says they should look the same.

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

This is a case that I think the working group has discussed, but I'm not sure if
there was a decision.  I think that the definition of 'auto' height on floats
should make it include other floats within its "formatting context".  I may have
discussed this a bit at the Redmond City F2F.

Comment 3

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

See http://lists.w3.org/Archives/Member/w3c-css-wg/2001OctDec/0264.html (no
decision, though)

Comment 4

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Also, the second and particularly the third of "the whiteboard notes" listed in:
http://lists.w3.org/Archives/Member/w3c-css-wg/2001OctDec/0289.html

Comment 5

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

What does WinIE do here?  (And MacIE?)

Comment 6

[Eric A. Meyer (dead account)]

IE/Mac and IE/Win do what Mozilla does now, but if it's not in the spec, hasn't
been errata'd, and the WG hasn't reached a decision, then I'm pretty sure we'd
be wrong to violate the specification.  (Or has that idea changed of late? 
Because if so, I have a few things I want to see change in Mozilla, like yesterday.)

Comment 7

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

It seemed like we pretty much agreed on it, although it wasn't formally decided
since it's not really appropriate for errata -- but rather, for changes in css3.

Speaking of which, did we formally agree on the "lost errata" for clip?  They're
not in the errata either.  The errata are hopelessly out of date, and many
things that we decide aren't formally noted as consensus and/or never get done.

Comment 8

[Hixie (not reading bugmail)]

I seem to recall a decision in Oslo whereby a float containing only another
float, where the outer float had auto dimensions, and the inner float had a non
zero fixed size, would result in the outer float overlapping the outer-most
in-flow text, and the outer float having zero size.

Comment 9

[Chris Waterson]

This is probably because of the changes for bug 52242, no?

Comment 10

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

This is clear in CSS 2.1, which should reach CR quite soon.  (The version in the
public draft is clear but slightly garbled.  The CR should be better.)