215555 - float with percent width cause odd wrap effect

Status: RESOLVED WONTFIX

(Core :: Layout: Floats, defect)

Description

[basic]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5b) Gecko/20030807 Mozilla
Firebird/0.6.1+

a float on a table cell with a percent width seems to cause an odd wrap effect.
It used to not have this effect. I'm not sure if this bug is valid or Moz was
doing it wrong before... testcase coming up

Comment 1

[basic]

Attachment: testcase

Comment 2

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

Attachment: simpler testcase

This is a simpler testcase that doesn't trigger the anonymous frame
construction described in:
http://www.w3.org/TR/CSS21/visuren.html#dis-pos-flo
http://www.w3.org/TR/CSS21/tables.html#anonymous-boxes
and is essentially equivalent to your first testcase.  Probably the cause of
the "regression" is that we're now implementing the rules at the first of those
two URLs correctly.  I don't see why you want to float the table cell -- it
doesn't make sense.

Comment 3

[Johannes Buchner]

Attachment: Testcase outlined, XHTML 1.0 Strict.

Comment 4

[Johannes Buchner]

Attachment: screenshot of the testcase

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20060531 Minefield/3.0a1
Personally, I don't like the child elements and text to be outside the parents.

Comment 5

[Daniel.S]

(In reply to comment #4)
> Personally, I don't like the child elements and text to be outside the parents. 

The layout of the testcase is undefined per CSS 2.1: The 30% width of the float is dependent on the width of the table cell, which itself is dependent on the width of it's content (in this case, the float).

I think these steps are used to calculate the layout:

1. Calculate the shrink-to-fit width of the floating element.
2. Set the width of the containing block (here: the table cell) to the shrink-to-fit-width of the float inside.
3. Then, recalculate the float's width by calculating the % of width of the containing block.

The result is the same in all four major browsers. A change would harm interoperability, so this is a Won'tFix.