63785 - [quirks]P has no margins in DIV in TABLE

Status: VERIFIED FIXED

(Core :: Layout, defect)

Description

[Eugene Savitsky]

1. Open the page in moz and in NN 4.x or other browser.
2. There should be spaces between <p>, but not in moz.

First seen ~2weeks ago.

20002504-talkback

Comment 1

[Fabian Guisset]

Do you think it would be possible to make a small testcase?
Thanks!

Comment 2

[Eugene Savitsky]

OK. I'll attach a test case.

The first table is with <div align='justify'>, the second witout.

Comment 3

[Eugene Savitsky]

Attachment: test case

Comment 4

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

This is quirks-mode only.  I will attach a simpler testcase.

Comment 5

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

Attachment: simpler testcase

Comment 6

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

This seems because karnaze%netscape.com added in revision 1.11 of quirk.css
(Dec. 21), the rule:

td div p {
  margin: 0px;
}

to "fix" bug 33784.  This rule is clearly wrong.  In the future perhaps it would
be better if Ian or I reviewed quirk.css and html.css changes?  (Bug 33784, by
the way, is a very complicated issue that I've investigated in detail in the
past when I was working on [MARGIN-C] bugs.  There are other bugs on it (search
for bugs with [MARGIN-C] in the subject).  I eventually came to the conclusion
that there was no good way to fix it without breaking other things or without
extensive modifications that might not work anyway.)

Chris, could you back out this change?  I'm surprised this got past the block
regression tests.

Comment 7

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

If you want to extend the current :first-node, etc. hack to DIV > P as well, you
could add the selector "td > div:first-node > :first-node" to the rule at line
77 of quirk.css and the selector "td > div:last-node > :last-node" to the rule
at line 81 of quirk.css.  (And you could do it for 2 or 3 divs also.)  But
that's still not a general solution to the problem.

Comment 8

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

I added the testcase I attached above to the block regression tests.

Comment 9

[Mats Palmgren (inactive)]

*** Bug 63983 has been marked as a duplicate of this bug. ***

Comment 10

[karnaze (gone)]

I backed out the patch to bug 33784.

Comment 11

[Loco]

VERIFIED on Mac builds 01-02-08 and 01-04-08.  Poof, all gone!

Changing Plat to All/All since this bug was.