29626 - residual style bug? after </font> and <font>, </i> ignored and italics run on

Status: VERIFIED FIXED

(Core :: CSS Parsing and Computation, defect, P3)

Description

[ekrock's old account (dead)]

Using Commercial 2/28 (which reports itself as Build ID 1999122808) on WinNT 4.0 
SP4, a </i> is ignored, causing the italic text property to propagate all the 
way to the end of the document.

I suspect this is caused by an erroneous residual style correction. The source 
HTML is a mess. Here's the key section from the simplified test case 
I'll attach:

<font face="Arial, Helvetica" size="2">
<p>where
<ul>
<li>
<i></font><tt>EventTarget</tt><font face="Arial, Helvetica" size="2"></i> is a 
reference</li>

Somehow it's forgetting about that </i>, causing "is a reference" (and the 
entire rest of the original document) to be italicized. This doesn't happen on 
Nav4.

Comment 1

[ekrock's old account (dead)]

Attachment: simplified testcase

Comment 2

[ekrock's old account (dead)]

Marking [TESTCASE] and testcase keyword (as BugAThon's in transition).

Comment 3

[Pierre Saslawsky]

This bug appeared in build 2000-02-28-08. It did not exist in 2000-02-25-12 where 
the rendering was identical to Nav4 (ie. nothing in italic).

Reassigned to rickg to re-evaluate his checkins from 02/25 22:14 in htmlparser, 
especially the one in CNavDTD.cpp:
http://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&
subdir=mozilla/htmlparser/src&command=DIFF_FRAMESET&file=CNavDTD.cpp&rev1=3.267&
rev2=3.268&root=/cvsroot

Note that I did not try to back out this checkin in my local tree to see if it 
would fix the problem. It's just a guess.

Comment 4

[rickg]

Hmm -- let me take a look.

Comment 5

[rickg]

The problem sequence is: <b><i></b>text<u></i>text. Arg.

Comment 6

[rickg]

I have a fix in hand, for what it's worth.

Comment 7

[rickg]

PDT team granted PDT+ status to this on Thursday; a=jar.

Comment 8

[rickg]

Actually -- I meant to imply that this is a dupe of 3944, which the PDT team 
approved on thursday.

Comment 9

[rickg]

Actually -- I meant to imply that this is a dupe of 3944, which the PDT team 
approved on thursday.

Comment 10

[rickg]

Fixed as side effect of bug 3944 -- improvement to cnavdtd::closecontainersTo().

Comment 11

[Christine Hoffman]

Using 08_09_08_m18 build, verified fixed.