118380 - <td ... > appears in table as text

Status: VERIFIED DUPLICATE of bug 117441

(Core :: DOM: HTML Parser, defect, P3)

Description

[Randell Jesup [:jesup] (needinfo me)]

Mozilla 2002010103

  <td width="6844" align="center" bgcolor="#707070">Systems

appears in a table on this page.  The HTML looks fine:

<table border="0" cellpadding="0" cellspacing="0" width="618">
<tr>
<td width="1" bgcolor="#ffffff">
<img
src="/images/en_ideamall/co_pcmall/div_pcmall/div_interface/home/spacer.gif"
width="1" height="17" border="0"></td>
<td width="6844" align="center" bgcolor="#707070">
<a class="nav" href="/pcmall/shop/vendor_spotlight/" onMouseOver="menu_enter
(0);" onMouseOut="menu_exit (0);">Brands</a></td>
<td width="1" bgcolor="#ffffff">
<img
src="/images/en_ideamall/co_pcmall/div_pcmall/div_interface/home/spacer.gif"
width="1" height="17" border="0"></td>
<td width="6844" align="center" bgcolor="#707070">
<a class="nav" href="/pcmall/shop/category.asp?eStore=26757000"
onMouseOver="menu_enter (1);" onMouseOut="menu_exit (1);">Systems</a></td>

Comment 1

[Mats Palmgren (inactive)]

WFM, 2002-01-04-03 (trunk) on Windows 98.

Comment 2

[Randell Jesup [:jesup] (needinfo me)]

rechecked with 2002010403 WinXP

Bug still is there, however I noticed the bug appears around 50% of the time.  I
also saw another hanging </td> in one of the tables.

Comment 3

[harishd]

We need a reproducable test case. I've seen this kind of a behavior was never
successful in reproducing it. Without a good test case it's almost impossible to
fix.

Comment 4

[Randell Jesup [:jesup] (needinfo me)]

I tried it a few times on that page; I didn't get any hits.

The problem is that this is an apparent timing (or packetization) issue.  That
makes it tough at best to get a repeatable testcase, and perhaps impossible for
it to be repeatable.  When I reported it (and when I retested), I was getting
>50% hits, though.  I have seen this bug randomly all over the net, rarely
repeatable though.  I've seen it for many months.

My suspicion is that due to some particular packetizing, the parser is deciding
that a tag is actually text - perhaps where the "<" comes in at the end of one
packet, and the "td>" comes in in the next.  Or quite possibly it's more complex
than that, given it doesn't happen on most pages.  (perhaps only when there's a
certain state - actually, that's quite likely).  Something like "we're inbetween
<tr> and <td>", etc, etc.

I have seen this dozens of times.  Apparently others (harishd) have seen it too.

Comment 5

[Randell Jesup [:jesup] (needinfo me)]

Looks like 117441 finally will fix this random behavior (this bug should have
been duped against it before; symptoms match).

*** This bug has been marked as a duplicate of 117441 ***

Comment 6

[Moied]

Verified dup