Pure HTML & CSS code that crashes IE7/8

#1 (**permalink**) 07-01-2009, 07:20 PM

The following code will crash Internet Explorer. Although clearer proper HTML
would not cause this issue, such could should at worst result in a poorly
formatted page, not a full program crash.
****
<html><head>
<style type="text/css">
#a {
margin:0 10px 10px;
}
#b {
width:100%;
}
</style>
<title>IE Crasher</title>
</head>
<body>
<table><tr><td>
<div id="a">
<form id="b">
<input type="text" name="test"/>
</div>
</td><td width="1"></td></tr></table>
</body></html>

#2 (**permalink**) 07-01-2009, 08:30 PM

"Eric Caron" wrote ...
> The following code will crash Internet Explorer. Although clearer proper
> HTML
> would not cause this issue, such could should at worst result in a poorly
> formatted page, not a full program crash.
> ****
> <html><head>
> <style type="text/css">
> #a {
> margin:0 10px 10px;
> }
> #b {
> width:100%;
> }
> </style>
> <title>IE Crasher</title>
> </head>
> <body>
> <table><tr><td>
> <div id="a">
> <form id="b">
> <input type="text" name="test"/>
> </div>
> </td><td width="1"></td></tr></table>
> </body></html>

Doesn't "crash" when you provide the missing: </form>

"Crash" in my test was IE6 displaying a blank page with hourglass cursor.
Easily closed with the window close [X] button.

#3 (**permalink**) 07-01-2009, 09:40 PM

"Richard Crowley" wrote:
> Doesn't "crash" when you provide the missing: </form>
>
> "Crash" in my test was IE6 displaying a blank page with hourglass cursor.
> Easily closed with the window close [X] button.
>

Which is why I prefaced the post with "Although clearer proper HTML
would not cause this issue". Removing the CSS, the div, the table, etc all
cause this code to not crash IE. (BTW, it crashes IE6 politely - but IE8
crashes really hard). This code started from a crash ifrom a >800 line HTML
file and I narrowed it down to being this combination of structures that
produces a crash.

I'm not looking for a solution on how to fix this - I thought the IE dev
team deserved to know a bug (not relying on JS or ActiveX) that would cause
the browser to break.

#4 (**permalink**) 07-02-2009, 02:50 AM

Eric wrote on Wed, 1 Jul 2009 21:32:05 -0700:

> "Richard Crowley" wrote:
>> Doesn't "crash" when you provide the missing: </form>

>> "Crash" in my test was IE6 displaying a blank page with hourglass
>> cursor.
>> Easily closed with the window close [X] button.

> Which is why I prefaced the post with "Although clearer proper HTML
> would not cause this issue". Removing the CSS, the div, the table, etc
> all cause this code to not crash IE. (BTW, it crashes IE6 politely -
> but IE8 crashes really hard). This code started from a crash ifrom a
> >800 line HTML file and I narrowed it down to being this combination
> of structures that produces a crash.

> I'm not looking for a solution on how to fix this - I thought the IE
> dev team deserved to know a bug (not relying on JS or ActiveX) that
> would cause the browser to break.

IE8 on mine did the same as IE6 on Richard's PC - blank screen with
hourglass. No hard crash, but it did hang.

Adding the </form> and </div> that are missing cures it - there's no need to
remove anything, you just need to create valid HTML instead of lazy junk
with missing tags. A few seconds with a HTML validator would have sorted
that out.

While it would be nice for IE to not get stuck like this, it would be much
better for HTML authors to take a little more time and not put badly
formatted junk into their sites.

--
Dan

#5 (**permalink**) 07-02-2009, 04:30 AM

I see one div and one /div. Where is the second div?
--
Ron Sommer

"Daniel Crichton" <msnews@worldofspack.com> wrote in message
news:eXBftZv#JHA.3612@TK2MSFTNGP04.phx.gbl...
snipped
>
> Adding the </form> and </div> that are missing cures it - there's no need
> to remove anything, you just need to create valid HTML instead of lazy
> junk with missing tags. A few seconds with a HTML validator would have
> sorted that out.
>
> While it would be nice for IE to not get stuck like this, it would be much
> better for HTML authors to take a little more time and not put badly
> formatted junk into their sites.
>
> --
> Dan
>

#6 (**permalink**) 07-02-2009, 04:40 AM

Outlook Express was hiding the closing div, sorry.

Dan

Ron wrote on Thu, 2 Jul 2009 05:55:07 -0500:

> I see one div and one /div. Where is the second div?
> --
> Ron Sommer

> "Daniel Crichton" <msnews@worldofspack.com> wrote in message news:eXBftZv#JHA.3612@TK2MSFTNGP04.phx.gbl...
> snipped

>> Adding the </form> and </div> that are missing cures it - there's no
>> need to remove anything, you just need to create valid HTML instead
>> of lazy junk with missing tags. A few seconds with a HTML validator
>> would have sorted that out.

>> While it would be nice for IE to not get stuck like this, it would be
>> much better for HTML authors to take a little more time and not put
>> badly formatted junk into their sites.

>> --
>> Dan

#7 (**permalink**) 07-02-2009, 05:30 AM

OE, mishandling improperly nested tags, quasi-freezes system
~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-
1) On Thu 02 Jul 2009 04:32:05 GMT, "Eric Caron" wrote:
« Which is why I prefaced the post with... »

I totally support you on this. You have all sorts of people:

- the most careful use to write properly AND to be fault-tolerant to
others; being careful takes a lot of time, so these people POST VERY
LITTLE; this is just what you are doing;
- the most reckless use to write casually hence improperly AND to be
intolerant to others' real OR SUPPOSED faults; being casual uses
little time, so these people POST A LOT;
- the most numerous are inbetween, more on the careful side but little
involved in the particular issue, which make them POST VERY LITTLE OR
NOT AT ALL.

The above thread perfectly illustrates this - and should NOT dilute
and slow down the efforts you are making to find solutions tolerant to
others' faults:
« I thought the IE dev team deserved to know a bug (not relying on JS
or ActiveX) that would cause the browser to break ».
I hope your report will be forwarded to the dev team and finally get
corrected.

2) I complete with my test, under Windows XP Pro 32bits (2GB RAM,
Pentium M760 2GHz) with the http://blogs.msdn.com/tonyschr/archi...25/421923.aspx
"Desktop Heap Limit" changed from 3,072 (default) to 8,192:

- the sample, automatically completed by OE (closing FORM added AFTER
the closing TABLE), still fails: IE6 slows *not all but a vast number
of* my open IE and OE windows; even Task Manager is slowed nearly to
freeze; after a moment the faulty IE window's Taskbar icon lets me
Right-Click it and close it, then after another delay it does close
and immediately everything is OK.
- that OE-completed sample, once corrected (just the "<FORM id=b>"
being moved BEFORE the OPENING "TABLE" tag), works OK.

I did no test in other browsers (lack time right now)

Versailles, Thu 2 Jul 2009 14:25:40 +0200

#8 (**permalink**) 07-02-2009, 05:40 AM

By "OE" I meant OE *or IE* depending which moment and which program is
using MSHTML (mostly OE was used to edit and IE crashed when
attempting to render).
Versailles, Thu 2 Jul 2009 14:34:30 +0200

#9 (**permalink**) 07-02-2009, 06:10 AM

The previous message was posted Thu 2 Jul 2009 12:25:40 GMT by me "Michel Merlin",
under the only name Google could give me ("Merlin1940"),
on the group http://groups.google.com/group/micro...plorer.general "microsoft.public.internetexplorer.general" > http://groups.google.com/group/micro...a2645dc598a4b2
"Pure HTML & CSS code that crashes IE7/8",
as message http://groups.google.com/group/micro...40692c2b09a004
"OE, mishandling improperly nested tags, quasi-freezes system".

Sorry, Google despite the lot it already does, can't so far avoid splitting one's assets and groups.

Versailles, Thu 2 Jul 2009 14:52:00 +0200

----- Parent Message -----
From: "Merlin1940" <merlin1940******.com>
Newsgroup: news:msnews.microsoft.com/microsoft.public.internetexplorer.general
Message: news:msnews.microsoft.com/d522c9c2-20d6-45c4-a411-6f59d835b655@s9g2000yqd..googlegroups.com
Sent: Thu 2 Jul 2009 05:25:40 -0700 (12:25:40 GMT)
Subject: Re: Pure HTML & CSS code that crashes IE7/8

OE, mishandling improperly nested tags, quasi-freezes system
~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-
1) On Thu 02 Jul 2009 04:32:05 GMT, "Eric Caron" wrote:
« Which is why I prefaced the post with... »

I totally support you on this. You have all sorts of people:

- the most careful use to write properly AND to be fault-tolerant to
others; being careful takes a lot of time, so these people POST VERY
LITTLE; this is just what you are doing;
- the most reckless use to write casually hence improperly AND to be
intolerant to others' real OR SUPPOSED faults; being casual uses
little time, so these people POST A LOT;
- the most numerous are inbetween, more on the careful side but little
involved in the particular issue, which make them POST VERY LITTLE OR
NOT AT ALL.

The above thread perfectly illustrates this - and should NOT dilute
and slow down the efforts you are making to find solutions tolerant to
others' faults:
« I thought the IE dev team deserved to know a bug (not relying on JS
or ActiveX) that would cause the browser to break ».
I hope your report will be forwarded to the dev team and finally get
corrected.

2) I complete with my test, under Windows XP Pro 32bits (2GB RAM,
Pentium M760 2GHz) with the http://blogs.msdn.com/tonyschr/archi...25/421923.aspx
"Desktop Heap Limit" changed from 3,072 (default) to 8,192:

- the sample, automatically completed by OE (closing FORM added AFTER
the closing TABLE), still fails: IE6 slows *not all but a vast number
of* my open IE and OE windows; even Task Manager is slowed nearly to
freeze; after a moment the faulty IE window's Taskbar icon lets me
Right-Click it and close it, then after another delay it does close
and immediately everything is OK.
- that OE-completed sample, once corrected (just the "<FORM id=b>"
being moved BEFORE the OPENING "TABLE" tag), works OK.

I did no test in other browsers (lack time right now)

Versailles, Thu 2 Jul 2009 14:25:40 +0200

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
IE7 crashed with this HTML+CSS code	Maya	Internet Explorer	2	03-28-2009 04:31 PM
View HTML Source Code in IE7	Ilya	Internet Explorer	5	03-28-2009 11:25 AM
Publish as pure HTML	hochprior	Microsoft OneNote	3	10-30-2007 01:40 AM
Where to post non-html code?	daveekelly1@embarqmail.com	Linux	3	10-29-2007 05:30 PM
Embedding and Copying code to HTML	Henery Ray Ray	Microsoft Office	1	04-16-2007 01:16 AM

New To Technology Questions?	Do You Need Help with Your Computer or Device?	Do You Need Help with this site?
Register to Participate Introduce Yourself Search the Site Terms and Conditions Privacy Policy	Manufacturers Hardware Software Gaming	Website Issues and Feedback Report Spam Did you forget your password? Mark Forums Read