OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
Re: legitimate use of recursive document loading into IE4
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: legitimate use of recursive document loading into IE4


  • To: NTBUGTRAQLISTSERV.NTBUGTRAQ.COM
  • Subject: Re: legitimate use of recursive document loading into IE4
  • From: "Joan L. Grove Brewer" <redroseAA.NET>
  • Date: Sat, 21 Mar 1998 10:34:25 -0800
  • Reply-To: "Joan L. Grove Brewer" <redroseAA.NET>
  • Sender: Windows NT BugTraq Mailing List <NTBUGTRAQLISTSERV.NTBUGTRAQ.COM>

>[in reference to recursive bug in IE4]
>
>>Seems to me that these are just obvious coding bugs that do not detect a
>>recursive case without an exit. I wonder if it is possible to write a
>>useful web page using this technique,
>
>I think so, if you were using javascript. 2 or more Documents could be
>combined into one, and the different parts could be displayed using
>'document.write()'s to form the different parts of the page depending on
>what the page's parent URL, or frame name, etc. was.

-----

> Maybe if the recursion case was looked for only after the document had
> been completely loaded (i.e. after the javascript had been executed).

i think this bug might turn out to be un-fixable - i believe it's known
as the halting problem in computer science circles - it's not possible
to tell whether a given (JavaScript) program will ever terminate, so you
can't ever catch all possible recursive nastiness that can be put into
a web page. the only thing to do is to make sure IE handles running out
of memory gracefully.

&rew.
--
        andrew kennedy : adkpresence-it.co.uk
        presence IT limited : death before unconciousness


---------- My Comments on Recursion -------

I've been doing recurision in graphics for a long time and it's doable.:-)
Andrew is quite right for the most part.  Recursion is very hard to manage.
This is the POINT. It must be carefully managed. With the INTERNET, you are
dealing with different systems and configuration, making this near impossible.

The recursion problem in IE is really in Windows itself.  The JAVA
interpreter is written in MS C++ (I'm guessing) and that compiler has always
been a stack hog--as is Windows.  I've tested CPU Mask and compilers.
When you look at the machine code output and see the difference in the
over head between MS C and other compilers its pretty scary. Add another
layer of interpreter code which also sucks up the stack and you are looking
at a major problem. This is the problem the game programmers have also
and why MS finally did a Game SDK after 5 years of complaints. Most Application
Errors are usually due to this HEAP and STACK problem.  We can't even
manually control the HEAP, by removing from memory disgarded graphics in
a multimedia or game program. There is nothing like having your machine
come to a complete stop so that the RAM memory and Virtual HEAP can swap
out.... Depending on how much RAM you are using and the complexity of the
HEAP it can really take a lonnnnng time. With all these itty bitty weeb icon
files flying around the server it's a mess to sort out. If the hard drive is
thrashing, you haven't crashed but most people think they are locked up or
something. At least we can't blame MS for designing HTTP and HTML! :-)

In Windows, every time you call the function from within itself, all the
previous data has to be stacked (somewhere). Here in lies the real problem.
Microsoft Windows is data intense and eats up memory this way. There is
also a pointer problem that may be either Intel or Microsoft. After all,
Windows 95 is running better and more stable on the MAC G3 than Pentium now.
One of the reasons VB is so limited in dynamic memory (link list etc) is due
to this problem.

Because Microsoft cuts and pastes code between their various group (and
systems) the design problems just keep replicating themselves.  Yep,
Microsoft does recursion better than anyone and is about to blow their
proverbial stack any time.:-)  If there is a bug in MS code, it just seems to
bounce all over the place. Finding all incidence of it in all their
applications
is not an easy task.:-)

There is hope on the horrision.  I just read that HP has re-engineered the
Sun JAVA interpreter for small machines using "state design==machine
coding technique" which can crunch 1M into 100K.   Some people think
this is a new competitive JAVA version. That's not true. It's SUN Java
standard.Perhaps MS will implement it in more than just Win CE. :-)


J~ BS CSE



>