Greg Mayman wrote:

. > So if M$ software at the server end is putting in that extra ";", how 
. > does it manage to handle it when it gets it back from a M$ client? Does
. > the M$ client software remove the extra ";" before returning the cookie?

I had a situation a couple of years ago where a source code had, "../../".  
When I removed one of the "../", the page worked as it should.  (This was at 
the alumni web site for the University of Arizona.)  I wrote the webmaster, 
who also happened to be a computer science professor, and told him of the 
problem.  His response was that *he* taught students how to create web 
browsers and that is one of his class assignments each year, and *if* he 
removed one of the "../", it would not work.  Since his mind was made up and 
he wouldn't be swayed by facts, I dropped the subject and manually removed 
one of the "../" each time I visited the web site.  More recently, things 
appear to have changed and the site is accessible by Arachne without any 
manual editing of the source page.

If a double ";" is not standard in a cookie, it should be easy to counter.  
Just do a search and replace in a cookie for two ;'s and replace it with 
one.  But Arachne would have to do this as it receives cookies, while M$ 
only has to do it when it reads cookies.

Wouldn't it be nice if we could forward all of these M$ specific tricks to 
the judge that oversaw the M$ trial?

Roger Turk
Tucson, Arizona

Reply via email to