if ($ENV{'HTTP_USER_AGENT'}=~/^googlebot\/2\.1|slurp|some_other_bot_na
+me$/){

#put any other routines you want to run on the request -> like logs, e
+tc here or just...

print "Location: "http:\/\/www\.example\.org\/$ENV{'PATH_INFO'}";
}

# And if not in your bot-list, do what you originally planned...

else{
your original script's body here
}
[download]

it probably be easier to use the whole googlebot\2.1 stuff in a $variable associated with some array - than putting too many | or statements into that regex... but i'm being lazy and non-thinking at the moment.

one word of caution before using something like this-> any and all $ENV variables need to be cleaned up; One needs to ensure that they have no evil characters in them. As a thought for the directories you have there: $ENV{'PATH_INFO'}=~/^([a-zA-Z_0-9]+\/?)*$/

one can't rely on the $ENVironmental variables too much, but in this case, it probably would work.

But isn't fixing this at the level which it is generated in better? I mean, think of the impact this will have if you had a slight error, or if somebody changes the underlying code without updating the filter.

I would look at simply denying login/session keeping to bot useragents, or using crafty robots.txt files to simply keep them out of wherever you don't want them.

Good luck!

Another possibility might be to have a PerlTransHandler check the user agent on incoming requests and munge the session ID out of the URL before it gets processed further down the request chain.