Technical Forum
Ask questions. Discover Answers.
Showing results for 
Search instead for 
Did you mean: 

Looking to separate out URL from URI

Does anyone have some advice on how to pull out just the URL from a full URI? I've attempted to use HTTP::path but that pulls out both the URL and URN. An example may help to illustrate this.



I have several directories on a webserver




server1 is accessible from the internet, I have an external NAT setup to point to the F5 which we're using as a reverse proxy and I want to limit web requests to only go to "" and not /abc/admin. That should be easy enough right? ... Well they share common .css and .js files in /abc




So what I'm looking to accomplish is allow requests with "/abc/user", "/abc/." but not "/abc/admin" or "/abc/*/."



Hopefully that notation makes sense... basically the files in /abc are fair game but additional directories aren't.




This is the basic start i have already, obviously it doesn't quite accomplish what i'm looking to do yet. If I checked again /abc/user as the path in the if statement I'd lose all formatting on the page.








if { not ([string tolower [HTTP::path]] starts_with "/abc") } {


HTTP::path "/abc/user"








Historic F5 Account
Use HTTP::uri it does what you are looking for



then setup a switch statement



switch [HTTP::uri] {


"/abc/user" {


use pool * }


"/abc/" {


use pool * }


default {


HTTP::redirect ""







That should work for you






Nailed it finally....



HTTP::path -- that just returns the path without the query string


HTTP::uri -- returns the path + the query string



I actually wanted to just use HTTP::path afterall. here's how I did it...






set host [HTTP::host]


set uri_list [split [string tolower [HTTP::path]] /]


if { [lindex $uri_list 2] equals "admin"} {


HTTP::respond 302 Location "https://$host/user"










Basically it just redirect the client to the "/user" directory if they try and access the "/admin" directory. The real key item here is the use of TCL lists and "split". I got in this mentality while working with python and using their lists frequently to solve problems like this.



"lindex" just returns the item at that index location and acts the exact same way as you'd expect "list[2]" to work in C or C++






This is also another valuable little snippet so that you can get the size of the list (all the different locations in the path + the filename) and then dump the actual filename in the path.




set list_size [llength [split [HTTP::uri] /]]


if{ $list_size > 1 } {


log local0. "Last item in the list: [lindex $uri_list [expr $list_size -1]]"