squid url_regex not matching characters

Question

I have a url_regex acl: ^http(s)://bitbucket.org/example/*

However, it's matching the url: http(s)://bitbucket.org/example_test/*

This seems to not consider the "/" at the end of ^http(s)://bitbucket.org/example/*

I'd like to match anything with subfolder matching example. Eg: bitbucket.org/example/case_1 bitbucket.org/example/case_2 But not, bitbucket.org/example_bad/case_3

It seems like a special blend of regex and its not considering the "/" –

Can anyone see what is wrong here ?

Answer 1

As steeldriver said, it’s matching for the same reason car*t matches cat  —  * means any number of the preceding character, including zero.  r* matches the zero r’s between a and t, and your /* is matching the zero slashes between example and _test/*.

Do you want to match http(s)://bitbucket.org/example (without any subdirectory)?  If not, it’s easy; just use

^http(s)://bitbucket.org/example/

or

^http(s)://bitbucket.org/example/.*$

If you do want to match http(s)://bitbucket.org/example, it’s a bit trickier.  You need example to be either at the end of the URL, or followed by a slash.  Some regex engines will let you say

^http(s)://bitbucket.org/example(/|$)

where the | inside (…) means “or”.  For example, grep -E supports this.  I don’t know whether this will work in Squid.

---

Does Squid let you specify multiple regexes?  If it does, specify

^http(s)://bitbucket.org/example$

and

^http(s)://bitbucket.org/example/