Home / server / Questions / 149896
In Process
Alex F
Asked: 2010-06-10 17:27:32 +0800 CST

negative regexp in Squirm (for Squid). Possible?

  • 772

Did someone achieve to do negative regexp (or part of) with Squirm ?
I tried negative lookahead things and ifthenelse regexps, but Squirm 1.26 fails to understand them.

What i want to do is simply:
* If the url begins by 'http://' and contains 'account', then rewrite/redirect to 301:https://
* It the url begins by 'https://' and does NOT contains 'account, then rewrite/redirect to 301:http://

So far, i do that using 2 lines of perl, but squirm redirectors would take less memory

squid

1 Answers

  1. Thanks to Chris Foote, the Squirm owner, i got it working with his trick. Before it did not occur to me that squirm would stop the rewriting once a pattern match

    regexi ^(https:\/\/.*\/account.*)$ \1 account
regexi ^http:\/\/(.*\/account.*)$ 301:https://\1 account
regexi ^https:\/\/(.*)$ 301:http://\1 https

    .
    Chris explained me how to deal in this situation:

    What I've done previously to bypass a general catch-all rewrite rule is to place a rule matching the thing you don't want redirected beforehand, just replaying the original URL without a 301 redirect. e.g.:

    regexi (some_pattern) \1       # let through URL unwritten
regexi (http://mydomain/all-to-one-page\.html) \1 # allow \
      direct access to redirect page in the rule below.
regexi .* 301:http://mydomain/all-to-one-page.html  # all \
      other requests get redirected to a single page.
    • 0