Re: regex - in a file?

"Peter Bowyer" <[email protected]> Thu, 9 Sep 2004 06:43:49 +0100
Newsgroups gmane.mail.exim.exiscan.user
Message-ID <0a8001c4962f$fc115480$0a46a8c0@pbdesktop>
Lawrence Ong <[email protected]> wrote:
> Hi,
>
> We are thinking of having multiple regexes.  Below is simlar to what
> we currently have.
>
> # deny messages matching a particular regex not in our network
> deny  message = This message matches a blacklisted regular expression
>       ($regex_match_string) condition = ${if
>       <{$message_size}{20k}{1}{0}} regex = http.//4drugs123.com : etc
> : etc2
>
> The problem that we have now is that the regex string is getting very
> long.  Not only so, continually modifying the exim.conf file is not
> feasible for us.
>
> Is it actually possible to have the 'regex' read from a file?   We
> tried using lsearch, and it seems that the regex line is not able
> to understand that?  Or did we make a syntax error somewhere?

Have you tried Exim's $readfile expansion?

regex = ${readfile {/path/to/file}{:}}

should work for a newline-delimited file.

http://www.exim.org/exim-html-4.40/doc/html/spec_11.html#IX831

Although I should poiunt out here that SpamAssassin is extremely good at
this kind of processing, and there are URIBLs which collect spamvertised
URIs which SA can test against - no updating needed from you at all.

Peter