Toen ik David W. Tamkin kietelde, kwam er dit uit:
Ruud:
But if you want to find a word throughout the whole message:
:0
* HB ?? \<some-word\>
some-action
The condition should be
* HB ?? ()\<some-word\>
or (not necessarily less expensive)
:0
* 9876543210^0 ^Subject:.*\<some-word\>
* 9876543210^0 B ?? \<some-word\>
some-action
The second condition should be
* 9876543210^0 B ?? ()\<some-word\>
OK, thanks.
I don't want to do checks on body-contents at all, I have
always considered that as a no-go area.
A check that might be useful is to score the occurance of
(certain) 8-bit codes in the first lines of the body.
:0
* B ?? ^^\/.*$*.*$*.*$*.*$*\
.*$*.*$*.*$*.*$*\
.*$*.*$*.*$*.*$*\
.*$*.*$*.*$*.*$*.*
{ B_chunk = $MATCH }
:0 D
* -1^1 B_chunk ?? .
* 2^1 B_chunk ?? =[0-7][0-9A-F]
* 20^1 B_chunk ?? [\x80-\xff]
* 22^1 B_chunk ?? =[89A-F][0-9A-F]
suspect
The \xHH's should be replaced by the actual characters.
Other scoring condition-lines can of course be added.
That regex will trip the 'Exceeded LINEBUF' message,
now and then. A better way puts at most 1000 bytes
(for example) of the body into B_chunk.
--
Affijn, Ruud
_______________________________________________
procmail mailing list
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail