On Tue, 2009-02-24 at 12:47 -0500, Gene Heskett wrote: > On Tuesday 24 February 2009, Karsten Bräckelmann wrote:
> > > Would this work? > > > > > > :0: > > > *^*no To-header on input* > > > /dev/null > > > > Nope, it wouldn't. Procmail uses REs, not shell-style globbing. > > I never claimed to understand regex's. I know the ^ anchors the start of the > search to the start of the line, and that the first * is needed to into a > recipe, but how does one go about allowing it to search the whole line for > the given character sequence, triggering on finding it at some arbitrary > location in that line? If grep can do it, why can't procmail? *sigh* OK, slow down a bit. You're mixing procmail and REs here. Yes, the very first * denotes a condition in the procmail recipe. Also yes, ^ anchors your RE at the beginning of the string. Procmail is no different than grep in this regard. grep undisclosed raw-message-headers-only.file Limiting to the headers only, which is the procmail default. Just to prevent grep matches in the body. Now that trivial grep will check ALL headers -- for example, including the Subject. If you want to limit your grep to a specific header, you need to anchor your RE. Exactly the same with procmail. (With the notable exception of procmail properly treating multi-line headers, which grep of course doesn't. Add formail to the mix for that. ;) So you want to match a specific header (read: the beginning of a given line), and match a string that might appear anywhere in that header. That's basically "whatever there is between the two". In RE, that's "any char, any number of times". The good old /.*/, and the RE becomes: ^Header: .*Catch Phrase Procmail, just like grep, is line oriented -- whereas SA rules make the syntax for dealing with that easier. For reference, and in a vain attempt to get back on-topic: header MY_CATCH_PHRASE Header =~ /Catch Phrase/ The rule is limited to header Header, but the RE itself is not anchored to the beginning, thus we don't need to care about stuff between them. Sorry, Gene -- this is most *basic* stuff. > IMO the Docs suck a deep space quality vacuum in re these details. If there > exists a decent tut on this subject, please point me at it. man procmailex For RE docs, see http://perldoc.perl.org/perlre.html in particular the Quick-Start and Tutorial introduction mentioned early. Note though, that procmail REs are quite different from that in anything even mildly advanced. > > If you don't want to anchor your condition REs at the beginning of the > > line, don't. IMHO you'd better do though, for multiple reasons -- speed, > > and not to match any arbitrary header but the To header only. > > Are you saying that if I remove the ^ and second *, then it will search the > whole header? Testing that now... Yes. ALL headers, since you didn't specify a header. Can appear anywhere in the line. See above for properly limiting to a specific header. FWIW, the "second *" (actually the first one in your RE) is just bad. I'd expect procmail to complain about that. > > That said, I do agree with Martin and John. The absence of a real > > recipient in the To header is NOT sufficient to silently discard mail. > > Even more so, since the POP3 server appears to have rewritten that > > stuff. > > If I was an ISP, maybe. But I'm just sick of junk mail & if I miss a free > offer for 20 boxes of viagra, well... :) Or the wedding of a friend. [1] Coincidentally, I sent some mail to Bcc only recipients the other day. You would discard it. Good one... Again, I *strongly* suggest to score it a point or two in SA, if you are so inclined. Leaving at least a chance for legitimate mail to survive it. guenther [1] No, not me. :) -- char *t="\10pse\0r\0dtu...@ghno\x4e\xc8\x79\xf4\xab\x51\x8a\x10\xf4\xf4\xc4"; main(){ char h,m=h=*t++,*x=t+2*h,c,i,l=*x,s=0; for (i=0;i<l;i++){ i%8? c<<=1: (c=*++x); c&128 && (s+=h); if (!(h>>=1)||!t[s+h]){ putchar(t[s]);h=m;s=0; }}}
