Re: [bug-anubis] add header[] puts headers in body

[Paolo]

On Wed, Sep 01, 2004 at 01:21:20PM +0300, Sergey Poznyakoff wrote:
...
> to make mutt talk directly to anubis.
> 
> See http://www.deez.info/sengelha/projects/mutt/libesmtp/ for more
> information on the patch, including downloads.

Thank you, that's a better solution indeed - and I'm gonna use it, once 
I get mutt+patch compile ...

Meanwhile, pls see attached corrected patch for msg2smtp.pl:

 - Mutt puts actual destinee as non opt args to 'sendmail' command, ie
   $(sendmail) -- address@hidden address@hidden ...
   which are discarded by current msg2smtp.pl.
   So I added -M opt as 'mutt mode', ie get destinees from nonopt args only. 
   (I fooled myself before, as I was bouncing myself a msg with
    To: myself and did not detect the problem ...)
 - Mutt doesn't put a Bcc: for Bcc: destinee, hence current msg2smtp.pl
   won't send to those addresses.
   OTOH, if msg has a Bcc: line, current msg2smtp.pl would get that address
   as well, and would send that line to others, which shouldn't happen.
   Then if MTA doesn't strip off Bcc:, that'd be like a Cc:
   Patch adds code to strip away Bcc: headers anyway
 - Local mail may have non-fq addresses, so '@localhost' is added as default
   if '@' isn't found in address.

I've been using it since a couple of days, and WFM so far.

There's quite a bit of overhead due to perl start-up on msg send, if not 
already cached, so that's another good point to the native ESMTP approach.

---><---

I've stumbled on another couple issues, for 'modify' command:

1. I'm having troubles to convince 'modify body' work as I understood it.

I have a RULE section like

---BEGIN RULE---
if header[X-Domain] "[ ]*mydomain.it"
  modify body :extended [".*X-Cruft:.*"]
fi
---END---

but that doesn't work:
if I write it like in the .info:

  modify body ["X-Cruft:"]

I get from anubis -c:

  .anubisrc:121: missing replacement value

regardless of :flag, and rc file is ignored. Well, makes sense, command 
is 'modify' hence .info needs correction.

Then if I put an "" as VALUE:

  modify body ["X-Cruft:"] ""

anubis -c is happy, but 'modify' does nothing - ie 'X-Cruft:' in text is 
not deleted.

Then if I say it's an :re :

  modify body :re [".*X-Cruft:.*"] ""

anubis -c is happy, but I get in trace file:

  /.anubisrc:121: MODIFY BODY [.*X-Cruft:.*] [NULL]

and mutt/msg2smtp.pl reports err9 and nothing is sent.
After several tests, seems that there must be an \(\) enclosing _at least_
one re, ie all following work:

  modify body :re [".*X-Cruft:\(.*\)"] ""
  modify body :re ["\(.*\)X-Cruft:.*"] ""
  modify body :re ["\(.*\)X-Cruft:\(.*\)"] ""

I think there's a bug in the [regex] parser, as '.*X-Cruft:.*' is correct.

2. Next issue: how to remove a piece from an header line? ie I'd need the 
equivalent of

  echo "Subject: Re: [X-Cruft] test anubis" | \
     sed "s/\(.*\)\[X-Cruft\]\(.*\)/\1\2/I"

I tried eg

  modify header :re [Subject] [Subject] "\(.*\)\\[brandbox.it\\]\(.*\)" "\1\2"

but got parse error (yes, according to .info that line is wrong) - I could
not find an obvious way to do that as I'd do with sed(1).
Is there a way to have Anubis do that, or is this a 'feature request'?

Thanks

-- paolo
 
 GPG/PGP id:0x21426690 kfp:EDFB 0103 A8D8 4180 8AB5  D59E 9771 0F28 2142 6690

msg2smtp.pl.diff
Description: Text document