Resending

Resending

[David K}gedal]

This patch fixes the most serious bugs in the resending code in 0.18,
making it almost useful...

It also fixes a bug that made Gnus always insert a Sender header, even
when it was not supposed to.

I still get the headers "Resent-Sender" and "Resent-Message-ID", which
I'm not too fond of, at least not the first.

*** lisp/ChangeLog	Tue Dec  5 06:11:09 1995
--- lisp/ChangeLog	Tue Dec  5 10:44:13 1995
***************
*** 1,3 ****
--- 1,10 ----
+ Tue Dec  5 10:35:51 1995  David K}gedal  <davidk@lysator.liu.se>
+ 
+ 	* gnus-msg.el (gnus-inews-insert-headers): Use cadr of the result
+ 	from gnus-extract-address-components instead of the car.
+ 	(gnus-summary-resend-message): Do not call mail-setup, and use
+ 	fewer headers.
+ 
  Tue Dec  5 03:22:37 1995  Lars Magne Ingebrigtsen  <larsi@narfi.ifi.uio.no>
  
  	* nnbabyl.el (nnbabyl-close-server): Restore buffer mode on exit.
*** lisp/gnus-msg.el	Tue Dec  5 06:10:53 1995
--- lisp/gnus-msg.el	Tue Dec  5 10:44:05 1995
***************
*** 1182,1193 ****
        (if (and from 
  	       (not (gnus-check-before-posting 'sender))
  	       (not (string=
! 		     (downcase (car (gnus-extract-address-components from)))
  		     (downcase (gnus-inews-real-user-address))))
  	       (or (null sender)
  		   (not 
  		    (string=
! 		     (downcase (car (gnus-extract-address-components sender)))
  		     (downcase (gnus-inews-real-user-address))))))
  	  (progn
  	    (goto-char (point-min))    
--- 1182,1193 ----
        (if (and from 
  	       (not (gnus-check-before-posting 'sender))
  	       (not (string=
! 		     (downcase (car (cdr (gnus-extract-address-components from))))
  		     (downcase (gnus-inews-real-user-address))))
  	       (or (null sender)
  		   (not 
  		    (string=
! 		     (downcase (car (cdr (gnus-extract-address-components sender))))
  		     (downcase (gnus-inews-real-user-address))))))
  	  (progn
  	    (goto-char (point-min))    
***************
*** 1512,1527 ****
      (let (resent)
        ;; We first set up a normal mail buffer.
        (nnheader-set-temp-buffer " *Gnus resend*")
!       (gnus-mail-setup 'new address)
        ;; Insert our usual headers.
        (gnus-inews-narrow-to-headers)
!       (let ((headers '(From Date To Message-ID Organization)))
  	(gnus-inews-insert-headers headers))
        (goto-char (point-min))
        ;; Rename them all to "Resent-*".
!       (while (re-search-forward "^" nil t)
  	(insert "Resent-"))
        (widen)
        (delete-region (point) (point-max))
        ;; Insert the message to be resent.
        (insert-buffer-substring gnus-original-article-buffer)
--- 1512,1536 ----
      (let (resent)
        ;; We first set up a normal mail buffer.
        (nnheader-set-temp-buffer " *Gnus resend*")
!       ;; This code from sendmail.el
!       (insert "To: ")
!       (let ((fill-prefix "\t")
! 	    (address-start (point)))
! 	(insert address "\n")
! 	(fill-region-as-paragraph address-start (point-max)))
!       ;;
!       (insert mail-header-separator "\n")
        ;; Insert our usual headers.
        (gnus-inews-narrow-to-headers)
!       (let ((headers '(From Date To)))
  	(gnus-inews-insert-headers headers))
        (goto-char (point-min))
        ;; Rename them all to "Resent-*".
!       (while (re-search-forward "^[A-Za-z]" nil t)
! 	(forward-char -1)
  	(insert "Resent-"))
        (widen)
+       (forward-line)
        (delete-region (point) (point-max))
        ;; Insert the message to be resent.
        (insert-buffer-substring gnus-original-article-buffer)


-- 
David Kågedal     Lysator Academic Computer Society       davidk@lysator.liu.se
http://www.lysator.liu.se/~davidk/                              +46-13 17 65 89

Re: Resending

[Sten Drescher]

davidk@lysator.liu.se (David K}gedal) said:

DK> This patch fixes the most serious bugs in the resending code in
DK> 0.18, making it almost useful...

DK> It also fixes a bug that made Gnus always insert a Sender header,
DK> even when it was not supposed to.

DK> I still get the headers "Resent-Sender" and "Resent-Message-ID",
DK> which I'm not too fond of, at least not the first.

	Resent-Message-ID is in RFC 822 as a reference field, and it is
a Good Thing - for example, I use procmail/formail to save 10KB of
Message-IDs, and drop duplicates.  The main use for this is in mailing
lists - like right now.  If you drop duplicates, you will see this once.
If you don't, you'll see it twice - once from me, and once from the
list. If you bounce a message to me, without alteration, I'd want it to
be dropped if I've already received it elsewhere.  Also, Resent-Sender
is also in the RFC as an originator field, although discouraged then it
matches Resent-From (which it would usually be in this case).

-- 
#include <disclaimer.h>				/* Sten Drescher */
To get my PGP public key, send me email with your public key and
	Subject: PGP key exchange
Key fingerprint =  90 5F 1D FD A6 7C 84 5E  A9 D3 90 16 B2 44 C4 F3

Re: Resending

[David K}gedal]

dreschs@mpd.tandem.com (Sten Drescher) writes:

>	Resent-Message-ID is in RFC 822 as a reference field, and it is
>a Good Thing - for example, I use procmail/formail to save 10KB of
>Message-IDs, and drop duplicates.  The main use for this is in mailing
>lists - like right now.  If you drop duplicates, you will see this once.
>If you don't, you'll see it twice - once from me, and once from the
>list. If you bounce a message to me, without alteration, I'd want it to
>be dropped if I've already received it elsewhere.

Yes, I use Gnus to read this, so I only have to see it once, but
aren't you talking about Message-ID? Where does Resent-Message-ID come
into all this?

>Also, Resent-Sender
>is also in the RFC as an originator field, although discouraged then it
>matches Resent-From (which it would usually be in this case).

Well, I think that Gnus should at least hide this header, as it does
for Sender.

-- 
David Kågedal     Lysator Academic Computer Society       davidk@lysator.liu.se
http://www.lysator.liu.se/~davidk/                              +46-13 17 65 89

Re: Resending

[Sten Drescher]

davidk@lysator.liu.se (David K}gedal) said:

DK> dreschs@mpd.tandem.com (Sten Drescher) writes:
>> Resent-Message-ID is in RFC 822 as a reference field, and it is a
>> Good Thing - for example, I use procmail/formail to save 10KB of
>> Message-IDs, and drop duplicates.  The main use for this is in
>> mailing lists - like right now.  If you drop duplicates, you will see
>> this once.  If you don't, you'll see it twice - once from me, and
>> once from the list. If you bounce a message to me, without
>> alteration, I'd want it to be dropped if I've already received it
>> elsewhere.

DK> Yes, I use Gnus to read this, so I only have to see it once, but
DK> aren't you talking about Message-ID? Where does Resent-Message-ID
DK> come into all this?

	It's in RFC 822.  But rereading the RFC, it doesn't work quite
the way I thought it did (the original Message-ID should be untouched).
Still, putting the Resent-Message-ID field in (which, like Message-ID,
is optional in the RFC) is prolly a good thing, since the Resent-
headers are supposed to parallel the normal headers.  But a parallel
question: goes Gnus know to supercede existing Resent- headers, either
by prepending them with Also- or Old- (I prefer Also-), or simply
stripping them out?  I can easily imagine a message getting resent more
than once, and:

rfc822>      4.2.  FORWARDING

[...]

rfc822> Note: In general, the "Resent-" fields should be treated as
rfc822> containing a set of information that is independent of the set
rfc822> of original fields.  Information for one set should not
rfc822> automatically be taken from the other.  The interpretation of
rfc822> multiple "Resent-" fields, of the same type, is undefined.

-- 
#include <disclaimer.h>				/* Sten Drescher */
To get my PGP public key, send me email with your public key and
	Subject: PGP key exchange
Key fingerprint =  90 5F 1D FD A6 7C 84 5E  A9 D3 90 16 B2 44 C4 F3

Re: Resending

[Lars Magne Ingebrigtsen]

dreschs@mpd.tandem.com (Sten Drescher) writes:

> But a parallel question: goes Gnus know to supercede existing
> Resent- headers, either by prepending them with Also- or Old- (I
> prefer Also-), or simply stripping them out?  I can easily imagine a
> message getting resent more than once,

September 0.23 will add an "Also-" prefix to all old headers that
start with "Resent-" and "Also-Resent-".  So if you resend mail a lot,
you'll get "Also-Also-Also-Also-Also-Resent-From:".  Whee!

-- 
Home is where the cat is.

Resending

[Lars Magne Ingebrigtsen]

`S D r' will now prompt you for an address and resend the current
message to that address.  A slew of "Resend-*"-headers will be added
to the message; other than that the message won't be changed at all. 

(Took longer to write the documentation than writing the function. :-) 

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Ingebrigtsen