Re: "patch" to smtpmail.el

Re: "patch" to smtpmail.el

[Simon Josefsson]

Could someone please commit this?  It was acknowledged to work.  It
also fixes :version of an older patch.

Is it possible to get CVS write permission, btw?  How do I get that?

2002-05-07  Simon Josefsson  <jas@extundo.com>

	* mail/smtpmail.el (smtpmail-send-it): Use user-mail-address from
	calling buffer.
	(smtpmail-auth-credentials): Bump :version to 21.4 (21.3 will not
	have the patch).

Index: smtpmail.el
===================================================================
RCS file: /cvsroot/emacs/emacs/lisp/mail/smtpmail.el,v
retrieving revision 1.41
diff -u -p -u -w -r1.41 smtpmail.el
--- smtpmail.el	26 Apr 2002 18:59:08 -0000	1.41
+++ smtpmail.el	7 May 2002 17:09:05 -0000
@@ -165,7 +165,7 @@ looks like `user@realm'."
 		       (string  :tag "Username")
 		       (choice (const :tag "Query when needed" nil)
 				       (string  :tag "Password")))))
-  :version "21.3"
+  :version "21.4"
   :group 'smtpmail)
 
 (defcustom smtpmail-starttls-credentials '(("" 25 "" ""))
@@ -218,6 +218,7 @@ This is relative to `smtpmail-queue-dir'
 	(case-fold-search nil)
 	delimline
 	(mailbuf (current-buffer))
+	(user-mail-address user-mail-address)
 	(smtpmail-code-conv-from
 	 (if enable-multibyte-characters
 	     (let ((sendmail-coding-system smtpmail-code-conv-from))

Simon Josefsson <jas@extundo.com> writes:

> Gareth Owen <usenet@gwowen.freeserve.co.uk> writes:
>
>> This makes the process-buffer inherit user-mail-address from the message
>> buffer.  Useful to stop your "real" email address leaking out in headers 
>> like X-From and X-Return-Path headers added by the SMTP agent.
>>
>> BCC'd to maintainer.
>
> Does this patch work instead?
>
> Index: smtpmail.el
> ===================================================================
> RCS file: /cvsroot/emacs/emacs/lisp/mail/smtpmail.el,v
> retrieving revision 1.41
> diff -u -p -u -w -r1.41 smtpmail.el
> --- smtpmail.el	26 Apr 2002 18:59:08 -0000	1.41
> +++ smtpmail.el	7 May 2002 16:13:26 -0000
> @@ -218,6 +218,7 @@ This is relative to `smtpmail-queue-dir'
>  	(case-fold-search nil)
>  	delimline
>  	(mailbuf (current-buffer))
> +	(user-mail-address user-mail-address)
>  	(smtpmail-code-conv-from
>  	 (if enable-multibyte-characters
>  	     (let ((sendmail-coding-system smtpmail-code-conv-from))

Re: "patch" to smtpmail.el

[Eli Zaretskii]

On Tue, 7 May 2002, Simon Josefsson wrote:

> Is it possible to get CVS write permission, btw?

Yes.

> How do I get that?

Go to http://savannah.gnu.org/account/register.php and register there.  
Then tell me your savannah login name, and I will set you up for write 
access.  The rest is explained on the Emacs CVS page (go to 
http://savannah.gnu.org/projects/emacs and follow the link to "CVS", then 
read the part about "Developers' access to CVS").

Re: "patch" to smtpmail.el

[Richard Stallman]

I think this is a safer, cleaner way to write that change.
Does it work?

*** smtpmail.el.~1.42.~	Wed May  8 09:17:18 2002
--- smtpmail.el	Wed May  8 22:50:36 2002
***************
*** 220,225 ****
--- 220,226 ----
  	(case-fold-search nil)
  	delimline
  	(mailbuf (current-buffer))
+ 	(address user-mail-address)
  	(smtpmail-code-conv-from
  	 (if enable-multibyte-characters
  	     (let ((sendmail-coding-system smtpmail-code-conv-from))
***************
*** 260,266 ****
  	    ;; they put one in themselves.
  	    (goto-char (point-min))
  	    (if (not (re-search-forward "^From:" delimline t))
! 		(let* ((login user-mail-address)
  		       (fullname (user-full-name)))
  		  (cond ((eq mail-from-style 'angles)
  			 (insert "From: " fullname)
--- 261,267 ----
  	    ;; they put one in themselves.
  	    (goto-char (point-min))
  	    (if (not (re-search-forward "^From:" delimline t))
! 		(let* ((login address)
  		       (fullname (user-full-name)))
  		  (cond ((eq mail-from-style 'angles)
  			 (insert "From: " fullname)

Re: "patch" to smtpmail.el

[Simon Josefsson]

I committed something similar recently (rev 1.43).  Or were you
thinking about the variable name?  I chosed mail-address instead.

I don't understand why the previous idea didn't work though.
Shouldn't letting a variable to a variable take priority over the
buffer-local or global variable (as the case may be)?  Or is it
nonsensical to let a variable to itself just to work around its buffer
localness?  O well.

Richard Stallman <rms@gnu.org> writes:

> I think this is a safer, cleaner way to write that change.
> Does it work?
>
> *** smtpmail.el.~1.42.~	Wed May  8 09:17:18 2002
> --- smtpmail.el	Wed May  8 22:50:36 2002
> ***************
> *** 220,225 ****
> --- 220,226 ----
>   	(case-fold-search nil)
>   	delimline
>   	(mailbuf (current-buffer))
> + 	(address user-mail-address)
>   	(smtpmail-code-conv-from
>   	 (if enable-multibyte-characters
>   	     (let ((sendmail-coding-system smtpmail-code-conv-from))
> ***************
> *** 260,266 ****
>   	    ;; they put one in themselves.
>   	    (goto-char (point-min))
>   	    (if (not (re-search-forward "^From:" delimline t))
> ! 		(let* ((login user-mail-address)
>   		       (fullname (user-full-name)))
>   		  (cond ((eq mail-from-style 'angles)
>   			 (insert "From: " fullname)
> --- 261,267 ----
>   	    ;; they put one in themselves.
>   	    (goto-char (point-min))
>   	    (if (not (re-search-forward "^From:" delimline t))
> ! 		(let* ((login address)
>   		       (fullname (user-full-name)))
>   		  (cond ((eq mail-from-style 'angles)
>   			 (insert "From: " fullname)

Re: "patch" to smtpmail.el

[Richard Stallman]

    I committed something similar recently (rev 1.43).  Or were you
    thinking about the variable name?  I chosed mail-address instead.

Yes, your later change solved the same problem.
I had not seen it yet.  I'm always a day behind.

    I don't understand why the previous idea didn't work though.
    Shouldn't letting a variable to a variable take priority over the
    buffer-local or global variable (as the case may be)?

No, it doesn't.