.signature vs message.el

.signature vs message.el

[Robert Pluim]

Hmm,

message.el is very nice, but it conflicts with the way I normally do
mail signatures. By default, I don't want a signature added to my
outgoing mail, but I want to be able to add one using C-c C-w . If I
set message-signature to nil, the first happens, but then I can't do
the second. If I set it to t, then it gets added automatically every
time I send a message, which I don't want either. Any suggestions?

Robert

PS I'm using sgnus-0.60

Re: .signature vs message.el

[Jack Vinson]

>>>>> "RP" == Robert Pluim <pluim@xylint.co.uk> writes:

RP> message.el is very nice, but it conflicts with the way I normally do
RP> mail signatures. By default, I don't want a signature added to my
RP> outgoing mail, but I want to be able to add one using C-c C-w . If I
RP> set message-signature to nil, the first happens, but then I can't do
RP> the second. If I set it to t, then it gets added automatically every
RP> time I send a message, which I don't want either. Any suggestions?

The way this is done in sendmail is 
a) if mail-signature is 't, then insert the contents of mail-signature-file
   into the *mail* buffer before the user starts composing the message.
b) if mail-signature is non-nil, it must be a string that is the
   signature.  Insert that.
c) Otherwise, do nothing.

This is handled by mail-setup which is now message-setup.

The function mail-signature, run by C-c C-w, simply inserted the contents of
the mail-signature-file into the buffer at the end of the buffer, or at the
current point given a prefix argument.

The problem is that Lars has changed message-setup to call
message-insert-signature, rather than writing similar code twice.  

If this is to remain the case, we need another variable:
message-auto-signature or something which we can set 't when we want
message to insert our message-signature for us and 'nil when we want to do
it ourselves with C-c C-w.

If you want to use it now, the patch would be something like this.  There
may be a line number or two messed up.  I did not test this, but the main
change is one line in message-setup which only calls
message-insert-signature if message-auto-signature is non-nil.
message-signature needs to be non-nil if ANY signature is to be inserted.
This makes sense; a nil message-signature indicates there is no signature
to be inserted.

*** ChangeLog~	Thu Mar 28 01:25:09 1996
--- ChangeLog	Thu Mar 28 14:28:03 1996
***************
*** 1,3 ****
--- 1,8 ----
+ Thu Mar 28 14:27:54 1996  Jack Vinson  <jvinson@cheux.ecs.umass.edu>
+ 
+ 	* message.el (message-auto-signature): New Variable.
+ 	(message-setup): use it.
+ 
  Thu Mar 28 06:12:28 1996  Lars Magne Ingebrigtsen  <larsi@aegir.ifi.uio.no>
  
  	* message.el (message-send-mail): Remove Gcc header.

*** message.el~	Thu Mar 28 01:25:01 1996
--- message.el	Thu Mar 28 14:26:03 1996
***************
*** 253,263 ****
    "*String to be inserted at the and the the message buffer.
  If t, the `message-signature-file' file will be inserted instead.
  If a function, the result from the function will be used instead.
! If a form, the result from the form will be used instead.")
  
  (defvar message-signature-file "~/.signature"
    "*File containing the text inserted at end of mail buffer.")
  
  (defvar message-distribution-function nil
    "*Function called to return a Distribution header.")
  
--- 253,268 ----
    "*String to be inserted at the and the the message buffer.
  If t, the `message-signature-file' file will be inserted instead.
  If a function, the result from the function will be used instead.
! If a form, the result from the form will be used instead.
! If nil, no signature will be inserted.")
  
  (defvar message-signature-file "~/.signature"
    "*File containing the text inserted at end of mail buffer.")
  
+ (defvar message-auto-signature t
+   "*Indicates whether Message should automatically insert a signature into 
+ outgoing messages.  See the  documentation of `message-signature.'")
+ 
  (defvar message-distribution-function nil
    "*Function called to return a Distribution header.")
  
***************
*** 1826,1832 ****
        (message-generate-headers message-required-mail-headers))
      (insert message-default-mail-headers))
    (insert message-header-separator "\n")
!   (message-insert-signature)
    (message-set-auto-save-file-name)
    (save-restriction
      (message-narrow-to-headers)
--- 1830,1836 ----
        (message-generate-headers message-required-mail-headers))
      (insert message-default-mail-headers))
    (insert message-header-separator "\n")
!   (if message-auto-signature (message-insert-signature))
    (message-set-auto-save-file-name)
    (save-restriction
      (message-narrow-to-headers)

-- 
Jack Vinson                       jvinson@cheux.ecs.umass.edu
"I Spit on Your Grave"            -- double feature at the drive-in as seen
"I Thumb Through Your Magazines"     from the Qwik-E mart

Re: .signature vs message.el

[Lars Magne Ingebrigtsen]

Jack Vinson <jvinson@cheux.ecs.umass.edu> writes:

> The problem is that Lars has changed message-setup to call
> message-insert-signature, rather than writing similar code twice.  
> 
> If this is to remain the case, we need another variable:
> message-auto-signature or something which we can set 't when we want
> message to insert our message-signature for us and 'nil when we want to do
> it ourselves with C-c C-w.

I've now made `C-c C-w' "force" a mail signature when called
interactively.  That is, if `mail-signature' is nil, then
`mail-signature-file' will be inserted when you do `C-c C-w'.

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

Re: .signature vs message.el

[Robert Pluim]

>>>>> "Lars" == Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:

    Lars> I've now made `C-c C-w' "force" a mail signature when called
    Lars> interactively.  That is, if `mail-signature' is nil, then
    Lars> `mail-signature-file' will be inserted when you do `C-c
    Lars> C-w'.

Brilliant! Thanks for the fix guys.

Robert

Re: .signature vs message.el

[Mark Borges]

>> On 29 Mar 1996 05:41:52 +0000,
>> Lars Magne Ingebrigtsen(lars) wrote:
lars> Jack Vinson <jvinson@cheux.ecs.umass.edu> writes:
>> The problem is that Lars has changed message-setup to call
>> message-insert-signature, rather than writing similar code twice.  
>> 

lars> I've now made `C-c C-w' "force" a mail signature when called
lars> interactively.  That is, if `mail-signature' is nil, then
lars> `mail-signature-file' will be inserted when you do `C-c C-w'.

I'm still having some difficulty getting the effect I want since the
great message.el switch.

Before the switch, I had a cheesy function defined on `\C-c\C-w' in
mail-mode-map that would insert the contents of "~/.sig" normally, but
if given a prefix argument, would insert the contents of
"~/.signature" instead.

What is the cleanest way to get this behavior back?

Thanks,
 -mb-

Re: .signature vs message.el

[Mark Borges]

>> On 09 Apr 1996 00:12:52 -0600,
>> Mark Borges(mb) wrote:
mb> Before the switch, I had a cheesy function defined on `\C-c\C-w'
mb> in mail-mode-map that would insert the contents of "~/.sig"
mb> normally, but if given a prefix argument, would insert the
mb> contents of "~/.signature" instead.

mb> What is the cleanest way to get this behavior back?

Well, this is what I came up with. Can anyone see anything cleaner?

Can message-mode-map somehow inherit from mail-mode-map (or
vice-versa), so I don't have to replicate code for VM?

--------------------------------------------------
; is there a better hook -- one that only runs once, after message has
; done its thing, e.g., message-init-hook? 

(add-hook 'message-setup-hook 'my-mail-setup)

(defun my-mail-setup ()
  (setq
   message-signature nil
   message-indent-citation-function nil
   message-citation-line-function nil
   message-cite-function 'sc-cite-original
   )
; ...other stuff...
  (define-key mail-mode-map [(control x) n] 'abbrev-hacking-next-line)
  (define-key mail-mode-map [(meta >)] 'abbrev-hacking-end-of-buffer)
  (define-key message-mode-map [(control c) (control w)] 'mdb:signature)
; so VM can see it too
  (define-key mail-mode-map [(control c) (control w)] 'mdb:signature))

;;; ******************** 
(defun mdb:signature (&optional long-sig)
  "Sign letter with contents of `mail-signature-file'.
Prefix arg means use long signature."
  (interactive "P")
  (require 'message)
    (if long-sig
	(setq message-signature-file "~/.signature")
	(setq message-signature-file "~/.sig"))
	(message-insert-signature 1)
)
--------------------------------------------------

-- 
  -mb-

Re: .signature vs message.el

[d. hall]

// thus on 09 Apr 1996 00:12:52 -0600, Mark virtually wrote:

Mark> What is the cleanest way to get this behavior back?

(setq message-signature 'my-mail-insert-signature)

This is probably the cleanest way.

Right now, my function replaces the current signature with another random
signature (or else places one at the end of the message buffer), if I don't
like the one put there initially (there is one placed there when the
message buffer is first formed).  Also with a prefix argument, it'll use
that as a reference in the vector to gather the appropriate signature.

d.
-- 
         ``It's a death row pardon... two minutes, too late''

                                ~ alanis morissette, ironic