Mail-Followup-To interaction with gnus-summary-followup

Mail-Followup-To interaction with gnus-summary-followup

[Paul J Collins]

I'm unsure of the semantics of Mail-Followup-To, so perhaps this isn't
a Gnus problem at all.

I'm on a mailing list that I filter into a separate group, and one of
the posters has a Mail-Followup-To header with the list's address in
it.  I had already set the group's parameters for to-list and
to-address to the address of list.

When I type `f' to follow up to this person's posts, my new message is
given a To field with the value of to-address, and a Cc field with the
value of the original poster's Mail-Followup-To.  (I can tell which is
which as we have different things in the non-mail address portion of
the address; they are both of the form: "smart comment
<list@server.tld>").

I'm running a CVS pull of Oort from this morning; 5.8.8, 5.8.7 and
5.8.3 did not previously display this behaviour.

By the way, how is Oort Gnus able to display a bitmap of the Gnus logo
in GNU Emacs 20.7?

-- 
Paul Collins <sneakums@sto-kerrig.org> - - - [         ]
 GPG: 0A49 49A9 2932 0EE5 89B2  9EE0 3B65 7154 8131 1BCD
"This one's optimistic, this one went to market;
 this one just came out of the swamp."

Apologies (Re: Mail-Followup-To interaction with gnus-summary-followup)

[Paul J Collins]

Aargh.  Sorry about the duplicate.

-- 
Paul Collins <sneakums@sto-kerrig.org> - - - [         ]
 GPG: 0A49 49A9 2932 0EE5 89B2  9EE0 3B65 7154 8131 1BCD
"This one's optimistic, this one went to market;
 this one just came out of the swamp."

Re: Mail-Followup-To interaction with gnus-summary-followup

[Paul J Collins]

>>>>> "PJ" == Paul Jarc <prj@po.cwru.edu> writes:

    PJ> Ok, here's the fix (against current CVS).  For real, this
    PJ> time. :)

That works great, thank you.

-- 
Paul Collins <sneakums@sto-kerrig.org> - - - [         ]
 GPG: 0A49 49A9 2932 0EE5 89B2  9EE0 3B65 7154 8131 1BCD
"This one's optimistic, this one went to market;
 this one just came out of the swamp."

Re: Mail-Followup-To interaction with gnus-summary-followup

[ShengHuo ZHU]

prj@po.cwru.edu (Paul Jarc) writes:

> Paul J Collins <sneakums@sto-kerrig.org> writes:
> > I'm still seeing this problem, where the Mail-Followup-To header's
> > value is being used on `f', even when message-use-followup-to is nil.
> 
> Ok, here's the fix (against current CVS).  For real, this time. :)

Committed.

ShengHuo

-- 
(setq gnus-posting-styles 
      '((".*" (signature (format "(setq gnus-posting-styles 
      '%S)" gnus-posting-styles)))))

Re: Mail-Followup-To interaction with gnus-summary-followup

[Paul Jarc]

[-- Attachment #1: Type: text/plain, Size: 265 bytes --]

Paul J Collins <sneakums@sto-kerrig.org> writes:
> I'm still seeing this problem, where the Mail-Followup-To header's
> value is being used on `f', even when message-use-followup-to is nil.

Ok, here's the fix (against current CVS).  For real, this time. :)


paul

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: message.patch --]
[-- Type: text/x-patch, Size: 1143 bytes --]

--- message.el~	Sun Feb 18 18:09:07 2001
+++ message.el	Sun Feb 18 18:08:39 2001
@@ -3921,7 +3921,8 @@
 	  mct (message-fetch-field "mail-copies-to")
 	  reply-to (message-fetch-field "reply-to")
 	  mrt (message-fetch-field "mail-reply-to")
-	  mft (message-fetch-field "mail-followup-to"))
+	  mft (and message-use-followup-to
+                   (message-fetch-field "mail-followup-to")))
 
     ;; Handle special values of Mail-Copies-To.
     (when mct
@@ -3933,15 +3934,14 @@
 		 (equal (downcase mct) "poster"))
 	     (setq mct (or mrt reply-to from)))))
 
-    (if (and (or (not message-use-followup-to)
-                 (not mft))
+    (if (and (not mft)
              (or (not wide)
                  to-address))
 	(progn
 	  (setq follow-to (list (cons 'To (or to-address mrt reply-to from))))
-	  (when (and (and wide (or mft mct))
-		     (not (member (cons 'To (or mft mct)) follow-to)))
-	    (push (cons 'Cc (or mft mct)) follow-to)))
+	  (when (and (and wide mct)
+		     (not (member (cons 'To mct) follow-to)))
+	    (push (cons 'Cc mct) follow-to)))
       (let (ccalist)
 	(save-excursion
 	  (message-set-work-buffer)

Re: Mail-Followup-To interaction with gnus-summary-followup

[Paul J Collins]

>>>>> "SZ" == ShengHuo ZHU <zsh@cs.rochester.edu> writes:

    SZ> prj@po.cwru.edu (Paul Jarc) writes:
    SZ> [...]

    >> Changelog entry:
    >> 2001-02-17  Paul Jarc <prj@po.cwru.edu>
    >> * message.el: Fix bug with Mail-Followup-To/to-address interaction.
    SZ> Committed.

I'm still seeing this problem, where the Mail-Followup-To header's
value is being used on `f', even when message-use-followup-to is nil.
Strangely, if message-use-followup-to is t, everything works as
expected.

This patch gives me the correct behaviour, but I cannot say from my
cursory reading of `message.el' if it's the Right Way To Do It:

--- message.el  Sun Feb 18 16:37:03 2001
+++ message.el.new      Sun Feb 18 17:00:51 2001
@@ -3941,7 +3941,7 @@
          (setq follow-to (list (cons 'To (or to-address mrt reply-to from))))
          (when (and (and wide (or mft mct))
                     (not (member (cons 'To (or mft mct)) follow-to)))
-           (push (cons 'Cc (or mft mct)) follow-to)))
+           (push (cons 'Cc (or (and message-use-followup-to mft) mct)) follow-to)))
       (let (ccalist)
        (save-excursion
          (message-set-work-buffer)

-- 
Paul Collins <sneakums@sto-kerrig.org> - - - [         ]
 GPG: 0A49 49A9 2932 0EE5 89B2  9EE0 3B65 7154 8131 1BCD
"This one's optimistic, this one went to market;
 this one just came out of the swamp."

Re: Mail-Followup-To interaction with gnus-summary-followup

[Per Abrahamsen]

Paul J Collins <sneakums@sto-kerrig.org> writes:

> Okay, just to be perfectly clear (and perhaps slightly anal); if I
> want `f' to go to the list, I need to set to-address; 

Yep.

> if I want `a' to
> go to the list, I need to set to-list.

Or to-address, either one will work.

> Are there any subtleties that will screw up a strategy where I set
> both to-address and to-list to the address of the list?

It should be unnecessary, to-address alone will (or should) do the
trick.

Re: Mail-Followup-To interaction with gnus-summary-followup

[Paul J Collins]

>>>>> "SZ" == ShengHuo ZHU <zsh@cs.rochester.edu> writes:

    >> Changelog entry:
    >> 2001-02-17  Paul Jarc <prj@po.cwru.edu>
    >> * message.el: Fix bug with Mail-Followup-To/to-address interaction.
    SZ> Committed.

Okay, just to be perfectly clear (and perhaps slightly anal); if I
want `f' to go to the list, I need to set to-address; if I want `a' to
go to the list, I need to set to-list.

Are there any subtleties that will screw up a strategy where I set
both to-address and to-list to the address of the list?

        Paul.

-- 
Paul Collins <sneakums@sto-kerrig.org> - - - [         ]
 GPG: 0A49 49A9 2932 0EE5 89B2  9EE0 3B65 7154 8131 1BCD
"This one's optimistic, this one went to market;
 this one just came out of the swamp."

Re: Mail-Followup-To interaction with gnus-summary-followup

[ShengHuo ZHU]

prj@po.cwru.edu (Paul Jarc) writes:

[...]

> Changelog entry:
> 2001-02-17  Paul Jarc <prj@po.cwru.edu>
> 
> 	* message.el: Fix bug with Mail-Followup-To/to-address interaction.

Committed.

ShengHuo

-- 
(setq gnus-posting-styles 
      '((".*" (signature (format "(setq gnus-posting-styles 
      '%S)" gnus-posting-styles)))))

Re: Mail-Followup-To interaction with gnus-summary-followup

[Paul Jarc]

Paul J Collins <sneakums@sto-kerrig.org> writes:
> The problem in this case is that Mail-Followup-To is not being set the
> the list software, just on an ad-hoc basis by a couple of users.

Mailing lists aren't supposed to set Mail-Followup-To; that would
defeat its purpose.

> Having updated Gnus from CVS just now, I note that it's asking me
> whether to obey Mail-Followup-To after I type `f'.  If I say y, it
> uses the poster's Mail-Followup-To, and I get no Cc.

As it should be, I think.

> If I say n, I get the posters own address in To: and Cc: is set to
> their value of Mail-Followup-To.

That is indeed broken.

> I would really like things to work as they did previously; typing `f'
> with message-use-followup-to set to `nil' or a response of `n' to the
> query results in the To: field being set to the value of to-list.

to-list isn't used for followups, to-address is.  Anyway, here's the
fix.  (Sorry it's not a proper patch, but I haven't learned CVS yet,
so it wouldn't be a patch against an up-to-date version anyway.)  In
message.el, in message-get-reply-headers, look for:
    (if (or (not wide)
            to-address)
and change it to:
    (if (and (or (not message-use-followup-to)
                 (not mft))
             (or (not wide)
                 to-address))
or some logical equivalent.  Can someone commit this?  Here'a a
Changelog entry:
2001-02-17  Paul Jarc <prj@po.cwru.edu>

	* message.el: Fix bug with Mail-Followup-To/to-address interaction.


paul

Re: Mail-Followup-To interaction with gnus-summary-followup

[Paul J Collins]

>>>>> "PJ" == Paul Jarc <prj@po.cwru.edu> writes:

Sorry for the delay in following up.

    PJ> Paul J Collins <sneakums@sto-kerrig.org> writes:
    >> I'm unsure of the semantics of Mail-Followup-To,

    PJ> <URL:http://cr.yp.to/proto/replyto.html>

    >> I'm on a mailing list that I filter into a separate group, and
    >> one of the posters has a Mail-Followup-To header with the
    >> list's address in it.  I had already set the group's parameters
    >> for to-list and to-address to the address of list.

    PJ> I think having to-address makes to-list unnecessary.

I think so too.  I added the to-list after the to-address, and forgot
to take the latter out.

    >> When I type `f' to follow up to this person's posts, my new
    >> message is given a To field with the value of to-address, and a
    >> Cc field with the value of the original poster's
    >> Mail-Followup-To.

    PJ> What is the value of message-use-followup-to?

I think I tried it with t anf nil, to no effect.

    >> I want to disable this behaviour.  Honouring Mail-Followup-To
    >> in the case where the mail is already going to that address
    >> doesn't make much sense to me.

    PJ> I think to-address is intended for a world without
    PJ> Mail-Followup-To, and vice versa.  If the message you're

The problem in this case is that Mail-Followup-To is not being set the
the list software, just on an ad-hoc basis by a couple of users.

Having updated Gnus from CVS just now, I note that it's asking me
whether to obey Mail-Followup-To after I type `f'.  If I say y, it
uses the poster's Mail-Followup-To, and I get no Cc.  If I say n, I
get the posters own address in To: and Cc: is set to their value of
Mail-Followup-To.  In neither case is the old behaviour present.  If I
set message-use-followup-to to `nil'.  I get the latter behaviour.

I would really like things to work as they did previously; typing `f'
with message-use-followup-to set to `nil' or a response of `n' to the
query results in the To: field being set to the value of to-list.

        Paul.

-- 
Paul Collins <sneakums@sto-kerrig.org> - - - [         ]
 GPG: 0A49 49A9 2932 0EE5 89B2  9EE0 3B65 7154 8131 1BCD
"This one's optimistic, this one went to market;
 this one just came out of the swamp."

Re: Mail-Followup-To interaction with gnus-summary-followup

[Paul Jarc]

Paul J Collins <sneakums@sto-kerrig.org> writes:
> I'm unsure of the semantics of Mail-Followup-To,

<URL:http://cr.yp.to/proto/replyto.html>

> I'm on a mailing list that I filter into a separate group, and one of
> the posters has a Mail-Followup-To header with the list's address in
> it.  I had already set the group's parameters for to-list and
> to-address to the address of list.

I think having to-address makes to-list unnecessary.

> When I type `f' to follow up to this person's posts, my new message is
> given a To field with the value of to-address, and a Cc field with the
> value of the original poster's Mail-Followup-To.

What is the value of message-use-followup-to?

> I want to disable this behaviour.  Honouring Mail-Followup-To in the
> case where the mail is already going to that address doesn't make
> much sense to me.

I think to-address is intended for a world without Mail-Followup-To,
and vice versa.  If the message you're following up to already tells
you where the followup should go, you don't need to-address.  Since
you can edit the addresses in the message either way, I think it's
best to prefer Mail-Followup-To (per-message, more specific) over
to-address (per-group, more general).  So message.el should be changed
around 19 lines into message-get-reply-headers:
    (if (or (not wide)
	    to-address)
should be:
    (if (and (not mft)
             (or (not wide)
                 to-address))

You can blame me for this, BTW.  I wrote the MFT code, but didn't
quite understand to-address at the time, I guess.


paul

Re: Mail-Followup-To interaction with gnus-summary-followup

[Kai Großjohann]

On 11 Feb 2001, Paul J. Collins wrote:

> By the way, how is Oort Gnus able to display a bitmap of the Gnus
> logo in GNU Emacs 20.7?

set-stipple-pixmap or similar is the magic word.

kai
-- 
Be indiscrete.  Do it continuously.

Re: Mail-Followup-To interaction with gnus-summary-followup

[Paul J Collins]

>>>>> "SZ" == ShengHuo ZHU <zsh@cs.rochester.edu> writes:

    SZ> Paul J Collins <sneakums@sto-kerrig.org> writes:
    >> When I type `f' to follow up to this person's posts, my new
    >> message is given a To field with the value of to-address, and a
    >> Cc field with the value of the original poster's
    >> Mail-Followup-To.  (I can tell which is which as we have
    >> different things in the non-mail address portion of the
    >> address; they are both of the form: "smart comment
    >> <list@server.tld>").

    SZ> MFT is new in oGnus.

I didn't actually come out and say it, but I want to disable this
behaviour.  Honouring Mail-Followup-To in the case where the mail is
already going to that address doesn't make much sense to me.

    >> By the way, how is Oort Gnus able to display a bitmap of the
    >> Gnus logo in GNU Emacs 20.7?
    SZ> Do you mean Gnus splash message? It is displayed when gnus.el is
    SZ> loaded by a gnus command except gnus-other-frame.

Well, normally it's made out of hyphens, but oGnus seems be able to
somehow display a bitmap on GNU Emacs 20.7, which doesn't have any
kind of graphical capability, to my knowledge.

-- 
Paul Collins <sneakums@sto-kerrig.org> - - - [         ]
 GPG: 0A49 49A9 2932 0EE5 89B2  9EE0 3B65 7154 8131 1BCD
"This one's optimistic, this one went to market;
 this one just came out of the swamp."

Re: Mail-Followup-To interaction with gnus-summary-followup

[Peter Makholm]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 763 bytes --]

ShengHuo ZHU <zsh@cs.rochester.edu> writes:

> > When I type `f' to follow up to this person's posts, my new message is
> > given a To field with the value of to-address, and a Cc field with the
> > value of the original poster's Mail-Followup-To.

> MFT is new in oGnus.

Well, probally time to change to ognus.

But the above behaviour is wrong IMHO. Shouldn't it only use the
content of the MFT-header? Just like a Follow-up header in a news
message don't get reply sent to both the group the message get read in
and the content of the Folloe-Up header.

-- 
  GNU sutter vel i det hele taget på mange måder når man først er
  holdt op med at falde i svime over at det er frit. Synd vi ikke 
  har andet...                                   - Hanne Munkholm

Re: Mail-Followup-To interaction with gnus-summary-followup

[ShengHuo ZHU]

Paul J Collins <sneakums@sto-kerrig.org> writes:

> When I type `f' to follow up to this person's posts, my new message is
> given a To field with the value of to-address, and a Cc field with the
> value of the original poster's Mail-Followup-To.  (I can tell which is
> which as we have different things in the non-mail address portion of
> the address; they are both of the form: "smart comment
> <list@server.tld>").
> 
> I'm running a CVS pull of Oort from this morning; 5.8.8, 5.8.7 and
> 5.8.3 did not previously display this behaviour.

MFT is new in oGnus.

> By the way, how is Oort Gnus able to display a bitmap of the Gnus logo
> in GNU Emacs 20.7?

Do you mean Gnus splash message? It is displayed when gnus.el is
loaded by a gnus command except gnus-other-frame.

ShengHuo

-- 
(setq gnus-posting-styles 
      '((".*" (signature (format "(setq gnus-posting-styles 
      '%S)" gnus-posting-styles)))))

Mail-Followup-To interaction with gnus-summary-followup

[Paul J Collins]

I'm unsure of the semantics of Mail-Followup-To, so perhaps this isn't
a Gnus problem at all.

I'm on a mailing list that I filter into a separate group, and one of
the posters has a Mail-Followup-To header with the list's address in
it.  I had already set the group's parameters for to-list and
to-address to the address of list.

When I type `f' to follow up to this person's posts, my new message is
given a To field with the value of to-address, and a Cc field with the
value of the original poster's Mail-Followup-To.  (I can tell which is
which as we have different things in the non-mail address portion of
the address; they are both of the form: "smart comment
<list@server.tld>").

I'm running a CVS pull of Oort from this morning; 5.8.8, 5.8.7 and
5.8.3 did not previously display this behaviour.

By the way, how is Oort Gnus able to display a bitmap of the Gnus logo
in GNU Emacs 20.7?

-- 
Paul Collins <sneakums@sto-kerrig.org> - - - [         ]
 GPG: 0A49 49A9 2932 0EE5 89B2  9EE0 3B65 7154 8131 1BCD
"This one's optimistic, this one went to market;
 this one just came out of the swamp."