Does Mail-Followup-To generation?

Does Mail-Followup-To generation?

[Nevin Kapur]

Following the directions in the manual, I've set the group parameter
of my ding group as

 (to-address . "Ding <ding@gnus.org>")
 (to-list . "Ding <ding@gnus.org>")
 (subscribed . t)

I tried this with another list that I am subscribed to and the
Mail-Followup-To header was not generated.

(If this message has the correct Mail-Followup-To, then I'm an idiot,
otherwise consider this a plea for help.)

Any reason why the subscribed group parameter does not appear in `G c'?

-- 
Nevin

Re: Does Mail-Followup-To generation work?

[Nevin Kapur]

Nevin Kapur <nevin@jhu.edu> writes:

> (If this message has the correct Mail-Followup-To, then I'm an idiot,
> otherwise consider this a plea for help.)

It looks like it is not enough to set the subscribed parameter to t.
One actually needs to set message-subscribed-address-functions?

Is this a documentation bug?

-- 
Nevin

Re: Does Mail-Followup-To generation work?

[Nevin Kapur]

Nevin Kapur <nevin@jhu.edu> writes:

> (If this message has the correct Mail-Followup-To, then I'm an idiot,
> otherwise consider this a plea for help.)

It looks like it is not enough to set the subscribed parameter to t.
One actually needs to set message-subscribed-address-functions?

Is this a documentation bug?

-- 
Nevin

Re: Does Mail-Followup-To generation?

[Kai Großjohann]

Nevin Kapur <nevin@jhu.edu> writes:

>  (to-address . "Ding <ding@gnus.org>")
>  (to-list . "Ding <ding@gnus.org>")
>  (subscribed . t)

I think you should set either to-address, or to-list, but not both.
I think if you set the subscribed parameter (but see also the
variable message-subscribed-addresses), then you don't need to set
to-list or to-address.

(Hm, maybe to-list is useful for the `a' key.)

kai
-- 
Silence is foo!

Re: Does Mail-Followup-To generation?

[Kai Großjohann]

Nevin Kapur <nevin@jhu.edu> writes:

>  (to-address . "Ding <ding@gnus.org>")
>  (to-list . "Ding <ding@gnus.org>")
>  (subscribed . t)

I think you should set either to-address, or to-list, but not both.
I think if you set the subscribed parameter (but see also the
variable message-subscribed-addresses), then you don't need to set
to-list or to-address.

(Hm, maybe to-list is useful for the `a' key.)

kai
-- 
Silence is foo!

Re: Does Mail-Followup-To generation?

[Nevin Kapur]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> Nevin Kapur <nevin@jhu.edu> writes:
>
>>  (to-address . "Ding <ding@gnus.org>")
>>  (to-list . "Ding <ding@gnus.org>")
>>  (subscribed . t)
>
> I think you should set either to-address, or to-list, but not both.
> I think if you set the subscribed parameter (but see also the
> variable message-subscribed-addresses), then you don't need to set
> to-list or to-address.

But the manual gives the impression that setting the subscribed
parameter to t will ensure that the to-list and to-address will be
considered in the generation of MFT.  My cursory examination of the
code indicates that this is not the case.

One needs to set message-subscribed-address-functions to '(gnus-

> (Hm, maybe to-list is useful for the `a' key.)

That's precisely why I have it in there.

-- 
Nevin

Re: Does Mail-Followup-To generation?

[Nevin Kapur]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> Nevin Kapur <nevin@jhu.edu> writes:
>
>>  (to-address . "Ding <ding@gnus.org>")
>>  (to-list . "Ding <ding@gnus.org>")
>>  (subscribed . t)
>
> I think you should set either to-address, or to-list, but not both.
> I think if you set the subscribed parameter (but see also the
> variable message-subscribed-addresses), then you don't need to set
> to-list or to-address.

But the manual gives the impression that setting the subscribed
parameter to t will ensure that the to-list and to-address will be
considered in the generation of MFT.  My cursory examination of the
code indicates that this is not the case.

One needs to set message-subscribed-address-functions to '(gnus-

> (Hm, maybe to-list is useful for the `a' key.)

That's precisely why I have it in there.

-- 
Nevin

Re: Does Mail-Followup-To generation?

[Paul Jarc]

Nevin Kapur <nevin@jhu.edu> wrote:
> Following the directions in the manual, I've set the group parameter
> of my ding group as
>
>  (to-address . "Ding <ding@gnus.org>")
>  (to-list . "Ding <ding@gnus.org>")
>  (subscribed . t)

You also need this:
(setq message-subscribed-address-functions '(gnus-find-subscribed-addresses))

Either the Group Parameters node of the Gnus manual or the docstring
for gnus-find-subscribed-addresses probably ought to link to the
Mailing Lists node of the Message manual.  (I'd do it if I knew how.)

WIBNI all docstrings had links to the corresponding info nodes
automatically added, like how server parameters have a note appended
to theirs?  The right info nodes could be found automatically via the
index (I think); this would also make it easy to catch variables and
functions that are missing index entries.

> Any reason why the subscribed group parameter does not appear in `G c'?

Maybe because no one had thought of that yet, I guess.


paul

Re: Does Mail-Followup-To generation?

[Paul Jarc]

Nevin Kapur <nevin@jhu.edu> wrote:
> Following the directions in the manual, I've set the group parameter
> of my ding group as
>
>  (to-address . "Ding <ding@gnus.org>")
>  (to-list . "Ding <ding@gnus.org>")
>  (subscribed . t)

You also need this:
(setq message-subscribed-address-functions '(gnus-find-subscribed-addresses))

Either the Group Parameters node of the Gnus manual or the docstring
for gnus-find-subscribed-addresses probably ought to link to the
Mailing Lists node of the Message manual.  (I'd do it if I knew how.)

WIBNI all docstrings had links to the corresponding info nodes
automatically added, like how server parameters have a note appended
to theirs?  The right info nodes could be found automatically via the
index (I think); this would also make it easy to catch variables and
functions that are missing index entries.

> Any reason why the subscribed group parameter does not appear in `G c'?

Maybe because no one had thought of that yet, I guess.


paul

[PATCH] Re: Does Mail-Followup-To generation work?

[Nevin Kapur]

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

> Nevin Kapur <nevin@jhu.edu> wrote:
>> Following the directions in the manual, I've set the group parameter
>> of my ding group as
>>
>>  (to-address . "Ding <ding@gnus.org>")
>>  (to-list . "Ding <ding@gnus.org>")
>>  (subscribed . t)
>
> You also need this:
> (setq message-subscribed-address-functions '(gnus-find-subscribed-addresses))

> Either the Group Parameters node of the Gnus manual or the docstring
> for gnus-find-subscribed-addresses probably ought to link to the
> Mailing Lists node of the Message manual.  (I'd do it if I knew how.)

Please consider the following patch which makes the MFT behavior works
as advertised in the manual.

cvs server: Diffing lisp
Index: lisp/ChangeLog
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/ChangeLog,v
retrieving revision 6.1267
diff -u -r6.1267 ChangeLog
--- lisp/ChangeLog	2002/03/01 14:01:25	6.1267
+++ lisp/ChangeLog	2002/03/01 16:53:12
@@ -1,3 +1,11 @@
+2002-03-01  Nevin Kapur  <nevin@jhu.edu>
+
+	* gnus.el (gnus-find-subscribed-addresses): Strip quoted names
+	from to-address and to-list.
+
+	* message.el (message-subscribed-address-functions): Give a
+	default value of '(gnus-find-subscribed-addresses).
+
 2002-03-01  ShengHuo ZHU  <zsh@cs.rochester.edu>
 
 	* mm-view.el (mm-view-pkcs7-verify): New function. A bogus
Index: lisp/gnus.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/gnus.el,v
retrieving revision 6.105
diff -u -r6.105 gnus.el
--- lisp/gnus.el	2002/02/27 17:03:59	6.105
+++ lisp/gnus.el	2002/03/01 16:53:18
@@ -2403,8 +2403,9 @@
     (dolist (entry (cdr gnus-newsrc-alist))
       (setq group (car entry))
       (when (gnus-group-find-parameter group 'subscribed)
-	(setq address (or (gnus-group-fast-parameter group 'to-address)
-			  (gnus-group-fast-parameter group 'to-list)))
+	(setq address (mail-strip-quoted-names
+		       (or (gnus-group-fast-parameter group 'to-address)
+			   (gnus-group-fast-parameter group 'to-list))))
 	(when address
 	  (push address addresses))))
     (list (mapconcat 'regexp-quote addresses "\\|"))))
Index: lisp/message.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/message.el,v
retrieving revision 6.208
diff -u -r6.208 message.el
--- lisp/message.el	2002/02/21 15:42:28	6.208
+++ lisp/message.el	2002/03/01 16:53:30
@@ -445,7 +445,8 @@
 		 (const use)
 		 (const ask)))
 
-(defcustom message-subscribed-address-functions nil
+(defcustom message-subscribed-address-functions
+  '(gnus-find-subscribed-addresses)
   "*Specifies functions for determining list subscription.
 If nil, do not attempt to determine list subscribtion with functions.
 If non-nil, this variable contains a list of functions which return


-- 
Nevin

[PATCH] Re: Does Mail-Followup-To generation work?

[Nevin Kapur]

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

> Nevin Kapur <nevin@jhu.edu> wrote:
>> Following the directions in the manual, I've set the group parameter
>> of my ding group as
>>
>>  (to-address . "Ding <ding@gnus.org>")
>>  (to-list . "Ding <ding@gnus.org>")
>>  (subscribed . t)
>
> You also need this:
> (setq message-subscribed-address-functions '(gnus-find-subscribed-addresses))

> Either the Group Parameters node of the Gnus manual or the docstring
> for gnus-find-subscribed-addresses probably ought to link to the
> Mailing Lists node of the Message manual.  (I'd do it if I knew how.)

Please consider the following patch which makes the MFT behavior works
as advertised in the manual.

cvs server: Diffing lisp
Index: lisp/ChangeLog
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/ChangeLog,v
retrieving revision 6.1267
diff -u -r6.1267 ChangeLog
--- lisp/ChangeLog	2002/03/01 14:01:25	6.1267
+++ lisp/ChangeLog	2002/03/01 16:53:12
@@ -1,3 +1,11 @@
+2002-03-01  Nevin Kapur  <nevin@jhu.edu>
+
+	* gnus.el (gnus-find-subscribed-addresses): Strip quoted names
+	from to-address and to-list.
+
+	* message.el (message-subscribed-address-functions): Give a
+	default value of '(gnus-find-subscribed-addresses).
+
 2002-03-01  ShengHuo ZHU  <zsh@cs.rochester.edu>
 
 	* mm-view.el (mm-view-pkcs7-verify): New function. A bogus
Index: lisp/gnus.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/gnus.el,v
retrieving revision 6.105
diff -u -r6.105 gnus.el
--- lisp/gnus.el	2002/02/27 17:03:59	6.105
+++ lisp/gnus.el	2002/03/01 16:53:18
@@ -2403,8 +2403,9 @@
     (dolist (entry (cdr gnus-newsrc-alist))
       (setq group (car entry))
       (when (gnus-group-find-parameter group 'subscribed)
-	(setq address (or (gnus-group-fast-parameter group 'to-address)
-			  (gnus-group-fast-parameter group 'to-list)))
+	(setq address (mail-strip-quoted-names
+		       (or (gnus-group-fast-parameter group 'to-address)
+			   (gnus-group-fast-parameter group 'to-list))))
 	(when address
 	  (push address addresses))))
     (list (mapconcat 'regexp-quote addresses "\\|"))))
Index: lisp/message.el
===================================================================
RCS file: /usr/local/cvsroot/gnus/lisp/message.el,v
retrieving revision 6.208
diff -u -r6.208 message.el
--- lisp/message.el	2002/02/21 15:42:28	6.208
+++ lisp/message.el	2002/03/01 16:53:30
@@ -445,7 +445,8 @@
 		 (const use)
 		 (const ask)))
 
-(defcustom message-subscribed-address-functions nil
+(defcustom message-subscribed-address-functions
+  '(gnus-find-subscribed-addresses)
   "*Specifies functions for determining list subscription.
 If nil, do not attempt to determine list subscribtion with functions.
 If non-nil, this variable contains a list of functions which return


-- 
Nevin

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Josh Huber]

Nevin Kapur <nevin@jhu.edu> writes:

> +	* gnus.el (gnus-find-subscribed-addresses): Strip quoted names
> +	from to-address and to-list.

This is good -- I guess I never thought to put a comment/name in the
to-list/to-address parameters. :)

> +	* message.el (message-subscribed-address-functions): Give a
> +	default value of '(gnus-find-subscribed-addresses).

When I first wrote the MFT code, I had this but later removed it
becuase message.el is not supposed to have any dependancies on the
gnus code AFAIUI...

ttyl,

-- 
Josh Huber

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Josh Huber]

Nevin Kapur <nevin@jhu.edu> writes:

> +	* gnus.el (gnus-find-subscribed-addresses): Strip quoted names
> +	from to-address and to-list.

This is good -- I guess I never thought to put a comment/name in the
to-list/to-address parameters. :)

> +	* message.el (message-subscribed-address-functions): Give a
> +	default value of '(gnus-find-subscribed-addresses).

When I first wrote the MFT code, I had this but later removed it
becuase message.el is not supposed to have any dependancies on the
gnus code AFAIUI...

ttyl,

-- 
Josh Huber

Re: Does Mail-Followup-To generation?

[Kai Großjohann]

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

> Either the Group Parameters node of the Gnus manual or the docstring
> for gnus-find-subscribed-addresses probably ought to link to the
> Mailing Lists node of the Message manual.  (I'd do it if I knew how.)

How about adding a link from the explanation of the "subscribed"
group parameter to gnus-find-subscribed-addresses and/or
message-subscribed-address-functions?  Then people who do "G c" on a
group will see what they need to do.

kai
-- 
Silence is foo!

Re: Does Mail-Followup-To generation?

[Kai Großjohann]

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

> Either the Group Parameters node of the Gnus manual or the docstring
> for gnus-find-subscribed-addresses probably ought to link to the
> Mailing Lists node of the Message manual.  (I'd do it if I knew how.)

How about adding a link from the explanation of the "subscribed"
group parameter to gnus-find-subscribed-addresses and/or
message-subscribed-address-functions?  Then people who do "G c" on a
group will see what they need to do.

kai
-- 
Silence is foo!

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Nevin Kapur]

Josh Huber <huber@alum.wpi.edu> writes:

> Nevin Kapur <nevin@jhu.edu> writes:
>
>> +	* message.el (message-subscribed-address-functions): Give a
>> +	default value of '(gnus-find-subscribed-addresses).
>
> When I first wrote the MFT code, I had this but later removed it
> becuase message.el is not supposed to have any dependancies on the
> gnus code AFAIUI...

How about something like 
(if (featurep 'gnus) '(gnus-find-subscribed-addresses)) ?

Or something in Gnus that could modify
message-subscribed-address-functions when Gnus is loaded?   I guess
one has to be careful about this, not overwriting a user configuration.

-- 
Nevin

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Nevin Kapur]

Josh Huber <huber@alum.wpi.edu> writes:

> Nevin Kapur <nevin@jhu.edu> writes:
>
>> +	* message.el (message-subscribed-address-functions): Give a
>> +	default value of '(gnus-find-subscribed-addresses).
>
> When I first wrote the MFT code, I had this but later removed it
> becuase message.el is not supposed to have any dependancies on the
> gnus code AFAIUI...

How about something like 
(if (featurep 'gnus) '(gnus-find-subscribed-addresses)) ?

Or something in Gnus that could modify
message-subscribed-address-functions when Gnus is loaded?   I guess
one has to be careful about this, not overwriting a user configuration.

-- 
Nevin

Re: Does Mail-Followup-To generation?

[Christoph Conrad]

    > WIBNI all docstrings had links to the corresponding info nodes
    > automatically added, like how server parameters have a note

OH YES!

Re: Does Mail-Followup-To generation?

[Christoph Conrad]

    > WIBNI all docstrings had links to the corresponding info nodes
    > automatically added, like how server parameters have a note

OH YES!

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Kai Großjohann]

Nevin Kapur <nevin@jhu.edu> writes:

> Or something in Gnus that could modify
> message-subscribed-address-functions when Gnus is loaded?   I guess
> one has to be careful about this, not overwriting a user configuration.

The default value for message-subscribed-address-functions is nil, so
something like the following during Gnus startup might help:

(setq message-subscribed-address-functions
      (or message-subscribed-address-functions
          '(gnus-find-subscribed-addresses)))

Thoughts?

kai
-- 
Silence is foo!

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Kai Großjohann]

Nevin Kapur <nevin@jhu.edu> writes:

> Or something in Gnus that could modify
> message-subscribed-address-functions when Gnus is loaded?   I guess
> one has to be careful about this, not overwriting a user configuration.

The default value for message-subscribed-address-functions is nil, so
something like the following during Gnus startup might help:

(setq message-subscribed-address-functions
      (or message-subscribed-address-functions
          '(gnus-find-subscribed-addresses)))

Thoughts?

kai
-- 
Silence is foo!

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Paul Jarc]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) wrote:
> Nevin Kapur <nevin@jhu.edu> writes:
>> Or something in Gnus that could modify
>> message-subscribed-address-functions when Gnus is loaded?   I guess
>> one has to be careful about this, not overwriting a user configuration.
>
> The default value for message-subscribed-address-functions is nil, so
> something like the following during Gnus startup might help:
>
> (setq message-subscribed-address-functions
>       (or message-subscribed-address-functions
>           '(gnus-find-subscribed-addresses)))

If the user explicitly sets it to nil, Gnus shouldn't override that.
We could change the default to '(ignore), which the user wouldn't do,
and then Gnus could change it if it's set to '(ignore).  I suggested
this earlier, possibly for a different variable, but it doesn't seem
to be in use now, so maybe there was a reason for removing it.


paul

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Paul Jarc]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) wrote:
> Nevin Kapur <nevin@jhu.edu> writes:
>> Or something in Gnus that could modify
>> message-subscribed-address-functions when Gnus is loaded?   I guess
>> one has to be careful about this, not overwriting a user configuration.
>
> The default value for message-subscribed-address-functions is nil, so
> something like the following during Gnus startup might help:
>
> (setq message-subscribed-address-functions
>       (or message-subscribed-address-functions
>           '(gnus-find-subscribed-addresses)))

If the user explicitly sets it to nil, Gnus shouldn't override that.
We could change the default to '(ignore), which the user wouldn't do,
and then Gnus could change it if it's set to '(ignore).  I suggested
this earlier, possibly for a different variable, but it doesn't seem
to be in use now, so maybe there was a reason for removing it.


paul

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Kai Großjohann]

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

> If the user explicitly sets it to nil, Gnus shouldn't override that.

Hm.

But on the other hand, why not change the MFT code to always look at
the group parameters?  Surely that can't be so harmful?  After all,
people who don't want to use the group parameter are unlikely to set
it: not setting the group parameter is enough to have Gnus behave
like before.  Right?

Thoughts?

kai
-- 
Silence is foo!

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Kai Großjohann]

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

> If the user explicitly sets it to nil, Gnus shouldn't override that.

Hm.

But on the other hand, why not change the MFT code to always look at
the group parameters?  Surely that can't be so harmful?  After all,
people who don't want to use the group parameter are unlikely to set
it: not setting the group parameter is enough to have Gnus behave
like before.  Right?

Thoughts?

kai
-- 
Silence is foo!

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Paul Jarc]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) wrote:
> But on the other hand, why not change the MFT code to always look at
> the group parameters?

Because it's in message.el and thus shouldn't depend on Gnus group
parameters, I guess.  But it could look at the group parameters only
if available, as well as the variables it uses now.  Would that
violate any encapsulation principle here?


paul

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Paul Jarc]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) wrote:
> But on the other hand, why not change the MFT code to always look at
> the group parameters?

Because it's in message.el and thus shouldn't depend on Gnus group
parameters, I guess.  But it could look at the group parameters only
if available, as well as the variables it uses now.  Would that
violate any encapsulation principle here?


paul

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Kai Großjohann]

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

> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) wrote:
>> But on the other hand, why not change the MFT code to always look at
>> the group parameters?
>
> Because it's in message.el and thus shouldn't depend on Gnus group
> parameters, I guess.

Oh.  Hm.

> But it could look at the group parameters only if available, as well
> as the variables it uses now.  Would that violate any encapsulation
> principle here?

I hope not.  Hm.  Maybe it could call gnus-find-subscribed-addresses
if that function is bound.

kai
-- 
Silence is foo!

Re: [PATCH] Re: Does Mail-Followup-To generation work?

[Kai Großjohann]

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

> Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) wrote:
>> But on the other hand, why not change the MFT code to always look at
>> the group parameters?
>
> Because it's in message.el and thus shouldn't depend on Gnus group
> parameters, I guess.

Oh.  Hm.

> But it could look at the group parameters only if available, as well
> as the variables it uses now.  Would that violate any encapsulation
> principle here?

I hope not.  Hm.  Maybe it could call gnus-find-subscribed-addresses
if that function is bound.

kai
-- 
Silence is foo!

Re: Does Mail-Followup-To generation?

[Lars Magne Ingebrigtsen]

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

> WIBNI all docstrings had links to the corresponding info nodes
> automatically added, like how server parameters have a note appended
> to theirs?  The right info nodes could be found automatically via the
> index (I think); this would also make it easy to catch variables and
> functions that are missing index entries.

And Custom already makes all the user-visible variables group, so
you'd just need a mapping from group->info file.  (And then use the
variable index to find the thing you're looking for.)

Sounds like a nice addition to Custom.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen