Re: Forwarding mails

Re: Forwarding mails

[Reiner Steib]

[ Including ding@gnus, as it's becoming (only) Gnus related.  Feel
  free to strip emacs-devel on reply. ]

On Sun, Jan 20 2008, Glenn Morris wrote:

> Reiner Steib wrote:
>> Didn't C-d (`gnus-summary-enter-digest-group') work for you?
>
> Yes (in that it sends me to a buffer with two messages, one empty with
> subject (none) from nobody, one the original message); 

I can't reproduce this.  After one `C-d', I get a summary buffer like
this:

| Martin Svenson        bugfix: python.el: completion in inferior python

With another `C-d', I get:

| m  n/a Martin Svenson        <* mixed> bugfix: python.el: completion [...]
|    n/a Martin Svenson        <1 text>
|    n/a Martin Svenson        <2 text> "python-el-completion-fix.patch"

> but I didn't know about that command.

I didn't know that `gnus-summary-enter-digest-group' can handle it,
but after someone mentioned some rmail undigest command, I just tried
it.  ;-)

> PS given that Gnus can discard duplicate messages, and the mailing
> list can be set to not send you duplicates anyway, 

I'm not subscribed to emacs-devel, I read it via Gmane, so I don't get
duplicates in the first place.

> do you really need to insert a Mail-Followup-To header?

I split Cc-ed mailing list message to special Gnus group, so it's no
big deal to receive Cc's, but they are not necessary for me.  If it is
annoying for you, I try to avoid it in the future.

> It means Gnus starts composing my followup as to: rms, which makes
> little sense, since I'm replying to you.

Does `message-dont-reply-to-names' include your name or address?  In
this case, Gnus shouldn't reply to you, I think (I'm to lazy to look
into the code).

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Forwarding mails

[Glenn Morris]

Reiner Steib wrote:

> Does `message-dont-reply-to-names' include your name or address?  In
> this case, Gnus shouldn't reply to you, I think (I'm to lazy to look
> into the code).

It does. Gnus doesn't reply to me. With your mail:

 To: rgm
 Cc: rms, emacs-devel
 Mail-Followup-To: rgm, rms, emacs-devel

hitting `F' produces a mail:

 To: rms
 Cc: emacs-devel

This doesn't make sense, because I'm not replying to rms, I'm replying
to you. So I have to edit the headers to just be To: emacs-devel.
Ideally, what I'd want Gnus to be able to do is, when replying to
someone who has M-F-T set to exclude their own address, compose a mail
To: the mailing list address, Cc: any other recipients (subject to
message-dont-reply-to-names).

I used to use Mail-Followup-To myself, but someone made the exact same
complaint to me, and while I didn't really get their point initially,
over time I've come to agree. It's not particularly a big deal, but it
does seem a bit of a flawed system.

Probably Gnus has a way to handle it, but Gnus is too complex for me.
:)

Re: Forwarding mails

[Miles Bader]

Glenn Morris <rgm@gnu.org> writes:
>  To: rgm
>  Cc: rms, emacs-devel
>  Mail-Followup-To: rgm, rms, emacs-devel
>
> hitting `F' produces a mail:
>
>  To: rms
>  Cc: emacs-devel
>
> This doesn't make sense, because I'm not replying to rms, I'm replying
> to you.

Why doesn't it make sense?  The MFT sort of says "don't reply to me in
followups, reply to these other people instead", and as you're following
up, Gnus did what was requested (after removing your own address from
that list).  Maybe MFT isn't an ideal mechanism, but until there's some
widely used, reasonable alternative (and there's not AFAIK), it's what
we have.

I'm not sure how Gnus decides what exactly to put in To:, and what in
CC: in the presence of MFT, but given that it _can't_ put the usual
thing (the From: address of the message being replied to) in To:,
grabbing the first address from CC:  seems as good as anything...

-Miles
-- 
Quotation, n. The act of repeating erroneously the words of another. The words
erroneously repeated.

Re: Forwarding mails

[Sven Joachim]

On 2008-01-21 10:14 +0100, Miles Bader wrote:

> I'm not sure how Gnus decides what exactly to put in To:, and what in
> CC: in the presence of MFT, but given that it _can't_ put the usual
> thing (the From: address of the message being replied to) in To:,
> grabbing the first address from CC:  seems as good as anything...

For mailing lists that insert a List-ID header, it would make sense to
put the address from that header into the To: field.  Only if it the
list is included in the MFT field as well, of course.

Sven

Re: Forwarding mails

[Elias Oltmanns]

Sven Joachim <svenjoac@gmx.de> wrote:
> On 2008-01-21 10:14 +0100, Miles Bader wrote:
>
>> I'm not sure how Gnus decides what exactly to put in To:, and what in
>> CC: in the presence of MFT, but given that it _can't_ put the usual
>> thing (the From: address of the message being replied to) in To:,
>> grabbing the first address from CC:  seems as good as anything...
>
> For mailing lists that insert a List-ID header, it would make sense to
> put the address from that header into the To: field.

List-ID is not required to contain an email address and, indeed, it
typically isn't either. Perhaps you mean that Gnus should evaluate the
List-Post header?

Also, it might be worthwhile to compare the addresses in the MFT to a
potential value of the To Address customisation entry of the current group.

Regards,

Elias

Re: Forwarding mails

[Sven Joachim]

On 2008-01-21 21:27 +0100, Elias Oltmanns wrote:

> Sven Joachim <svenjoac@gmx.de> wrote:
>> On 2008-01-21 10:14 +0100, Miles Bader wrote:
>>
>>> I'm not sure how Gnus decides what exactly to put in To:, and what in
>>> CC: in the presence of MFT, but given that it _can't_ put the usual
>>> thing (the From: address of the message being replied to) in To:,
>>> grabbing the first address from CC:  seems as good as anything...
>>
>> For mailing lists that insert a List-ID header, it would make sense to
>> put the address from that header into the To: field.
>
> List-ID is not required to contain an email address and, indeed, it
> typically isn't either. Perhaps you mean that Gnus should evaluate the
> List-Post header?

Perhaps; I just looked up the headers in a random message on this list,
and it seems that ding@gnus.org does not have a List-Post header. :-/

> Also, it might be worthwhile to compare the addresses in the MFT to a
> potential value of the To Address customisation entry of the current group.

Yes, interesting idea.

Sven

Re: Forwarding mails

[Jason L Tibbitts III]

>>>>> "SJ" == Sven Joachim <svenjoac@gmx.de> writes:

SJ> Perhaps; I just looked up the headers in a random message on this
SJ> list, and it seems that ding@gnus.org does not have a List-Post
SJ> header. :-/

This list has probably existed for a decade longer than the standard
regulating List-* headers.  I could add one if people think there's
actually any benefit.

 - J<

Re: Forwarding mails

[Reiner Steib]

On Mon, Jan 21 2008, Elias Oltmanns wrote:

> Sven Joachim <svenjoac@gmx.de> wrote:
>> On 2008-01-21 10:14 +0100, Miles Bader wrote:
>>> I'm not sure how Gnus decides what exactly to put in To:, and what in
>>> CC: in the presence of MFT, but given that it _can't_ put the usual
>>> thing (the From: address of the message being replied to) in To:,
>>> grabbing the first address from CC:  seems as good as anything...

I tend to agree.

>> For mailing lists that insert a List-ID header, it would make sense to
>> put the address from that header into the To: field.
>
> List-ID is not required to contain an email address and, indeed, it
> typically isn't either. Perhaps you mean that Gnus should evaluate the
> List-Post header?

When replying to a message sent to you (and to the list), this won't
work, if you reply to your copy.

> Also, it might be worthwhile to compare the addresses in the MFT to a
> potential value of the To Address customisation entry of the current group.

Same problem as above.  Additionally, I don't have set the to-address
for the vast majority of my groups.

[Cc-ing Glenn, as I don't know if he's reading ding.]

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Forwarding mails

[Greg Troxel]

Miles Bader <miles.bader@necel.com> writes:

> Glenn Morris <rgm@gnu.org> writes:
>>  To: rgm
>>  Cc: rms, emacs-devel
>>  Mail-Followup-To: rgm, rms, emacs-devel
>>
>> hitting `F' produces a mail:
>>
>>  To: rms
>>  Cc: emacs-devel
>>
>> This doesn't make sense, because I'm not replying to rms, I'm replying
>> to you.

You aren't replying, you are following up (err. "wide reply").

> Why doesn't it make sense?  The MFT sort of says "don't reply to me in
> followups, reply to these other people instead", and as you're following
> up, Gnus did what was requested (after removing your own address from
> that list).  Maybe MFT isn't an ideal mechanism, but until there's some
> widely used, reasonable alternative (and there's not AFAIK), it's what
> we have.
>
> I'm not sure how Gnus decides what exactly to put in To:, and what in
> CC: in the presence of MFT, but given that it _can't_ put the usual
> thing (the From: address of the message being replied to) in To:,
> grabbing the first address from CC:  seems as good as anything...

The above example seems incorrect.  MFT has three entries, so the To:
should contain those three; I don't see any justification for putting
some in To and some in Cc.  So I think that should have been:

To: rgm, rms, emacs-devel

or if rgm's MUA if configured not to reply to self:

To: rms, emacs-devel

It would be nice to have gnus-summary-followup-broken-follow-to on S B f
that's like gnus-summary-followup but ignores Mail-Followup-To:.  I use
'S B r' all the time to send private replies on lists that insert a
Reply-To:

gnus-summary-reply-broken-reply-to S B r
  Command: Like `gnus-summary-reply' except removing reply-to field.