Mail-Followup-To (was Re: IDEA License)

Julian H. Stacey jhs at berklix.com
Sat Mar 30 03:20:12 CET 2013


Peter Lebbing wrote:
> On 27/03/13 14:40, Julian H. Stacey wrote:
> > I created it, as far as I recall, from my copy direct from Ulrich, 
> > which had no Mail-Followup-To
> 
> Correct, the problem originated when you replied[1] to Werner's mail[2].
> Werner's mail had the following header:
> 
> Mail-Followup-To: "Julian H. Stacey" <jhs at berklix.com>, gnupg-users at gnupg.org
> 
> The difference between that line and a simple Reply-to-All is that Werner would
> be in the recipient list with the Reply-to-All, and not with the
> Mail-Followup-To. Your reply should have only had gnupg-users at gnupg.org and your
> manually added CC to Ulrich as recipients, since your MUA would conclude that
> you don't need to CC yourself :).
> 
> > I'm familiar with Reply-to:  Not familar with Mail-Followup-To:
> > What's the difference ?
> 
> Because Reply-To didn't really work out in practice for mailing lists, DJB came
> up with two "non-canon" mail headers to remove ambiguity from the meaning of the
> Reply-To header. He describes it in [3]. Not everybody agrees with his
> view/solution, though.

The quoted [3]<http://cr.yp.to/proto/replyto.html> contains:
	News: The following list is obsolete. Daniel Faber has
	collected a newer list of Mail-Followup-To implementations
	at http://www.leptonite.org/mft/software.html.
which contains refs to claws mail etc ...
	http://www.thewildbeast.co.uk/claws-mail/bugzilla/show_bug.cgi?id=1441
		"Status: 	RESOLVED WONTFIX" ... 2007 2008 ... 
		Colin Leroy 2008-07-05 15:52:44 CEST 
		I'm marking this WONTFIX.
So
	Claws-mail project have no interest to implement Mail-Followup-To ..
	& Claws-mail is a modern mailer (a friend who used to use
	EXMH reckons claws-mail is slicker/ better/ more modern than
	exmh he used & I still use)


http://larve.net/people/hugo/2000/07/ml-mutt
	"It is not a standard .. a hack that can potentially do
	more harm than good"

http://www.ietf.org/rfc/rfc2822.txt
	Includes reply-to
	Does NOT include Followup-To

http://www.ietf.org/proceedings/43/I-D/draft-ietf-drums-mail-followup-to-00.txt
	The ''Mail-Followup-To'' header
	November 1997 ... Internet-Draft

http://tools.ietf.org/html/rfc2076
	3.5 Response control
	...  "ambiguous, since" ... controversial ...  RFC 822 RFC 1036
	author

Reply-to:
	Works fine on lists I run with majordomo on berklix.org
	seems to help lots of people running a variety of MUAs on 
	Microsoft & Unix etc do better than they did before.

Peter off list sent me a PS:
	> Oh, and BTW, I couldn't easily find whether EXMH supports
	> Mail-Followup-To (which makes me lean towards: no, it
	> doesn't, because you'd expect documentation to show up if
	> it did).

I looked (after doing a'make patch' to extract
source trees on latest FreeBS current ports)

	cd /pri/FreeBSD/branches/-current/ports/mail/exmh2
	find . -type f -exec grep -l -i Followup-To {} \;
	
	find . -type f -exec grep -l -i Reply-To {} \;
	./work/exmh-2.8.0/exmh.CHANGES
	./work/exmh-2.8.0/exmh.README
	./work/exmh-2.8.0/exmh.TODO
	./work/exmh-2.8.0/lib/html/exmh-faq.html
	./work/exmh-2.8.0/lib/html/exmh.CHANGES.txt
	./work/exmh-2.8.0/lib/html/reference.html
	./work/exmh-2.8.0/lib/thread.tcl
	./work/exmh-2.8.0/misc/mhthread
	./work/exmh-2.8.0/misc/mhthread-manpage.html
	
	cd /pri/FreeBSD/branches/-current/ports/mail/nmh
	find . -type f -exec grep -l -i Followup-To {} \;
	./work/nmh-1.5/docs/DIFFERENCES
	./work/nmh-1.5/docs/FAQ
	./work/nmh-1.5/docs/TODO
	./work/nmh-1.5/etc/replgroupcomps
	
	find . -type f -exec grep -l -i Reply-To {} \;
	./work/nmh-1.5/ChangeLog
	./work/nmh-1.5/docs/ChangeLog_MH-3_to_MH-6.6
	./work/nmh-1.5/docs/ChangeLog_MH-6.7.0_to_MH-6.8.4.html
	./work/nmh-1.5/docs/DIFFERENCES
	./work/nmh-1.5/docs/FAQ
	./work/nmh-1.5/docs/MAIL.FILTERING
	./work/nmh-1.5/docs/TODO
	./work/nmh-1.5/etc/digestcomps
	./work/nmh-1.5/etc/replcomps
	./work/nmh-1.5/etc/replgroupcomps
	./work/nmh-1.5/man/mh-format.man
	./work/nmh-1.5/test/forw/test-forw-digest
	./work/nmh-1.5/test/repl/test-multicomp
	./work/nmh-1.5/test/repl/test-trailing-newline
	./work/nmh-1.5/uip/forwsbr.c
	./work/nmh-1.5/uip/mhlsbr.c
	./work/nmh-1.5/uip/post.c
	./work/nmh-1.5/uip/rcvdist.c
	./work/nmh-1.5/uip/replsbr.c
	./work/nmh-1.5/uip/slocal.c
	./work/nmh-1.5/uip/spost.c

Conclusion: I will ignore/ forget Followup-To & stick to Reply-To.


Werner wrote:

> To: Peter Lebbing <peter at digitalbrains.com>
> Cc: "Julian H. Stacey" <jhs at berklix.com>, gnupg-users at gnupg.org
> 
> On Wed, 27 Mar 2013 19:27, peter at digitalbrains.com said:
> 
> > Whether you like the headers Bernstein created or not, it would seem Werner
> > didn't want to be on the recipient list, which is why I brought it up
> 
> The thing is that I put most mailing lists I am subscribed to on Gnu's
> message-subscribed-addresses list.  This list takes care of maintaining
> a MFT header.  Gnus will do that only if it can be sure that everyone
> agrees to this.  Thus in most cases you will see an explicit CC anyway.
> MFT works only for those folks with full support of MFT and if they
> maintain their list of subscribed addresses well.  Given that the bad
> habit of sending text+html alternative mails seems to be impossible to
> expunge [1];

Yup, horrible (as also is quoted printable, usually not needed) 

>  I consider missing MFT handling a micro annoyance.
> 
> I any case, I consider it a good idea to explicitly add a To: header to
> notify the addressee that this particular mail gains his attention.

Yup
> 
> BTW, exmh is a nice MUA I used a long time ago and only stopped using it
> because back then a remote X connection was not really usable (and I
> didn't want to use plain mh).

Not sure what remote problems you had, but:
Even localy EXMH reply key does not work right unless one starts
from ttys with xdm & uses xauth. Starting with the ttys login xhost
+ route fails.

A person at my site regularly uses an EXMH on a slow X display
started from xdm, with AMD + NFS ~/mail/ on a faster server, works fine.

Yesterday I was just testing a new EXMH, both with DISPLAY= local laptop screen,
& my tower display, but in both cases exmh running on laptop,
with NFS+AMDsupporting ~/mail , with 493 sub dirs (`find . -type d | wc -l`)
	It took minutes to start. Unusable really, I need to solve that.

I assume one could use ssh to support a tunnel for X for EXMH, but
not tried that as I dont need it.

> Shalom-Salam,
> 
>    Werner
> 
> [1] If you often send mails to Outlook users, you may want to use the
>     X-message-flag header to tell them about this problem.

I run lists with 100s of people, mostly clueless MS users, running
every MUA one can dream of. Less of a dream than a nightmare.

Cheers,
Julian
-- 
Julian Stacey, BSD Unix Linux C Sys Eng Consultant, Munich http://berklix.com
 Reply below not above, like a play script.  Indent old text with "> ".
 Send plain text.  No quoted-printable, HTML, base64, multipart/alternative.



More information about the Gnupg-users mailing list