">" -vs- "Re:" -- is this guy crazy?

">" -vs- "Re:" -- is this guy crazy?

[Matt Armstrong]

I had never seen ">" used instead of "Re:" -- has anyone else?  It
breaks Gnus threading (as readers of the ntemacs list will have
noticed).

I know Gnus is customizable in this regard, but if ">" is as common as
this guy says it is, perhaps Gnus should handle it by default?

The fact that this guy insulted my mail reader is just _barely_
mitigated by the fact that he uses Emacs. :-)

-- 
matta


------- Start of forwarded message -------
Date: Tue, 24 Feb 1998 16:14:19 -0500
From: David Biesack <sasdjb@unx.sas.com>
To: matta@geoworks.com
Subject: > > > Getting jde/cc-mode/custom to work

> From: Matt Armstrong <matta@geoworks.com>
> Date: 24 Feb 1998 11:13:42 -0800
> 
> Might I ask what mail program you're using?
> 
> Its use of ">" instead of "Re:" is confusing my reader (which thinks
> your messages are a new thread instead of a reply to something).
> 
> -- 
> matta

I use Emacs RMAIL to read mail and emacs sendmail to reply to mail. "> "
is a standard way of responding to mail in a subject, probably as common
as "Re: ". I'd say your mail reader is broken :-)

-- 
David J. Biesack                                    sasdjb@unx.sas.com
Object Programming Technology                  (919) 677-8000 ext 7771
SAS Institute Inc. SAS Campus Drive Cary, NC 27513  http://www.sas.com
<<<<   Join the Coalition Against UCE at http://www.cauce.org/    >>>>

------- End of forwarded message -------

Re: ">" -vs- "Re:" -- is this guy crazy?

[Karl Kleinpaste]

Matt Armstrong <matta@geoworks.com> writes:
> I had never seen ">" used instead of "Re:" -- has anyone else?

I've been hacking mailers and news agents since RFC733 specified the
ARPANET mail standard, and I can't recall ever once in almost 20 years
having seen ">" used in a Subject: line as a replacement for "Re: ".

Whatever it is he's smoking, with a stash that large, he should share.

Re: ">" -vs- "Re:" -- is this guy crazy?

[Wes Hardaker]

>>>>> On 24 Feb 1998 13:41:38 -0800, Matt Armstrong <matta@geoworks.com> said:

MA> I had never seen ">" used instead of "Re:" -- has anyone else?  It
MA> breaks Gnus threading (as readers of the ntemacs list will have
MA> noticed).

My personal favorite:  -Reply at the end of the subject, regardless if 
there is one already present...  Needless to say, the subject lines
can get pretty long...

-- 
"Ninjas aren't dangerous.  They're more afraid of you than you are of them."

Re: ">" -vs- "Re:" -- is this guy crazy?

[SL Baur]

[To answer the question in the Subject:, yes]

Matt Armstrong <matta@geoworks.com> writes:

> I had never seen ">" used instead of "Re:" -- has anyone else?

No.  I've never seen it.  There is a (broken) Microsoft mailer that
uses `FW: '.  There is a (broken) Novell mailer that appends a
`-Reply'.  There are various (broken) Unix mailers that use `Re[2]: '
`Re[3]: ', etc.  VM doesn't add anything by default.

The only acceptable thing to do is to prepend a `Re: ' if one is not
already there, otherwise don't add anything.

> It breaks Gnus threading (as readers of the ntemacs list will have
> noticed).

It shouldn't if the References: header is properly formed.

> I know Gnus is customizable in this regard, but if ">" is as common as
> this guy says it is, perhaps Gnus should handle it by default?

Since when has Gnus been customizable in this area?

> The fact that this guy insulted my mail reader is just _barely_
> mitigated by the fact that he uses Emacs. :-)

Re: ">" -vs- "Re:" -- is this guy crazy?

[Matt Armstrong]

SL Baur <steve@xemacs.org> writes:

> [To answer the question in the Subject:, yes]
> 
> Matt Armstrong <matta@geoworks.com> writes:
> 
> > I had never seen ">" used instead of "Re:" -- has anyone else?
> 
> No.  I've never seen it.

Yeah, apparently RMAIL lets you customize your reply prefix.  He had
it set to "> " to match the behavior he saw from nn.  He also said he
has had it that way for 12 years without a single comment.

He has switched back to the default RMAIL "Re:" prefix.


> > It breaks Gnus threading (as readers of the ntemacs list will have
> > noticed).
> 
> It shouldn't if the References: header is properly formed.

His posts didn't have a References: header.  Perhaps another RMAILizm.
I think so, because I remember posts from RMS breaking threading as
well -- he uses RMAIL last I heard.


> > I know Gnus is customizable in this regard, but if ">" is as common as
> > this guy says it is, perhaps Gnus should handle it by default?
> 
> Since when has Gnus been customizable in this area?

I was talking about things like gnus-simplify-ignored-prefixes, and
gnus-simplify-subject-fuzzy-regexp.  Hell, there is even a
gnus-simplify-subject-functions list of simplify functions.


-- 
matta

Re: ">" -vs- "Re:" -- is this guy crazy?

[Per Abrahamsen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=US-ASCII, Size: 496 bytes --]

Matt Armstrong <matta@geoworks.com> writes:

> He had it set to "> " to match the behavior he saw from nn.

NN never _posted_ "> ".  It _recognized_ "Re: " and _displayed_ it as
"> ".  That is probably why David Biesack came to believe "> " was
common.

Early versions of NN would _post_ replies to replies as "Re^2: ", and
replies to replies to replies as "Re^3: ", etc., and _display_ these
as ">> " and ">>> ".  Obviously, the NN author was flamed
appropriately, and changed to just "Re: ".

Re: ">" -vs- "Re:" -- is this guy crazy?

[Dave Love]

>>>>> "Matt" == Matt Armstrong <matta@geoworks.com> writes:

 Matt> His posts didn't have a References: header.  Perhaps another RMAILizm.
 Matt> I think so, because I remember posts from RMS breaking threading as
 Matt> well -- he uses RMAIL last I heard.

FWIW rms' replies do (now?) have References:.  Those from Cygnus
people typically don't :-(.

Re: ">" -vs- "Re:" -- is this guy crazy?

[Per Abrahamsen]

Dave Love <d.love@dl.ac.uk> writes:

> FWIW rms' replies do (now?) have References:.  Those from Cygnus
> people typically don't :-(.

RMAIL 20.1 maintains the references header in replies.  Versions
before that didn't.  

But the old RMAIL did put the right message-id in the "In-Reply-To:"
header, so the messsages do thread well in Gnus.

in-reply-to: threading [was Re: ">" -vs- "Re:" -- is this guy crazy?]

[Dave Love]

>>>>> "Per" == Per Abrahamsen <abraham@dina.kvl.dk> writes:

 Per> But the old RMAIL did put the right message-id in the
 Per> "In-Reply-To:" header, so the messsages do thread well in Gnus.

That's not what I see :-(.  (Although I don't know with what all the
messages were composed, they do have In-Reply-To: which
gnus-summary-refer-parent-article groks).  Here's a summary fragment
from the sort of high-volume list for which one wants Gnus:

O  [ 365: H.J. Lu             ] HELP: egcs 1.0.2 and linux kernel
E      <  23: H.J. Lu             > 
E      <  17: Jeffrey A Law       > 
E      <  14: Jeffrey A Law       > 
E      <  28: H.J. Lu             > 

Numbering the `E' articles from the top, they should form the chain 1 ->
2 -> 4 ->3.  I have '(gnus-simplify-whitespace
gnus-simplify-subject-re) on gnus-simplify-subject-functions to cope
with Law's trailing blank, sigh.  Sample headers:

In-Reply-To: <29794.888557423@hurl.cygnus.com> from "Jeffrey A Law" at Feb 26, 98 10:30:23 pm
In-reply-to: Your message of Thu, 26 Feb 1998 22:08:37 MST.
             <m0y8IyL-00058nC@ocean.lucon.org> 

I haven't tried to figure out what's happening other than a bit of
RTFM.  Is there something obvious one should (not?) be doing to get
this right?

[Due respect to Ken, Jason, et al, but there's a substantial weighting
factor involved :-).]

Re: in-reply-to: threading [was Re: ">" -vs- "Re:" -- is this guy crazy?]

[Lars Magne Ingebrigtsen]

Dave Love <d.love@dl.ac.uk> writes:

> In-Reply-To: <29794.888557423@hurl.cygnus.com> from "Jeffrey A Law" at Feb 26, 98 10:30:23 pm
> In-reply-to: Your message of Thu, 26 Feb 1998 22:08:37 MST.
>              <m0y8IyL-00058nC@ocean.lucon.org> 

Well, what are the Message-ID headers of the articles and stuff?

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

Re: in-reply-to: threading [was Re: ">" -vs- "Re:" -- is this guy crazy?]

[Per Abrahamsen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Dave Love <d.love@dl.ac.uk> writes:
> 
> > In-Reply-To: <29794.888557423@hurl.cygnus.com> from "Jeffrey A Law" at Feb 26, 98 10:30:23 pm
> > In-reply-to: Your message of Thu, 26 Feb 1998 22:08:37 MST.
> >              <m0y8IyL-00058nC@ocean.lucon.org> 
> 
> Well, what are the Message-ID headers of the articles and stuff?

Correct.  As far as I can see, the references field in the
egcs/list/.overview file is empty.  Shouldn't it contain the
message-id from the In-reply-field?

Consequently, evaluating this

	(setq nnml-nov-is-evil t)

repairs the threading.

Re: in-reply-to: threading [was Re: ">" -vs- "Re:" -- is this guy crazy?]

[Dave Love]

>>>>> "Per" == Per Abrahamsen <abraham@dina.kvl.dk> writes:

 Per> As far as I can see, the references field in the
 Per> egcs/list/.overview file is empty.  Shouldn't it contain the
 Per> message-id from the In-reply-field?

I sent this to the Towers, but probably should have copied here in
case anyone else follows the clue.  After it, a nov regen does good
things and I finally have Law and order.

1998-03-03  Dave Love  <d.love@dl.ac.uk>

	* nnheader.el (nnheader-parse-head): Fix in-reply-to code. Return
	nil consistently if not found.

*** nnheader.el	1998/03/03 23:01:13	1.1
--- nnheader.el	1998/03/03 23:23:18
***************
*** 241,248 ****
  		       (setq ref2 (substring in-reply-to (match-beginning 0)
  					     (match-end 0)))
  		       (when (> (length ref2) (length ref))
! 			 (setq ref ref2))))
! 		 "")))
  	   ;; Chars.
  	   0
  	   ;; Lines.
--- 241,249 ----
  		       (setq ref2 (substring in-reply-to (match-beginning 0)
  					     (match-end 0)))
  		       (when (> (length ref2) (length ref))
! 			 (setq ref ref2)))
!                      ref)
! 		 nil)))
  	   ;; Chars.
  	   0
  	   ;; Lines.

Re: ">" -vs- "Re:" -- is this guy crazy?

[Ken Raeburn]

Dave Love <d.love@dl.ac.uk> writes:

> FWIW rms' replies do (now?) have References:.  Those from Cygnus
> people typically don't :-(.

Hey!  Some of us do use decent mailreaders. :-)
Even if we still do gripe about performance issues. :-(

The mail agents that annoy me are the ones that don't even provide
in-reply-to lines either, so even if we use hacks to extract
message-ids from such lines, we still get no useful threading data on
these people's mail.  It's like they're trying to make you pay close
attention or something.  :-)

Ken

Re: ">" -vs- "Re:" -- is this guy crazy?

[lantz moore]

>>>>> "sb" == SL Baur <steve@xemacs.org> writes:

>> It breaks Gnus threading (as readers of the ntemacs list will have
>> noticed).

sb> It shouldn't if the References: header is properly formed.

this is a sketchy assumption given that the guy is using ">" instead of
"Re". 

:)

-- 
lantz moore, contigo software                         lantz_moore@contigo.com

Re: ">" -vs- "Re:" -- is this guy crazy?

[Christian Nybø]

SL Baur <steve@xemacs.org> writes:

[...]
> No.  I've never seen it.  There is a (broken) Microsoft mailer that
> uses `FW: '.  

Friends from school working in corporate environments sometimes answer
mail with 'SV: ' instead of 'Re: '. SV are the first two letters in
the norwegian word 'svar', which means reply. I have seen german
variations as well, I think.

gnus to the rescue: I think setting 'gnus-cite-reply-regexp' could help
you correspond  with these not-so-well-informed beings.

-- 
chr

Re: ">" -vs- -- is this guy crazy?

[jari.aalto]

 98-02-25 chr@mediascience.no (Christian Nybø) list.ding
| SL Baur <steve@xemacs.org> writes:
| 
| > No.  I've never seen it.  There is a (broken) Microsoft mailer that
| > uses `FW: '.  
| 
| Friends from school working in corporate environments sometimes answer
| mail with 'SV: '

It's definitely some MS product that uses "SV:", because I have seen
it too many times.  In addition MS Explorer uses "Vs:" for substitute
of "Re:", Japanese MUA Denshin 8 Go V321.1b7 uses Re^2 and and
some other MUAs seem to use  "Re2:" or "Re[2]"

I got tired of seeing 

    Subject: Re^2: Re[32]: FW: Re: Re[15]: Sv: Re[9]: test message

So I made a procmail plug-in module to convert message subjects like
this into

    Subject: Re: test message

jari

Re: ">" -vs- "Re:" -- is this guy crazy?

[Steinar Bang]

>>>>> chr@mediascience.no (Christian Nybø):

> SL Baur <steve@xemacs.org> writes:
> [...]
>> No.  I've never seen it.  There is a (broken) Microsoft mailer that
>> uses `FW: '.  

> Friends from school working in corporate environments sometimes answer
> mail with 'SV: ' instead of 'Re: '. SV are the first two letters in
> the norwegian word 'svar', which means reply. I have seen german
> variations as well, I think.

I have the followin rule in my ~/.procmailrc to deal with these.  What
*really* makes this interesting, is that occasionally these beasties
will appear in RFC 2047 encoded subject fields.

Everything in the hairy perl regexps below, is based on actually
received examples that escaped what I thought was a waterproof
solution.  It was soon too complex for sed, so I had to move on to
perl.

#
# Fix Subject RE/SV/AW line for MSExchange/MSMail/MSOutlook
# No X-Mailer field to rely on, unfortunately!
# Remove MSE quotes of type "'Real Name'"
#
:0 fhw
| /opt/pub/bin/perl -ne "s/^(Subject:)\s+(SV:\s*|AW:\s*|R[eE](\(\d+\))?:\s*)+/\1 Re: /g;s/^(Subject:)\s+(=\?((iso|ISO)-8859-1|us-ascii)\?Q\?)((SV|AW|R[eE](\(\d+\))?)(:|=3[Aa])_*)+/\1 \2Re:_/g; s/\042\'([^\']*)\'\042/\042\1\042/g ; print"

Outlook is so incredibly broken, in so many ways, it simply boggles
the mind...

(For the record: the first regexp also matches Re(2):, Re(3)
etc. which can't be blamed on Outlook)

Re: ">" -vs- "Re:" -- is this guy crazy?

[Justin Sheehy]

SL Baur <steve@xemacs.org> writes:

> No.  I've never seen it.  There is a (broken) Microsoft mailer that
> uses `FW: '.  There is a (broken) Novell mailer that appends a
> `-Reply'.  There are various (broken) Unix mailers that use `Re[2]: '
> `Re[3]: ', etc.  VM doesn't add anything by default.
> 
> The only acceptable thing to do is to prepend a `Re: ' if one is not
> already there, otherwise don't add anything.

I notice that you nicely avoid directly saying that VM is broken in
this regard, even though it is obvious from the second sentence above
that you agree that it is.  Correct me if I've misread.

I'll come out and say it: this behavior of VM is broken.

> > It breaks Gnus threading (as readers of the ntemacs list will have
> > noticed).
> 
> It shouldn't if the References: header is properly formed.

Lots of mail clients don't create a References: header, or create bad
ones.

One of my current pet peeves: the way that mh breaks threading.

If you reply/followup to a message, it does the right thing with Re:
but then appends a space to the Subject: as well.  

I know that Gnus can be customized to ignore the space, but that does
not change that it is broken behavior on mh's part.

I had hoped that this would be fixed in nmh, but I have seen similarly
broken messages coming from nmh users.

Feh.

> Since when has Gnus been customizable in this area?

See the variables gnus-simplify-*

-- 
Justin Sheehy

In a cloud bones of steel.
  

Re: ">" -vs- "Re:" -- is this guy crazy?

[Steinar Bang]

>>>>> Justin Sheehy <justin@linus.mitre.org>:

> One of my current pet peeves: the way that mh breaks threading.

> If you reply/followup to a message, it does the right thing with Re:
> but then appends a space to the Subject: as well.  

> I know that Gnus can be customized to ignore the space, but that does
> not change that it is broken behavior on mh's part.

> I had hoped that this would be fixed in nmh, but I have seen similarly
> broken messages coming from nmh users.

It can be fixed with an appropriate replcomps file.  Something like
this: 

%(lit)%(formataddr %<{reply-to}%?{from}%?{sender}%?{return-path}%>)\
%<(nonnull)%(void(width))%(putaddr To: )\n%>\
%(lit)%(formataddr{to})%(formataddr{cc})%(formataddr(me))\
%<(nonnull)%(void(width))%(putaddr cc: )\n%>\
%<{subject}%(trim)Subject: Re: %(putstr)\n%>\
%<{date}In-reply-to: Your message of "\
%<(nodate{date})%{date}%|%(pretty{date})%>."%<{message-id}
             %{message-id}%>\n%>\
%<{message-id}References: \
%<{references}%(void{references})%(trim)%(putstr) %>\
%(void{message-id})%(trim)%(putstr)\n%>\
--------

No space at the end of the Subject, and inserts a references header

Re: ">" -vs- "Re:" -- is this guy crazy?

[Justin Sheehy]

Steinar Bang <sb@metis.no> writes:

> >>>>> Justin Sheehy <justin@linus.mitre.org>:
> 
> > One of my current pet peeves: the way that mh breaks threading.

> It can be fixed with an appropriate replcomps file.  Something like
> this: 

[snip]

I know this; before Gnus, I used mh for my mail and I did something
similar.  Each and every user should not have to unbreak their
software.

mh is broken.

-- 
Justin Sheehy

In a cloud bones of steel.
  

Re: ">" -vs- "Re:" -- is this guy crazy?

[Steinar Bang]

>>>>> Justin Sheehy <justin@linus.mitre.org>:

> Steinar Bang <sb@metis.no> writes:
>> >>>>> Justin Sheehy <justin@linus.mitre.org>:
>> 
>> > One of my current pet peeves: the way that mh breaks threading.

>> It can be fixed with an appropriate replcomps file.  Something like
>> this: 

> [snip]

> I know this; before Gnus, I used mh for my mail and I did something
> similar.  Each and every user should not have to unbreak their
> software.

> mh is broken.

Yup.  But it is "unbreakable" at the installation configuration level
(which is what I did.  The users I installed mh for, never even knew
that their MH worked correctly...;-) )

Re: ">" -vs- "Re:" -- is this guy crazy?

[Richard Coleman]

> > One of my current pet peeves: the way that mh breaks threading.
> 
> > If you reply/followup to a message, it does the right thing with Re:
> > but then appends a space to the Subject: as well.  
> 
> > I had hoped that this would be fixed in nmh, but I have seen similarly
> > broken messages coming from nmh users.

This is the first I've heard of this problem.  Rather than just having
the default replcomps remove that trailing space, I would prefer to
figure out where is it coming from.  I'll look into it, and try to
fix it for nmh.

--
Richard Coleman (nmh maintainer)
coleman@math.gatech.edu

Re: ">" -vs- "Re:" -- is this guy crazy?

[Jason R Mastaler]

Richard Coleman <coleman@math.gatech.edu> writes:

> This is the first I've heard of this problem.  Rather than just having
> the default replcomps remove that trailing space, I would prefer to
> figure out where is it coming from.  I'll look into it, and try to
> fix it for nmh.

But how about having the default replcomps add a References header?
That would help too.

Re: ">" -vs- "Re:" -- is this guy crazy?

[David S. Goldberg]

> But how about having the default replcomps add a References header?
> That would help too.

Especially since it's not all that difficult to do.  I built a
replcomps based on advice from the O'Reilly book (or, possibly, email
from the author of that book -- I don't recall) that allowed me to do
that.  Of course I don't actually use it anymore...  If anyone's
interested I can mail it to them.  It seems to me that fixing MH is
not really the right discussion for this list.

Dave Goldberg
Post: The Mitre Corporation\MS B305\202 Burlington Rd.\Bedford, MA 01730
Phone: 781-271-3887
Email: dsg@mitre.org

Re: ">" -vs- "Re:" -- is this guy crazy?

[SL Baur]

Justin Sheehy <justin@linus.mitre.org> writes:

> SL Baur <steve@xemacs.org> writes:
>> VM doesn't add anything by default.

> I notice that you nicely avoid directly saying that VM is broken in
> this regard, even though it is obvious from the second sentence above
> that you agree that it is.  Correct me if I've misread.

You are correct.

> I'll come out and say it: this behavior of VM is broken.

I've complained more than once.  Kyle's response is that since `Re: '
is a netnews requirement and not an RFC822 one VM isn't broken.

>> Since when has Gnus been customizable in this area?

> See the variables gnus-simplify-*

"this area" == customizing the `Re: ' string.  Sorry for the ambiguous 
wording.

Re: ">" -vs- "Re:" -- is this guy crazy?

[Lars Balker Rasmussen]

SL Baur <steve@xemacs.org> writes:
> I've complained more than once.  Kyle's response is that since `Re: '
> is a netnews requirement and not an RFC822 one VM isn't broken.

Somebody please confiscate his white hat.
-- 
Lars Balker Rasmussen, Software Engineer, Mjolner Informatics ApS
lbr@mjolner.dk

Re: ">" -vs- "Re:" -- is this guy crazy?

[Aaron M. Ucko]

SL Baur <steve@xemacs.org> writes:

> No.  I've never seen it.  There is a (broken) Microsoft mailer that
> uses `FW: '.  There is a (broken) Novell mailer that appends a
> `-Reply'.  There are various (broken) Unix mailers that use `Re[2]: '
> `Re[3]: ', etc.  VM doesn't add anything by default.

FWIW, another mailer (the headers don't seem to say which) uses 'RE> ',
which is of course also borken.  Anyway, whatever happened to being
liberal in what one accepts?

-- 
Aaron M. Ucko, KB1CJC <amu@mit.edu> (finger amu@monk.mit.edu)

Re: ">" -vs- "Re:" -- is this guy crazy?

[Steinar Bang]

>>>>> amu@MIT.EDU (Aaron M. Ucko):

> SL Baur <steve@xemacs.org> writes:
>> No.  I've never seen it.  There is a (broken) Microsoft mailer that
>> uses `FW: '.  There is a (broken) Novell mailer that appends a
>> `-Reply'.  There are various (broken) Unix mailers that use `Re[2]: '
>> `Re[3]: ', etc.  VM doesn't add anything by default.

> FWIW, another mailer (the headers don't seem to say which) uses 'RE> ',
> which is of course also borken.  Anyway, whatever happened to being
> liberal in what one accepts?

Did you see my procmail rule, with the complicated and ugly regexp?
Growing more complicated and ugly for each Outlook-produced message I
encounter that the regexp doesn't cover...

There is a limit as to how much effort one should put into smoothing
over the sender's errors.