Path: utzoo!utgpu!water!watmath!clyde!att!ucbvax!THINK.COM!taylor
From: taylor@THINK.COM
Newsgroups: comp.mail.mh
Subject: Re: In-Reply-To header field
Message-ID: <8810240131.AA17114@odin.think.com>
Date: 24 Oct 88 01:31:29 GMT
References: <5886@killer.DALLAS.TX.US>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 62

<Date: 23 Oct 88 18:02:52 GMT
<From: Bill Wisner <killer!wisner@eddie.mit.edu>
<
<The appropriate portions of RFC822:
<
<optional-field =
<            /  "In-Reply-To"       ":"  *(phrase / msg-id)
<phrase      =  1*word                       ; Sequence of words
<word        =  atom / quoted-string
<atom        =  1*<any CHAR except specials, SPACE and CTLs>
<
<An in-reply-to field can, therefore, consist of just about anything,
<if you interpret it as being a string of atoms.

But, specials includes ".", ":", and "," among other things.

And therefore it looks like the In-Reply-To field that I showed is not
in compliance.  But, it's fairly similar to the In-Reply-To field
generated by other mailers -- such as GNU Emacs's Rmail.

For reference, the In-Reply-To field is:

<In-Reply-To: Your message of Mon, 10 Oct 88 15:44:20 -0400.
<             <8810101944.AA10760@balder.think.com> 

[and was generated by the default replcomps file supplied with MH 6.6]

and since ",", ":", and "." are `specials', they can't appear in atoms.

Now, when I look at the outputs of other mailers, I find some variation.

Some just have the msg-id field.  That is, they have:

In-Reply-To: <8810101944.AA10760@balder.think.com> 

Others render the above as:

In-Reply-To: Your message of Mon, 10 Oct 88 15:44:20 EDT,
             <8810101944.AA10760@balder.think.com> 

That is, they use `EDT' instead of the `-0400' and a `,' instead of a
`.'.

And others render it as:

In-Reply-To: Your message of Mon, 10 Oct 88 15:44:20 EDT <8810101944.AA10760@balder.think.com>

That is, as one long line -- and don't seem to care if it's longer than
80 characters.

And I'm sure I could find yet more variations if I tried.

Now, looking at the date stuff in rfc822, `-0400' and `EDT' are equally
valid.  But, it appears to me, that both the `.' and the `,' are illegal
--  they're not legal constituents of atoms and the string is not
quoted.

So, it looks like it's illegal.  But, commonly violated.  And, I was
wondering -- am I overlooking something?  Or is there a later standard?
Or an addendum to rfc822?  Or, is it illegal, just as it appears?

David