Re: FWS problem

["Charles Lindsey" <chl@xxxxxxxxxxxxxxxx>]

In <41CAC8D6.1D57@xxxxxxxxxxxxxxxxx> Frank Ellermann <nobody@xxxxxxxxxxxxxxxxx> writes:

>Charles Lindsey wrote:

>> USEFOR already says that we take only the syntax from
>> section 3 of RFC 2822, not that from section 4.

>Okay, then the FWS problem is reduced to 2*7 occurences
>in 7 productions of usefor-02, and about 2*7 [CFWS].  For
>[FWS] I get:

>1: unstructured    = *WSP 1*utext *([FWS] 1*utext) *WSP

>Very dubious, it's essentially any ASCII minus NUL, CR, or LF,
>but allowing CRLF.  And again all those funny NO-WS-CTL, sigh.

Indeed, but politically impossible to change it unilaterally.

>2: msg-id          = *WSP msg-id-core *WSP
>3: newsgroup-list  = *WSP newsgroup-name
>                     *( [FWS] "," [FWS] newsgroup-name ) *WSP
>4: path-list       = *WSP
>                     *( path-identity [FWS] path-delimiter [FWS] )
>                     tail-entry *WSP
>5: poster-text     = *WSP %d112.111.115.116.101.114 *WSP
>6: control         = "Control:" SP control-message CRLF
>   control-message = *WSP verb *( [FWS] argument ) *WSP

Move <control> down to your [CFWS] cases. I think that is a bug in USEFOR
which I have aksed Ken to fix.

>7: dist-list       = *WSP dist-name *( "," [FWS] distaname ) *WSP

>For newsgroup-list you allow [FWS] before and after the comma,
>for dist-list it's only after the comma, is this what you want ?

No, I think you just spotted another bug in USEFOR (draft-13 allowed the
[FWS] on either side of the ",").

Yes, all that syntax works, but that is not to say that we should change
it.

>And now the seven [CFWS] in "our" headers, defining "our" by
>"already not the same as in RfC 2822 for some other reasons"):

But your next lot of syntax does not work, because you have forbidden, for
example:

Supersedes:  (a comment)
    <abcd@xxxxxxxxxxx>
    (another comment)

So you would need some syntax of the form:

    WSPC = *WSP [ comment [CFWS] ]
    CWSP = [ [CFWS] comment ] *WSP


>1: references      = "References:" SP msg-id-list CRLF
>   msg-id-list     = *WSP msg-id-core *([CFWS] msg-id-core) *WSP

>Is that really what we want, no separator or only a comment as
>separator is okay ?  s-o-1036 has apparently 1*WSP as separator.

No, you just spotted another bug in USEFOR.

>2: supersedes      = "Supersedes:" SP *WSP msg-id-core *WSP CRLF
>3: injection-info  = "Injection-Info:" SP *WSP path-identity
>                     *( [CFWS] ";" [CFWS] inj-info-para )
>                     *WSP CRLF

>Dubious.  I've added [CFWS] before and after the semicolon, your
>syntax allowed "Injection-Info:" SP CFWS path-identity CFWS CRLF
>among other things.

Yes, that is right. Optional CFWS is allowed on either side of any MIME
<parameter>, and that needs to be made clear in USEFOR.

>4: lines           = "Lines" ":" SP *WSP 1*DIGIT *WSP CRLF

>Who needs a comment in a deprecated header, let alone folding ?

It is a fundamental principle of RFC 2822 that comments and folding are
allowed in _all_ headers. We stuck out our necks rather in the case of
Newsgroups, Message-ID, etc where there are performance issues, but there
are no grounds to make an exception for Lines.

But I am still not convinced that your syntax has forbidden folded lines
containing only WSP. All you need is some situation where two CFWS can
appear adjacent, and you will get it. And there are plenty of places in
RFC 2822 where that can happen. Moreover, you have got now to deal with
all those things we take from RFC 2822, such as <mailbox-list> in the From
header. That has [CFWS] at either end, and you can't start rewriting all
the syntax inside RFC 2822. So you still need wording to forbid folded
lines with WSP only, and header content 1st lines with WSP only, so all
your modified syntax has still not solved the problem it set out to solve,
so what have you gained?


>P.S. and OT:  Sometimes something on your side adds ***SPAM-3***
>              to the subject, and I forget to remove it.

Yes, that is added as a warning by my Spamassassin filter when the score
is 3 (anything with a score of 4 or more goes straight to devnull).

-- 
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133   Web: http://www.cs.man.ac.uk/~chl
Email: chl@xxxxxxxxxxxxxxxx      Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9      Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5