ietf-822
[Top] [All Lists]

RE: Funny disruption of subject lines

1999-03-11 17:41:16
On Thu, 11 Mar 1999 17:36 -0600, Pete Resnick wrote:

On 3/11/99 at 3:19 PM -0800, Dan Wing wrote:

But this isn't the correct way to unwrap?  Unwrapping should yield one
space between characters, not multiple spaces.

Says who?

I'll cite implementations, but I doubt that is what you were really
asking.

Netscape V4.5, PMDF's UA (V5.1-7), MultiNet's SMTP, and MX (a VMS mailer)
have the space reduction behavior to varying degrees.

Actually, PMDF used to, but the current version (V5.2-31) doesn't any
more. We changed it to match DRUMS. (And we've had some interoperability
problems as a result, but that's another story...)

Netscape compresses
everything to one space, PMDF seems to compress only the wrapped portions,

It depends on the header field. Subject lines and other free-form fields
used to only compress at wrapping points, and now don't compress at all.

When structured fields are rewritten, however, the entire field ends up
getting compressed.

MultiNet and (I believe) MX only compresses the wrapped portions.

The rule (which is clear in the DRUMS rewrite of 822) is that you
remove only the CRLF. You do not do any "space reduction".

Agreed for unfolding.

But for folding, doesn't this ABNF:

  FWS             =       ([*WSP CRLF] 1*WSP) /   ; Folding white space
                          obs-FWS

allow me to place an arbitrary number of spaces (or tabs) on the folded
line?

I believe this is correct.

                                Ned