Re: Missing ABNF terms in 2821bis?

Pete Resnick wrote:

> Should we allow control characters in quoted-pair inside 
> domain-literal?

No.  
 
> Should we allow quoted-pair at all in domain-literal?

No, please get rid of it.  \[, \\, and \] are unnecessary
in domain-literals (and incompatible with STD 66).  

> Remember that they all appear in the obsolete (i.e., 
> must be able to handle for interpretation) syntax.

I don't understand why.  There never was a domain-literal
with control characters and / or quoted pairs in the wild,
what's the idea ?

It's not the same issue as for the local part (LHS).

> John, you are saying that we should leave NO-WS-CTL in 
> quoted-string?

That would be about quoted-pair and qtext, and it affects
the 2821 <Mailbox> (similar to the 2822 <addr-spec>).  If
an "implementation and interoperability report" would show
that it anyway doesn't work, can it be moved to obsolete ?

> And the answer to whether or not we should leave NO-WS-CTL
> in these is....?

ID.klensin-net-utf8 section 2 bullet 3:

| Other than CR, LF, Space (SP, U+0020), and Form Feed (FF,
| U+000C), control characters (U+0000 to U+001F and U+007F
| to U+009F) SHOULD generally be avoided. 

For e-mail better s/FF/HT/.  Of course you can't reference
this Internet-Draft, but its general idea "stay away from
unnecessary control characters" is fine.

 Frank