Re: The state I'm in - rfc 822 parsing.

[Sam Roberts]

Quoting Alain Magloire <address@hidden>, who wrote:
> > 
> > > > If I find the awful comment, I put it where I would have
> > > > put a phrase.
> > > 
> > > Seems you actually remove comments totaly.
> > > Some clients may want them right.  Especially in the "Received:"
> > 
> > Uh, GetMailBox() grabs the comment and puts in the personal,
> > doesn't it? I don't have the source in front of me, but
> > if what I sent you doesn't do that I sent the wrong diff.
> 
> # ./addr
> alain(the squirrel fighter)@qnx.com
> alain(the squirrel fighter)@qnx.com=> pcount 1
> 1 email <address@hidden>
>    local-part <alain> domain <qnx.com>

That's awful. Is that really desireable? It's explicitly non-compliant.
The code parses rfc822 syntax, with the specific exception of also
grabbing:

address@hidden (the soon to be victorious squirrel fighter)

*One* comment, *after* the addr-spec. If you really think
it's useful to read

alain(the)@( squirrel)qnx.( fighter)com ==
alain(the squirrel fighter)@qnx.com ==
address@hidden (the squirrel fighter) ==
the squirrel fighter <address@hidden>

then we'll talk after I've done the other clean ups
which actually impact parsing of *RFC822* email addresses,
as opposed to "whatever somebody thought would be cool"
email addresses.

> Actually, I'm off topic with the Received field, since I do 
> not think anyone will use it in address_create().

I've got a date-time parser as well in Mail++, I might slap
that and a Received header field parser in next.

> > > > - <>, valid in SMTP "rcpt from:", should I parse and ignore it?
> > > 
> > > I vote strip. Since I want to reuse the email part to build the "From "
> > > string separator for Unix Mbox.
> > 
> > By strip, you mean parse and ignore? If I parse and ignore, this
> > would be a list with two addresses: address@hidden, <>, address@hidden
> 
> Yes, which sounds reasonnable, no?
> for exemple:
>       To: Alain M. <address@hidden>
> I would expect address_get_email () --> "address@hidden".

Oops, I guess that wasn't clear.

I really meant an address of "<>". This is a legal rfc821
address, it's mandated for returning undeliverable messages, among
other things, and is used by vmail, for one, in the smtp dialog when
you bounce a message, i.e.

mail from:<>
rcpt to:<address@hidden>
data
..

I started the cleanup last week, I'll send you more patches after
I've got a bunch farther, rather than little diffs. Just so
you know I haven't abandoned this.

Sam

p.s. I thought you were hunting squirrels, not wooing them!

-- 
Sam Roberts <address@hidden>