On Thu, Apr 20, 2000 at 01:14:17PM +0100 , Philip Hazel wrote:
> On Thu, 20 Apr 2000, Petr Cech wrote:
> > > generates* conform to RFC 822.
> >
> > by using '?'. ewww
I should have put ':))' at the end of the line above
> That code is among the earliest that I wrote for Exim, back in 1995. I
> didn't know much about character sets in those days. Was the escaping by
> text such as "=?iso-8859-2?Q?Petr_=C8ech?=" even defined then? I didn't
don't know.
> expect to find non-printing characters in password files, and, as I
> said, I did expect the MUAs to generate the From: lines, so this was
> "backstop" code. Converting to ? was just a cheat for what I thought
> would almost never happen.
:) There are plenty of problems with non-ascii characters in various programs.
> Obviously I was wrong. This isn't the first time!!
Sorry, didn't want to rant too much.
> > > Noted. It currently says, in section 46.14, "In all cases the user name
> > > is made to conform to RFC822 by quoting all or parts of it if
> >
> > it's not quoting them, it's mangling them. That is, what I don't like.
>
> I know! I know! I do see your point.
good :))
> I will put this on the Wish List - but won't look at it till I've
> finished the book. However, to escape like "=?iso-8859-2?Q?Petr_=C8ech?="
> Exim would need to know what the character code was. How can it find
> that out?
That's the hard part. It could be either system configurable, because the
message must originate on localhost or it could be escaped using some general
approach. I know almost nothing about, what should be done with unknown
charsets. I would think that there is some general method of escaping 8bit
characters of unknown charset. Quoted-printable perhaps? (don't beat me).
So, I'm adding custom header to mutt and disabling Sender: generation.
Good luck with the book
Petr Cech
--
Debian GNU/Linux maintainer - www.debian.{org,cz}
cech@???
Resistance is futile. You all will be packaged
