On Mon, 25 Sep 2000, Phil Pennock wrote:
> strip_trailing_dots_local - like strip_trailing_dots, but only for mails
> submitted locally, not via SMTP.
But you can submit mails locally via SMTP. Several MUAs do. Or do you
mean "submitted locally not via SMTP"? (Note subtle difference.) <grin>
> Thoughts? And should it be a default? Or default if "-t"?
It would be tedious to implement. The current facility is just a single
flag which affects the address parser. Having to make it conditional
would complicate things.
I don't like the idea of making -t any more special that it already is.
> Rational:
> trailing dots are valid in DNS, are used in example A.2.6. in RFC822,
I think that's a typo. There's no such example in the draft of the revision.
> and follow the "be liberal in what you accept from humans principle,
> even if you're a pedantic git when dealing with machine-machine
> protocols".
IMHO it always causes trouble in the long run if you are liberal in what
you accept. From anywhere. As far as humans are concerned, if you don't
complain, they never learn. :-)
Can you tell I don't like this idea?
> So the MUA->MTA interface should handle them. So Exim
> should strip them.
Why Exim and not the MUA? Because MUAs are sometimes human? :-)
I tried this with Pine. It tried to send the address to Exim over SMTP.
Exim complained. Pine showed me the error and highlighted the failing
address. Seemed pretty reasonable.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
