On Tue, 3 Feb 2004, Peter D. Gray wrote:
> With sendmail, -io means "IgnoreDots" and the docs say
> "The IgnoreDots option tells sendmail to treat any line
> that contains only a single period as ordinary text, not
> as an EOF indicator." (page 1005, 3rd edition).
>
> However, that is not what sendmail actually does.
Why am I not surprised to read that? :-)
> What it does do is control whether sendmail treats leading dots as per
> RFC821. If set, dots are left alone, but the default behaviour is to
> convert .. to . if at the beginning of line.
Sigh. What do other MTAs do? I've recently been sent the Postfix book,
but I can't find a reference to this in it. The old smail 3.1.29
documentation I have says
Do not allow a single '.' to end an incoming message. Otherwise, a dot
on a line by itself will end a message.
> It appears exim does what sendmail says it does, rather than
> what it actually does.
This is not the first time this has happened. There was a similar
problem with the -t option.
> MUA's which call sendmail without the ignoredots option use RFC821 dot
> handling and double dots if thay occur at the beginning of line
> knowing that sendmail will convert the double dot to a single dot.
Are you sure that *all* MUAs do this? I haven't got any MUAs installed
that operate in this way, that is, do not use SMTP for local submission
and don't set -oi. (I tried "mailx" on Solaris, but I cannot be sure it
did not set -oi or -i, though neither of these strings is in the
binary. Pine uses local SMTP transmission.)
> I have such a MUA and it assumes sendmail behaviour.
> This could be considered a bug in the MUA.
I don't consider it a bug, since it is performing as documented.
However, if what you say is true, it is an incompatibility with
Sendmail.
> So the question is, should exim do what sendmail does
> or what it says it does?
Anybody else have information or views on this? Any data about other MUA
behaviour? Other MTA behaviour? I can't believe this is too urgent,
since it's been this way since Exim was first implemented.
Philip
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
Get the Exim 4 book: http://www.uit.co.uk/exim-book
