Re: Bugs in tmda-ofmipd?

On Fri, Nov 10, 2006 at 08:11:23PM -0700, Jason R. Mastaler wrote:

> Is it really a "bug" though?  Todd is testing with telnet.  People

It's a bug if it isn't properly implementing the SMTP protocol, or if it
is fragile when presented with real-world data. A real SMTP server can
be tested with telnet--in fact, you can send all your email that way if
one is so inclined. tmda-ofmipd is supposed to be a SMTP server with
limited capabilities, but it should still provide a real and non-fragile
subset of the SMTP protocol.

For testing purposes, if someone wants to tell me *what* it's choking
on, I can code around it in my testing. But I don't really think that's
a substitute for real robustness. A program can complain or
validate/reject data, but it shouldn't *die* or close unexpectedly
without providing a way to correct the problem. A failure to handle a
use case *is* a bug by definition, isn't it?

FWIW, the same test works properly against postfix. Granted, tmda-ofmipd
isn't trying to be postfix, but it *does* show that my test is a valid
one, and that tmda-ofmipd isn't doing what it should.

> used tmda-ofmipd with (except tbird) properly detects when a
> connection is dropped, raises a flag for the user, and no mail is
> lost.

But what happens if someone is testing with a new or non-standard
client? If they don't know why it's failing, they have no way to
determine if the problem is their program or not. That way lies madness.
:)

-- 
Unabashedly littering the information superhighway with detritus like
this for over 15 years now.