Re: configuration inconsistency. MX usage forbidden..

[mea@nic.funet.fi]

...
> Damn..busy ;)
> 
> Ok, behaviour I see, is that after a succesfull smtp connection (ehlo,
> mail from, rcpt to, data, quit) the mail is sent to the errorchannel with
> info:

	  Huh ??  That is, at first it presents a SUCCESS report,
	and then it presents a FAILURE report ?
	But it has done transmission ?

	If all stack frames work ok, that report can happen only
	when you have '-x' option in your smtp channel command line.
	(That's why I suspect gcc ...)

> 19990412183904 /opt/zmailer/bin/ta/smtp: 77288/89//error2 smtp; 500
> (configuration inconsistency. MX usage forbidden, no address in the DNS:
> 'nl.uu.net')
> 
> The address that is not in the DNS is the part after @ in the rcpt to
> command. It happens always, unless the domain after @ is a hostname and
> accepts mail.

	Yes, and because I know that in my own case the emergence of
	this report was with previously working CONFIGURATION at a host
	which got installed with a new compiler, and I happened to want
	to upgrade the version a bit, thus ending up recompiling the beast...

	That is, when with unchanged config (without '-x'), I got that
	error, became highly suspicuous about GCC-2.8.0 at SPARC (Solaris).
	In the mean time the same source was working just fine at my PPro200,
	and at my Alpha...

... 
> Switching back to 2.7.2.3 works, but it isn't handy that gcc 2.8.0 doesn't
> do it right....Now how to fix that....

	egcs-1.1.2 ?

	At this I do happen to believe at Bazar approach on the
	software development. gcc-2.8.* has practically stopped
	from developing in recent times.

> > 	I think it relates to optimization at SPARCs (or the particular
> > 	model of SPARC we have), and compiling without "-O" might produce
> > 	working binary.  Could you care to try ?
> 
> Doesn't make any difference.

	Huh...  (It could be the headers too)

> > 	Anyway, I use gcc-2.7.2, or egcs-1.1.2, or some vendor specific
> > 	compiler -- depending the platform...
> 
> 2.7.2.3 works.

	Given that egcs-1.1.2 does all kinds of magic flow optimization
	things pretty well, I would like to recommend it instead of the
	gcc-2.7.2.*

> Mark
> Mark Visser / Calslaan 26-31              | E-mail: mark@dnd.utwente.nl
> 7522 MC Enschede / Telephone: 053-4895038 | SNT-mail: mark@snt.utwente.nl

/Matti Aarnio <mea@nic.funet.fi>