[comp.protocols.tcp-ip] question about SMTP MX records

robin@csuchico.edu (Robin Goldstone) (10/18/90)

I am trying to send a message to someone@applelink.apple.com.  This
host has no TYPE A record, only an MX record.  My mailer currently
cannot resolve MX records.  As a workaround, I thought I would just
send to someone%applelink.apple.com@apple.com.  It is my (limited) 
understanding that addresses are parsed from right to left, so this
message would be sent to apple.com, who would then be able to forward
it to applelink.apple.com.

Some questions:
1) is the syntax of the address I am trying to use valid?
2) am I violating any network rules by routing my message through
another host?
3) should this message be getting delivered?  
I have sent several test messages that have disappeared into a black hole...

Thanks in advance for any help you can provide!

Robin Goldstone, Systems Software Specialist
California State University, Chico Computing Services
robin@csuchico.edu

craig@bbn.com (Craig Partridge) (10/19/90)

In article <1990Oct18.164200.5699@ecst.csuchico.edu> robin@csuchico.edu (Robin Goldstone) writes:
>...  As a workaround, I thought I would just
>send to someone%applelink.apple.com@apple.com.  It is my (limited) 
>understanding that addresses are parsed from right to left, so this
>message would be sent to apple.com, who would then be able to forward
>it to applelink.apple.com.

>1) is the syntax of the address I am trying to use valid?

    Yes the syntax is correct.

>2) am I violating any network rules by routing my message through
>another host?

    No, though doing this sort of thing frequently (like sending all
your mail via another system because your system doesn't support MX)
is considered rude.

>3) should this message be getting delivered?  

    Yes.  Apple.com is the MX for applelink.apple.com, so it should
    accept mail for applelink.apple.com.  Note that if Apple.com was
    not the MX for applelink.apple.com, then all bets are off.  You
    should not assume that via j random host using the %-hack is
    safe or reasonable.

    You mention your mail is going into a black hole, that's definitely
    a problem.  Mail should not vanish without a trace...

Craig

oberman@rogue.llnl.gov (10/20/90)

In article <1990Oct18.164200.5699@ecst.csuchico.edu>, robin@csuchico.edu (Robin Goldstone) writes:
> I am trying to send a message to someone@applelink.apple.com.  This
> host has no TYPE A record, only an MX record.  My mailer currently
> cannot resolve MX records.  As a workaround, I thought I would just
> send to someone%applelink.apple.com@apple.com.  It is my (limited) 
> understanding that addresses are parsed from right to left, so this
> message would be sent to apple.com, who would then be able to forward
> it to applelink.apple.com.
> 
> Some questions:
> 1) is the syntax of the address I am trying to use valid?
> 2) am I violating any network rules by routing my message through
> another host?
> 3) should this message be getting delivered?  
> I have sent several test messages that have disappeared into a black hole...

The use of the '%' hack is not a part of any standard. But it usually works. So
the syntax of the address is probably OK. APPLE.COM is the mail exchanger for
APPLE and APPLELINK, so it SHOULD work, but only by convention. It does not
violate any "rules", such as they aren't. But this type of routing is quite
undesireable--but very common.

Mail to APPLELINK is staged on APPLE.COM for delivery, so maybe the route
between APPLELINK and APPLE is down.

You should really try to get a mailer that handles MX records some day.

					R. Kevin Oberman
					Lawrence Livermore National Laboratory
					Internet: oberman@icdc.llnl.gov
   					(415) 422-6955

Disclaimer: Don't take this too seriously. I just like to improve my typing
and probably don't really know anything useful about anything.

tim@ggumby.cs.caltech.edu (Timothy L. Kay) (10/21/90)

oberman@rogue.llnl.gov writes:

>You should really try to get a mailer that handles MX records some day.

Fine, but how do we get Silicon Graphics to provide a mailer that supports
MX records?  Is there a way of putting pressure on them, such as being able
to claim that their machines violate internet standards or somesuch?

Tim

craig@bbn.com (Craig Partridge) (10/22/90)

In article <tim.656457948@ggumby> tim@ggumby.cs.caltech.edu (Timothy L. Kay) writes:
>Fine, but how do we get Silicon Graphics to provide a mailer that supports
>MX records?  Is there a way of putting pressure on them, such as being able
>to claim that their machines violate internet standards or somesuch?

    Any host which does not support MX RR's in the mailer is not
Host Requirements conformant.  See page 65 of RFC 1123.

barns@GATEWAY.MITRE.ORG (10/26/90)

Ouch, Craig, I hope I didn't hear you say that every host X that is
an MX for another host Y is required to support the percent hack,
that is, treat user%Y@X equivalently to user@Y?  It works in most
places, probably, but I don't remember this ever being raised as
even a proposed requirement...

Bill Barns / MITRE-Washington / barns@gateway.mitre.org

craig@NNSC.NSF.NET (Craig Partridge) (10/26/90)

> Ouch, Craig, I hope I didn't hear you say that every host X that is
> an MX for another host Y is required to support the percent hack,
> that is, treat user%Y@X equivalently to user@Y?

RFC 1123 doesn't explicitly require it, but it suggests that support
for the %-hack is expected (the discussion in Section 5.2.16 being
my primary source).

Craig

imp@marvin.Solbourne.COM (Warner Losh) (10/29/90)

In article <9010270002.AA00258@ucbvax.Berkeley.EDU> craig@NNSC.NSF.NET (Craig Partridge) writes:
>RFC 1123 doesn't explicitly require it, but it suggests that support
>for the %-hack is expected (the discussion in Section 5.2.16 being
>my primary source).

However, in existing practice there is at least one mailer that uses
the % for a gateway between internet and decnet.  The address of the
form "imp%node.decnet@twg.com" will cause mail to be sent to NODE::IMP
from the internet host TWG.COM.  This being the case, mail sending
programs should not assume that <a%b@c> is the same as "<@c:a@b>".

Warner
--
Warner Losh		imp@Solbourne.COM
How does someone declare moral bankruptcy?