[comp.sys.dec] Summary of VAX/VMS<==>SUN Mail bridge postings

roddy@miclon.UUCP (roddy) (10/25/88)

About 2 months ago, I posted an article requesting information about mail
links between UNIX and VAX/VMS machines. Our particular configuration
was SUNs running OS version 4.0, and most of the replies received reflect this.

This (rather long) posting summarises the replies. For those who don't want
to read them all, we ended up taking a package called 'dnamail2' from the
SUNspots archives at 'rice', which fitted our needs perfectly. It bolts
on to a SUN-supplied DECNET interface package and has two modules; one
is an outgoing mailer which is called by sendmail when it detects that
a message is destined for a VAX/VMS host. The other is a daemon which waits
on DECNET mail requests from remote hosts, receives the mail and passes them
on to sendmail. Only one machine is required to run the SUN DNI package, and
can act as a host to which other non-DNI SUNs can forward DECNET mail. 

One area of confusion: there are two packages called 'dnamail' in the SUNspots
archives. The above version is written by Darin Johnson,
(leadsv!laic!darin@pyramid.pyramid.com). There is another version which does
not include the incoming mail daemon, by another author. I make no comparisons
as to quality of code - I had no problem with either - but the daemon makes
the mail system much easier to operate, allowing mail in both directions
between SUN and VAX/VMS

Now for the Mail received - My thanks to all senders.


=========================



> From: Dan Trottier <maccs!dan>
> Subject: Re: VMS <==> UNIX mail interfacing
> In-Reply-To: <110@miclon.UUCP>
> Organization: McMaster U., Hamilton, Ont., Can.
> 
> This may not be what you want but we have had a lot of success doing
> it this way.
> 
> All changes involve running new software on your VMS machines. Get a
> TCP/IP package for your VMS machines. I believe that the University of
> Washington has a PD one. This will give you the ability to send mail
> from your Unix machines to the VMS machine. If you want to send from
> the VMS machine to Unix machines you will have to buy a new mailer for
> the VMS machines. We bought PMDF (Pascal Memo Distribution Facility)
> which has worked quite well.
> 
> -- 
> Dan Trottier                                            dan@maccs.McMaster.CA
> Dept of Computer Science                       ...!uunet!utai!utgpu!maccs!dan
> McMaster University                                      (416) 525-9140 x3444

We don't have any VMS hackers on this site, but we do have several people
with UNIX experience, so this was not an easy option for us, but may be
better for those with little UNIX knowledge. Also, does not require DECNET
support for SUNs (a SUN supplied option).


> From: Marius Olafsson <marius@rhi.hi.is>
> Subject: Re: VMS <==> UNIX mail interfacing
> 
> You do not mention how your VMS machines are networked to the SUN's but
> if you have an Ethernet I suggest you use CMU-TEK to get TCP/IP for VMS
> and PMDF (Pascal Memo Distr. Fac.) to do SMTP on VMS. 
> We use this setup here to make our VMS machines nearly
> transparanet wrsp mail. Unix users address VMS users exactly as if they
> were on other Unix machines, and VMS users use VMS-MAIL (but have to
> use slightly strange addressing). Reply works in both directions.
> 
> PMDF is public-domain, and CMU-TEK req. a licence (but is cheap). I
> am sure that many people will give you the same advice, but feel free
> to contact me for more info if you get nothing better (or nearer :-)
> 
> 
> --
> Marius Olafsson                 Internet: marius@rhi.hi.is
> University of Iceland           Non-MX:   marius%rhi.hi.is@uunet.uu.net
> 								UUCP:     {mcvax,enea,uunet}!hafro!rhi!marius

Once again, this would probably be appealing to VMS hackers.


> From: Darin Johnson <taurus!darin>
> Subject: VMS <-> UNIX mail
> 
> I have a program that sends mail easily between VMS and Sun (both ways).
> 
> The only requirement is that you have Sunlink/DNI for the Suns (DECnet
> package).  If you want to use it transparently from mailtool, etc.,
> then you need to do some minor sendmail.cf hacking.
> 
> If you are interested, send me mail, and I will send you the package.
> 
> Darin Johnson
> leadsv!laic!darin@pyramid.pyramid.com
> 

This was perfect! The only changes required were that SUN has altered its
package name (it used to be called sunlink/DNA not it's SUNLINK/DNI), so
the source has to be hacked to change every instance of dna to dni. Also,
we were running  SUN/386i machines, and there is a small bug where the
source checks a VMS status value as (if status == 0x01000000), which is
a VAX/VMS '1' in 68000 byte ordering. The fix requred by us INTEL users
is to check against '1' because the VAX and 386 byte ordering is the same.

It is available from the sunspots archives at 'rice', called 'dnamail2.shar'.
Full documentation comes with it.

My thanks to A. Mossberg <amossb@umbio.miami.edu>  and 
Keith Moore <moore%utkcs2.cs.utk.edu%mcvax@cygnusx1.cs.utk.edu>
who also pointed me in this direction.

Finally, I was sent a previous posting on this subject by Dieter Woerz,
UUCP:   ...{uunet!unido, pyramid}!iaoobel!woerz. This was about
300 lines of informationin a summary similar to this,
but I hesitate to add it here as it was aimed
more at users not using SUNs/sunlink. However, I have kept a copy, and will
mail it to interested parties.

Once again my thanks to all those who assisted,

Roderick Manzie, Micrognosis International, London.

...ukc!uk.co.mic.lon!roddy
...ukc!{pyrltd,qmc-cs}!miclon!roddy