fred@mot.UUCP (Fred Christiansen) (09/12/85)
it seems some of this domain addressing is not working to hotly.
i recently received a msg from ..!seismo!decuac!decuac.UUCP!user
???????????
and also ..!tove.ARPA!user instead of ..!user@tove.ARPA
?????????
the former creations are unusable for returning mail.
--
<< Generic disclaimer >>
Fred Christiansen ("Canajun, eh?") @ Motorola Microsystems, Tempe, AZ
UUCP: {seismo!terak, trwrb!flkvax, utzoo!mnetor, ihnp4!btlunix}!mot!fred
ARPA: oakhill!mot!fred@ut-sally.ARPA AT&T: 602-438-3472avolio@decuac.UUCP (Frederick M. Avolio) (09/14/85)
In article <263@mot.UUCP>, fred@mot.UUCP (Fred Christiansen) writes: > it seems some of this domain addressing is not working to hotly. > i recently received a msg from ..!seismo!decuac!decuac.UUCP!user > ??????????? > and also ..!tove.ARPA!user instead of ..!user@tove.ARPA > ????????? > > the former creations are unusable for returning mail. Ahem.... As both came through my site I thought I'd explain. The From: line as it passes through here is unchanged. If everyone did that you'd get mail from user@decuac.UUCP and user@tove.ARPA on your From: lines. The From lines looked like the above. While it *is* a bit ugly, and I hope to see it "fixed," it *is* surely usable for reply. Your mailer should send to the first system on the path and so on down the line until it finally gets to the system which gave you the "funny" address. We understand what to do with a domain. Now if other sites broke the address after us I can't help that. But the "Path" you saw was a true path and quite usable.m And host1!host2!host3!uset@host.dom is ambiguous as it is *normally* handled differently on UUCP systems and ARPA systems. --- Fred @ DEC -- ULTRIX Applications Center
dave@uwvax.UUCP (Dave Cohrs) (09/14/85)
> i recently received a msg from ..!seismo!decuac!decuac.UUCP!user > and also ..!tove.ARPA!user instead of ..!user@tove.ARPA > the former creations are unusable for returning mail. Personally, I don't see why this affects your return mail. If decuac knows how to handle the address 'decuac!decuac.UUCP!user' and if whichever site (seismo probably) that generated the '...!tove.ARPA!user' can understand it (seismo can, I know), it shouldn't affect your return mail UNLESS your mailer trys to be "smart" (read dumb) and not use the route provided to it. Mailers should *not* take addresses and chop them up. What I mean is, if you get mail from ...!seismo!decuac!decuac.UUCP!user you shouldn't change this into decuac.UUCP!user and try to optimize the routing. If you do this, don't complain about the hostname. Your site wasn't meant to understand it, decuac was. If your mailer insists on editing routing info, then you should add the code to understand all possible things people can and will do to their addresses. -- Dave Cohrs (608) 262-1204 ...!{harvard,ihnp4,seismo,topaz}!uwvax!dave dave@wisc-romano.arpa
west@cs.reading.UUCP (Jerry West) (09/18/85)
Sender: Xpath: reading cs gateway.cs In article <263@mot.UUCP> fred@mot.UUCP (Fred Christiansen) writes: >it seems some of this domain addressing is not working to hotly. >i recently received a msg >[from] ..!tove.ARPA!user instead of ..!user@tove.ARPA > ????????? >[which is] unusable for returning mail. Hmm, they may well be unusable for you to return mail (and for us for all I know :-}) but this form of addressing is the ONLY way a site with a brain-damaged mailer (eg reading.am.uts - UTS 1.0 (NOT the sys5 clone)) can pass mail to (eg) the ARPAnet. We simply forward the request to mail to the ARPAnet to some smart guy who knows whats going on and let them figure it out. So to mail to user@tove.ARPA, we send to ...smartguy!tove.ARPA!user and let smartguy figure it out. Surprisingly, it works very well. Fortunately, we receive ARPA mail by a recognised gateway to the UK JANET system (ie our primary network is JANET, though we do UUCP stuff and get USENET news etc). Before the flames pour in, let me add that this is not given as our return address (which is in domain format) and does not munge the header apart from the obvious To line. > [also from] ..!seismo!decuac!decuac.UUCP!user I believe this is to do with subdomains, etc. My official UUCP address is, I believe, ...!reading!cs.reading.UUCP!west (ie west@cs.reading.UUCP). -------- As well as the usual disclaimers, may I add that I am not concerned with running this system, and am, therefore, talking thru my xxx. Jerry
hokey@plus5.UUCP (Hokey) (09/18/85)
In article <308@uwvax.UUCP> dave@uwvax.UUCP (Dave Cohrs) writes: >Mailers should *not* take addresses and chop them up. What I >mean is, if you get mail from ...!seismo!decuac!decuac.UUCP!user you >shouldn't change this into decuac.UUCP!user and try to optimize the >routing. If you do this, don't complain about the hostname. Your >site wasn't meant to understand it, decuac was. If your mailer insists >on editing routing info, then you should add the code to understand >all possible things people can and will do to their addresses. You changed horses. Mailers should not chop up *addresses*. Chopping up *routes* is a different story. Especially when the addressee uses a domain specification. Routes are useful in UUCP because, amongst other things, they provide an implicitly rooted path. Domain addresses are explicitly rooted. It would be prefectly reasonable for seismo to have sent the message along with ">From seismo!decuac.UUCP!user", stripping off the extraneous *routing* information (!decuac!). Ideally, that routing information would be placed in a Received: line, but that is another issue. Note also that seismo could strip off !decuac! *even if it doesn't know how to talk to decuac.UUCP* as long as seismo knows how to route or talk to a .UUCP nameserver. -- Hokey ..ihnp4!plus5!hokey 314-725-9492
richl@lumiere.UUCP (Rick Lindsley) (09/19/85)
In article <263@mot.UUCP> fred@mot.UUCP (Fred Christiansen) writes: >it seems some of this domain addressing is not working to hotly. >i recently received a msg from ..!seismo!decuac!decuac.UUCP!user > ??????????? >and also ..!tove.ARPA!user instead of ..!user@tove.ARPA > ????????? > >the former creations are unusable for returning mail. Afraid I must disagree ... as long as they are usable to the site before them they should not bother you at all. Unless ... you aren't trying to rewrite them are you??? Then of course you must be very wizardly to figure out the implied routing. We routinely send out things with ...!tektronix!site.domain!further-path if it is going to make it easier on us to return it. We interpret site.domain and figure it out for you, don't worry. (Even if it is to tell you we can't get there from here.) In fact, all uucp mail goes out with all addresses converted to ! format (if we can decipher the address as handed to us) because some of our neighbors dislike @'s, :'s, and %'s in all their beastly forms. Lastly, >>> it is a BAD idea to mix !'s and @'s <<< this cannot be repeated enough. Rick Lindsley Postmaster
honey@down.FUN (Peter Honeyman) (09/24/85)
we're getting two stories here. some people say ...!decuac!decuac.DEC serves to assist rerouting on replies. (this is from rebuttals to my criticisms of such routes as "ugly, stupid, and wrong".) others now contend that this and all other routes should be left alone, arguing that decuac will do the right thing. if that's the case, why resort to the redundancy at all? peter