[comp.mail.uucp] Some questions about pathalias and smail

bill@twwells.uucp (T. William Wells) (12/05/88)

I have just started to do automatic processing of the mail maps so
that I don't have to figure out mail paths for myself.

Anyway, I am a bit dissatisfied with the output of pathalias; I
believe that it is less than optimal, and may also violate some mail
rules. The following line illustrates both:

	3b5node proxftl!novavax!uflorida!gatech!cs.utexas.edu!romp!3b5node!%s

First, I can route my messages to domain names through uflorida, as in:

	3b5node proxftl!novavax!uflorida!cs.utexas.edu!romp!3b5node!%s

I believe that the Internet has a chance to better route the message
than I; I'd rather send my messages that go to the Internet to
uflorida (or uunet, when I get my modem and get connected) than some
further site.

The other thing wrong with this is that it seems to route a message
*through* the Internet, something I understand to be wrong. I had used
the -D option on pathalias, hoping to discourage this; it didn't work
(unless there was no other path?).

So, here are my questions:

    1) Is there a way to tell pathalias to route a subset of domain
       addresses through one site?  I don't want to route all of them
       there, as I don't think it understands all domains that are in
       use; at least I've had trouble routing stuff to the .uk domain
       through there. I could use smart-host, but that does a little
       more than I want.

    2) Does my example violate any mail rules?

    3) If it does, how do I prevent pathalias from generating such
       paths?

    4) Before starting to use these maps (with smail 2.5), what else
       should I be aware of that is likely to cause problems?

If these are already overly-discussed issues, please reply by e-mail
rather than posting.

---
Bill
{uunet|novavax}!proxftl!twwells!bill

honey@mailrus.cc.umich.edu (peter honeyman) (12/05/88)

the texas maps include links involving cs.utexas.edu.  as far as
pathalias is concerned, that is a valid host name; pathalias is
entirely unaware of any domain interaction.  (domain handling
applies to names starting with dot, not to names with a dot in
the middle.)  so most of the questions you ask are moot.

	peter

dhesi@bsu-cs.UUCP (Rahul Dhesi) (12/05/88)

The following is supposed to be illegal:

     uucp-site -> Internet-site -> uucp-site

It used to be that UUCP sites had names ending with .UUCP and Internet
sites didn't.  But now many UUCP sites have domain names, and many
Internet sites are also UUCP sites.  So how could you possibly avoid
routings of the form shown above?
-- 
Rahul Dhesi         UUCP:  <backbones>!{iuvax,pur-ee}!bsu-cs!dhesi

vixie@decwrl.dec.com (Paul A Vixie) (12/05/88)

[Wells]
# I believe that the Internet has a chance to better route the message than
# I; I'd rather send my messages that go to the Internet to uflorida (or
# uunet, when I get my modem and get connected) than some further site.

The trick, then, is to keep a file around which contains your local "hints"
for pathalias.  Stuff you want pathalias to see but which is not in your map
entry and also not in the comp.mail.maps data.  I call this file "glue.local"
and I put into it things like:

->	dead sites which are still mentioned in lots of map entries: cbosgd
	and seismo are/were examples of this;
		dead {seismo}
		dead {cbosgd}

->	rerouters that I want to pretend are dead; I have a small list of
	these which I will publish in this group from time to time: rutgers
	is an example;
		dead {rutgers}

->	sites that bungle From: lines such that mail through them is often
	unreplyable: sun is an example;
		dead {sun}

->	remote links which I know about but which are not for use by the
	public: ubvax!vixie used to be an example;
		ubvax vixie(????)
		vixie ubvax(????)

->	internet wormholes like the one you are trying to use.
		uflorida .edu, .com, .mil, .net, .org

Note that using the internet as a carrier for non-internet traffic is frowned
upon -- though anybody with a domain name is "on the internet" for the purpose
of this rule.  So you could mark your uflorida wormhole as terminal, i.e.,
		uflorida <.edu>, <.com>, <.mil>, <.net>, <.org>
...it's really not going to change anything.

These probably _are_ overly discussed issues according to many, but I get a
lot of mail from sideline lurkers who are completely dazed by the complexity
of it all and I think that a public answer can do only good.

Good luck.
--
Paul Vixie
Work:    vixie@decwrl.dec.com    decwrl!vixie    +1 415 853 6600
Play:    paul@vixie.sf.ca.us     vixie!paul      +1 415 864 7013

dhesi@bsu-cs.UUCP (Rahul Dhesi) (12/07/88)

In article <4957@bsu-cs.UUCP> dhesi@bsu-cs.UUCP I carelessly wrote:
>The following is supposed to be illegal:
>
>     uucp-site -> Internet-site -> uucp-site

Oops, I've been corrected by email.  I should have typed:

     uucp-site -> Internet-site -> Internet-site -> uucp-site
-- 
Rahul Dhesi         UUCP:  <backbones>!{iuvax,pur-ee}!bsu-cs!dhesi

pat@orac.pgh.pa.us (Pat Barron) (12/08/88)

In article <4957@bsu-cs.UUCP> dhesi@bsu-cs.UUCP (Rahul Dhesi) writes:
>The following is supposed to be illegal:
>
>     uucp-site -> Internet-site -> uucp-site
>
>It used to be that UUCP sites had names ending with .UUCP and Internet
>sites didn't.  But now many UUCP sites have domain names, and many
>Internet sites are also UUCP sites.  So how could you possibly avoid
>routings of the form shown above?

Actually, what's illegal is:

       uucp site -> DDN site -> uucp site

The DDN consists of the ARPANET, MILNET, and random other networks controlled
by DCA.  It's perfectly OK to hop through some University's campus network,
or maybe even a regional network, if those networks permit it.

As to your point about how to avoid routing through the Internet now that
UUCP sites are getting real domain names ... Repeat after me:  "Names are
not routes.  Routes are not names."

--Pat.
-- 
Pat Barron
Internet:  pat@orac.pgh.pa.us  - or -   orac!pat@gateway.sei.cmu.edu
UUCP:  ...!uunet!apexepa!sei!orac!pat  - or -  ...!pitt!darth!orac!pat

bill@twwells.uucp (T. William Wells) (12/15/88)

In article <158@orac.pgh.pa.us> pat@orac.pgh.pa.us (Pat Barron) writes:
: As to your point about how to avoid routing through the Internet now that
: UUCP sites are getting real domain names ... Repeat after me:  "Names are
: not routes.  Routes are not names."

Repeat after me: "UUCP doesn't know about names, only routes...."

---

After considering what I've received, I've decided that I'm going to
let the maps have their way. My reasoning? If the sites didn't want
to be routed through, they should have made sure that their map entry
reflected this. If they didn't? Not my problem, I'm not their mommy.

Anyone have a good reason for not doing this? If so, USE E-MAIL to
tell me.

---
Bill
{uunet|novavax}!proxftl!twwells!bill