🚀 go-pugleaf

In <1992Jul27.021919.4356@cs.su.oz.au> piers@basser.cs.su.oz.au (Piers Lauder) writes:

| The routing loop is a SUNIII problem, not an ACSnet one :-)

We know that...

| (some MHSnet sites seem to
| have "archie.au" as a map for "plaza.aarnet.oz.au" -- maybe plaza exported
| this alias at one time or another?).

Perhaps, I didn't set it up, but I don't recall that, we had decided it
should be archie.oz.au well before MHSnet was installed there.

In any case, archie itself (plaza) does NOT believe its name is "archie.au"
on ACSnet, so if some site were to send a message intended for "archie.au"
to archie via MHSnet it would be bounced, or something ...

plaza# _ubin/netmap -k | grep archie | sed 's/^.................//'
			  9=archie.2=oz.0=au ==> plaza.aarnet.oz.au


The "sed" nonsense is just to make the output fit in 80 columns, all that
was deleted was spaces.   Plaza (archie) knows only one alias that
contains "archie" and that's for "archie.oz.au", not "archie.au".

If some MHSnet has an archie.au alias, they should delete it.

I don't have enough mhsnet knowledge to actually work out what
plaza/archie do with an MHSnet message sent to archie.au - if Piers
is right, I guess it would be routed to basser, and if basser believes
that archie.au is an alias for plaza, I guess it would route it right
back, which would seem to create a nice routing loop.

kre