PE> I doubt that there's any BBS where if you send mail to a point,
PE> it doesn't get there, but if you had sent it to the boss, it would.

RS> Thats not what is being discussed. What is being discussed
RS> is the situation where you cannot specify a point address
RS> but you can specify the boss address. Hardly surprising that
RS> that happens when the point system was a later addition.

PE> Oh, that's what the story is.  Must be pretty old software
PE> to have that problem.

Not that old, the mailers have had some problems with the point
addressing even quite recently. And its remarkable how long some
people keep using some stuff, there are still some Opus boards
around even now.

PE> I would be happy to write a utility that did a scan of the
PE> BBS netmail area and printed out a list of all messages that
PE> weren't routed though.

RS> Whats the point of that instead of automating the mapping
RS> when a new point is added ? The remap then works without
RS> human intervention forever. Anything else is worse.

RS> It should a quite trivial matter to just take your point list
RS> and auto produce the remap list from that. Presumably a few
RS> line Awk can do it.

PE>> I have temporarily put the Remaps back in, because of other reasons.

RS> I cant see any good reason to not have them forever myself.

PE> Because with the remap in, if anyone sends a message to 3:711/934.10
PE> (which happens to be a point running a BBS, just like Alexander
PE> Watson Law), with a name of "Jeff Green", which happens to be a
PE> common name, it will get remapped to 3:711/934.17, or whatever
PE> Jeff's node number is.

I wasnt talking about that sort of remapping. The problem we were
talking about was the situation where someone addresses a netmail
to the boss node number and doesnt include the point address. In
that case it makes sense for the remapping to add the point address.

I dont see any real value in remapping more than that. For example
it doesnt make a lot of sense to try to fix bad point addresses.
Or try fancy stuff which sends stuff to a full nother node with
the right point address if its netmailed to your system. If that
point doesnt bother to pick up his mail from you directly, thats
his problem.

PE> I don't want to be overriding the addressing information, for
PE> the same reason that Dave Hatch doesn't override the addressing
PE> info for a "Paul Edwards".

I think you are getting gloriously confused about what we are
talking about. We are talking about ONLY fixing incoming netmail
if its addressed to your node address only, no point number included.
Because there are some systems out there which cant send netmail to
a point address. And because its the fail safe approach if someone
who can send the netmail to the point address but stuffs it up.

PE> Also if I get two points with the same name I have to take away
PE> the remap from both of them.

I think this theoretical stuff is too unlikely to worry about too much.
The much more likely possibility of netmail arriving without the point
number on it is far more important. Particularly important that it fails
safe because otherwise it yet another black hole for mail. The netmail
system in Fido is far too unreliable already, the last thing we need is
to add another black hole to it.

PE> Which isn't the level of automation I wanted.  Seems very dodgy to me.

I think its too trivial a possibility to worry about. And your auto
point rego system can handle it any if you really do care. In fact
it might well be worthwhile as a safety check for the situation where
the same person bumbles around and manages to rego himself twice.

PE> It would help if Squish only did the remap to names going to the
PE> BBS address instead of to the point addresses.  I put a message
PE> in TUB about that but didn't get a reply.

You say these things so damned cryptically. If you are saying that you
cant currently just add the point address to an incoming netmail which
has the point address missing, then IMO the only sensible approach is
to continue to do the remapping you have done in the past an hope squish
is enhanced. Basically because those relatively remote possibilitys you
list are rather less serious than netmail ending up in a black hole.

The fundamental thing which is wrong with so much of Fido technology
is that it doesnt fail safe. Thats a fucked way to implement something
where reliability is important.

--- PQWK202