RE: [coldsync-hackers] inetd netsync

["Rob Bloodgood" <robb at empire2 dot com>]

>    today i was finally able to use coldsync from inetd.

WOO-HOO!

>  The funny thing is that i was able to sync without using any
>  pi-csd or the equivalent coldsync udp code.
>
>  Anyone knows _when_ pi-csd is required? It's my Palm Desktop
>  totally faulty?

The UDP daemon (pi-csd) was NOT required in my testing on a localhost or
same subnet lan.  It was ABSOLUTELY required to connect to a NAT'd coldsync
server from the Internet via Palm modem.  I think it's needed whenever there
is routing (eg the network addresses of the Palm and the coldsync server
aren't the same).  Hmmm now that I think of it, I never did figure out if
setting netmasks or leaving it blank made any difference on that score.

<speculating>
But when communicating with my Nat'd host, I'm inclined to *guess* (don't
actually know) that the netsync protocol includes the server's (and/or
client's) addresses as payload vs envelope.  Meaning that the NAT which
re-addresses the envelop will show inconsistent results with the payload.

So when my Palm dials up as 210.xxx.xxx.xxx
and tries to connect to 207.xxx.xxx.xxx
and gets 192.168.1.50 back as the server address,

I see pi-csd send essentially a redirect, via udp,
to say, "No, go to THIS host for network communications"
(which was the same as the host it was talking to, but had a different
envelope address).

pi-csd would probably not ever be required on a small intranet on a single
subnet, and quite possibly (experimentation required) not on an intranet
with multiple subnets that never routes externally.
</speculating>

L8r,
Rob


-- 
This message was sent through the coldsync-hackers mailing list.  To remove
yourself from this mailing list, send a message to majordomo@thedotin.net
with the words "unsubscribe coldsync-hackers" in the message body.  For more
information on Coldsync, send mail to coldsync-hackers-owner@thedotin.net.