UUCP problem

[Jack F. Vogel]

In article <609 at g4lzv.co.uk> keith at g4lzv.co.uk (Keith Brazington) writes:
>In article <483 at ispi.UUCP>, jbayer at ispi.UUCP (Jonathan Bayer) writes:
>> Permission file entry:
>> 
>> MACHINE=libcmp LOGNAME=libcmp \
>> 	COMMANDS=uucp:rmail \
>> 	READ=/usr/spool/uucppublic:/usr/tmp \
>> 	WRITE=/usr/spool/uucppublic:/usr/tmp \
>> 	SENDFILES=yes REQUEST=yes 
>> 
>> System file entry:
>> 
>> libcmp Never ACU 19200 p0000 ogin:--ogin:--ogin: nuucp
>
>The problem with this is the LOGNAME entry. It should be LOGNAME=nuucp
>assuming that the remote uses this as the login id to get into the system.
>Remember that MACHINE is the ID of the machine, and that LOGNAME is the ID used
>to gain access to uucico.
 
This is not true, I do not remember exactly what the problem of the original
poster was but LOGNAME does not need to be equal to nuucp. I have a number of
sites that login as nuucp but their LOGNAME is equal to their real systemid.
"nuucp" is the id passed through getty to login and identifies the 'shell' to
run as uucico, LOGNAME is determined during the uucp handshaking between the
two systems after uucico is started up, when it gets that id it looks in the
Permissions file for a LOGNAME equal to that id. Also it is not true that
MACHINE is the id of the incoming system, rather it is the id used when your
system makes an outbound call, at that time uucico again checks Permissions,
this time for the MACHINE id rather than for LOGNAME.

I realize this does not resolve whatever the original problem was, but then
incorrect information helps even less.


-- 
Jack F. Vogel
Turnkey Computer Consultants, Westchester,CA
UUCP: ...{nosc|uunet|gryphon}!turnkey!jack 
Internet: jack at turnkey.TCC.COM or lcc!jackv at cs.ucla.edu