Mailing List Archive

7.1.0 released
Hello all!

We'll, I'm finally packaging up version 7.1.0. There are two important
things to realize about this version. First, there have been a lot of
changes, enhancements, and a couple bug fixes (some very cool stuff -
check out the list below). Second, to upgrade, you *must* upgrade both
the client and server - this version is not backward compatible with
previous versions. I hated to do that, but it was necessary since
there were some fairly nasty bugs that couldn't be fixed without it.
Here's the info about upgrading from the INSTALL file:

- The client/server protocol has changed. You *MUST* use a
7.1.0 client with a 7.1.0 and above server. A 7.1.0 client
is *not* backward compatible with a pre-7.1.0 server.

- Some of the flags in the client (-d, -D, and -r) and server
(-n) have been given new identities to make the client and
server flags more uniform.

- The conserver.passwd file now uses the first username match
to determine access rights - if you have multiple instances
of a username in an existing password file, they must be
combined into one to continue to work.

Here's a list of systems that have been known to successfully compile
conserver 7.1.0:

- Solaris 2.5.1 thru 8 (sparc/x86), gcc
- BSDI 3.X, gcc
- MacOS X
- Linux 2.2.18 (x86), gcc
- Linux 2.4.2 (x86), gcc
- FreeBSD 4.2 (x86), gcc
- cygwin (w2k),gcc 2.95.3
- DEC Tru64 4.0, gcc
- DEC Tru64 4.0/5.1, DEC cc
- HP-UX 10.20, gcc

And, as always, here's the list of changes. Enjoy!

version 7.1.0 (Jul 26, 2001):
- Hostname in access list that began with a digit was treated as
an IP address - only strings using [0-9./] are considered CIDR
notation and they must be of the form a.b.c.d[/n]
- Fixed SIGHUP always opening logfile(-L) - shouldn't without -d
- Fixed 'make install' problem under MacOS X and cygwin
- Client -l option totally broken in 7.0.3 - patch by Daniel E.
Singer <des@cs.duke.edu>
- Client now accepts IP addresses from server and server now sends
the IP of the socket instead of its hostname to the client,
fixing part of the multi-interface problem
- Client no longer passes hostname to server along with username
- Client now turns off IEXTEN so stuff like ^V passes through
- Server now pauses one second when reopening a TCP-based console
to give the terminal server a chance to clean up
- Master server process now fork()s when accepting clients
- timestamp-spec can be numeric only ('m' default)
- Ambiguous console name error now shows ambiguous list
- Console list in conserver.passwd can now have whitespace
- Access lists in conserver.cf can now use ',' as a separator
- Added special '*any*' username in conserver.passwd
- Username match (real or '*any*') now stops conserver.passwd file
processing - allowing you to lock out users
- Added -u flag to server to enable "unloved" output - this is the
opposite of the old -n flag, which now does nothing
- Added -7 flag to client and server for stripping the high bit
of data on either side (--disable-8bit removed)
- Added -b to server to specify the base port number for the
secondary communication channel - based on code from Enrik
Berkhan <Enrik.Berkhan@planb.de>
- Changed -d and -D flags to -r and -R in client - now -D enables
debugging mode in the client (and -d is unused)
- Changed -r to -G in client so -r could be used for the above
- Client now uses getpassphrase() (if available) for > 8 char
passwords - pointed out by Gregory Bond <gnb@itga.com.au>
- Improved signal handling
- Improved process handling (POSIX waitpid() now necessary)
- Significant rework of STREAMS-based pseudo-tty code
- TCP connections ('!' config syntax) now have some Telnet protocol
knowledge, removing the "noise" seen when conneting to a terminal
server - reported by Todd Stansell <todd@stansell.org>
- Various code and message "pretty-printing" done
- Client and server both run in Windows (tested with Windows
2000) under the cygwin environment thanks to a debug-fest with
Emmett Hogan <hogan@gnac.com>
- Using autoconf version 2.50
- Man pages reworked and conserver.passwd page created by Dave
Stuit <djs@gnac.com> - the pain he suffered documenting
the program (and pointing out many of the bugs fixed above)
should help endless others

Bryan Stansell