Re: Moving telnet/telnetd from base to pkgsrc

[David Holland <dholland-tech%netbsd.org@localhost>]

On Sun, Dec 16, 2018 at 07:02:27PM +0000, maya%netbsd.org@localhost wrote:
 > On Sun, Dec 16, 2018 at 01:54:24PM -0500, Christos Zoulas wrote:
 > > On Dec 16,  6:05pm, dholland-tech%netbsd.org@localhost (David Holland) wrote:
 > > -- Subject: Re: Moving telnet/telnetd from base to pkgsrc
 > > 
 > > | On Sat, Dec 15, 2018 at 04:37:30PM +0000, Christos Zoulas wrote:
 > > |  > I have already started fixing the telnet client code. There is not
 > > |  > so much of it...
 > > | 
 > > | Good luck with that.
 > > | 
 > > | You want the diffs from my attempts twenty years ago or would you
 > > | rather repeat that work? :-/
 > > 
 > > Send them over.
 > > 
 > > christos

Here is the diff:
   https://www.netbsd.org/~dholland/tmp/netkit-telnet-0.18pre.diff

12000 lines...

 > They're also hosted here:
 > 
 > 0.17 release:
 > https://launchpad.net/ubuntu/+archive/primary/+sourcefiles/netkit-telnet/0.17-41.1/netkit-telnet_0.17.orig.tar.gz
 > patches from debian:
 > https://launchpad.net/ubuntu/+archive/primary/+sourcefiles/netkit-telnet/0.17-41.1/netkit-telnet_0.17-41.1.debian.tar.xz
 > patches from fedora:
 > https://src.fedoraproject.org/cgit/rpms/telnet.git/tree/

0.17 wasn't the latest; there was an 0.18pre1 and apparently an
0.18pre1a, but I don't remember the circumstances and they may not
have got pushed out very far.

I have found the 0.18pre1 tarballs if anyone wants them.

Note that the 0.18pre README for telnet says

   The attempt to clean up telnet failed utterly. While this code base
   will still be maintained, features or major changes will not be
   accepted. (Unless they're substantial cleanups... but don't waste your
   time.)

   If you feel the urge to hack the telnet protocol, write a new
   implementation. The world will thank you.


"will still be maintained" did not turn out to be very true, although
apparently brave folks from debian and redhat still have been to some
extent.

Anyone doing anything with this should also pull in the existing
debian and redhat patches.

Plus while looking at it over the last couple days I noticed that
ringbuf_printf in ring.c does the wrong thing on snprintf overflow and
sends trash to the other end, which probably leaks local addresses.
That should get fixed as part of any kind of merge.

Also carray.[ch] is an old version of code we now have elsewhere in
the tree and probably still has at least one bug that was found in
that code since 2002...

-- 
David A. Holland
dholland%netbsd.org@localhost