Call for testing: OpenSSH-6.2

Kevin Brott kevin.brott at
Wed Mar 6 03:42:36 EST 2013

Looks like that did the trick - my old 11.11 build/test host did have GNU
make on it (required for some other packages I build), so I never ran into
this one before.  The loaner I'm running this test on is pretty 'stock'
except for the gcc install, et voila!

openssh-SNAP-20130306.tar.gz seems to be building cleanly across all my
hosts - I'll report back if anything else blows up.

Oops .... spoke too soon ... I did say this was a stock host: so no ssh
already installed ... and I don't have root here ...

$ make tests

        test "no" != yes || \
        /var/tmp/ssh/openssh/ssh-keygen -Bf
/var/tmp/ssh/openssh/regress//t9.out > /dev/null
run test ...
Privilege separation user sshd does not exist
FATAL: sshd_proxy broken
*** Error exit code 1

*** Error exit code 1


Sooo ... it looks like I ought to be able get around this with:
  ./configure --with-zlib=$TLIB --with-privsep-user=nobody

But noooo ...
"/var/tmp/ssh/empty must be owned by root and not group or world-writable."

Just wondering - is everybody just configuring and compiling as root these
days?  *licks gums* "Why in my day, we rolled everything by hand and then
used root just to place files" :p

Guess I'll have to bug the admin to see if I can schmooze him into changing
the root password temporarily so I can run the bloody tests all the way out
... (damn bloody regional territoriality anyway).

More information about the openssh-unix-dev mailing list