Pending OpenSSH release, call for testing.

[Jason McCormick]

  I believe I've found the cause of the failing tests in 
dynamic-foward.sh.  Through a bunch of testing, I've discovered that 
the testing script wasn't giving the ssh forwarding command time to 
tear down from the protocol 1 tests after the -HUP to the process 
before trying to establish the protocol 2 tests.  All of the protocol 1 
tests always ran successfully and the collision occurred when the 
protocol 2 tests tried to start.  The script found there was already an 
ssh running on port 4243, moved on and then died as soon as it tried to 
connect to it to run the tests since the listener on port 4243 was 
reaped by then.  Adding a sleep step to the script always allows this 
to complete properly.

  Why this is different between Fedora and Gentoo I'm not sure.  My one 
box is Gentoo running ~x86 and exhibits this problem.  My Fedora box is 
Core 2 and does not exhibit this behavior.  Both are Linux 2.6, GNU 
coreutils 5.2.1, gcc 3.3, etc.. .  The only difference that comes to 
mind is that the FC2 box is a 400Mhz PIII vs 1.3Ghz Athlon.  

  I've attached a patch to this post and will also file the patch to 
Bugzilla if there's no objections.  It fixes the regression test on 
Gentoo and doesn't seem to break anything in Fedora.

Thoughts?  Peter can you try this patch?

-- 
Jason McCormick 
jason at devrandom.org
GPG Key ID: 96D6CF63
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dynamic-forward.sh.patch
Type: text/x-diff
Size: 271 bytes
Desc: not available
Url : http://lists.mindrot.org/pipermail/openssh-unix-dev/attachments/20040816/99fb106e/attachment.bin