Call for release testing

[Darren Tucker]

Corinna Vinschen wrote:
> I've just tested against the latest from CVS (checked out 10:00 UTC)
> and I got a new regression failure which looks pretty bogus to me.

You're right, it's bogus.  It's a problem with the test itself.

> In banner.sh, the `ssh -q' test fails.  The reason is that banner.out
> is not empty:
> 
> $ cat banner.out
> debug1: permanently_set_uid: 18/0
> 
> 18 is the uid of the SYSTEM account on Windows so the above is equivalent
> to `debug1: permanently_set_uid: 0/0' on other systems.
> 
> What's the debug message doing in the output here?

I bumped the loglevel in the regression tests to make easier (or even 
possible :-) to figure out the cause of failing tests.  Normally that 
debug output is sent to /dev/null unless being specifically logged.

That message is from the session (via the client).  Not sure why it's 
only causing problems on Cygwin, though.  Will look at it tomorrow.

-- 
Darren Tucker (dtucker at zip.com.au)
GPG key 8FF4FA69 / D9A3 86E9 7EEE AF4B B2D4  37C9 C982 80C7 8FF4 FA69
     Good judgement comes with experience. Unfortunately, the experience
usually comes from bad judgement.