selftest on win32

[Martin Pool]

On Sat, Oct 25, 2008 at 10:36 AM, Mark Hammond
<mhammond at skippinet.com.au> wrote:
> Any more thoughts on how we can reduce the test noise on Windows?  We are very lucky that Alexander noticed, then took the time to ask about this issue while he was working on something entirely unrelated.  In this case the regression appears to be limited to the test suite, but we shouldn't rely on a chain of luck like this...

(Hm, kind of old thread, I see after writing the reply.)

There is certainly a lot of noise.  I ran it a while ago and saw the
warnings but I don't recall the specifics.  I would say that in
general, the suite should be quiet aside from reporting specific
results.  If anything is printed out, we should change it to either
success, failure, or a known failure.  If you want to talk about
particular messages, post them here (or refer to older threads.)

-- 
Martin <http://launchpad.net/~mbp/>