[Twisted-Python] Twisted 2.0 prerelease (close!!)

[Stephen Waterbury]

Christopher Armstrong wrote:
> So I did a Twisted 2.0 prerelease....
> ....  Please
> download anything relevant to you and give it a try, and be sure to
> post your experience to this thread even if it went smoothly.

Hey Chris,

I'd like to heartily second Mike's "Thanks for all the
work"!!  I remember at last year's PyCon the discussion of
potentially gnarly issues with the Great Refactorization,
so I know you and all the minions have dealt with a lot, but
the new Twisted is looking just fine!

Since I'm now running Debian unstable on my dev machines, which
I suspect you are too, it's probably no surprise that I had
no problems installing and running TwistedSumo.  All the
PanGalactic tests ran fine (good news because you don't even
wanna know how long it takes to reboot a galaxy ... heh :).

But I noticed one little hiccup that I hadn't seen before
in my server log.  Here's an excerpt (which will probably get
wrapped and look like shit, but anyway):

-------------------------------------------------------------
2005/03/09 16:53 EST [-] pgercred.BasicAuthResource request: <POST /RPC2 HTTP/1.0>
2005/03/09 16:53 EST [admin:PgerXmlrpcService] yo
2005/03/09 16:53 EST [-] 127.0.0.1 - - [09/Mar/2005:21:53:28 +0000] "POST /RPC2 HTTP/1.0" 200 128 "-" "xmlrpclib.py/1.0.1 (by www.pythonware.com)"
2005/03/09 16:53 EST [HTTPChannel,2,127.0.0.1] /usr/lib/python2.3/site-packages/twisted/internet/tcp.py:216: exceptions.UserWarning: SSL connection shutdown possibly unreliable, please upgrade to ver 0.XX
-------------------------------------------------------------

I had never seen that warning before, and it only seems to appear in
this one place (client "login"), but *all* the client/server interactions
use ssl and never get this, so maybe it's just be an over-zealous warning
triggered by a very short interaction:  in the login interaction that
corresponds to these log messages, the client just sends a "yo"
(in xmlrpc) to the server, which then sends a "yo" back to the client,
and that's it (the client then decides that it's "logged in").  I looked
at tcp.py and noticed that some comments around that warning code seemed
to express something less than complete confidence in the appropriateness
of the test that raises that exception, so ...

Thoughts?

Cheers,
Steve