Mailman 3 April 2006 - Twisted

[Twisted-Python] Re: Minimal example of Words/OSCAR?
by Jason A. Lefkowitz April 11, 2006

April 11, 2006

Moe -- this one works perfectly! Thanks for putting up with my newbie-ism :) -- Jason

1 0

[Twisted-Python] Re: Minimal example of Words/OSCAR?
by Jason A. Lefkowitz April 11, 2006

April 11, 2006

Moe, I've been trying to use your oscardemo.minimal.py, but it doesn't seem to work for me -- I launch it, and it asks for my AIM username & password, but after I provide them I don't see my AIM user show up on my buddy list (I'm logged in via IM client as another user, to test the script). When I run oscardemo.py, though, it connects fine. Is there something I have to do to oscardemo.minimal.py before it's runnable? Or am I misunderstanding the way Words works (i.e. it is connected, but you need to do something more to have the bot show up on people's buddy lists)? -- Jason

2 1

[Twisted-Python] marking 3 problem tests on Windows 'todo'
by glyph＠divmod.com April 11, 2006

April 11, 2006

Currently there are 3 tests which are causing problems on the select reactor on win32, as detailed by this build report: http://twistedmatrix.com/buildbot/win32-select/builds/445/step-default/2 I think they should all be marked 'todo'. Although ideally all the functionality these tests cover should be supported on Windows, the tests themselves are somewhat suspect, and marking them as 'todo' conditionally for the platform would turn the Windows buildbot green, allowing us to notice much more quickly when other, more serious things have gone wrong. I am writing out a big rationale here becuase normally .todo'ing tests which are passing is a big no-no, and I think that this is something of a special case due to the long period of neglect for the platform involved. The test in question are: * twisted.pb.test.test_negotiate.CrossfireReverse.test1 I've talked to brian about this test being todo'd because it is intermittently failing on _all_ platforms; windows just seems to have a far greater chance of failing. * twisted.test.test_udp.MulticastTestCase.testJoinFailure I'm not really sure what is going on here. Some nuance of error reporting in Python? 127.0.0.1 _is_ considered a multicast address on Windows? This seems like it should be TODO'd conditionally, and there should be a ticket to conclusively diagnose it. * twisted.web2.test.test_http.SSLServerTest.testLingeringClose Buffering 'lo' on Windows is weird. I don't think that there is a proper way to fix this; IMHO the test is slightly buggy. It's worth noting that the thing that's blowing up is a direct call to openssl's send(), _not_ a Twisted API. While I think that this test is sort of suspicious, and it should figure out a way to be more lenient to its subprocess, I think that it can be todo'd for win32 since the behavior being tested isn't platform dependent, only the client APIs involved. I believe that the ticket here should use Twisted in the subprocess so that if there is a portability issue with buffering, we can properly abstract it. * twisted.test.test_ssl.ConnectionLostTestCase.testFailedVerify This is reporting the wrong error, but it is correctly dropping the connection. There is a small security risk here, which bothers me, but there may be no way to expose the proper information without fixing bugs in openssl. Considering that this passes on UNIX I believe it should also be conditionally marked on Windows. The ticket here is simply to get the test to pass, since I think the test itself is OK in this circumstance. In conclusion: one of these tests should be todo'd on all platforms, the rest should be todo'd conditionally if isWindows(). Any objections?

3 3

[Twisted-Python] PB v XMLRPC v CORBA v SOAP
by Brendan Simon April 10, 2006

April 10, 2006

I have the following application in mind. 1. Using a Configuration Management System that only runs on UNIX. 2. Write an application for Windows to remotely connect to the UNIX host and perform commands. 3. Optionally add a GUI (wxPython) to the application. The CMS has a command to output XML of parts of its internal database which will make it easy to display the info in a client app :) For me the question is how do I get the info to/from the CMS on a remote server. I think writing a small server app to interpret my requests and relay the XML output to the client is the answer. I want to use SSH which is how I got onto Twisted (Conch) though I do realise that Paramiko is available too. I also want to make use of rsync. Reading some Python books, suggested technologies are Twisted's Perspective Broker, XMLRPC, CORBA and also SOAP is mentioned. CORBA seems like overkill for my simple server. SOAP I don't know enough about yet. XMLRPC seems very appropriate. PB seems easy enough too. One big question for me is portability, such that other clients written in other languages could access the same data. I'm thinking that all technologies would have this property _except_ for PB. Is that correct? i.e. could a C, perl or ruby application access my server if the server was written using Twisted PB ??? If not would anyone recommend SOAP or XMLRPC instead?? Also, it would be nice if I could use any transport mechanism to send the commands. eg. http, https, ssh, rsh. Can Twisted be used to provide such a generic transport mechanism? If not would, SOAP or XMLRPC be better choices, or is the question of transport irrelevant. I was thinking of creating my own XML files to request information from my server. Is this the way to go, or am I recreating the wheel? Thanks, Brendan.

3 3

[Twisted-Python] Re: Minimal example of Words/OSCAR?
by Jason A. Lefkowitz April 10, 2006

April 10, 2006

Thanks, Moe! Very helpful. -- Jason

1 0

[Twisted-Python] Minimal example of Words/OSCAR?
by Jason A. Lefkowitz April 10, 2006

April 10, 2006

Hey -- I'm just getting started with Twisted, and I'm particularly interested in getting up to speed with Twisted Words. I'm curious if there is a minimalistic demo example anywhere of a bot that connects to the OSCAR/AIM network. I've been looking over the included oscardemo.py and it's been helpful, but it seems to include a lot of functionality (handling chat-room requests, change of status of buddies, etc.) and I'm having trouble parsing out what is absolutely necessary just to establish a connection with the OSCAR network. Does anyone know of an example that just shows this? Thanks in advance! -- Jason Lefkowitz

3 2

Re: [Twisted-Python] PB v XMLRPC v CORBA v SOAP
by Oliver Scheel April 10, 2006

April 10, 2006

Hi, > Reading some Python books, suggested technologies are Twisted's > Perspective Broker, XMLRPC, CORBA and also SOAP is mentioned. > > CORBA seems like overkill for my simple server. > SOAP I don't know enough about yet. > XMLRPC seems very appropriate. > PB seems easy enough too. I was exactly in the same situation last month :-) My thoughts: CORBA ist quite nice, wide spreaded, but more work to implement. PB ist quite nice, at least twisted only spreaded, but easy to implement. XML-RPC is included with Python. SOAP is much better the way data types and other things are handled, but not part of Python. Both are easy to use. Twisted favour is SOAPpy and that is build the (pythonic) way like "xmlrpclib". I use both, but I built the services and the clients that way, that I can change between XML-RPC and SOAP. I need XML-RPC, because one service is located on a Windows 98 machine and I want to reduce complexity out there :-) Oliver

1 0

[Twisted-Python] Re: [Twisted-commits] r16582 - Remove util.wait and all uses.
by Jean-Paul Calderone April 10, 2006

April 10, 2006

On Sun, 09 Apr 2006 21:21:34 -0600, Jonathan Lange <jml(a)wolfwood.twistedmatrix.com> wrote: >Author: jml >Date: Sun Apr 9 21:21:33 2006 >New Revision: 16582 > >Modified: > trunk/twisted/conch/test/test_filetransfer.py > trunk/twisted/conch/test/test_telnet.py > trunk/twisted/lore/nevowlore.py > trunk/twisted/mail/test/test_mail.py > trunk/twisted/mail/test/test_pop3client.py > trunk/twisted/names/test/test_names.py > trunk/twisted/protocols/ftp.py > trunk/twisted/test/test_adbapi.py > trunk/twisted/test/test_defgen.py > trunk/twisted/test/test_ftp.py > trunk/twisted/test/test_reflector.py > trunk/twisted/trial/runner.py > trunk/twisted/trial/test/test_util.py > trunk/twisted/trial/unittest.py > trunk/twisted/trial/util.py > trunk/twisted/web2/test/test_stream.py >Log: >Remove util.wait and all uses. > > [snip] > >Modified: trunk/twisted/protocols/ftp.py >============================================================================== >--- trunk/twisted/protocols/ftp.py (original) >+++ trunk/twisted/protocols/ftp.py Sun Apr 9 21:21:33 2006 >@@ -660,7 +660,7 @@ > if cmd == 'PASS': > return self.ftp_PASS(*params) > else: >- return BAD_CMD_SEQ >+ return BAD_CMD_SEQ, "PASS required after USER" > > elif self.state == self.AUTHED: > method = getattr(self, "ftp_" + cmd, None) >@@ -672,7 +672,7 @@ > if cmd == 'RNTO': > return self.ftp_RNTO(*params) > else: >- return BAD_CMD_SEQ, "Blah" >+ return BAD_CMD_SEQ, "RNTO required after RNFR" > > > def ftp_USER(self, username): > Hmmm. I don't see any changes related to wait() in these hunks. ;) > [snip] >Modified: trunk/twisted/test/test_ftp.py >============================================================================== >--- trunk/twisted/test/test_ftp.py (original) >+++ trunk/twisted/test/test_ftp.py Sun Apr 9 21:21:33 2006 > [snip] >@@ -86,7 +82,10 @@ > # Connect a client to it > portNum = self.port.getHost().port > clientCreator = protocol.ClientCreator(reactor, ftp.FTPClientBasic) >- self.client = wait(clientCreator.connectTCP("127.0.0.1", portNum)) >+ d = clientCreator.connectTCP("127.0.0.1", portNum) >+ def gotClient(client): >+ self.client = client >+ return d.addCallback(gotClient) > > def tearDown(self): > # Clean up sockets While it doesn't appear that this change *introduced* a new race condition, it does seem as though it has jostled things so as to cause an existing one to start going the wrong way. This seems to be a pretty common one in the test suite, so everyone pay attention. Just because one side of a TCP connection has had its connection-established-successfully callback fired (ie, buildProtocol on the Factory) does not mean that both sides of the TCP connection have had this callback fired. Quite often, in fact, this is not the case (particularly on platforms you forgot to run the unit tests on ;). This ftp setUp needs to be made to wait for both the server and the client to become connected before proceeding. The failure to do so can be seen as a new failing test on the reactors buildslave in the Qt step. (Of course, the Qt step was already red, because a previous branch uncovered still another race condition.) Please revert the branch or fix the race post-haste. If someone would take a look at the failing tcp half-close test on that reactor, I would appreciate it, otherwise I'll try to get to it sometime in the next couple days. Jean-Paul

2 1

[Twisted-Python] Weekly Bug Summary
by exarkun＠twistedmatrix.com April 9, 2006

April 9, 2006

Bug summary ______________________ Summary for 2006-04-02 through 2006-04-09 Bugs opened: 10 Bugs closed: 5 Total open bugs: 580 (+5) |== Type Changes |== Priority Changes |== Component Changes |Defect: +1 |Highest: -2 |Core: +5 |Enhancement: +5 |High: +0 |Names: +1 |Normal: +8 |Pb: +0 |Trial: -2 |Web: +1 |Web2: +1 New / Reopened Bugs ______________________ ===== High ===== [#1618] twisted.pb.test.test_negotiate.CrossfireReverse.test4 fails intermittently (opened by jml) (CLOSED, duplicate) defect pb http://twistedmatrix.com/trac/ticket/1618 ===== Normal ===== [#1619] Reloading html page with unchanged CSS generates traceback (opened by jacob) defect web2 http://twistedmatrix.com/trac/ticket/1619 [#1620] don't break an usage.Options error message when it contains encoded big5 (opened by radix) enhancement core http://twistedmatrix.com/trac/ticket/1620 [#1621] Memory leak: twisted.web.http.Request objects should call _cleanup() on lost connection (opened by jon) defect web http://twistedmatrix.com/trac/ticket/1621 [#1622] twisted.names.client.Resolver.__init__.__doc__ error (opened by Nerdbeard) defect names http://twistedmatrix.com/trac/ticket/1622 [#1623] wxreactor/wxpython does not work well together - GUI "hangs" (opened by robertckcheung) enhancement core http://twistedmatrix.com/trac/ticket/1623 [#1624] A way to specify time format on the twistd commandline (opened by moof) (CLOSED, wontfix) enhancement core http://twistedmatrix.com/trac/ticket/1624 [#1625] Integrate Mike Orr's finger tutorial additions (opened by hypatia) enhancement core http://twistedmatrix.com/trac/ticket/1625 [#1626] Integrate Nicola Larosa's collection of interesting mailing list posts into the documentation (opened by hypatia) enhancement core http://twistedmatrix.com/trac/ticket/1626 [#1627] Integrate Mike Fletcher's blog posts into the docs (opened by hypatia) enhancement core http://twistedmatrix.com/trac/ticket/1627 Closed Bugs ______________________ ===== Highest ===== [#1470] test_reporter.TestErrorReporting.test_timing uses clock and tests many things (opened by jml, closed by jml, fixed) defect trial http://twistedmatrix.com/trac/ticket/1470 [#1538] Trial timing reporter doesn't work (opened by jml, closed by jml, fixed) defect trial http://twistedmatrix.com/trac/ticket/1538 ===== High ===== [#1618] twisted.pb.test.test_negotiate.CrossfireReverse.test4 fails intermittently (opened by jml, closed by exarkun, duplicate) defect pb http://twistedmatrix.com/trac/ticket/1618 [#1354] [PATCH] t.p.usage error messages mangle option case (opened by Tv, closed by radix, fixed) defect core http://twistedmatrix.com/trac/ticket/1354 ===== Normal ===== [#1624] A way to specify time format on the twistd commandline (opened by moof, closed by exarkun, wontfix) enhancement core http://twistedmatrix.com/trac/ticket/1624

1 0

[Twisted-Python] Cleanly shutting down Service
by apocalypznow April 8, 2006

April 8, 2006

How do I cleanly shut down a Service? I want the shut down to make the stopService() method get called.

6 7