Mailman 3 November 2010 - Twisted

Re: [Twisted-Python] Announcing Twisted 10.2.0pre3!
by exarkun＠twistedmatrix.com 29 Nov '10

29 Nov '10

On 24 Nov, 10:57 pm, eric.faurot(a)gmail.com wrote: >On Wed, Nov 24, 2010 at 9:28 PM, Glyph Lefkowitz ><glyph(a)twistedmatrix.com> wrote: >>Hi Eric, >> >>Thanks for participating in the release process! It's very >> satisfying to actually get some feedback in response to a prerelease >> ;). >> >>If OpenBSD were a supported platform, a bug like this would cause >>the release process to stop until it could be addressed; as it is, >>we will try to merge a bugfix into the release branch if it is >>available and reviewed in time (before next Monday). Being a >>supported platform means that there's a buildbot, and that it shows >>up on this page: >><http://buildbot.twistedmatrix.com/boxes-supported>. Of course, if >>OpenBSD were a supported platform, this bug never would have been >>introduced in the first place, because we would have noticed failing >>tests immediately! >> >>Does anyone feel like contributing an OpenBSD buildslave, and >>hopefully also some more expertise for fixing bugs on that platform? >>Or, for that matter, any other *BSDs? > >Sure. I have a couple of test machines for OpenBSD (amd64 and sparc64) >that are mostly idling, and that could be used for that. I just need >to know the procedure for setting up a buildbot, and I'll see what I >can do. That would surely help me with the port maintainance. That'd be great. The machines need to be online basically 100% of the time, and be able to make an outgoing TCP connection. Other than that, they'll stay idle except when someone commits to trunk or wants to try out a branch before merging. You can find more detailed instructions on the wiki: http://twistedmatrix.com/trac/wiki/ContinuousIntegration Jean-Paul

2 2

[Twisted-Python] Weekly Bug Summary
by exarkun＠twistedmatrix.com 27 Nov '10

27 Nov '10

1 0

[Twisted-Python] Announcing Twisted 10.2.0pre3!
by Glyph Lefkowitz 24 Nov '10

24 Nov '10

Despite a brief and unsuccessful attempt by Canadian forces to wrest control of the Twisted release process, it continues. The francophone domination of North America may be inevitable, but it won't happen today. Coming to you from an ominous tower above Cambridge, Massachusetts, Twisted 10.2.0, pre-release 3, is here! You may - and should! - retrieve it from < http://twistedmatrix.com/users/glyph/10.2.0pre3/>. To avoid any further delays in getting this release out, we will be keeping to a tight schedule. Unless there are any undiscovered, unfixed regressions from 10.1.0,* the final release of Twisted 10.2.0 will be made on November 29, 2010. That is **one week from today.* **Developers, please install the release now and test with your applications, so you won't be caught off guard when the final release is made. This new pre-release includes important bugfixes and several exciting new features. Just since the previous prerelease: - A new, simpler, *substantially* more robust CoreFoundation reactor. http://tm.tl/1833 - A plug-in API for adding new types of endpoint descriptions. http://tm.tl/4695 Other highlights, since the 10.1 release: - Improvements to the implementation of Deferred which should both improve performance and fix certain runtime errors with long callback chains. http://tm.tl/411 - Deferred.setTimeout is (finally) gone. http://tm.tl/1702 - NetstringReceiver is substantially faster. http://tm.tl/4378 Thanks to everyone who worked on the release automation tools, for making this release possible. I can guarantee that if I had to spend a moment longer on it, it wouldn't have gotten done.

2 3

Re: [Twisted-Python] [Twisted] #4747: tw.inet.test.ProcessTestsBuilder.test_openFileDescriptors is broken in many ways
by Eric Faurot 24 Nov '10

24 Nov '10

On Wed, Nov 24, 2010 at 3:28 PM, Twisted <trac(a)twistedmatrix.com> wrote: > #4747: tw.inet.test.ProcessTestsBuilder.test_openFileDescriptors is broken in many > ways > ----------------------------+----------------------------------------------- > Reporter: soyt | Owner: soyt > Type: regression | Status: new > Priority: normal | Milestone: Twisted 10.2 > Component: core | Keywords: > Branch: | Branch_author: > Launchpad_bug: | > ----------------------------+----------------------------------------------- > Changes (by exarkun): > > * owner: glyph => soyt > * keywords: review => > > > Comment: > > Thanks very much for the patch. We don't have any OpenBSD machines for > testing, or I suppose we would have noticed this sooner. :) > > To answer the question of what functionality it provides that's worth > testing... `_listOpenFDs` is used by the `reactor.spawnProcess` support > to find out which file descriptors need to be closed so that they aren't > inherited by the child process. So, it must work right otherwise the > child may inherit file descriptors you did not want it to inherit. This > is an optimization over trying to close (roughly) every integer in the > range (3, RLIMIT_NFILES), where RLIMIT_NFILES is often around 1024. So in my opinion, the function should really return the list of actually opened files in all cases, and not fallback to a list of possible list, otherwise the test case makes no sense. Of course this means trying all fds as last resort on platforms that don't provide a simpler way to do it. Doing it there or doing it by ignoring EBADF on close() later would be the same, as far as "suboptimization" is concerned. Another (nicer?) possibility to achieve the "close non standard fds" would be to use closefrom(). > I don't think the patch is quite correct, since ''at most'' the one extra > `os.listdir` file descriptor should be in the result set. Anything more > than that indicates the desired functionality isn't being provided. The > more correct test assertion would be that the output is ''either'' > range(3) or range(4). However, I'm not sure what you mean by the "uses a > pair of interconnected pipes" comment. Uses it for what? To implement > `os.listdir`? No, it's much worse :) While trying to fix the issue, I was surpised to notice that python had 5 opened fds by default here... After digging a bit further, I found out it is actually due to python being linked with libpthread, which installs a pair of pipe at init... > It would be nice if we didn't have to scan and close all of > range(RLIMIT_NFILES) on OpenBSD though. Do you know if there is any way > to get this information on that platform? For the general case no, but for the intended purpose I'd say use closefrom() Eric.

2 2

[Twisted-Python] Weekly Bug Summary
by exarkun＠twistedmatrix.com 20 Nov '10

20 Nov '10

1 0

[Twisted-Python] TextEntry
by RSGames Support 20 Nov '10

20 Nov '10

Hi, I have been using the following code to show the user a TextEntryDialog, that would allow users to enter data in, and press return. Since I've started using Twisted, this doesn't work on Mac OS X. The dialog pops up okay, but you can't actually enter any data into the box. The cursor does not appear. When I click inside the box to try to make it appear, it quickly disappears again. def handle_input(data): prompt=data.replace("+INPUT ","") new=wx.TextEntryDialog(global_vars.frame, "", prompt) new.Raise() if new.ShowModal() == wx.ID_OK: socks.sockObj.sendData(new.GetValue()) else: socks.sockObj.sendData("+~!?CANCEL?!~") new.Destroy() Thank you. ____________________________________________________________ Send your photos by email in seconds... TRY FREE IM TOOLPACK at http://www.imtoolpack.com/default.aspx?rc=if3 Works in all emails, instant messengers, blogs, forums and social networks.

4 4

[Twisted-Python] signalfd
by Neal Becker 17 Nov '10

17 Nov '10

On modern linux, signalfd can be used to convert sigchld into an event on a file descriptor. Looks like just what is wanted for processProtocol. There is also a python interface here: https://launchpad.net/python-signalfd/+download This test seems to work for me: import signalfd import signal import subprocess import select signalfd.sigprocmask (signalfd.SIG_BLOCK, [signal.SIGCHLD]) fd = signalfd.signalfd (-1, [signal.SIGCHLD]) child = subprocess.Popen (['sleep', '5']) while True: a,b,c, = select.select ([fd], [], []) print 'a:', a, 'b:', b, 'c:', c if child.poll() != None: break

3 4

[Twisted-Python] wxReactor Issues
by RSGames Support 16 Nov '10

16 Nov '10

Hello, My Twisted Python application with wxReactor crashes every time the user exits the application. The OnClose event calls a method that does reactor.stop(). I've ran this through gdb, and have noticed that this is because 2 threads aren't getting closed. What I find strange though, that is if the server closes the connection to the user, the 2 threads exit, and you are able to close the application our properly. I have tried running reactor.disconnectAll() before the reactor.stop() in my OnClose event, with no luck. The client closes the connection to the server, but the threads do not exit. I have also this patch found here (http://twistedmatrix.com/trac/ticket/3948): if wxreactor.WxReactor.callFromThread is not None: oldStop = wxreactor.WxReactor.stop def stopFromThread(self): reactor.callFromThread(oldStop) wxreactor.WxReactor.stop = stopFromThread This problem occurs when I do and don't call the MainLoop() on the wx.App(). If you all could provide some insight on this, I would be very grateful. Thank you. ____________________________________________________________ Send any screenshot to your friends in seconds... Works in all emails, instant messengers, blogs, forums and social networks. TRY IM TOOLPACK at http://www.imtoolpack.com/default.aspx?rc=if2 for FREE

3 15

[Twisted-Python] sample unit testing for xml-rpc's
by anusha k 16 Nov '10

16 Nov '10

Hello all, I want to know how to test xml-rpc functions written using python-twisted. I came across trial framework. But cant find proper information regarding that. Can anyone provide info regarding this or sample unit test-case how to test that. Thanks in advance -- Njoy the share of Freedom :) Anusha Kadambala

2 1

[Twisted-Python] Error in twisted-python while running nox-core
by Ramana Reddy 15 Nov '10

15 Nov '10

Hi Everyone, I got the following error after running the following command: ./nox_core -i ptcp:6633 switch switch_management hostip hosttracker monitoring jsonmessenger NOX 0.9.0(zaku)~full~beta (nox_core), compiled Nov 15 2010 21:36:36 Compiled with OpenFlow 0x01 00001|nox|ERR:Cannot change the state of 'python' to INSTALLED: 'python' ran into an error: Unable to construct a Python component: Traceback (most recent call last): File "./nox/coreapps/pyrt/pyoxidereactor.py", line 364, in instance return pyoxidereactor(ctxt) File "./nox/coreapps/pyrt/pyoxidereactor.py", line 133, in __init__ signal.signal(signal.SIGCHLD, self._handleSigchld) AttributeError: 'pyoxidereactor' object has no attribute '_handleSigchld' Thanks, Ramana

2 1