Mailman 3 September 2007 - Twisted

[Twisted-Python] waiting for data before shutdown
by Andreas Poisel Sept. 25, 2007

Sept. 25, 2007

Hi! My client is supposed to wait for two seconds to receive data when a shutdown event is triggered. In buildProtocol() the trigger is added: self.sdt= reactor.addSystemEventTrigger('before', 'shutdown', protocol.shutdown) The method shutdown() is supposed to wait for 2 seconds before closing the connection: def shutdown(self): self.doSomeCleanup() # works self.sendSomeData() # works d= defer.Deferred() reactor.callLater(2.0, d.callback, None) return d The shutdown() method gets called, cleanup and sending of data work as expected. But the connection is closed by my client _before_ the two seconds are over. Isn't the shutdown process supposed to be halted until the deferred returned by shutdown() is fired? Thank you! -- Bye, Andreas

1 0

[Twisted-Python] Weekly Bug Summary
by exarkun＠twistedmatrix.com Sept. 23, 2007

Sept. 23, 2007

Bug summary ______________________ Summary for 2007-09-16 through 2007-09-23 Bugs opened: 3 Bugs closed: 5 Total open bugs: 1035 (-2) |== Type Changes |== Priority Changes |== Component Changes |Defect: -2 |Highest: -3 |Conch: -2 |Task: +0 |Normal: +1 |Core: +1 |Words: -1 New / Reopened Bugs ______________________ ===== Highest ===== [#2835] _pollingfile calls pauseProducing on registered Pull producers (opened by Peaker) defect core http://twistedmatrix.com/trac/ticket/2835 ===== Normal ===== [#2833] deprecate twisted.internet.gtkreactor (opened by exarkun) task core http://twistedmatrix.com/trac/ticket/2833 [#2834] Logging should deal nicely with problems logging Failures (opened by itamarst) defect core http://twistedmatrix.com/trac/ticket/2834 Closed Bugs ______________________ ===== Highest ===== [#1884] Document adjustment to review policy (opened by exarkun, closed by exarkun, invalid) task core http://twistedmatrix.com/trac/ticket/1884 [#2099] manhole telnet server broken (opened by exarkun, closed by exarkun, invalid) defect conch http://twistedmatrix.com/trac/ticket/2099 [#1967] twisted.conch.telnet dies in conjunction with PuTTY (opened by sam, closed by exarkun, fixed) defect conch http://twistedmatrix.com/trac/ticket/1967 [#2809] 99.9% cpu load when using the epoll reactor (opened by mvw, closed by exarkun, fixed) defect core http://twistedmatrix.com/trac/ticket/2809 ===== Normal ===== [#2828] xmlstream.IQ shouldn't create timeout deferreds for every IQ (opened by jack, closed by ralphm, invalid) defect words http://twistedmatrix.com/trac/ticket/2828 Ticket Lifetime Stats ______________________ Oldest open ticket - [#38] Should be a way to create documentation aliases (since 2003-07-12 16:08:03). Newest open ticket - [#2835] _pollingfile calls pauseProducing on registered Pull producers (since 2007-09-22 00:36:54). Mean open ticket age: 524 days, 9:14:54.799999. Median: 422 days, 23:27:03.859903. Standard deviation: 401 days, 17:45:04.268446. Interquartile range: 503 days, 18:33:50. Mean time between ticket creation and ticket resolution: 121 days, 7:14:56.715258. Median: 14 days, 10:59:04. Standard deviation is 224 days, 16:56:44.223591. The interquartile range is 127 days, 8:25:41. Mean time spent in review: 65 days, 4:49:15.983357. Median: 4 days, 5:45:01. Standard deviation: 143 days, 15:51:23.365689. Interquartile range: 29 days, 10:44:36. Mean number of times a ticket is reviewed: 1.94600938967. Median: 1 Standard deviation: 1.58466700636. Interquartile range: 1. Contributor Stats ______________________ In the last 4 weeks, 16 unique ticket reporters 5 unique ticket reviewers 3 unique ticket resolvers In the last 24 weeks, 78 unique ticket reporters 14 unique ticket reviewers 14 unique ticket resolvers In the last 48 weeks, 155 unique ticket reporters 19 unique ticket reviewers 26 unique ticket resolvers

1 0

Re: [Twisted-Python] Web Proxy that does something different once a day
by Fred Trotter Sept. 21, 2007

Sept. 21, 2007

> It looks like this is just for the web browsers > (i.e., HTTP protocol). If so, then at step 2 > you would make your test. If you decide A should > go to the "required reading" URL instead of the > requested URL, simply skip steps 3-5 and return > a valid HTTP document with a "moved temporarily" > code (30x, I think). The browser will then request > the other URL. The one catch is that, depending on > the browser, the user may get the same page if they > request the original URL again in the same browser > window/session. If that's the case, they would need > to open a new window to get to the original URL. That sounds reasonable and good. Any clue how you would implement that in Twisted? -FT -- Fred Trotter http://www.fredtrotter.com

1 0

Re: [Twisted-Python] Web Proxy that does something different once a day
by Samuel Reynolds Sept. 21, 2007

Sept. 21, 2007

>Hoping someone else chimes in with how to stick a redirect into a proxy, Nothing to do with Twisted, per se. A proxy is a two-way conduit: 1. Receive a request from A 2. Massage and/or log the request as necessary 3. Make a request to B. 4. Receive the response from B. 5. Massage and/or log the response as necessary 6. Return the response to A. It looks like this is just for the web browsers (i.e., HTTP protocol). If so, then at step 2 you would make your test. If you decide A should go to the "required reading" URL instead of the requested URL, simply skip steps 3-5 and return a valid HTTP document with a "moved temporarily" code (30x, I think). The browser will then request the other URL. The one catch is that, depending on the browser, the user may get the same page if they request the original URL again in the same browser window/session. If that's the case, they would need to open a new window to get to the original URL. - Sam

1 0

AW: Re: [Twisted-Python] Can a deferred catch a segmentation fault ?
by Andreas Kostyrka Sept. 21, 2007

Sept. 21, 2007

Setting up such signal handler wouldn't be trivial (as in unsafe). The "normal" thing in such a situation is to longjmp (in C) or to let the process to die (which is the safe thing to do), Now presuming that you cannot fix the segfaults (which as sad it sounds is quite common when using commercial 3rd party DLLs. Practice shows that being even a huge customer often get a quick turn around on bugs especially if its a hard to reproduce random crash. Driver DLLs for hardware costing $10K upwards come to mind, *shudder*), your only safe way is to isolate the driver usage into a seperate process. (If you are unlucky not even that is enough because some hardware gets stuck in a state that needs manual cold reseting, but let's be optimistic here). With a seperate process, you can restart it should it die. Andreas -- Ursprüngl. Mitteil. -- Betreff: Re: [Twisted-Python] Can a deferred catch a segmentation fault? Von: Jean-Paul Calderone <exarkun(a)divmod.com> Datum: 20.09.2007 16:34 On Thu, 20 Sep 2007 09:14:11 -0700, "Gerald John M. Manipon" <geraldjohn.m.manipon(a)jpl.nasa.gov> wrote: >Hi, > >I'm using deferToThread to run some python code which may or may not >be safe. In particular, the code may load some modules that could >seg fault. In this case, deferToThread won't work and the python >process will bomb so I was wondering what Defer strategy I could >use to handle this situation: Deferreds don't do anything which will directly help in this situation. If your process receives a signal for which the action is to exit, then it will exit. If you want, you can change the action associated with certain signals (even SIGSEGV). What action you *do* cause to happen in such a case is independent of Deferreds. You could, conceivably, turn a segfault into an exception and then errback a Deferred. However, I wouldn't actually recommend this. I would recommend not using software which segfaults, fixing segfault bugs you find in software you are using, and having comprehensive unit tests so that if something is going to segfault, it does so during the course of development, not on whatever server or client machines your software is deployed to. Jean-Paul _______________________________________________ Twisted-Python mailing list Twisted-Python(a)twistedmatrix.com http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python

4 3

[Twisted-Python] Web Proxy that does something different once a day
by Fred Trotter Sept. 21, 2007

Sept. 21, 2007

Hello, I am new to python and I am potentially interested in writing a new protocol with Twisted. However I wanted to start with a simpler problem that I have been working on, so that I can understand better how twisted works. A particular client wants to enforce a home page for the first internet connection each day. They have a intranet portal that gives messages to employees and they want to be sure that each employee, no matter how they have their homepage set, will see the messages the first time they connect. After the first connection they want the proxy to simply forward requests to the internet without filtering. Since I need to understand this issue for my other project too, how do I get the python loop do one thing based on a "new day" for an IP address and then do something else after that action has been completed. Example code that addresses my problem would be great, but I would be happy to have links to examples that do this... Thanks!! -- Fred Trotter http://www.fredtrotter.com

3 3

[Twisted-Python] Twisted Server and Objective-C Client
by Stefan Arentz Sept. 21, 2007

Sept. 21, 2007

I need to let Objective-C (OS X / Cocoa) clients speak to a Twisted server. The protocol needs to be a pretty simple request/response style protocol on a persistent TCP connection. Possible over SSL. Before I start inventing a wheel, does anyone know if there is something simple available already? Would the Perspective Broker be a good start? I haven't looked at the internals but maybe it is not too difficult to create a non-python client for it? Just thinking out loud :) S.

2 1

[Twisted-Python] Can a deferred catch a segmentation fault?
by Gerald John M. Manipon Sept. 20, 2007

Sept. 20, 2007

Hi, I'm using deferToThread to run some python code which may or may not be safe. In particular, the code may load some modules that could seg fault. In this case, deferToThread won't work and the python process will bomb so I was wondering what Defer strategy I could use to handle this situation: from twisted.internet import threads, reactor def segFaultFunc(): import binning #this will seg fault return True def myCb(res): print "Got", res reactor.stop() d = threads.deferToThread(segFaultFunc) d.addCallback(myCb) reactor.run() Any help on this is greatly appreciated. Thanks, Gerald

2 2

[Twisted-Python] Who's using Gtk reactor?
by Jean-Paul Calderone Sept. 20, 2007

Sept. 20, 2007

Greetings, As part of the work I've been doing to fix a bug in epoll reactor's handling of "unclean" tcp disconnection (the case which gives ConnectionLost instead of ConnectionDone), I've noticed that gtk reactor also doesn't handle this case correctly. Anyone who has looked at gtkreactor.py probably knows this (there's a big comment in _readAndWrite, so whoever wrote it in the first place realized what they were doing). Before I fixed epoll reactor, I added a test for this case. Gtk reactor won't be able to pass this test, since it doesn't provide the correct functionality. I'm not going to fix gtk reactor, since glib2 or gtk2 reactor is really what people should be using these days. However, if someone else is interested in fixing gtk reactor, the fix would be happily accepted (the hard part is done, since the unit test is written already ;). If no one is interested in fixing it, though, then I think this is a good time to start thinking about deprecating gtk reactor and eventually removing it. Anyone have any thoughts on this? Jean-Paul

2 2

[Twisted-Python] twisted.trial and test discovery
by Drew Smathers Sept. 19, 2007

Sept. 19, 2007

Maybe a dead horse, but there was some prior discussion on nose vs. trial in Feb: http://twistedmatrix.com/pipermail/twisted-python/2007-February/014771.html I like trial a lot (both for Twisted and non-Twisted applications - heck I've even tested [gasp] Java applications with it). The only somewhat interesting feature of nose (vs. trial) I've found is test discovery. Is test discovery a planned feature for trial? -- \\\\\/\"/\\\\\\\\\\\ \\\\/ // //\/\\\\\\\ \\\/ \\// /\ \/\\\\ \\/ /\/ / /\/ /\ \\\ \/ / /\/ /\ /\\\ \\ / /\\\ /\\\ \\\\\/\ \/\\\\\/\\\\\/\\\\\\ d.p.s

2 3