Mailman 3 May 2015 - Twisted

[Twisted-Python] trial internals help?
by Wilfredo Sánchez Vega 19 May '15

19 May '15

I took a stab at porting trial to use twisted.logger (http://tm.tl/7863) but I have a test failing: twisted.trial.test.test_reporter.TestErrorReporting.test_hiddenException I don’t quite know what that test is doing. I added some additional output on my branch (branches/log-trial-7863) but that didn’t make me any smarter. Is anyone familiar with this code able to explain how I may have broken it? -wsv

2 1

[Twisted-Python] Twisted on Windows, parts 2 and 3
by Christian Long 19 May '15

19 May '15

I added to my series of posts about running Twisted on Windows. Part 2 is about virtual service accounts http://www.christianlong.com/blog/twisted-on-windows-2015-edition-part-2.ht… Part 3 is about permissions http://www.christianlong.com/blog/twisted-on-windows-2015-edition-part-3.ht… Let me know if you have any corrections or improvements. Christian

4 4

[Twisted-Python] Weekly Bug Summary
by trac＠twistedmatrix.com 17 May '15

17 May '15

1 0

[Twisted-Python] Need advice on how to move to the new logging system
by Eeshan Garg 13 May '15

13 May '15

Hello everyone! I am aware that I have sent an email with a similar subject line before, but this time this involves a particular ticket (< https://twistedmatrix.com/trac/ticket/7897>) that I have submitted for review. The ticket is about logging the log events in twisted.internet.protocol.Factory using the new logging system, and it is one of the first tickets about modifying existing logging statements in the Twisted code base to use the new system. While working on this ticket, I ran into some issues that I have summarized in a comment on the same ticket < https://twistedmatrix.com/trac/ticket/7897#comment:2> and it would mean a lot to me if someone could take a look. I recall that HawkOwl had some concerns regarding moving to the new logging system. And so I thought that it would be better if I share my thoughts on the mailing list as well and not just on the particular ticket, so that anybody who is interested in sharing their concerns can do so. Regards, Eeshan Garg

1 0

[Twisted-Python] Twisted 15.2 Prerelease Announcement
by HawkOwl 12 May '15

12 May '15

Hello everyone! It's about time that the owl-themed release management simulation kicked into high-gear, and as such, I delivereth to you all the first prerelease of Twisted 15.2. Bringing not only headlining features but also a lot of incremental improvements, this release has got plenty to like: - twisted.logger has landed! This is a brand-new, feature-rich logging framework. - Python 3.4 is now a supported platform for all the Py3 ported modules. - twisted.trial.unittest.TestCase's assertEqual, assertTrue, and assertFalse methods now pass through the standard library's more informative failure messages. - twisted.python.filepath.FilePath now supports Unicode (text) paths properly, and includes as{Bytes,Text}Mode methods for interacting with APIs that require a text/bytes-only FilePath. - twisted.mail.smtp.sendmail now supports ESMTP and provides a high-level interface for sending mail. - The following parts of Twisted are now ported to Python 3: - twisted.internet.process - twisted.cred.credentials - twisted.python.modules - twisted.internet.kqreactor - twisted.internet.endpoints.ProcessEndpoint - twisted.web.static As usual, tarballs can be found at http://twistedmatrix.com/Releases/pre/15.2.0pre1/ , and the full NEWS file can be found at http://twistedmatrix.com/Releases/pre/15.2.0pre1/NEWS.txt . On a personal note, I'd like to thank Tavendo for sponsoring a majority of my Py3 porting work, and everyone who reviewed tickets for getting it landed. Please test out the new prerelease :) Twisted regards, Amber "Hawkie" Brown

3 2

[Twisted-Python] INCOMPATIBLE CHANGE: Removing PAM support from Twisted
by HawkOwl 11 May '15

11 May '15

As per https://twistedmatrix.com/trac/wiki/CompatibilityPolicy#ProcedureforExcepti…: Twisted's PAM support is reliant on a library which a) doesn't materially exist anymore, b) is blocking, c) is uninstalled on all our buildbots if I'm correct and so therefore hasn't been tested for ages, and d) requires us to do insecure things (like setting euid as root). Since it's basically uninstallable (I can't even find a source tarball newer than 1999) and almost certainly doesn't work on any Python versions we support, I propose outright removal, rather than emitting deprecation warnings that literally nobody will see. I have prepared a patch at https://github.com/twisted/twisted/compare/trunk...remove-pamauth-3728-2 . Under the deprecation policy's exclusions rule, this branch is given for people to make sure that their code does not break. Three other committers will need to also pitch support for this (although I don't think that'll be an issue ;) ). This patch not only removes PAM, but all of its (unusedness) in Conch. The ticket is available at https://twistedmatrix.com/trac/ticket/3728 and will be put in review shortly. The buildbot results can be seen at https://buildbot.twistedmatrix.com/boxes-supported?branch=/branches/remove-… . - Amber

3 2

[Twisted-Python] LoopingCall.withCount countCallable called with 0
by Dario Vinella 10 May '15

10 May '15

Hi all, I found an odd behavior using a LoopingCall.withCount with a very small interval (0.02 sec): sometimes the countCallable is called with 0 as argument and it looks like the now - _realLastTime is smaller than the looping interval. Here you can find the very simple code I use to reproduce the "anomaly", with a main loopingCall and another one that prints its internals https://gist.github.com/dvinella/31d94f9dd0e47586658d and here an extract of the logs it produces https://gist.github.com/dvinella/8cfdce0b9f5fdda7b375 The doc for the withCount states that the argument should be normally 1, higher if the reactor blocks for some reason, but it doesn't say anything about 0. The beahavior is quite misleading since the loop seems to anticipate the interval timeout, saying "I've been called, but I should have not been called" I must add that the anomaly doesn't appear right at the start, but after tens of seconds, mostly when the system is under medium/heavy load. So is it only a documentation issue for an exceptional case, due to timers imprecision, that should be handled at higher levels, or is it a real issue? dario

3 4

[Twisted-Python] Weekly Bug Summary
by trac＠twistedmatrix.com 10 May '15

10 May '15

1 0

[Twisted-Python] New Supported Platforms - Python 3.3 and Python 3.4 on Debian 7
by HawkOwl 08 May '15

08 May '15

Hi everyone! I'm pleased to announce that Twisted will have one more new supported Python version (3.4) as of Twisted 15.2. The debian7-py33 and debian7-py34 builders will shortly be moving from 'unsupported' to 'supported'. - Hawkie

4 4

[Twisted-Python] Crochet 1.4.0: use Twisted anywhere!
by Itamar Turner-Trauring 06 May '15

06 May '15

Crochet is an MIT-licensed library that makes it easier to use Twisted from regular blocking code: from __future__ import print_function from twisted.names import client from crochet import setup, wait_for setup() @wait_for(timeout=5.0) def gethostbyname(name): """Use the Twisted DNS library.""" d = client.lookupAddress(name) d.addCallback(lambda result: result[0][0].payload.dottedQuad()) return d if __name__ == '__main__': # Application code using the public API - notice it works in a normal # blocking manner, with no event loop visible: import sys name = sys.argv[1] ip = gethostbyname(name) print(name, "->", ip) New in 1.4.0 is Python 3.4 support as well as support for using Crochet with uWSGI, and additional documentation on workarounds for various limitations of Crochet. https://pypi.python.org/pypi/crochet for downloads. https://crochet.readthedocs.org/ for documentation. --Itamar Turner-Trauring

2 1