pmxbot IRC Log Viewer

[02:15:50] <aclark> hah

[04:20:04] <aclark> DCPython_Hubot: ascii me happy mailman day

[04:20:05] <DCPython_Hubot> _ _ _ _

[04:20:05] <DCPython_Hubot> | | (_) | | |

[04:20:05] <DCPython_Hubot> | |__ __ _ _ __ _ __ _ _ _ __ ___ __ _ _| |_ __ ___ __ _ _ __ __| | __ _ _ _

[04:20:05] <DCPython_Hubot> | '_ \ / _` | '_ \| '_ \| | | | | '_ ` _ \ / _` | | | '_ ` _ \ / _` | '_ \ / _` |/ _` | | | |

[04:20:05] <DCPython_Hubot> | | | | (_| | |_) | |_) | |_| | | | | | | | (_| | | | | | | | | (_| | | | | | (_| | (_| | |_| |

[04:20:06] <DCPython_Hubot> |_| |_|\__,_| .__/| .__/ \__, | |_| |_| |_|\__,_|_|_|_| |_| |_|\__,_|_| |_| \__,_|\__,_|\__, |

[04:20:06] <DCPython_Hubot> | | | | __/ | __/ |

[04:20:07] <DCPython_Hubot> |_| |_| |___/ |___/

[04:21:03] <aclark> barry: i bet you never imagined you'd have an entire day named after your software!

[05:57:23] <craigkerstiens> aclark: ping

[06:11:55] <aclark> craigkerstiens: sup!

[06:12:18] <craigkerstiens> aclark: wanted to get some feedback from you around the january meetup

[06:12:27] <aclark> craigkerstiens: k

[06:12:51] <craigkerstiens> i suspect there will be at least 2 of us there, maybe 3, but probably 2 that could split up a few parts of a talk

[06:12:59] <aclark> ok

[06:13:11] <craigkerstiens> would intro to heroku be worthwhile or skipping down to deeper/advanced stuff be good as well

[06:13:23] <craigkerstiens> we're pretty flexible, just want to make sure its most valuable to everyone

[06:14:06] <aclark> probably both, we have a wide range of folks from first timers to guru level like hazmat

[06:15:21] <aclark> i know i'm looking forward to the intro; if i'm lucky i'll be able to spin up a few apps before the meeting so i can ask questions

[06:16:21] <aclark> craigkerstiens: you guys support python 3, out of curiousity?

[06:16:44] <craigkerstiens> aclark: not yet, but can easily be done

[06:17:05] <craigkerstiens> our build packs are open sourced so can easily be extended

[06:17:11] <aclark> cool

[06:17:29] <craigkerstiens> so pypy/python3 could easily be done

[06:18:11] <aclark> CherryPy works in python3 IIRC

[06:19:44] <aclark> craigkerstiens: that plone stuff is quite a feat. people have been dancing around doing pip install plone for years without quite getting it right.

[06:20:10] <craigkerstiens> aclark: I was impressed to see it from day 1 of the launch actually

[06:20:14] <aclark> nice to see that working

[06:26:21] <craigkerstiens> actually, that could be fun to live add python 3 or pypy support

[06:26:31] <craigkerstiens> actually go into a bit of how it actually works

[06:26:53] <aclark> sounds good to me :-)

[06:27:07] <craigkerstiens> could be rather fun

[06:27:26] <craigkerstiens> might have to script it out slightly of course, but adding something like pypy support live shows some of the flexibility

[06:29:34] <craigkerstiens> another option is maximizing what you can do in a single process

[06:31:44] <aclark> oh?

[06:37:03] <craigkerstiens> concurrency, but also things like having a worker and web process in a single dyno

[06:37:04] <craigkerstiens> https://github.com/fdr/double-boiler

[06:37:13] <aclark> oh neat

[06:40:27] <craigkerstiens> but definitely helps to know to spread it from intro stuff to advanced stuff

[06:40:34] <craigkerstiens> we should be able to cover a full spectrum

[06:41:03] <aclark> yeah, cool

[07:06:17] <kennethreitz> craigkerstiens: or, you know, just a shell script

[07:06:23] <kennethreitz> :)

[07:06:38] <craigkerstiens> kennethreitz: eh?

[07:06:46] <kennethreitz> web: run.sh

[07:06:55] <kennethreitz> gunicorn blah blah &

[07:07:00] <craigkerstiens> kennethreitz: and scale it up to 100 processes?

[07:07:04] <kennethreitz> mange.py run blah blah

[07:07:16] <craigkerstiens> kennethreitz: my favorite is always saving to the file system, then heroku restart

[07:07:31] <craigkerstiens> then saving to the filesystem and heroku scale web=2

[07:07:33] <kennethreitz> oh does doubleboiler actually use forman?

[07:07:43] <craigkerstiens> it does

[07:07:49] <kennethreitz> nice

[07:07:51] <craigkerstiens> its heroku inside heroku

[07:15:38] <kennethreitz> yo, dawg...

[07:15:59] <kennethreitz> I would have clalled it xzibit

[14:32:57] <MattBowen> anyone know if chameleon is smart about escaping non-structure text? It's looking like it is, because I"m passing through strings that have & in them to tal:content without structure, and it's not coming out &amp;

[14:33:10] <MattBowen> (which is what i'd expect it to do)

[14:35:35] <hazmat> MattBowen, yeah.. it does escape by default

[14:35:53] <hazmat> MattBowen, you need the 'structure:' directive to unescape if your using tal syntax

[14:35:57] <MattBowen> hazmat: right, what's weird to me is if I pass in something that's already escaped, it doesn't escape all the &'s again

[14:36:22] <hazmat> MattBowen, its wierd to you that it doesn't double escape?

[14:36:35] <hazmat> isn't that a good thing?

[14:37:12] <MattBowen> hazmat: it is, it's jut not what I'd expect. Like, if I have tal:content="'Cats & Dogs'"

[14:37:27] <MattBowen> I'd expect it to see the & and say "need to escape that ot be "Cats &amp; Dogs"

[14:37:50] <barry> aclark: it always makes slogging through a bazillion bounce messages so much more enjoyable :)

[14:38:22] <hazmat> MattBowen, the escape impl is probably smart about not considering entities, instead of a simple string replace

[14:38:56] <MattBowen> hazmat: yeah, that's what it looks like. i was trying to find that documented somewhere, but for my purposes, I'll trust what i'm seeing

[14:39:32] <hazmat> heh barry you got a moment? i've been trying to track down a wierd problem with py2.7 only on precise and only on the buildd servers

[14:39:52] <barry> hazmat: sure, what's up?

[14:40:59] <MattBowen> hazmat: thank you

[14:41:28] <hazmat> barry, so juju launches another python process as part of its test suite, to support dev environments we basically set pythonpath env for the new process to ":".join(sys.path) the problem is that python process seems to die.. the only output is

[14:41:29] <hazmat> Failure: juju.machine.errors.UnitDeploymentError: /usr/lib/python2.7/dist-packages/zope/__init__.py:3: UserWarning: Module argparse was already imported from /usr/lib/python2.7/argparse.pyc, but /usr/lib/python2.7/dist-packages is being added to sys.path

[14:41:29] <hazmat> import pkg_resources

[14:41:44] <hazmat> https://launchpadlibrarian.net/86299197/buildlog_ubuntu-precise-i386.juju_0.5%2Bbzr424-1juju1~precise1_FAILEDTOBUILD.txt.gz

[14:41:57] <hazmat> hmm.. now that i saw it, its not clear thats enough to go on..

[14:42:02] <hazmat> its something that only happens on precise though

[14:42:25] <hazmat> and i haven't been able to reproduce it on an ec2 precise instance running the tests by hand

[14:42:42] <barry> looks like a warning getting turned into an error by juju?

[14:43:25] <hazmat> barry, its the stdout/stderr of the sub process being turned into an error

[14:43:37] <hazmat> the subprocess had a non zero exit code

[14:43:37] <hazmat> hmm

[14:43:55] <barry> i've seen those warnings before but i can't remember in which context

[14:44:22] <barry> ah, good ol' pkg_resources

[14:44:37] <hazmat> hm.. i think i see the problem

[14:44:53] <hazmat> we expect all the logs to a file

[14:45:02] <hazmat> and this is an import level stderr message

[14:45:24] <hazmat> and the parent process see the stderr output on process exit and assumes there's a problem

[14:45:42] <hazmat> which causes the exception

[14:46:03] <hazmat> barry, do you know why that warning is being generated?

[14:46:25] <hazmat> but only on precise

[14:46:41] <hazmat> so its argparse is in the stdlib as of 2.7 and we still have a separate argparse package

[14:47:36] <barry> hazmat: i do remember seeing it before but i don't remember the cause. :/ i think it doesn't come from python but from the import of pkg_resources

[14:47:53] <barry> (almost certainly also w/debian patch to use dist-packages)

[14:48:13] <barry> hazmat: i'd have to re-investigate. is that something you'd like some help with?

[14:48:29] <hazmat> barry, cool, i think i've got enough to go on, simple fix to avoid it is just don't have a dep python-argparse where python == 2.7

[14:48:50] <barry> yep

[14:48:54] <hazmat> barry, thats all i need, thanks

[14:49:12] <barry> hazmat: cool. i'll be around :)

[16:41:56] <djbpython1> anyone want to build something and get a piece of that 1/2 billion dollar venture fund?

[16:59:05] <aclark> barry: :-)

[16:59:12] <aclark> morning y'all

[18:07:07] <hazmat> djbpython1, which fund?

[18:18:49] <djbpython1> hazmat: http://www.washingtonpost.com/business/economy/leonsis-case-found-450-million-fund-to-aid-washington-area-start-up-firms/2011/11/30/gIQAtb1fEO_story.html

[18:22:21] <aclark> !slap aclark

[18:27:00] <papna> !respondtoctcprequests

Log file Viewer

Help | Karma | Search:

#dcpython logs for Thursday the 1st of December, 2011