#jython IRC Log (v0.9)

Index

IRC Log for 2014-04-03

Timestamps are in GMT/BST.

[3:44] * lheuer1 (~Adium@f048232201.adsl.alicedsl.de) has joined #jython
[3:47] * lheuer (~Adium@unaffiliated/lheuer) Quit (Ping timeout: 252 seconds)
[4:05] * robbyoconnor (~wakawaka@guifications/user/r0bby) has joined #jython
[4:16] * lheuer1 (~Adium@f048232201.adsl.alicedsl.de) Quit (Quit: Leaving.)
[6:02] * lheuer (~Adium@f048232201.adsl.alicedsl.de) has joined #jython
[6:02] * lheuer (~Adium@f048232201.adsl.alicedsl.de) Quit (Changing host)
[6:02] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[6:21] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Read error: Connection reset by peer)
[6:27] * robbyoconnor (~wakawaka@guifications/user/r0bby) has joined #jython
[8:50] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Excess Flood)
[8:51] * robbyoconnor (~wakawaka@guifications/user/r0bby) has joined #jython
[9:01] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Read error: Connection reset by peer)
[9:01] * robbyoconnor (~wakawaka@guifications/user/r0bby) has joined #jython
[10:51] * Arfrever (~Arfrever@apache/committer/Arfrever) Quit (Ping timeout: 264 seconds)
[12:08] * vext01_ (~edd@88-106-252-6.dynamic.dsl.as9105.com) has joined #jython
[12:12] * vext01 (~edd@host-78-148-191-172.as13285.net) Quit (Ping timeout: 265 seconds)
[12:34] * smaudet (~smaudet@206-51-157-254.nktelco.net) has joined #jython
[13:33] * clajo04 (~clajo04@pool-96-232-190-28.nycmny.fios.verizon.net) Quit (Quit: clajo04)
[13:37] * smaudet (~smaudet@206-51-157-254.nktelco.net) Quit (Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/)
[14:19] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Read error: Operation timed out)
[14:27] * whg (whg@unaffiliated/whg) has joined #jython
[14:40] * [Arfrever] (~Arfrever@apache/committer/Arfrever) Quit (Ping timeout: 245 seconds)
[14:41] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[15:19] <jimbaker> topi`, i have started taking a look at http://bugs.jython.org/issue2070
[15:19] <jimbaker> actually http://bugs.jython.org/issue1973
[15:19] <jimbaker> is better described and has the patch
[15:21] * robbyoconnor (~wakawaka@guifications/user/r0bby) has joined #jython
[15:23] <jimbaker> still need to make it work for the test Arfrever describes in http://bugs.jython.org/msg7467, apparently there's additional logic for such imports asfrom . import *
[15:25] <topi`> I've rarely seen this from . import *
[15:25] <topi`> and I only tested my patch in conjunction with those ordinary cases of from . import foobar
[15:26] <topi`> the patch enabled PyMySql to work with jython
[15:28] * r0bby (~wakawaka@guifications/user/r0bby) has joined #jython
[15:29] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Ping timeout: 240 seconds)
[15:43] * r0bby (~wakawaka@guifications/user/r0bby) Quit (Ping timeout: 240 seconds)
[15:49] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Read error: Connection reset by peer)
[16:12] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[16:12] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Client Quit)
[16:12] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[17:20] * robbyoconnor (~wakawaka@guifications/user/r0bby) has joined #jython
[17:23] * vext01_ (~edd@88-106-252-6.dynamic.dsl.as9105.com) Quit (Quit: upgrade server)
[17:41] * shashank (~shashank@208.91.2.1) has joined #jython
[18:02] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Read error: Connection reset by peer)
[18:03] * robbyoconnor (~wakawaka@guifications/user/r0bby) has joined #jython
[18:24] <jimbaker> topi`, got it
[18:25] <jimbaker> my plan is to fix from . import * as well - should be straightforward
[18:30] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Quit: Konversation terminated!)
[18:40] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[18:48] <agronholm> jimbaker: will you be coming to pycon?
[18:58] <jimbaker> agronholm, sadly, not this time around
[18:58] <agronholm> :(
[18:58] <agronholm> too far away?
[18:59] <jimbaker> no, my talk wasn't accepted (related to jython, on storm), so it was harder to get this set up
[19:00] <jimbaker> but i have better expectations of making it to europython
[19:00] <agronholm> I will not be going there :(
[19:00] <agronholm> I can only afford one pycon a year
[19:01] <agronholm> last year I couldn't even afford that
[19:01] <jimbaker> right, i had the same sort of thinking - if i was paying for myself, i might just go to europython
[19:01] <jimbaker> given i haven't gone since 2008
[19:01] <agronholm> last year the list of talks at europython was not that impressive
[19:01] <agronholm> I only found a couple talks I was interested in
[19:02] <jimbaker> but it's a bit sad because i have been to pycon every year since 2005
[19:02] <agronholm> last year I was actually in Florence at the time of Europython but I didn't attend
[19:03] <jimbaker> so it looks like we will have a large enough group to make for a good sprint at europython on jython
[19:04] <jimbaker> this could be a very good way to make progress against a jython 2.7 release
[19:05] <jimbaker> given that we should be prepared for beta 3 in perhaps 2 weeks, given that socket-reboot basically works
[19:05] <agronholm> what about beta 2?
[19:06] <jimbaker> beta2 is out right now
[19:06] <jimbaker> (although i haven't used it yet as such... i'm totally focused on trunk at this time!)
[19:06] <agronholm> is it? there is no announcement on jython.org
[19:06] <agronholm> and the download links still point to beta1
[19:07] <agronholm> and the topic too
[19:07] <agronholm> nor is there any mention of that on the users mailing lis
[19:07] <agronholm> t
[19:07] <jimbaker> hmmm, i don't believe i have been dreaming about it ;)
[19:08] <agronholm> well you haven't been announcing it either :)
[19:08] <jimbaker> tag here - http://hg.python.org/jython
[19:08] <jimbaker> hey, i was a reluctant gateway to beta 2
[19:08] <jimbaker> ;)
[19:08] <agronholm> who can update jython.org?
[19:09] <agronholm> is this on maven yet?
[19:09] <jimbaker> agronholm, that's my understanding
[19:09] <jimbaker> let me find the posting
[19:09] <jimbaker> http://markmail.org/message/ye6wzvydu5a5seuu
[19:10] <whg> Have they announced where PyCon will be after Montreal?
[19:10] <agronholm> ohh
[19:10] <agronholm> "soft launch"
[19:10] <agronholm> I get it now
[19:10] <agronholm> err
[19:10] <agronholm> waitaminute
[19:10] <agronholm> that post was from last week
[19:11] <agronholm> too bad frank is never on IRC these days
[19:11] <jimbaker> agronholm, he's super busy in his role at adconion doing stuff besides jython
[19:11] <jimbaker> but i did see him last week, when i was in san francisco
[19:13] <jimbaker> whg, http://pycon.blogspot.com/2013/03/looking-for-2016-2017-host-cities.html - my expectation is that they will be announcing at pycon
[19:13] <agronholm> I hope it'll be as easily accessible as Montr??al is
[19:13] <agronholm> Santa Clara wasn't :P
[19:13] <jimbaker> right now i'm in the same building as jesse noller
[19:14] <jimbaker> or van lingberg (fellow rackers, visting our san antonio location)
[19:17] * enebo (~enebo@67.138.38.246) has joined #jython
[19:17] <jimbaker> agronholm, big cities definitely have that advantage. i think austin texas would be a very good pick if they in fact proposed
[19:18] <jimbaker> just a quick connecting flight from major airport hubs in texas (dallas, houston)
[19:18] <agronholm> that would work
[19:18] <agronholm> Helsinki -> NYC -> Austin
[19:18] <agronholm> my next week's flight is Helsinki -> Munich -> Montr??al
[19:19] <jimbaker> i personally would like a pycon not in the middle of basically nothing
[19:19] <agronholm> the nice thing about santa clara was having a hotel room in the same building as the conference
[19:19] <agronholm> that was super convenient
[19:20] <jimbaker> austin is a very convenient airport. i should know - i'm in austin later tonight and will by flying back to denver from there
[19:20] <jimbaker> atlanta was perhaps the best recent venue for pycon - conference hotel, in the middle of downtown
[19:21] <whg> Yay ME!!!
[19:21] <whg> I helped set up wireless back then
[19:21] <whg> I'm glad we made a decent impression
[19:21] <jimbaker> whg, oh cool, yes you did
[19:21] <whg> Even though we're not a very walkable city
[19:21] <jimbaker> whg, there were still some very good places to walk to
[19:22] <whg> The problem is that most of the "best of" places are scattered all over
[19:22] <whg> But yeah, there's some nice stuff downtown these days
[19:22] <jimbaker> i liked the nearby olympic park, and there were several good pubs that could accommodate convention-sized crowds
[19:23] <jimbaker> we also had a great dinner for all of the core python implementers (cpython, pypy, ironpython, jython) at a southern-style restaurant maybe 0.5 mile walk from the hotel
[19:23] <whg> To switch topics with no warning: Is clamped the "best practice" way of using objects written in Python from Java code? And furthermore, can you use clamped to make it so Spring can manage Python objects
[19:24] <whg> Mary Mac's?
[19:24] <whg> Or the Gone With The Wind-themed one?
[19:25] <whg> (I don't get down there very often, but I try to keep track of the well-reviewed options, just in case)
[19:25] <jimbaker> whg, i think it was in fact mary macs - very nice
[19:26] <jimbaker> whg, back to the usual topic for this channel - although we never mind talking about other fun stuff here - yes, clamp is the best practice
[19:27] <jimbaker> but of course, it's still pretty early, but i think we can stop staying pre-alpha or whatever - the api has proven to be good so far
[19:27] <jimbaker> so we are not going to arbitrarily change that part
[19:28] <whg> OK, because I'm about to embark on a new API project at work
[19:28] <jimbaker> given that socket-reboot is basically done and in just clean up mode, i will be back to working on clamp
[19:28] <jimbaker> re spring - i think the key enabler will be adding annotation support
[19:28] <jimbaker> so that's definitely a feature i want to see happen. tell you what - you can be an early adopter and tell me what you need :)
[19:28] <agronholm> jimbaker: jython 2.7b2 is NOT in maven
[19:28] <whg> jimbaker: Heh
[19:29] <agronholm> so apart from being tagged, it hasn't "really" been released
[19:29] <whg> jimbaker: I wouldn't mind, but they jammed me up
[19:29] <jimbaker> agronholm, ok - i'm going to ping frank directly and see where it stands
[19:29] <whg> jimbaker: It needs to have a RESTful front-end, but the actual logic itself must be able to be used in-process from Java as well
[19:30] <whg> jimbaker: And we're using Spring to inject the current Java-equivalent stuff
[19:30] <whg> jimbaker: And there's a time crunch
[19:30] <jimbaker> whg, so here's what you can do right now
[19:30] <jimbaker> as i understand spring injection, it does require annotations
[19:31] <jimbaker> you will have to write some wrapper java objects that then pass through this info to the python code
[19:32] <whg> jimbaker: How much boilerplate are we talking?>
[19:32] <jimbaker> now given that clamped objects like like java, and could be imported as such, this should be reasonable, at least to a small amount
[19:32] <jimbaker> whg, i think it's just a question of construction right?
[19:33] <jimbaker> so it's more that you will need this for each clamped class
[19:33] <whg> jimbaker: We're not doing too much fanciness, I think: just a default constructor and then invoking the setters
[19:33] <jimbaker> whg, yeah, that should be just fine
[19:34] <jimbaker> whg, one more question: are you running on windows?
[19:34] <whg> No
[19:34] <whg> Praise be
[19:34] <jimbaker> because we know clamp is not quite correct there
[19:34] <jimbaker> whg, cool
[19:34] <whg> We dev on OS X and deploy on CentOS
[19:34] <jimbaker> yeah, that's absolutely fine
[19:34] <jimbaker> apparently we get / mixed with \
[19:34] <jimbaker> and that doesn't work for jars
[19:35] <jimbaker> although i suspect if one built on osx/linux/reasonable system, it would just run fine on windows
[19:35] <jimbaker> reminds me: need to find out about my msdn license...
[19:35] <whg> jimbaker: Man, I haven't thought about *that* in a long time
[19:36] <whg> jimbaker: OK, so you think with Jython 2.7 (socket-reboot version) we would be able to do standard web stack stuff (Flask, SQLAlchemy, etc.) and then just use some boilerplate Java fluff + clamped to embed the core logic in a Java process?
[19:37] <jimbaker> flask for sure. sqlalchemy presumably would fine. pjenvey might know... unit tests would tell us for sure
[19:38] <jimbaker> (i say flask should be fine because it's so simple i cannot imagine it has now stopped working)
[19:38] <jimbaker> and of course the usual java, because that's why one would run jython :)
[19:39] <jimbaker> whg, it seems that ssl has been the real roadblock for jython. so now... socket-reboot works and all the sudden stuff is working. i'm sure we will find more bits & pieces, but as long as it's not big runtime issues, we should be fine
[19:41] <whg> jimbaker: Sounds promising. Maybe I can get some time to do POC/Spikes on a couple of parts
[19:42] <whg> jimbaker: But I want to do the Python/Jython route so bad I can taste it
[19:42] <whg> New projects are the best time to switch stacks :-)
[19:43] <jimbaker> whg, absolutely
[19:43] <jimbaker> btw, i want to take a look at using fireside
[19:43] <agronholm> I would love JyNi to death if it worked on Python 3
[19:43] <jimbaker> https://github.com/jythontools/fireside
[19:43] <agronholm> fortunately there is still py4j
[19:43] <agronholm> all my new projects are python 3 based
[19:44] <whg> agronholm: Sadly, none of my Python has actually been related to data analysis
[19:44] <whg> (I'm a transactional man by day, and my ML course is in R)
[19:45] <agronholm> I didn't mention data analysis...
[19:45] <whg> agronholm: NumPy etc. are most commonly used on big datasets, right?
[19:45] <agronholm> I didn't mention NumPy either
[19:45] <agronholm> or big data
[19:45] <whg> agronholm: JyNi does
[19:45] <agronholm> oh
[19:46] <jimbaker> agronholm, but you do want jython 3.x :)
[19:46] <whg> agronholm: I realize it's for all C extensions, but NumPy and friends were getting top billing
[19:46] <agronholm> jimbaker: it would be preferable, yes
[19:46] <jimbaker> whg, one thing i have recently started exploring, enough to start a project around it - https://github.com/jythontools/jiffy/
[19:47] <jimbaker> and registered in pypi
[19:47] <agronholm> right now the project is a frankenstein monster of python and java code in the same process
[19:47] <agronholm> complete conversion has been held back due to a number of reasons
[19:48] <whg> Out of curiosity: how will Java 8 affect Jython? I haven't been following as closely as I used to, so I'm not sure how much of the Java The Language goodness led to new awesomeness on the JVM side
[19:49] <whg> agronholm: "held back due to a number of reasons" has a familiar ring to it, thanks to my day job
[19:49] <whg> agronholm: Hope it's not too frustrating in your case
[19:49] <jimbaker> whg, java 8 should support new goodies for us, but for more for users to access
[19:49] <jimbaker> i'm more interested in the opportunities in zippy
[19:49] <agronholm> whg: it's painful since the server side is stuck with jboss 4
[19:50] <agronholm> I would be a lot happier if I could advance the conversion far enough to kick that out
[19:50] <whg> agronholm: Oh jeez. I finally got out from under that a couple of years ago. I'm sorry.
[19:53] * robbyoconnor (~wakawaka@guifications/user/r0bby) has joined #jython
[20:13] <jimbaker> re jiffy - basically to get cffi support, we need to translate this file to use jffi instead of ctypes. and then, we should be done: https://bitbucket.org/cffi/cffi/src/e6c30206e1db130bb8768b69677b827e575291e3/cffi/backend_ctypes.py?at=default
[20:13] <jimbaker> so that's about 1000 lines - not too bad
[20:17] <whg> Well, that's pretty awesome
[20:19] <whg> That's actually far less horrifying than I expected
[20:49] <jimbaker> whg, the design for cffi is actually quite good - it assumes an absolute minimum of what the actual calls to c need to support, and instead does a fair amount of c source code analysis/generation in pure python
[20:50] <jimbaker> but since the later is (generally) done ahead of time, there's great performance
[20:51] <whg> Sounds pretty nice
[20:51] <jimbaker> in particular, cffi knows nothing about reference counting, or exposing python internals - it just call c and gets back resutls
[20:51] <whg> I vaguely remember doing some JNI back in the Java 1.2 days, and it's sort of scarred me
[20:52] <jimbaker> whg, so jffi is a much better interface to jni
[20:52] <whg> As far as interfaces to C goes, it was not fun
[20:52] <whg> go, even
[21:04] <whg> jimbaker: Then I'm excited about JFFI for NumPy
[21:08] <jimbaker> whg, sadly numpy doesn't plan to use cffi
[21:08] <jimbaker> there's a fork of numpy, numpypy which does
[21:09] <jimbaker> http://buildbot.pypy.org/numpy-status/latest.html
[21:09] <jimbaker> although i believe they use python as much as possible, which means this would require using zippy. but that's a better long term plan from a perf perspective
[21:10] <jimbaker> very much hopeful about jyni
[21:11] <whg> Why is NumPy dodging cffi?
[21:11] <whg> Just too much legacy? Or philosophical objections?
[21:11] <whg> I remember that the PyPy folks had to fork
[21:20] <whg> jimbaker: Do you have a packaged-up version of the socket-reboot work? I think I might do a spike using it, so I was hoping to just stuff it in my local maven repo and go to town on it
[21:20] <jimbaker> whg, sorry, i don't have that
[21:21] <whg> OK, I might clone your repo and see what I can do then
[21:21] <whg> jimbaker Thanks
[21:21] <jimbaker> but it will be part of beta 3, and that should come out in about 2 weeks - there's really no reason that i can think of that will hold up that
[21:21] <jimbaker> whg, cool!
[21:23] <jimbaker> whg, re numpy - i think there's a lot of people who like cython in numpy. but cython really wants support to the c ext api - the same reason why pypy doesn't want to go there, because it slows down their jit
[21:56] <jimbaker> whg, btw, you do need to use my branch of pip for it work - https://github.com/jimbaker/pip
[21:57] <jimbaker> but i think we had this conversation, since this is exactly what we should put into patois
[22:16] * enebo (~enebo@67.138.38.246) Quit (Quit: enebo)
[22:36] * lheuer (~Adium@unaffiliated/lheuer) Quit (Quit: Leaving.)
[22:52] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Ping timeout: 268 seconds)
[23:08] * Thev00d00 (~v00d00@gentoo/developer/TheV00d00) Quit (Ping timeout: 255 seconds)
[23:20] * Thev00d00 (~v00d00@89.206.161.76) has joined #jython
[23:20] * Thev00d00 (~v00d00@89.206.161.76) Quit (Changing host)
[23:20] * Thev00d00 (~v00d00@gentoo/developer/TheV00d00) has joined #jython

Index

These logs were automatically created by JythonLogBot_ on irc.freenode.net using a slightly modified version of the Java IRC LogBot (github).