#jython IRC Log

Index

IRC Log for 2013-09-20

Timestamps are in GMT/BST.

[0:05] * lheuer (~Adium@unaffiliated/lheuer) Quit (Quit: Leaving.)
[0:06] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Ping timeout: 246 seconds)
[1:11] * lopex (uid4272@gateway/web/irccloud.com/x-nhlpgpivpqcbufkz) Quit (Ping timeout: 268 seconds)
[1:55] * zz_whg is now known as whg
[2:13] * whg is now known as zz_whg
[2:53] * jimbaker (~jbaker@unaffiliated/jimbaker) has left #jython
[2:53] * jimbaker (~jbaker@unaffiliated/jimbaker) has joined #jython
[3:11] * lopex (uid4272@gateway/web/irccloud.com/x-qzelgrrvmahtxjdm) has joined #jython
[3:26] * lopex (uid4272@gateway/web/irccloud.com/x-qzelgrrvmahtxjdm) Quit (Ping timeout: 245 seconds)
[4:25] * lheuer (~Adium@f048036128.adsl.alicedsl.de) has joined #jython
[4:25] * lheuer (~Adium@f048036128.adsl.alicedsl.de) Quit (Changing host)
[4:25] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[4:27] * lheuer (~Adium@unaffiliated/lheuer) Quit (Client Quit)
[4:34] * lheuer (~Adium@f048036128.adsl.alicedsl.de) has joined #jython
[4:34] * lheuer (~Adium@f048036128.adsl.alicedsl.de) Quit (Changing host)
[4:34] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[5:24] * lopex (uid4272@gateway/web/irccloud.com/x-gatrwpiocbdkbtfg) has joined #jython
[6:59] * mcurve (~quassel@pop.nakinasystems.com) Quit (Read error: Operation timed out)
[6:59] * supersven__ (~sven@port-14866.pppoe.wtnet.de) has joined #jython
[7:00] * mcurve (~quassel@pop.nakinasystems.com) has joined #jython
[7:04] * supersven_ (~sven@port-30049.pppoe.wtnet.de) Quit (Ping timeout: 268 seconds)
[7:44] <topi`> jimbaker: interesting. what are the biggest changes needed?
[7:44] <topi`> as far as I understand, ssl.py already uses Java classes to do stuff
[7:44] <topi`> since PyOpenSSL is not available.
[8:38] <daenney> pyOpenSSL is a nightmare though
[8:58] <daenney> Most SSL libraries are. Too much of the terminology and how their bindings work shine through the implementation, it's difficult for non-savvy users to just pick it up and work with it
[11:02] * supersven__ (~sven@port-14866.pppoe.wtnet.de) Quit (Quit: Ex-Chat)
[13:37] * zz_whg is now known as whg
[13:46] * fwierzbicki1 (~Adium@99-106-170-162.lightspeed.sntcca.sbcglobal.net) has joined #jython
[13:47] * clajo05 (~clajo04@pool-108-46-137-135.nycmny.fios.verizon.net) has joined #jython
[13:48] * fwierzbicki (~Adium@99-106-170-162.lightspeed.sntcca.sbcglobal.net) Quit (Read error: Connection reset by peer)
[13:48] * clajo04 (~clajo04@pool-108-46-137-135.nycmny.fios.verizon.net) Quit (Read error: Connection reset by peer)
[13:48] * clajo05 is now known as clajo04
[13:49] * thereisnospoon (~thereisno@27-33-1-87.tpgi.com.au) Quit (Ping timeout: 245 seconds)
[13:49] * cdleonard (~Crestez_D@205.168.23.154) has joined #jython
[13:49] * thereisnospoon (~thereisno@27-33-1-87.tpgi.com.au) has joined #jython
[14:29] * supersven (~sven@port-14866.pppoe.wtnet.de) has joined #jython
[14:34] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[15:20] <jimbaker> topi`, ssl in trunk is completely broken
[16:47] <topi`> jimbaker: is it broken in 2.5.3 as well?
[16:47] <jimbaker> topi`, yes
[16:48] <jimbaker> this is one case where we should consider a backport
[16:55] <topi`> jimbaker: so far, you've done some work on the ssl.py
[16:56] <topi`> maybe I can help finish the work?
[16:56] <jimbaker> yes, i have something that proves that it works w/ pip/easy_install
[16:56] <topi`> I'm not an openssl guru, though
[16:56] <jimbaker> topi`, well, since we are going against java ssl, no need for that :)
[16:57] <jimbaker> more of a question of figuring how how to translate from cpython's wrapping of openssl to what we need to do in java ssl
[16:57] <topi`> what i've noticed so far is that the java ssl is a lot stricter about cert validation than python/pyopenssl
[16:57] <jimbaker> btw, probably don't want to do class mutation as i did in my branch
[16:57] <jimbaker> right, but since we have a SSLContext, we can control what that means
[16:58] <jimbaker> so give it any desired behavior, including of course supporting ca_certs - not certain what the file format of that needs to be vs what java takes fwiw
[16:59] <jimbaker> it might be worthwhile to provide an insecure mode to jython. or at least doc how to set up one w/ site.py
[16:59] <jimbaker> using that trust package we talked about
[16:59] <jimbaker> hypothetical pypi trust package :)
[16:59] <topi`> yeah, we talked about that :)
[17:00] <topi`> but perhaps rewriting the ssl.py is higher priority
[17:00] <jimbaker> sure, i think rewriting ssl.py is higher priority
[17:00] <jimbaker> i would like to see it eventually rewritten in java... but a python intermediate version is great
[17:01] <topi`> you'd use java classes from python as well, right?
[17:01] <jimbaker> we should not be moving bytes in tight loops w/in python code in jython
[17:01] <topi`> why not?
[17:01] <jimbaker> yes, so i might call that approach, pure jython - python code that uses java classes
[17:01] <jimbaker> topi`, i do care about performance
[17:01] <topi`> I think the biggest perf impact is the startup time
[17:02] <jimbaker> i know how to fix jython to make regular python code performant
[17:02] <topi`> once jython is running, it actually is quite performant (except when loading classes)
[17:02] <jimbaker> topi`, you may have saw whg's post yesterday about the ibm java 8 jvm
[17:02] <jimbaker> that looked quite exciting
[17:02] <topi`> I only saw a post about IBM's paper on sharing JVMs
[17:03] <jimbaker> that's it - and they profile jruby as part of it. i expect jython would see better performance, since we have a different compilation model
[17:03] <jimbaker> they first interpret, then only compile some classes
[17:03] <jimbaker> this reduces their classloading time
[17:04] <whg> jimbaker: Can't claim that one, sadly
[17:04] <whg> I think that was daenney
[17:07] <jimbaker> whg, ok, i mixed up your role at IBM as telling us about it. we can still appreciate this however :)
[17:10] <whg> jimbaker: No worries. Just didn't want daenney getting made at me :-)
[17:35] <fwierzbicki1> jimbaker: This looks good to me, but you have worked more on this issue: https://bitbucket.org/jython/jython/pull-request/5/issue1746-fix-issue-with-static-codecs/diff ??? if you think it looks good too I'm happy to do the final spot checking and pull it into trunk.
[17:35] <fwierzbicki1> I'm looking at Sven's skips now - I should be pulling that sometime today.
[17:36] <jimbaker> fwierzbicki1, this branch looks fantastic
[17:36] <fwierzbicki1> cool - I'll be sure to pull that in next
[17:36] <jimbaker> exactly what i have been hoping we do for PySystemState - banish statics like this
[17:38] <fwierzbicki1> As soon as I do I'll have to go find and apologize to Brandon - I asked him for it, he put it up immediately, and I dropped it on the floor
[17:38] * fwierzbicki1 is now known as fwierzbicki
[17:53] * purplefox (~purplefox@88-105-150-78.dynamic.dsl.as9105.com) Quit (Read error: Operation timed out)
[18:09] * purplefox (~purplefox@88-105-150-78.dynamic.dsl.as9105.com) has joined #jython
[20:58] * purplefox (~purplefox@88-105-150-78.dynamic.dsl.as9105.com) Quit (Ping timeout: 256 seconds)
[21:03] * whg is now known as zz_whg
[22:43] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)

Index

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