#jython IRC Log


IRC Log for 2015-04-03

Timestamps are in GMT/BST.

[0:12] * fwierzbicki (~Adium@99-106-169-5.lightspeed.sntcca.sbcglobal.net) Quit (Quit: Leaving.)
[0:19] * fwierzbicki (~Adium@99-106-169-5.lightspeed.sntcca.sbcglobal.net) has joined #jython
[1:01] * [Arfrever] (~Arfrever@apache/committer/Arfrever) Quit (Read error: Connection reset by peer)
[1:01] * [Arfrever] (~Arfrever@apache/committer/Arfrever) has joined #jython
[1:01] * sirdancealot (~koo5@236.152.broadband3.iol.cz) Quit (Ping timeout: 250 seconds)
[1:11] * sirdancealot (~koo5@236.152.broadband3.iol.cz) has joined #jython
[1:40] * xemdetia (xemdetia@nat/ibm/x-kpcqztxqzuhobnny) Quit (Ping timeout: 264 seconds)
[3:15] * mataug (~gautam@ has joined #jython
[3:15] * mataug (~gautam@ Quit (Remote host closed the connection)
[4:47] * lopex (uid4272@gateway/web/irccloud.com/x-mvncjifrlanprlrk) Quit (Ping timeout: 272 seconds)
[4:56] * lopex (uid4272@gateway/web/irccloud.com/x-rqbjneogkexzpvkq) has joined #jython
[5:34] * fwierzbicki (~Adium@99-106-169-5.lightspeed.sntcca.sbcglobal.net) Quit (Quit: Leaving.)
[5:36] * fwierzbicki (~Adium@99-106-169-5.lightspeed.sntcca.sbcglobal.net) has joined #jython
[6:40] * sirdancealot (~koo5@236.152.broadband3.iol.cz) Quit (Ping timeout: 264 seconds)
[6:42] * claudiupopa (~claudiupo@ has joined #jython
[7:55] * mbooth (~mbooth@cpc11-shef10-2-0-cust659.barn.cable.virginm.net) has joined #jython
[7:55] * mbooth (~mbooth@cpc11-shef10-2-0-cust659.barn.cable.virginm.net) Quit (Changing host)
[7:55] * mbooth (~mbooth@redhat/mbooth) has joined #jython
[8:44] * ebarrett (~edd@ Quit (Ping timeout: 255 seconds)
[8:44] <agronholm> fwierzbicki: anything on rc2?
[9:05] * ebarrett (~edd@ has joined #jython
[9:15] <agronholm> jimbaker: I investigated my event listener issue -- it was a problem in my code after all
[10:07] <agronholm> the tests are passing now, although I had to replace super() calls with direct calls to the superclass __init__ due to #1540
[10:18] * mbooth (~mbooth@redhat/mbooth) Quit (Ping timeout: 250 seconds)
[10:20] * ebarrett_ (~edd@vext01.default.vext01.uk0.bigv.io) has joined #jython
[10:23] * ebarrett (~edd@ Quit (Disconnected by services)
[10:23] * ebarrett_ is now known as ebarrett
[11:25] * sirdancealot (~koo5@236.152.broadband3.iol.cz) has joined #jython
[11:56] * siel (~siel@unaffiliated/motley) Quit (Remote host closed the connection)
[12:00] * siel (~siel@unaffiliated/motley) has joined #jython
[12:22] * sirdancealot (~koo5@236.152.broadband3.iol.cz) Quit (Ping timeout: 272 seconds)
[13:14] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[13:50] * trevorDAO (~Trevor@74-94-171-213-NewEngland.hfc.comcastbusiness.net) has joined #jython
[13:51] * xemdetia (xemdetia@nat/ibm/x-jxtghpzqcvkrqizt) has joined #jython
[15:44] <jimbaker> agronholm, re rc2 - i had done some earlier work on the installer, but for rc2 is when i really spent some time digging it
[15:45] <jimbaker> in it
[15:45] <agronholm> ok
[15:45] <agronholm> I'm prepping a new release of jython-swingutils
[15:45] <jimbaker> the installer has 3 different modes of operation - command line, console, and gui - each of which required modification to support jython.bat -> jython.exe as well as the new ensurepip support
[15:46] <agronholm> alright
[15:46] <jimbaker> annoying, but that's what it is
[15:46] <jimbaker> may be hard to DRY in this case
[15:47] <jimbaker> the more interesting issue is that i had assumed the following would work
[15:47] <jimbaker> #!/usr/bin/env python2.7 -E
[15:47] <jimbaker> for the new launcher. but that doesn't work on linux. (ouch.)
[15:47] <agronholm> erm, is that supposed to run jython?
[15:47] <jimbaker> so i changed that to #!/usr/bin/env python2.7
[15:48] <jimbaker> agronholm, yes, that's the top of bin/jython
[15:48] <jimbaker> https://hg.python.org/jython/file/0213400c518f/src/shell/jython.py
[15:49] <jimbaker> anyway, i'm concerned that losing -E may cause other issues
[15:49] <jimbaker> one alternative is for the launcher to write the necessary shebang line
[15:50] <jimbaker> so it can avoid /usr/bin/env
[15:50] <jimbaker> meanwhile this is all because the old bin/jython bash script has bugs, eg --profile for example does not work iirc
[15:51] <jimbaker> and consistency with jython.exe
[15:51] <agronholm> that makes jython dependent on python 2.7, doesn't it?
[15:51] <agronholm> cpython 2.7 that is
[15:52] <agronholm> I'm not sure it's installed by default on all non-windows OSes
[15:52] <agronholm> particularly arch linux which uses python 3 by default
[15:54] <jimbaker> the installer only does this if cpython 2.7 is available
[15:55] <jimbaker> otherwise it uses the bash launcher
[15:55] <agronholm> not sure how I feel about that
[15:55] <jimbaker> well, i suppose we could fix the buggy bash script...
[15:55] <jimbaker> but single source seems to be a good thing
[15:55] <agronholm> it would be preferable, but it's certainly not trivial
[15:56] <jimbaker> moving to jython.py will eventually allow us to offer nailgun style launches
[15:56] <jimbaker> with security
[15:57] <jimbaker> via domain sockets
[15:57] <jimbaker> so it certainly has long-term benefits. regardless, we now have native launcher support for windows and that's been the biggest benefit so far
[15:58] <agronholm> how does it work, in short?
[15:59] <jimbaker> jython.py remains a launcher; you can see this with bin/jython --print
[15:59] <agronholm> I mean the windows launcher
[16:00] <jimbaker> jython.exe is built by pyinstaller
[16:00] <jimbaker> http://pythonhosted.org/PyInstaller/
[16:00] <agronholm> is there any relation to jython.py?
[16:00] <jimbaker> it's the output of > pyinstaller jython.py
[16:01] <agronholm> oh.
[16:01] <jimbaker> which i built on my windows 8.1 vm
[16:02] <jimbaker> https://hg.python.org/jython/rev/2a4d36522fa6 shows the latest version
[16:02] <jimbaker> ;)
[16:02] <jimbaker> (i didn't bother rebuilding with the shebang line change, since that has no impact)
[16:11] <agronholm> btw, jython-swingutils had a twisted-like @inlineCallbacks thing which uses the AWT event loop instead of twisted's reactor. In 2.0, I've changed this to work with Futures instead of the custom AsyncToken class
[16:14] <agronholm> I'm now trying to decide whether to add a dependency to the futures backport or to provide my own version of the Futures api
[16:26] <jimbaker> maybe best to use the futures backport? after all with pip support, no need to worry about bundling
[16:27] <jimbaker> i'm hardly a packaging expert, but i would assume vendor lib type stuff should be reserved only for stuff like pip itself
[16:29] <agronholm> it's only a problem when you have to package futures separately in a jar
[16:29] <agronholm> for deployment on JEE servers, web containers or web start
[16:34] * mbooth (~mbooth@cpc11-shef10-2-0-cust659.barn.cable.virginm.net) has joined #jython
[16:34] * mbooth (~mbooth@cpc11-shef10-2-0-cust659.barn.cable.virginm.net) Quit (Changing host)
[16:34] * mbooth (~mbooth@redhat/mbooth) has joined #jython
[16:57] <jimbaker> agronholm, is this is a question of installation in site-packages vs some other directory?
[16:57] <agronholm> no
[16:57] <jimbaker> the same issue came up yesterday, i believe DanC_ raised it, re modjy
[16:58] <agronholm> if swingutils depends on futures, I have to package futures separately in a jar
[16:58] <agronholm> that's all
[16:58] <jimbaker> the futures code is in java?
[16:58] <agronholm> no
[16:58] <agronholm> but how would you package it?
[16:58] <agronholm> you can't exactly use wheels in a web start environment
[16:59] <jimbaker> not zipped wheels for the moment
[17:00] <jimbaker> what's the support of web start for class loaders and where code comes from?
[17:00] <agronholm> I'm not sure I understand the question
[17:01] <jimbaker> eg if i have a Lib/ directory full of $py.class files in my web start jar, i should be able to load them, right?
[17:01] <agronholm> how would it know to load them from the Lib directory?
[17:01] <jimbaker> it's only if i go through an additional extraction process, say because it's a zipped file, that it could not
[17:02] <jimbaker> agronholm, i'm assuming current standalone support
[17:02] <agronholm> did you modify the class loader to look in Lib/ too? I wasn't sure what exactly you did there
[17:02] <jimbaker> i may have used a webstart application, but i certainly have never looked at one as a dev
[17:02] <jimbaker> agronholm, no changes there
[17:03] <jimbaker> but we still have http://bugs.jython.org/issue2143
[17:03] <jimbaker> outstanding
[17:03] <agronholm> this is about the inability to load the standard library modules
[17:03] <agronholm> it doesn't work if they're not in the root of the jar
[17:04] <agronholm> that's why I had to create the modification script
[17:04] <jimbaker> ok, so it's definitely related
[17:04] <agronholm> web start works with a deployment descriptor. you give it a number of jars which it downloads and puts on your class path.
[17:04] <agronholm> let me show you mine
[17:04] <jimbaker> fwiw, clamp does it own rewriting
[17:05] <jimbaker> agronholm, i feel like this will be an interesting topic for the sprints
[17:05] <agronholm> https://bpaste.net/show/2dd796c45609
[17:05] <jimbaker> however, i'm going to have to back out now - i have an important meeting in a couple of hours, and i need to prep for it
[17:06] <agronholm> this is my production JNLP
[17:06] <agronholm> each python library is packaged as a separate jar
[17:06] <jimbaker> ack. looks interesting
[17:06] <jimbaker> so we could have site-packages still work here - each jar would be separate location for site-packages
[17:07] <jimbaker> let's definitely discuss next week when we are face-to-face
[17:07] <agronholm> I think it's a bad idea to try to stuff anything in site-packages in a JEE/web start deployment
[17:07] <agronholm> yeh
[17:07] <jimbaker> but how are we going to make it compatible with python practices?
[17:08] <jimbaker> that's the question
[17:08] <agronholm> what do you mean
[17:09] <jimbaker> i'm just saying that we could potentially support site.addpackage(JAR)
[17:09] <jimbaker> and it could just work
[17:09] <jimbaker> i have no idea
[17:09] <agronholm> eh?
[17:09] <agronholm> how would that work with web start?
[17:09] <agronholm> you have no idea where the jars are physically located, or if they're unpacked or not
[17:09] <jimbaker> i don't know. maybe we need a manifest as well on the python side?
[17:10] <jimbaker> (jython side)
[17:10] <agronholm> I can only ask: why?
[17:10] <jimbaker> but it's a very good brainstorming topic for the sprints
[17:10] <jimbaker> because... why this difference?
[17:10] <agronholm> what difference
[17:10] <jimbaker> i want to use pip, virtualenv to build out a package
[17:11] <jimbaker> and use it to work with java generally
[17:11] <jimbaker> it works (mostly) with clamp and singlejar
[17:11] <agronholm> you can put them all in one jar
[17:11] <jimbaker> maybe it could be use to split out stuff as well
[17:11] <agronholm> you just have to make sure the packages are directly on the jar root
[17:11] <jimbaker> any event, as i said - very cool topic
[17:11] <agronholm> yeh
[17:12] <jimbaker> but i have a seriously important meeting in < 2h
[17:12] <agronholm> ok
[17:52] * claudiupopa (~claudiupo@ Quit (Ping timeout: 244 seconds)
[18:07] * claudiupopa (~claudiupo@ has joined #jython
[18:10] * sirdancealot (~koo5@236.152.broadband3.iol.cz) has joined #jython
[18:10] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[18:13] * claudiupopa (~claudiupo@ Quit (Ping timeout: 244 seconds)
[18:55] * mbooth (~mbooth@redhat/mbooth) Quit (Ping timeout: 244 seconds)
[19:22] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[19:47] <agronholm> I've release jython-swingutils 2.0
[19:47] <agronholm> *released
[19:47] <agronholm> I'll drop an update once #1540 is fixed
[20:37] <jimbaker> agronholm, sounds good
[20:55] * Rondom (~rondom@nonmodosedetiam.net) Quit (Remote host closed the connection)
[20:56] * Rondom (~rondom@nonmodosedetiam.net) has joined #jython
[21:08] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[22:25] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[22:37] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[23:05] * claudiupopa (~claudiupo@ has joined #jython


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