#jython IRC Log


IRC Log for 2014-03-10

Timestamps are in GMT/BST.

[0:06] * lheuer (~Adium@unaffiliated/lheuer) Quit (Ping timeout: 252 seconds)
[1:14] * plankton (~plankton@ has joined #jython
[2:28] * plankton (~plankton@ Quit (Quit: Saindo)
[4:59] * siel (~siel@unaffiliated/motley) Quit (Ping timeout: 240 seconds)
[5:06] * thereisnospoon_ (~thereisno@113-61-86-28.static.qld.dsl.net.au) Quit (Ping timeout: 240 seconds)
[5:23] * thereisnospoon_ (~thereisno@113-61-86-28.static.qld.dsl.net.au) has joined #jython
[5:36] * agronholm (~agronholm@2001:1bc8:102:6f29:9c4:a015:ab5f:6ab) has joined #jython
[5:36] * ChanServ sets mode +o agronholm
[6:10] * ruserge (~ruserge@ Quit (Ping timeout: 244 seconds)
[6:33] * corsair (~corsair@ has joined #jython
[6:34] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[6:41] * Arfrever (~Arfrever@apache/committer/Arfrever) Quit (Ping timeout: 264 seconds)
[7:06] * Arfrever (~Arfrever@apache/committer/Arfrever) has joined #jython
[7:06] * corsair (~corsair@ Quit (Remote host closed the connection)
[11:18] * lheuer (~Adium@unaffiliated/lheuer) Quit (Ping timeout: 252 seconds)
[12:50] * whg (whg@unaffiliated/whg) has joined #jython
[13:47] * siel (~siel@unaffiliated/motley) has joined #jython
[15:31] * zimmermann (~user@178-24-50-233-dynip.superkabel.de) Quit (Read error: Connection reset by peer)
[15:32] * zimmermann (~user@178-24-50-233-dynip.superkabel.de) has joined #jython
[16:01] * zimmermann (~user@178-24-50-233-dynip.superkabel.de) Quit (Ping timeout: 265 seconds)
[16:37] <jimbaker> topi`, very interesting re http://bugs.jython.org/issue1973
[16:48] <jimbaker> i assume this is just the sort of thing that somehow missed any unit tests in python, but it looks very straightforward regardless
[17:12] <jimbaker> topi`, lastly, i just took a look at https://twistedmatrix.com/trac/ticket/3413#comment:40 - at the very least i can reproduce what you're seeing with respect to no reporters for trial. i suppose it might be possible to run twisted.python tests with another test runner however
[17:12] <jimbaker> just so this gordian knot can be cut
[17:39] * oscar_toro (~Thunderbi@h-17-170.a328.priv.bahnhof.se) has joined #jython
[18:24] <topi`> jimbaker: yup, I can't see why it would be so awfully important to support pluggable reporters for the sake of running unit tests
[18:25] <topi`> but I need to start wrapping up the day and head home
[18:25] <topi`> my next project is going to be to write a "twistd replacement" that would actually be a servlet that can be deployed as a .war file :)
[18:26] <topi`> then you can just deploy your twisted project bundled onto a .war, with jython.jar and everything ready, in-place
[18:26] <topi`> the idea would be to replicate the functionality of a "twistd -y foo.tac" invocation
[18:27] <topi`> twistd is, btw, a really sysop-friendly piece of sw, especially when running systemd (it can bind privileged ports without suid bits, because of systemd integration)
[18:28] <topi`> jimbaker: have you tried if IPython would work with the patch in #1973 in place?
[18:29] <topi`> I'm not a big IPython consumer, but I know many folks in the sciences *depend* on IPython
[18:29] <topi`> since there are a fair number of scientific frameworks written in Java, that would just make sense
[18:32] * smaudet (~smaudet@206-51-157-254.nktelco.net) has joined #jython
[18:36] <smaudet> jimbaker: hey, I think I may have found my issue from before - and yes the Java/Python bridge is throwing a silent error.
[18:36] <smaudet> Why its silent I'm not sure.
[18:36] <smaudet> but
[18:38] <smaudet> Perhaps we could go over areas in the code to look at?
[18:46] <jimbaker> hi, just was in a meeting
[18:48] <jimbaker> topi`, not certain how #1973 helps for ipython... i know it's being explored, need to follow up on that
[18:48] <jimbaker> smaudet, certainly let's discuss
[18:49] <jimbaker> topi`, one more thing - using war as a way of getting twistd like functionality sounds great
[18:53] <topi`> jimbaker: if I remember correctly, IPython did a number of relative star imports...
[18:53] <topi`> jimbaker: it sounds like a "management compatible" way of getting twisted services running ;)
[18:54] <topi`> anyhow, when I progress more with "twistd servlet", I'll let you know and test it out :)
[18:59] <smaudet> jimbaker: Ok, firstly I'm pretty sure I've been seing some Heisenbug behavior, I've intermittently managed to get TypeErrors to show up when using a debugger and not at any other point - and this is using the Proxies. So, just for starters sake, I'm just trying the Java -> Python bridge, not the whole thing. I can confirm, using the debugger, that only __builtins__ and exceptions are in the
[18:59] <smaudet> initial java path for python - is there any way perhaps to load the other modules in? And I fixed the code so that I have the entire of Lib inside my jar when building via clamp
[19:02] <jimbaker> smaudet, so i'm not certain what's happening here
[19:02] <jimbaker> in general - the key to get anything loaded in advance will be site.py
[19:03] <jimbaker> which you can certainly modify to try things out. ideally of course, it will go through site-packages
[19:03] <jimbaker> note: it's possible to turn off site.py loading
[19:03] <jimbaker> just in case you inadvertently did that
[19:05] <smaudet> jimbaker: line 732 of imp.jar in org.python.core I'm currenlty paused on an invocation of the sample library 'clamped', the 'modules' variable on line 731 has a table value with a length of '2', and the names of both keys within this table are __builtins__ and exceptions
[19:05] <smaudet> s/imp.jar/imp.java/
[19:06] <smaudet> how would I turn off site.py loading?
[19:06] <smaudet> (so that I can ensure that I'm not turning it off inadvertently, of course)
[19:07] <smaudet> And this is jython-ssl source code, not vanilla jython source code
[19:10] <smaudet> if it helps any, sys.meta_path is also empty
[19:10] <jimbaker> sure, we are using jython-ssl right now :)
[19:10] <jimbaker> *we are all*
[19:11] <jimbaker> although i finally started working on the jython-socket-reboot branch yesterday
[19:11] <jimbaker> going through all the python unit tests
[19:11] <smaudet> nice
[19:13] <smaudet> Also, sys.path is filled with three paths, one is the clamped/Lib directory, which is non-existent, the other two are __classpath__ and __pyclasspath__
[19:13] <smaudet> I suppose I'm mainly looking for direction at this point as to where these paths are all set, and when.
[19:14] <jimbaker> smaudet, so i assume you are doing something like $ CLASSPATH=`pwd`/clamped-0.1-single.jar java org.python.util.jython
[19:14] <smaudet> jimbaker: yes
[19:14] <smaudet> at the moment: 'java -Dpython.executable="C:\Users\se027348.US\bin\jython-ssl.bat" -Xdebug -Xrunjdwp:transport=dt_socket,server=y,address=8000,suspend=y -cp ..\clamped-0.1-single.jar;. UseClamped'
[19:15] * oscar_toro (~Thunderbi@h-17-170.a328.priv.bahnhof.se) Quit (Read error: Operation timed out)
[19:15] <jimbaker> next, when you run jar tf clamped-0.1-single.jar, do you see anything in Lib ? my assumption is no, based on what you just said
[19:16] <smaudet> jimbaker: so if you were to run jar xf clamped.jar , there would be a Lib directory extracted, with everything you'd expect to see in a normal jython/python install in there
[19:16] <jimbaker> smaudet, ok, that's much better then i suppose
[19:17] <jimbaker> this means setup.py singlejar in fact worked
[19:17] <smaudet> But the path that I'm seeing in the java variable 'sys.path' is referecing the directory where the jar was built
[19:17] <jimbaker> ahh, interesting
[19:17] <smaudet> jimbaker: after I fixed another broken fuction I found in clamp (I will submit a patch eventually) ;)
[19:17] <jimbaker> smaudet, patches gladly accepted :)
[19:18] <jimbaker> especially against github, which is so much easier to manage vs the near broken system we have to endure w/ hg.python.org
[19:20] <jimbaker> smaudet, so one thing to know about singlejar - sys.path actually is not meaningful. it should be, but the current classloader mechanism is sort of broken imho
[19:21] <jimbaker> that's why singlejar repacks everything into Lib and collapses all classes into a single jar
[19:21] <jimbaker> so because it's simpler, if not ideal, i would think it should work. especially since this is the mechanism that jython has used for a long long time for this case
[19:23] <smaudet> Let me try to make it happy for a bit.
[19:23] <jimbaker> the logic managing this can be seen in PySystemState around the test for isStandalone
[19:23] <smaudet> I'm going to stick the Lib directory where it thinks it should be
[19:24] <jimbaker> smaudet, always a good idea
[19:27] <smaudet> progress, it errored out, but on a different package (I gave it the wrong Lib folder) - but this behavior is still incorrect or at least unspecified.
[19:29] <smaudet> Hah ok now I have some fun error about the debugger being unable to 'escape' lol, but it finally ran the Java bridge
[19:29] <smaudet> well, ran the Java/Python/Java
[19:30] <smaudet> So where would it load that bad 'Lib' path from?
[19:30] <smaudet> And as you say if it usually just sticks it into the Jar, maybe there's a third broken method somewhere (or a broken part of my patches?)
[19:30] <smaudet> That puts the right path into the Jar, that is.
[19:32] <smaudet> jimbaker: the Python -> Java -> Python -> Java bridge is still not working though
[19:32] <jimbaker> smaudet, ok. i'm still trying to refresh my memory of which code is used in the case of standalone
[19:33] <jimbaker> i must have made a note of it...
[19:35] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[19:36] <jimbaker> smaudet, brb - i need to get some coffee
[19:38] <smaudet> np
[19:42] * clajo04 (~clajo04@pool-108-21-222-14.nycmny.fios.verizon.net) Quit (Quit: clajo04)
[19:45] * clajo04 (~clajo04@pool-108-21-222-14.nycmny.fios.verizon.net) has joined #jython
[20:09] <jimbaker> ok, coffee + short bike ride to cafe has refreshed. but still looking through code to find the specific place that standalone is really supported - PySystemState simply sets this up for subsequent usage
[20:10] <jimbaker> smaudet, ^^^
[20:10] <smaudet> ok
[20:23] <jimbaker> smaudet, it's possible that getJarFileNameFromURL doesn't quite work
[20:25] <jimbaker> given that sys.path didn't look right
[20:33] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[20:38] <jimbaker> smaudet, i double checked on getting a msdn license as a python core dev, so that should be ready in a few days. i'm going to have to look at this again later (hopefully today), but please do submit a pull request again clamp or a patch against jython as it comes up
[20:46] * lheuer1 (~Adium@ has joined #jython
[20:48] * lheuer (~Adium@unaffiliated/lheuer) Quit (Ping timeout: 252 seconds)
[21:06] <smaudet> jimbaker: ok, will bbl, I've got to go.
[21:15] * smaudet (~smaudet@206-51-157-254.nktelco.net) Quit (Ping timeout: 246 seconds)
[21:17] * lheuer1 is now known as lheuer
[21:41] * jimbaker (~jbaker@unaffiliated/jimbaker) Quit (Ping timeout: 252 seconds)
[21:44] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[21:46] * oscar_toro (~Thunderbi@h-17-170.a328.priv.bahnhof.se) has joined #jython
[21:55] * oscar_toro (~Thunderbi@h-17-170.a328.priv.bahnhof.se) Quit (Quit: oscar_toro)
[22:12] * jimbaker (~jbaker@c-75-71-81-32.hsd1.co.comcast.net) has joined #jython
[22:12] * jimbaker (~jbaker@c-75-71-81-32.hsd1.co.comcast.net) Quit (Changing host)
[22:12] * jimbaker (~jbaker@unaffiliated/jimbaker) has joined #jython
[23:43] * diametric (~diametric@2604:3400:dc1:43:216:3eff:fe27:bf9d) Quit (Quit: out.)
[23:45] * diametric (~diametric@2604:3400:dc1:43:216:3eff:fe27:bf9d) has joined #jython
[23:47] * thereisnospoon_ (~thereisno@113-61-86-28.static.qld.dsl.net.au) Quit (Ping timeout: 240 seconds)
[23:48] * diametric (~diametric@2604:3400:dc1:43:216:3eff:fe27:bf9d) Quit (Excess Flood)
[23:48] * ArcTanSusan (~susantan@50-203-159-62-static.hfc.comcastbusiness.net) has joined #jython
[23:54] * diametric (~diametric@2604:3400:dc1:43:216:3eff:fe27:bf9d) has joined #jython
[23:54] * lheuer (~Adium@ Quit (Quit: Leaving.)


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