#jython IRC Log

Index

IRC Log for 2014-10-01

Timestamps are in GMT/BST.

[0:24] * lheuer (~Adium@unaffiliated/lheuer) Quit (Ping timeout: 250 seconds)
[1:42] * ebarrett (~edd@host-92-20-142-254.as13285.net) Quit (Ping timeout: 244 seconds)
[1:44] * ebarrett (~edd@host-92-20-142-254.as13285.net) has joined #jython
[2:10] * robbyoconnor (~wakawaka@2600:1001:b119:4bfe:212d:61d8:aa2e:d914) has joined #jython
[2:10] * robbyoconnor is now known as Guest6749
[2:22] * Guest6749 (~wakawaka@2600:1001:b119:4bfe:212d:61d8:aa2e:d914) Quit (Max SendQ exceeded)
[2:23] * Guest6749 (~wakawaka@2600:1001:b119:4bfe:212d:61d8:aa2e:d914) has joined #jython
[2:34] * Guest6749 (~wakawaka@2600:1001:b119:4bfe:212d:61d8:aa2e:d914) Quit (Max SendQ exceeded)
[2:35] * Guest6749 (~wakawaka@2600:1001:b119:4bfe:212d:61d8:aa2e:d914) has joined #jython
[2:46] * Guest6749 (~wakawaka@2600:1001:b119:4bfe:212d:61d8:aa2e:d914) Quit (Max SendQ exceeded)
[2:47] * Guest6749 (~wakawaka@2600:1001:b119:4bfe:212d:61d8:aa2e:d914) has joined #jython
[2:54] * Guest6749 (~wakawaka@2600:1001:b119:4bfe:212d:61d8:aa2e:d914) Quit (Quit: Konversation terminated!)
[4:16] * robbyoconnor (~wakawaka@cpe-68-174-63-155.hvc.res.rr.com) has joined #jython
[4:38] * mgolisch_ is now known as mgolisch
[6:12] * a42niem (~niem@p578b7daf.dip0.t-ipconnect.de) has joined #jython
[7:08] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[8:00] * mbooth (~mbooth@redhat/mbooth) has joined #jython
[8:06] * ebarrett (~edd@host-92-20-142-254.as13285.net) Quit (Ping timeout: 272 seconds)
[8:06] * ebarrett (~edd@host-92-20-142-254.as13285.net) has joined #jython
[8:22] * lheuer (~Adium@unaffiliated/lheuer) Quit (Ping timeout: 250 seconds)
[8:37] * lheuer (~Adium@dyndsl-091-096-010-106.ewe-ip-backbone.de) has joined #jython
[8:37] * lheuer (~Adium@dyndsl-091-096-010-106.ewe-ip-backbone.de) Quit (Changing host)
[8:37] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[9:16] * fwierzbicki1 (~Adium@99-106-169-5.lightspeed.sntcca.sbcglobal.net) has joined #jython
[9:18] * fwierzbicki (~Adium@99-106-169-5.lightspeed.sntcca.sbcglobal.net) Quit (Ping timeout: 258 seconds)
[9:21] * lheuer (~Adium@unaffiliated/lheuer) Quit (Quit: Leaving.)
[9:33] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[10:36] <paolo> agronholm: thanks for your suggestion, py4j is amazing for my use case.
[10:36] <agronholm> glad it worked out for you
[10:45] <agronholm> I think I'm starting to understand a few things
[10:45] <agronholm> why my web start thing isn't working
[10:45] <agronholm> and why the standalone jar won't work with JWS
[10:45] <agronholm> it seems that JWS unpacks the jar
[10:46] <agronholm> so the standalone flag isn't set in PySystemState and so the Lib dir isn't added to the path
[10:46] <agronholm> but the worst problem is that the home directory is not set since there's no jar file
[11:56] * robbyoconnor (~wakawaka@cpe-68-174-63-155.hvc.res.rr.com) Quit (Ping timeout: 260 seconds)
[11:57] * lheuer1 (~Adium@host-091-097-064-066.ewe-ip-backbone.de) has joined #jython
[11:59] * lheuer (~Adium@unaffiliated/lheuer) Quit (Ping timeout: 250 seconds)
[12:58] * robbyoconnor (~wakawaka@2600:1017:b013:966c:347e:9652:4e7f:582b) has joined #jython
[12:58] * robbyoconnor is now known as Guest82239
[13:14] * Guest82239 (~wakawaka@2600:1017:b013:966c:347e:9652:4e7f:582b) Quit (Quit: Konversation terminated!)
[13:14] * Guest82239 (~wakawaka@2600:1017:b013:966c:347e:9652:4e7f:582b) has joined #jython
[13:14] * Guest82239 (~wakawaka@2600:1017:b013:966c:347e:9652:4e7f:582b) Quit (Client Quit)
[13:19] * xemdetia (xemdetia@nat/ibm/x-ivzdmeujgirpwhgf) has joined #jython
[15:43] * gthank (~gthank@unaffiliated/gthank) has joined #jython
[16:14] * Arfrever (~Arfrever@apache/committer/Arfrever) Quit (Ping timeout: 272 seconds)
[16:54] <jimbaker> @agronholm, ack - in general, we need to make this Lib finding more robust
[16:56] <jimbaker> paolo, py4j can be a great integration since it's a much simpler model. so just depends.
[16:57] <jimbaker> paolo, you might find this useful - demonstrates jython working with spring - https://github.com/rackerlabs/keystone-jvm
[16:57] <jimbaker> to run a non-trivial app
[16:57] <paolo> thanks jimbaker, i'll surely have a look
[16:58] <jimbaker> in our case, it was trivial to back keystone models with cassandra or mongodb or mysql, thanks to springdata
[16:58] <jimbaker> and this is just an example project that's distilled from some work we have done at rackspace
[17:06] <paolo> cool
[17:12] <agronholm> jimbaker: correction, the jar file does exist but it's named randomly and the resource uri is an http uri
[17:12] <agronholm> java doesn't let it figure out the local path, even with all permissions enabled
[17:12] <jimbaker> agronholm, what's the specific url?
[17:13] <agronholm> whatever the jars were downloaded from
[17:13] <agronholm> *wherever
[17:14] <agronholm> I'm not sure how this should be fixed
[17:15] <jimbaker> @agronholm, so in http://bugs.jython.org/issue2143, i was thinking that we should be able to support jar url references
[17:15] <agronholm> how would that help?
[17:16] <agronholm> the main problem is that sys.prefix does not get set
[17:16] <agronholm> supporting jar url references does not help if it's an HTTP url, and we already support local urls
[17:16] * Arfrever (~Arfrever@apache/committer/Arfrever) has joined #jython
[17:17] <agronholm> we have a code path that lets the prefix remain unset
[17:17] <agronholm> which is no good because without the prefix, site.py is not importable
[17:17] <agronholm> so the interpreter won't work anyway
[17:17] <agronholm> better to fail early
[17:18] <jimbaker> @agronholm, ok. it's possible that what i'm thinking about - and did some early implementation work on https://bitbucket.org/jimbaker/jython-singlejar-site-packages/ - is not sufficient for your case
[17:18] <agronholm> I package third party libs in their own jars
[17:18] <agronholm> after compilation
[17:19] <jimbaker> sure. this is for python code support
[17:20] <agronholm> meaning?
[17:20] <agronholm> I mean, I package python libraries as separate jars
[17:21] <jimbaker> agronholm, interesting
[17:21] <jimbaker> agronholm, so meaning that we support site-packages. the case i have been thinking about is for singlejar. but it would be very cool to support site-packages that are actually other jars, as you are doing
[17:22] <agronholm> what would that support involve then?
[17:22] <jimbaker> i'm pretty sure a generic jar url approach would make this work
[17:22] <agronholm> I mean, do we get new functionality?
[17:22] <jimbaker> support jar urls for file names
[17:22] <jimbaker> for readonly access
[17:23] <agronholm> I'm not sure I get it
[17:23] <agronholm> is this for debugging, or something else?
[17:24] <jimbaker> no, it's for packaging python code in jars. but they are of course files. which means code occasionally needs to introspect the file structure as well
[17:24] <agronholm> ok so you mean __file__ ?
[17:30] <jimbaker> @agronholm, really directory structure. so there's code out there that will look at the directory structure of site-packages to see what has been installed
[17:31] <agronholm> oh
[17:32] <jimbaker> so if packaged in a jar, need to support some standard os.path ops
[17:32] <jimbaker> or it will fail
[17:33] <jimbaker> given that jar urls can always tell this is in a jar, enabling them can be a simple way of providing this support
[17:33] <jimbaker> without breaking other things
[17:35] <agronholm> jimbaker: how should I fix the JWS issue of prefix == null? Set just any path, like / or C:\ ?
[17:36] <agronholm> does it matter at all what the prefix is?
[17:36] <jimbaker> agronholm, well, it should be a usable prefix... is the actual prefix being null causing jython to blow up?
[17:36] <agronholm> indeed
[17:37] <agronholm> sysconfig.py tries to normalize the path which blows up since None does not have .startswith()
[17:38] <jimbaker> got it. so that's pretty nasty
[17:38] <agronholm> everything is loaded from the class path, so I'm not sure what the prefix is needed for
[17:38] <jimbaker> at the very least we should let it be null/None, test that scenario in our unit tests, and then have at least a reasonable error message
[17:39] <jimbaker> well the prefix is used to find Lib... but obviously it's not robust in your case
[17:39] <agronholm> how would it find Lib in the jython jar file anyway
[17:40] <agronholm> right, it would have to set the standalone flag on
[17:41] <agronholm> but we have a check that will supposedly make that work
[17:41] <jimbaker> right, this is done in PySystemState through a complicated procedure/hack
[17:41] <agronholm> but it doesn't work with JWS
[17:41] <agronholm> the first import (site.py) fails
[17:41] <agronholm> I'm not sure it ever has worked
[17:41] <jimbaker> right, because that's now the default behavior for jython as of some commit
[17:42] <jimbaker> https://hg.python.org/jython/rev/51b28cc2c43d
[17:42] <agronholm> loading site.py is fine
[17:42] <agronholm> but the import failing is obviously not
[17:42] <jimbaker> so in the beta 3 cycle we changed that. you can disable with the corresponding option, importSite
[17:44] <agronholm> disabling it doesn't help me if it doesn't know how to import from Lib/ inside the jar
[17:45] <agronholm> the imports only worked if I moved Lib/ to the root of the jar
[17:48] <agronholm> jimbaker: I'm guessing that the security manager is throwing a wrench in the import gears
[17:50] * mbooth (~mbooth@redhat/mbooth) Quit (Ping timeout: 245 seconds)
[17:51] <jimbaker> agronholm, is it really a security manager issue?
[17:52] <agronholm> not sure
[17:52] <agronholm> but I know import site does not work
[17:52] <agronholm> from JWS
[17:53] <agronholm> the only solution so far has been to move the Lib directory to the root
[17:56] <jimbaker> agronholm, that certainly makes sense, given the limited find support that we implement. in any event i plan to spend some more time on http://bugs.jython.org/issue2143, even if it's a bit on the frustrating side to work on ;)
[17:56] <jimbaker> due to unnecessary complexity in many places
[18:00] * lheuer1 is now known as lheuer
[18:28] * mbooth (~mbooth@80.2.226.243) has joined #jython
[18:28] * mbooth (~mbooth@80.2.226.243) Quit (Changing host)
[18:28] * mbooth (~mbooth@redhat/mbooth) has joined #jython
[18:36] * Arfrever (~Arfrever@apache/committer/Arfrever) Quit (Ping timeout: 272 seconds)
[18:44] * Trundle (~andy@python/site-packages/trundle) has joined #jython
[18:52] * Arfrever (~Arfrever@apache/committer/Arfrever) has joined #jython
[19:48] * mbooth (~mbooth@redhat/mbooth) Quit (Ping timeout: 246 seconds)
[20:18] * mbooth (~mbooth@redhat/mbooth) has joined #jython
[20:28] * lheuer (~Adium@host-091-097-064-066.ewe-ip-backbone.de) Quit (Quit: Leaving.)
[20:40] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[20:42] * gthank (~gthank@unaffiliated/gthank) Quit (Ping timeout: 258 seconds)
[20:53] * gthank (~gthank@unaffiliated/gthank) has joined #jython
[21:12] * Trundle (~andy@python/site-packages/trundle) Quit (Remote host closed the connection)
[21:20] * mbooth (~mbooth@redhat/mbooth) Quit (Ping timeout: 244 seconds)
[21:34] * mbooth (~mbooth@redhat/mbooth) has joined #jython
[22:14] <agronholm> jimbaker: the ant issue is still a big mystery
[22:14] <agronholm> may I persuade you to try and invoke it? I'll make you a sample build.xml for it.
[22:25] * gthank (~gthank@unaffiliated/gthank) Quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz???)
[22:32] <agronholm> https://bpaste.net/show/164b0db7b24a
[22:35] * xemdetia (xemdetia@nat/ibm/x-ivzdmeujgirpwhgf) Quit (Ping timeout: 245 seconds)
[22:40] * a42niem (~niem@p578b7daf.dip0.t-ipconnect.de) Quit (Quit: Leaving.)
[23:54] * lheuer (~Adium@unaffiliated/lheuer) Quit (Ping timeout: 250 seconds)

Index

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