Timestamps are in GMT/BST.
[0:05] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[0:18] * Juneau001 (~Juneau@50-45-101-37.dklb.il.frontiernet.net) has joined #jython
[0:58] * Snova (~snova@unaffiliated/snova) has joined #jython
[1:44] * darjus (~darjus@96-26-62-132.war.clearwire-wmx.net) has joined #jython
[2:46] * Snova (~snova@unaffiliated/snova) has left #jython
[2:46] * Juneau001 (~Juneau@50-45-101-37.dklb.il.frontiernet.net) Quit (Quit: Juneau001)
[2:57] * msavy (~msavy@redhat/jboss/msavy) Quit (Remote host closed the connection)
[6:34] * darjus (~darjus@96-26-62-132.war.clearwire-wmx.net) Quit (Quit: Konversation terminated!)
[8:33] * jimbaker (~jbaker@canonical/jimbaker) Quit (Ping timeout: 264 seconds)
[8:41] * jimbaker (~jbaker@c-71-237-106-195.hsd1.co.comcast.net) has joined #jython
[8:41] * jimbaker (~jbaker@c-71-237-106-195.hsd1.co.comcast.net) Quit (Changing host)
[8:41] * jimbaker (~jbaker@canonical/jimbaker) has joined #jython
[9:03] * seletz (~seletz@HSI-KBW-149-172-100-49.hsi13.kabel-badenwuerttemberg.de) has joined #jython
[10:17] * purplefox (~purplefox@host-78-150-0-71.as13285.net) has joined #jython
[10:18] <purplefox> Does anyone know where I can find the repository tag for jython 2.5.3? There doesn't seem to be a tag here http://jython.svn.sourceforge.net/viewvc/jython/tags/
[11:34] * msavy (~msavy@redhat/jboss/msavy) has joined #jython
[11:38] * verterok (~ggonzalez@91.189.93.48) Quit (Ping timeout: 252 seconds)
[12:10] <agronholm> purplefox: you're looking in the wrong place
[12:11] <agronholm> hg.python.org is the official repository site
[12:11] <purplefox> agronholm: I tried hg too, i can successful find the 2.5.3 tag there
[12:11] <purplefox> agronholm: however when i try to build a full build (as specified in build.xml) it actually seems to pull in the source from SVN, and there isn't a 2_5_3 SVN tag :(
[12:12] <purplefox> agronholm: so I can't figure how to build 2.5.3
[12:12] <agronholm> it only pulls cpython sources (to get the cpythonlib)
[12:12] <purplefox> agronholm: from build.xml (in 2.5.3 branch of hg repo) {
[12:12] <purplefox> Use case 2: full build for a release (using svn checkout)
[12:12] <purplefox> ---------------------------------------------------------
[12:12] <purplefox> - make sure you have access to the Jython Subversion repository
[12:12] <purplefox> (https://jython.svn.sourceforge.net/svnroot/jython/trunk)
[12:12] <purplefox> - override svn.main.dir in ant.properties (if necessary)
[12:13] <purplefox> - call target 'full-build'
[12:13] * Juneau001 (~Juneau@fess-120727.dhcp.fnal.gov) has joined #jython
[12:13] * Juneau001 (~Juneau@fess-120727.dhcp.fnal.gov) Quit (Client Quit)
[12:14] <agronholm> I will try it myself now from scratch
[12:15] <purplefox> agronholm: so i don't know what to set for svn.main.dir in ant.properties
[12:18] <agronholm> purplefox: well, I only did hg clone http://hg.python.org/jython/
[12:18] <agronholm> and then ran ant
[12:18] <agronholm> and jython was successfully built
[12:18] <agronholm> what problems did you run into when doing that?
[12:19] <purplefox> I don't think ant creates the full release
[12:19] <agronholm> by default it creates a developer build
[12:19] <purplefox> according to build.xml you need ant full-build
[12:20] <purplefox> the developer build is no good for me since it does not contain the deps
[12:20] <agronholm> ant jar-complete should do the trick
[12:20] <purplefox> i need to create my own version of a release build
[12:22] <purplefox> that seems to work
[12:23] <purplefox> how does that differ from what the release makes?
[12:23] <agronholm> "the release"?
[12:27] <purplefox> so.. basically my code works fine with the jython-standalone-2.5.3.jar from Maven
[12:27] <purplefox> but if I build jython.jar myself from the 2.5.3 tag in hg, it doesn't work. so they must be different somehow
[12:27] <agronholm> you never said you wanted a standalone jar
[12:27] <purplefox> I haven't debugged yet to figure out _why_ it doesn't work but something is not the same
[12:27] <agronholm> standalone jar contains the standard library
[12:28] <agronholm> do "ant jar-standalone" then
[12:29] <purplefox> it works!
[12:29] <purplefox> thanks
[12:30] <purplefox> question, why does jython use hg and svn? i'm curious
[12:30] <purplefox> why not just hg?
[12:30] <agronholm> how do you check out only a part of a repository using hg?
[12:30] <purplefox> or better still.. git
[12:30] <agronholm> jython uses hg because cpython does
[12:31] <purplefox> agronholm: i'm thinking of forking jython to github so I can fix some memory leaks that have been outstanding for a long time, and are causing us a lot of problems in my project, Vert.x
[12:31] <purplefox> agronholm: which is why I'm going through this process
[12:32] <purplefox> agronholm: so basically trying to understand the jython build a bit better
[12:33] <agronholm> jython needs a copy of the cpython library, which is checked out with svn because at least at the time of jython's hg conversion, hg could not do partial checkouts and we didn't want to check out the entire cpython codebase
[12:34] <purplefox> agronholm: when I do 'ant jar-standalone' where does it get the cpython library from?
[12:35] <agronholm> the library is included in the hg repository
[12:35] <agronholm> but iirc a full build refreshes that with svn
[12:35] <purplefox> agronholm: ok
[12:36] <purplefox> agronholm: so I could feasible import what's in hg into git, and should be able to do ant jar-standlone without needing to contact svn?
[12:36] <agronholm> right
[12:36] <agronholm> converting between hg and git will likely bring its own troubles though
[12:37] <purplefox> agronholm: yeah, i don't want to lose history for one thing
[12:37] <purplefox> agronholm: I'll give it a go anyway
[12:38] <purplefox> agronholm: thanks for your help
[12:38] <agronholm> np
[12:40] * synasius (~syn@host19-66-static.243-95-b.business.telecomitalia.it) has joined #jython
[13:35] * verterok (~ggonzalez@91.189.93.88) has joined #jython
[13:54] * verterok (~ggonzalez@91.189.93.88) Quit (Quit: Coyote finally caught me)
[13:54] * verterok (~ggonzalez@91.189.93.88) has joined #jython
[14:05] * verterok (~ggonzalez@91.189.93.88) Quit (Changing host)
[14:05] * verterok (~ggonzalez@unaffiliated/verterok) has joined #jython
[14:31] * Juneau001 (~Juneau@131.225.24.178) has joined #jython
[16:16] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[16:37] * lheuer (~Adium@unaffiliated/lheuer) has joined #jython
[16:48] <fwierzbicki> purplefox, as of 2.5.3 we went ahead and did a full copy of the CPython libs to avoid needing both hg and svn - I'll double check to make sure we aren't still pulling svn in full-build
[16:49] <purplefox> fwierzbicki: thanks
[16:50] <purplefox> fwierzbicki: i've got it working now by using the jar-standalone target
[16:50] <fwierzbicki> purplefox, ah I see build.xml is still littered with svn directions
[16:50] <fwierzbicki> purplefox, cool - that does correspond with what ends up released
[16:50] <purplefox> fwierzbicki: yes I think so
[16:51] <purplefox> fwierzbicki: I got help from agronholm
[16:51] <purplefox> fwierzbicki: btw, fyi: fwierzbicki: btw, fyi I've forked jython here: https://github.com/purplefox/jython so I can fix these issues http://bugs.jython.org/issue1746 and http://bugs.jython.org/issue1327
[16:51] <fwierzbicki> we do a little dance to lock the version in place and do a full download. That process is older than me entering the project. I guess its to keep us from putting local changes into a real build
[16:51] <fwierzbicki> purplefox, that's great! I'd love to get those cleared up before 2.5.4 and 2.7
[16:52] <fwierzbicki> jimbaker, is too busy to finish those these days
[16:52] <purplefox> fwierzbicki: yeah, they were killing us since we start up and shutdown 1000s of PythonInterpreter instances
[16:52] <purplefox> fwierzbicki: in a JVM instance
[16:53] <fwierzbicki> ah. I'll make sure I'm nosy on those so when you can comment or send patches I'll see them
[16:53] <purplefox> fwierzbicki: my fix to http://bugs.jython.org/issue1746 is a bit hacky since I'm not sure what the correct approach is
[16:53] <fwierzbicki> oh right they have backwards compatibility issues - may only be good for 2.7. I'll have to think about it...
[16:54] <purplefox> fwierzbicki: I've commented out appending to the searchPath - it doesn't seem to break anything (for me anyway)
[16:54] <fwierzbicki> purplefox, ok - I'll try to examine them in the near future. I'm very busy with 2.7 stuff ATM
[16:54] <purplefox> fwierzbicki: fair enough
[16:54] <fwierzbicki> tricky stuff :)
[16:54] <purplefox> fwierzbicki: how is jython dev going? you are working on new stuff?
[16:55] <fwierzbicki> I am closing in on a 2.7 beta. We probably have enough for a 2.5.4 release as well
[16:55] <fwierzbicki> IIRC jimbaker did not want to fix the threading API stuff in 2.5 since it would be backwards incompatible
[16:56] <purplefox> fwierzbicki: any plans for invokeDynamic support?
[16:56] <fwierzbicki> post 2.7 definitely - probably with backporting to 2.7.x
[16:57] <purplefox> fwierzbicki: ace
[16:57] <fwierzbicki> When 2.7.0 comes out, I think I'll warn that there will be backwards compatibility changes in 2.7.1, 2.7.2, .... just because we will likely learn a bunch about invokedynamic and other things while working on 3.x
[16:57] <fwierzbicki> and calling it 2.8 would be too confusing
[16:58] <purplefox> fwierzbicki: well, I'm glad to hear jython dev is alive and well :)
[16:58] <fwierzbicki> we're doing much better lately. We have some very productive new blood as well.
[16:59] <fwierzbicki> things are definitely looking up
[16:59] <purplefox> fwierzbicki: are any companies currently paying devs to work on it, or is it all "in spare time" stuff?
[16:59] <fwierzbicki> I get an official 10% and can fill in on the clock when between projects. Other than that I don't think so
[17:00] <fwierzbicki> In the 3.x cycle we'll be getting help on the CPython side. The plan is to combine the standard lib
[17:00] <fwierzbicki> so it won't be a ton of work to update on every release
[17:01] <fwierzbicki> I believe the pypy folks are involved in a combined standard lib as well
[17:01] <purplefox> sounds good
[17:02] <fwierzbicki> I shudder to think of the amount of my life that has been spent merging newer .py files from the standard lib :)
[17:03] <purplefox> fwierzbicki: i bet
[17:05] <fwierzbicki> ok I'm nosy on both issues now. I know some major work was done on encoding recently, though I don't think it was memory leak related
[17:11] * synasius (~syn@host19-66-static.243-95-b.business.telecomitalia.it) Quit (Quit: Sto andando via)
[17:33] * msavy (~msavy@redhat/jboss/msavy) Quit (Remote host closed the connection)
[18:10] * synasius (~syn@79.143.113.124) has joined #jython
[18:14] <jimbaker> fwierzbicki, you mean ThreadState?
[18:14] <fwierzbicki> yeah I think that is the older bug
[18:14] <fwierzbicki> http://bugs.jython.org/issue1327
[18:15] <jimbaker> yes, that's the one
[18:15] <fwierzbicki> jimbaker, ^^^
[18:15] <jimbaker> it should be fixed
[18:15] <fwierzbicki> in 2.7?
[18:15] <fwierzbicki> or also in 2.5.x?
[18:15] <jimbaker> not in 2.5
[18:15] <fwierzbicki> ok that's what I though - so it is checked into 2.7
[18:15] <fwierzbicki> *thought
[18:16] <jimbaker> fwierzbicki, i have not done such work, but it doesn't imply someone else hasn't
[18:16] <jimbaker> although i suspect not
[18:16] <jimbaker> the solution, it is reasonably subtle
[18:16] <fwierzbicki> But the patch should be good if it is just inspected and applied
[18:17] <jimbaker> fwierzbicki, actually, that should be fine in this particular case
[18:17] <fwierzbicki> BTW I think I am closing in on the trouble in that compiler patch. I hope to have it in either this or next week
[18:17] <jimbaker> i guess i was thinking of where i wanted to take it
[18:17] <jimbaker> yes, let's apply that patch
[18:17] <jimbaker> let not the perfect stand in the way of the good
[18:17] <fwierzbicki> ok I'll make sure it gets in before 2.7 beta
[18:17] <jimbaker> thanks!
[18:19] <fwierzbicki> np!
[18:54] * seletz_ (~seletz@HSI-KBW-149-172-100-49.hsi13.kabel-badenwuerttemberg.de) has joined #jython
[18:56] * seletz_ (~seletz@HSI-KBW-149-172-100-49.hsi13.kabel-badenwuerttemberg.de) Quit (Read error: Operation timed out)
[18:57] * seletz (~seletz@HSI-KBW-149-172-100-49.hsi13.kabel-badenwuerttemberg.de) Quit (Ping timeout: 245 seconds)
[19:01] * jimbaker (~jbaker@canonical/jimbaker) has left #jython
[19:21] <agronholm> fwierzbicki: I'm close to merging bz2 support to 2.7, I'm just waiting for the original patch author to comment on my modifications to it
[19:21] <agronholm> some bz2-tarfile tests still fail but those can be fixed later
[19:21] <agronholm> all of the bz2 tests pass though
[19:22] <fwierzbicki> agronholm, great! That will really help w/ a bunch of project that should run on Jython but don't
[19:23] <fwierzbicki> I know I've had some "pip install" failures due to the lack of bz2
[19:23] <agronholm> yep
[19:23] <fwierzbicki> that I can work around but people shouldn't have to
[19:46] * synasius (~syn@79.143.113.124) Quit (Read error: No route to host)
[19:46] * synasius (~syn@79.143.113.76) has joined #jython
[20:03] * Juneau001 (~Juneau@131.225.24.178) Quit (Quit: Juneau001)
[20:37] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[21:01] * meritum (~jakub@79.97.179.17) has joined #jython
[21:16] <purplefox> fwierzbicki: quick question: Is an instance of PythonInterpreter save to access concurrently with multiple threads?
[21:16] <purplefox> fwierzbicki: /save/safe
[21:16] <purplefox> fwierzbicki: i.e. execute exec(..) on it concurrently from different threads
[21:21] * jimbaker (~jbaker@canonical/jimbaker) has joined #jython
[21:41] <fwierzbicki> purplefox, it isn't inherently unsafe -- but like any other multithreaded coding you'd need to be careful to lock any mutations that might conflict.
[21:42] <fwierzbicki> there are lots of caveats though... like everything in multithreaded coding - for example changing the internal PySystemState could goof things up
[21:43] <fwierzbicki> jimbaker (sorry to bug you again, but you've thought about this a bunch) - I think you where away but purplefox asked:
[21:43] <fwierzbicki> quick question: Is an instance of PythonInterpreter save to access concurrently with multiple threads?
[21:44] <fwierzbicki> and he meant "safe" instead of "save"
[21:44] <jimbaker> fwierzbicki, in general, yes
[21:45] <jimbaker> i guess i just echoed what you said above
[21:45] <jimbaker> it'
[21:45] <fwierzbicki> cool - I gave the usual disclaimers too :)
[21:46] <jimbaker> practical experience says that it's just fine. which makes sense, because threaded code in python tends to do this same work, if indirectly
[21:46] <jimbaker> but i'm sure we can come up with any number of examples of what not to do
[21:47] <fwierzbicki> purplefox, I'd still suggest plenty of testing of course
[21:48] <fwierzbicki> especially around accessing mutable state from multiple threads
[21:49] <jimbaker> purplefox, probably the biggest thing i can think of is how a python interpreter (via PySystemState -> ThreadState) can reference any number of classloaders, and corresponding classes
[21:50] <jimbaker> related to the earlier bug
[21:50] <jimbaker> but if you allow that to gc, then the classloaders go away
[21:50] <jimbaker> so those are the sort of real bugs we see
[21:54] * verterok` (~ggonzalez@91.189.93.48) has joined #jython
[21:54] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[22:05] * meritum (~jakub@79.97.179.17) Quit (Remote host closed the connection)
[22:25] * Taggnostr (~quassel@dyn57-146.yok.fi) has left #jython
[23:25] * xymox (lechuck@unaffiliated/contempt) Quit (Ping timeout: 260 seconds)
[23:28] * xymox (~xymox@unaffiliated/contempt) has joined #jython
[23:37] * synasius (~syn@79.143.113.76) Quit (Quit: Sto andando via)
These logs were automatically created by JythonLogBot_ on irc.freenode.net using a slightly modified version of the Java IRC LogBot (github).