#jython IRC Log


IRC Log for 2014-06-25

Timestamps are in GMT/BST.

[1:13] * siel (~siel@unaffiliated/motley) has joined #jython
[3:22] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Ping timeout: 240 seconds)
[3:22] * r0bby (~wakawaka@guifications/user/r0bby) has joined #jython
[3:24] * r0bby is now known as robbyoconnor
[11:21] * xemdetia (xemdetia@nat/ibm/x-ssykcwtrovqnmjdw) has joined #jython
[11:34] * plankton (~plankton@gw.grupoperetto.com.br) has joined #jython
[12:04] * xemdetia (xemdetia@nat/ibm/x-ssykcwtrovqnmjdw) Quit (Ping timeout: 245 seconds)
[12:19] * xemdetia (xemdetia@nat/ibm/x-eeijjtqwylkfqnyz) has joined #jython
[12:31] * xemdetia (xemdetia@nat/ibm/x-eeijjtqwylkfqnyz) Quit (Remote host closed the connection)
[12:36] * zz_whg is now known as whg
[12:37] * xemdetia (xemdetia@nat/ibm/x-qdplolpnbivhhufe) has joined #jython
[12:49] * whg is now known as zz_whg
[12:50] * hashcat (~hashcat@ has joined #jython
[12:57] * hashcat (~hashcat@ Quit (Quit: Leaving)
[12:57] * hashcat (~hashcat@ has joined #jython
[13:23] * plankton (~plankton@gw.grupoperetto.com.br) Quit (Quit: Saindo)
[13:55] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython
[14:15] * zz_whg is now known as whg
[14:19] * robbyoconnor (~wakawaka@guifications/user/r0bby) Quit (Excess Flood)
[14:19] * ebarrett (~edd@host-78-148-184-239.as13285.net) Quit (Excess Flood)
[14:20] * ebarrett (~edd@host-78-148-184-239.as13285.net) has joined #jython
[14:20] * r0bby (~wakawaka@guifications/user/r0bby) has joined #jython
[14:43] <agronholm> fwierzbicki: I am unable to connect to dinsdale.python.org with my current ssh key
[14:44] <agronholm> it won't accept my key
[14:45] <agronholm> my public key is AAAAB3NzaC1yc2EAAAABJQAAAIEAoAooLyA81RQSzJKuQ3r7XwFOFpuCQACW5pM7qrNDkwSJALxIFn/TIecGc94AecRyhQfzQThJVNGm7x4+W92tDb+LY8EGO+LiizgHwMFSF9SOosHLiuAklC/OFy7xAnfwEJbW41ZALsW+K84tDv4WseO2iKCqhwP+9Wi6ToV+9Nk=
[14:46] <agronholm> I'm not aware of having any other ssh keys
[14:47] * r0bby (~wakawaka@guifications/user/r0bby) Quit (Quit: Konversation terminated!)
[14:58] <fwierzbicki> agronholm: I sent that on - Benjamin Peterson looked into it, I just emailed him the above and cc'ed you.
[14:58] <agronholm> thanks
[15:06] <jimbaker> in https://bitbucket.org/jimbaker/jython-resource-leaks, i have removed all classes of resource leaks (to be verified more exhaustively, this is just applying code patterns) other than mutable public statics, as seen in PySystemState
[15:07] <jimbaker> one thing in particular is that ThreadStateMapping is implemented as weakkey/weakvalue concurrent map, courtesy of guava MapMaker
[15:09] <jimbaker> but since this slows things down a bit (about 15% on the pystone test, perhaps the one case where pystone is useful), i removed the use of recusion_limit checking, since this is something we get for free from the JVM anyway. sys.getrecursionlimit/sys.setrecursionlimit then become no-ops
[15:11] <jimbaker> (if we do restore this checking, we should follow what pypy has done and only make it approximate, by doing it PyBaseCode/PyTableCode, and not every __call__ receiver implementation)
[15:12] <jimbaker> with this change pystone is back to what we currently see
[15:13] <jimbaker> pjenvey, any thoughts from the pypy perspective?
[15:51] * int3__ (~int3__@ has joined #jython
[16:06] * xemdetia_ (xemdetia@nat/ibm/x-yarybtftamaxkaam) has joined #jython
[16:08] * xemdetia (xemdetia@nat/ibm/x-qdplolpnbivhhufe) Quit (Ping timeout: 244 seconds)
[16:45] * int3__ (~int3__@ Quit ()
[17:30] * xemdetia__ (xemdetia@nat/ibm/x-twyiwdrzqljznqhd) has joined #jython
[17:34] * xemdetia_ (xemdetia@nat/ibm/x-yarybtftamaxkaam) Quit (Ping timeout: 245 seconds)
[17:34] * xemdetia__ (xemdetia@nat/ibm/x-twyiwdrzqljznqhd) Quit (Remote host closed the connection)
[17:41] * xemdetia (xemdetia@nat/ibm/x-roassglcmbwfddst) has joined #jython
[18:22] * xemdetia (xemdetia@nat/ibm/x-roassglcmbwfddst) Quit (Remote host closed the connection)
[18:26] * xemdetia (xemdetia@nat/ibm/x-xforjdvlljmsyjhd) has joined #jython
[18:35] * hashcat (~hashcat@ Quit (Ping timeout: 255 seconds)
[19:27] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) Quit (Quit: enebo)
[20:01] <pjenvey> jimbaker - definitely want the JVM to do those checks if possible, though I remember adding a recursion limit check in the past for a test
[20:01] <pjenvey> don't really remember why it was necessary now
[20:02] <jimbaker> pjenvey, we definitely should support setting/getting, just a question of whether it's a no-op or not
[20:04] <jimbaker> however, we may want to check call depth anyway, so we can support http://bugs.jython.org/issue2142 - this would be nice to have, it just needs to be cleared out to avoid class loader leaks
[20:05] <jimbaker> so i'm thinking of supporting a hybrid solution, again at the callsite, not the receiver. let's see if i can make that work in practice
[20:27] * whg is now known as zz_whg
[21:51] <jimbaker> ok, i restored the recursion depth tracking as part of a general hybrid solution to add ThreadState as a thread local. so now getting 180373 pystones/second in my best run, about 10000 more than the approach in trunk
[21:51] <jimbaker> (pystone is a good test for this sort of thing)
[21:52] <jimbaker> (incidentally cpython 2.7 does 113480 pystones/second on my machine, a one year-old retina mac book pro. this is not a good way to compare, but hey look at how much better we are doing on this one test ;)
[21:53] <jimbaker> this also adds support for setting the class loader per thread to imp.getSyspathJavaLoader(), and should reliably clean it up, because of the call depth tracking
[22:20] <OmegaPhil> cool
[22:47] * lopex (uid4272@gateway/web/irccloud.com/x-yzfwlfaoreeadhjj) Quit (Ping timeout: 264 seconds)
[22:49] * lopex_ (uid4272@gateway/web/irccloud.com/x-exvgzujynriqdegu) has joined #jython
[23:13] * xemdetia (xemdetia@nat/ibm/x-xforjdvlljmsyjhd) Quit (Ping timeout: 264 seconds)
[23:37] * robbyoconnor (~wakawaka@guifications/user/r0bby) has joined #jython
[23:52] * enebo (~enebo@c-75-73-8-169.hsd1.mn.comcast.net) has joined #jython


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