#jython IRC Log (v0.9)


IRC Log for 2017-05-21

Timestamps are in GMT/BST.

[8:06] -niven.freenode.net- *** Looking up your hostname...
[8:06] -niven.freenode.net- *** Checking Ident
[8:06] -niven.freenode.net- *** No Ident response
[8:06] -niven.freenode.net- *** Couldn't look up your hostname
[8:06] * JythonLogBot (~PircBot@ has joined #jython
[8:06] * Topic is 'Jython 2.7 final released: http://tinyurl.com/muqfuul | This channel is logged: http://jython.extreme.st/irclogs/ | Please update the wiki: http://wiki.python.org/jython | Jython Book: http://jythonbook.com | Podcast: http://jython.org/jythonpodcast/'
[8:06] * Set by agronholm!~agronholm@nblzone-211-20.nblnetworks.fi on Sun May 03 05:10:52 UTC 2015
[9:42] * girish946 (~gsjoshi@ has joined #jython
[9:47] * Oti (~ohumbel@77-54-60-212.dyn.cable.fcom.ch) Quit (Quit: Oti)
[10:00] * girish946 (~gsjoshi@ Quit (Quit: Leaving)
[12:10] <topi`> so what's going to be new in the upcoming jython 2.7.1 release?
[13:50] * stewori (~stefan@ip-95-223-198-21.hsi16.unitymediagroup.de) has joined #jython
[16:48] * agronholm (~agronholm@2a00-1190-c00e-0000-021f-d0ff-fead-d185.v6.cust.suomicom.net) has joined #jython
[16:48] * ChanServ sets mode +o agronholm
[19:01] * agronholm (~agronholm@2a00-1190-c00e-0000-021f-d0ff-fead-d185.v6.cust.suomicom.net) has left #jython
[21:00] <topi`> jimbaker: I took a simple java algo of a mandelbrot calculation and converted it to python. import Mandel; Mandel.Mandel() runs in 0.2 seconds and import mand; mand.mand() runs in about 16 seconds
[21:03] <topi`> that is an order of magnitude difference. My wish is that in the future (jython 3.5) we could enable a certain optimization, that depends on type hints, that would allow to skip the normal type checks that are clearly slowing down the execution here
[21:04] <topi`> btw, do you remember how jython's Complex class is implemented? Is it backed by JDK stdlib or is there a custom impl in java?
[21:09] <stewori> it is implemented in org.python.core.PyComplex. So I guess it's what you refer to as a custom implementation
[21:09] <topi`> yeah
[21:09] <stewori> JDK stdlib features no complex AFAIK
[21:09] <stewori> I observed this "magnitudes of slower"-thing when testing sympy
[21:10] <topi`> interesting, I opened a file for writing and tried to write the contents of the buffer and I got java.lang.NullPointerException
[21:10] <topi`> I wonder what could happen there, the code looks innocent
[21:10] <stewori> I wasn't able to complete sympy tests. At some point I wasn't able to decide if it's hanging or just sooo slow
[21:11] <topi`> actually this code also bugs with python2.7
[21:11] <topi`> TypeError: argument 1 must be string or buffer, not list
[21:11] <stewori> topi, share the failing code at github or bugs.jython.org
[21:11] <topi`> but it's nasty that with jython you just get NullPointer
[21:11] <stewori> okay, then the cod emight be faulty...?
[21:12] <topi`> optimally it should say the same as CPython: argument must be string, not list
[21:12] <stewori> okay. Then its an error-msg bug. Maybe post at bugs.jython.org
[21:12] <topi`> I'll try to make a smallest code possible to tease out that bug
[21:13] <stewori> perfect
[21:13] <topi`> yeah, there it goes
[21:13] <topi`> lst = [1,2,3,4,5]
[21:13] <topi`> with open('out', 'wb') as f: f.write(lst)
[21:13] <stewori> I'm currently investigating what extlibs we can safely upgrade for 2.7.1
[21:14] <stewori> please post this to bugs.jython.org
[21:14] <stewori> so we can keep track. Right now I cannot look into this
[21:14] <topi`> let's see if I remember my account at jython.org :)
[21:17] <topi`> there it is :)
[21:20] <topi`> oops, wrong description. should be list, not buffer
[21:22] <topi`> now, I can fix my original code :)
[21:22] <topi`> the java version used a buffer, not a list of integers... my bad
[22:04] * stewori (~stefan@ip-95-223-198-21.hsi16.unitymediagroup.de) has left #jython


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