01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4641 [py3.6]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/6467 [py3.6]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/5245 [py3.6]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4421 [py3.6]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/6243 [py3.6]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7493 [py3.6]

02:04 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/5245 [py3.6]

02:19 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/6243 [py3.6]

03:38 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/6467 [py3.6]

03:46 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4421 [py3.6]

04:05 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7493 [py3.6]

05:00 <bbot2> Started: http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/2613

06:13 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4641 [py3.6]

06:52 <njs> what's the main python 3 branch now?

06:53 <njs> just noticed that I have CI still testing against the py3.5 branch and that seems out-of-date...

06:53 <mattip> py3.6

06:53 <njs> cool

06:55 <mattip> someday we may have a py3.7 or even py3.8

07:17 <bbot2> Success: http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/2613

08:57 <lesshaste> is most of the pypy work these days on getting 3.x compatibility?

08:57 <lesshaste> or are people still working on making numpy fast?

08:58 <mattip> lesshaste: mostly fixing bugs, you can see the hg history

08:58 <lesshaste> mattip, ah ok. That's a shame in a way as I thought your ideas for making numpy fast were exciting

08:59 <lesshaste> pypy devs tend to be very good at making things fast even when it appears impossible

09:11 <lesshaste> mattip, is it a manpower issue or did the numpy work hit a brick wall?

09:14 <mattip> the walls are getting higher, but I think there are ways around them.Money to pay developers always helps

09:15 <mattip> it would be nice if microsoft would discover PyPy and bring windows up to speed, or

09:15 <mattip> someone who cared about c-extensions would contribute to making those fast

09:17 <LarstiQ> there seems to be quite a python team in Microsoft, maybe talk to some of them?

09:19 <mattip> we need a compelling use case with data to back it up

09:20 <LarstiQ> well, if it was me I'd start with "are you aware of pypy and what do you think about it on windows?"

09:20 <LarstiQ> just having a conversation already helps

09:48 <mattip> they know about PyPy, some of them are core CPython devs and they are very aware of what is going on

09:49 <mattip> since they develop Visual Code or whatever it is called they also have user statistics

09:49 <mattip> for some cross-section of users

09:55 <rindolf> mattip: https://github.com/Microsoft/vscode/

12:07 <alawrence> mattip: Do we have any usage statistics for PyPy?

13:36 <alawrence> mattip: I have defined a module at the interpretter level called _testconsole that contains two methods. I would like to also expose the methods at the application level. Is this possible? If it is how do I go about doing this?

13:37 <arigato> alawrence: for usage in test files?

14:56 <alawrence> arigato: _testconsole is used in the lib-python tests. I was thinking I may also need it to write interpreter level tests but it could be that I don't do this and just put the module in lib_pypy.

16:33 <LarstiQ> mattip: I see. Do you know what they think of PyPy?

21:37 <kenaan> andrewjlawrence winconsoleio a5a68d94883d /: Further implementation of winconsoleio. Work in progress.