01:00 <bbot2> Started: http://buildbot.pypy.org/builders/build-pypy-c-jit-linux-armel/builds/2499 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/build-pypy-c-jit-linux-armhf-raspbian/builds/2260 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4015 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5740 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7023 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4171 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/6097 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/4863 [py3.5]

06:51 <kenaan> mattip unicode-utf8-py3 55b18ad54e86 /pypy/module/_codecs/test/test_locale.py: don't encode utf8 in tests

06:51 <mattip> down to 6 failing own tests on unicode-utf8-py3 for linux64

06:51 <mattip> this one pypy.module._multibytecodec.test.test_app_codecs AppTestCodecs.().test_encode_custom_error_handle

06:51 <mattip> does not pass with -A python=python3

06:57 <mattip> these tests pypy.module._multibytecodec.test.test_c_codecs test_decode_hz*

06:58 <mattip> seem problematic too, hz encoding seems not well defined on CPYthon

06:58 <mattip> for more info see http://buildbot.pypy.org/summary?branch=unicode-utf8-py3

12:44 <antocuni> https://talkpython.fm/episodes/transcript/113/dedicated-ai-chips-and-running-old-python-faster-at-intel

12:45 <antocuni> do any of us know David Stewart or Suresh Srinivas? (or maybe they are here in the channel?)

12:45 <antocuni> reading this interview, they seem to imply that Intel is optimizing PyPy for their processors and they are pushing contributions upstreams

12:45 <cfbolz> eh

12:45 <antocuni> but I am not aware of any of such contributions?

12:45 <cfbolz> mail them?

12:46 <antocuni> yeah, I still have to finish reading it

12:46 <antocuni> but e.g.: "eah, and so what's really cool, I think, is some of this work that you guys are doing is being pushed upstream to CPython, it's being pushed upstream to PyPy."

12:46 <cfbolz> antocuni: they mention openstack, I suppose they mean the work they did after the sprint

12:47 <antocuni> yes, this I remember; but AFAIK, there was never any direct contribution to PyPy?

12:48 <cfbolz> antocuni: there were some pull requests merged afterwards

12:48 <cfbolz> like the support for profile-guided optimization

12:49 <antocuni> ah, ok

12:49 <antocuni> btw, this is also interesting: http://languagesperformance.intel.com/

12:50 <antocuni> ah, but I cannot find any reference to what these benchmarks actually are

12:51 <cfbolz> who knows

15:51 <fijal> antocuni: their claims and contributions are very very strange

15:51 <fijal> I mean, if all they need is some form "yes, we are collaborating with OS projects!" then cool, they can have it as far as I'm concerned

15:54 <antocuni> yes, but a casual reader will think they are actively contributing on pypy

15:54 <antocuni> at least, that's what my collegues at gambit thought :)

15:58 <fijal> right

15:58 <fijal> pfff

15:58 <fijal> not sure what can be done to stop them from saying things like that

17:56 <pjenvey> only read the first few paragraphs but I'm reading it as maybe a thinko, David says "we were showing off a doubling of throughput of like OpenStack Swift with the PyPy contributions we had made" -- maybe meaning "pypy contributions we had made" as a confusing description of their efforts to run swift on it

17:57 <pjenvey> interviewer understandbly thinks they contributed to pypy but then David clarifies it later

22:38 <_aegis_> I'm building a very strict import graph and `from package import module` is messing me up (I just see the package import)

22:38 <_aegis_> 6 IMPORT_NAME 0 (user)

22:38 <_aegis_> 9 IMPORT_FROM 1 (std)

22:38 <_aegis_> so my import hook only sees the user

22:39 <_aegis_> is there some way to hook IMPORT_FROM and get the path `user.std`?

22:39 <_aegis_> my current plan is to step back a frame and look for IMPORT_FROM in the bytecode

22:39 <_aegis_> because I can't see another way besides shimming packages or something

22:43 <_aegis_> I'm going to double check sys.meta_path but I feel like there was a reason that wasn't working :(

22:56 <_aegis_> oh meta_path doesn't seem to work for repeated imports of the same module

22:56 <_aegis_> it only fires for the first import