arigato changed the topic of #pypy to: PyPy, the flexible snake (IRC logs: https://quodlibet.duckdns.org/irc/pypy/latest.log.html#irc-end ) | use cffi for calling C | mac OS and Fedora are not Windows
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
fryguybo1 has quit [Read error: Connection reset by peer]
lritter has quit [Ping timeout: 255 seconds]
fryguybob has joined #pypy
jcea has quit [Quit: jcea]
forgottenone has joined #pypy
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
jacob22 has quit [Read error: Connection reset by peer]
jacob22 has joined #pypy
dddddd has quit [Remote host closed the connection]
_whitelogger has joined #pypy
_whitelogger has joined #pypy
<mattip> is py3.6 really on stdlib 3.6.1? we should update that
<cfbolz> Yes
<cfbolz> mattip: is there documentation how to do that?
<mattip> "we have a branch called vendor/stdlib"
<cfbolz> mattip: but no explanation how that branch gets updated
<mattip> looking at the last work on that branch, it seems for python2.7 something like
<mattip> - update the stdlib in that branch to the latest version, clobbering what is there
<mattip> - start a new branch off it vendor/stdlib2.7.x where x is the latest version
<mattip> - merge default into that branch
<mattip> - cleanup, fix, eventually merge back to default
<mattip> it would be nice to document the process this time
<cfbolz> mattip: yes
<mattip> for instance, vendor/stdlib-3.5 is closed, but vendor/stdlib-3.5.1 is not, while stdlib-2.7.13 was closed-and-merged to default
<cfbolz> mattip: right
agronholm has quit [Ping timeout: 250 seconds]
agronholm has joined #pypy
dddddd has joined #pypy
<mattip> building _blake2 fails on linux32 https://paste.ubuntu.com/p/b74kkgzrNv
<mattip> it seems the chroot is picking up the wrong headers?
<mattip> something to do with -msse vs -msse2 ?
<mattip> windows also is failing to build blake2
<kenaan> cfbolz py3.6 1d06586fec74 /pypy/interpreter/: copy slightly lengthy test from cpython to find the (crashing) bug in set_lineno
<kenaan> cfbolz py3.5 eb6e70a9658e /: close 3.5 branch to focus on py3.6
<hruske> is there anything else you pypy folks would find interesting about python packages, besides if a package is using cython?
<hruske> obviously that was more of a one time question, but still
<bbot2> Started: http://buildbot.pypy.org/builders/own-win-x86-32/builds/1980 [mattip: force build, py3.6]
<mattip> weird, the test_rutf8 failure on rpython runs http://buildbot.pypy.org/summary?builder=rpython-linux-x86-64&branch=%3Ctrunk%3E
<mattip> shouldn't be there
<mattip> in that version
<mattip> the build is missing 213fc2573b4d somehow
<kenaan> mattip default 789d5650e3af /rpython/rlib/test/test_rutf8.py: fix
lritter has joined #pypy
vstinner has joined #pypy
<vstinner> hi. there is an ongoing discussion on the C API on python-dev: https://mail.python.org/pipermail/python-dev/2019-February/156344.html you may want to have a look
<fijal> vstinner: it's a typical very-long-discussion with no real consensus?
<simpson> I found it informative and useful for fine-tuning my ideas about why extension modules are bad.
<mattip> arigato: do you remember how you set up the chroot on bencher4? I want to do the same for a 32 bit chroot buildslave
<mattip> but a 32 bit chroot is a bit special, it needs more things
<mattip> I don't want to mess up bencher4 by installing the wrong things
nunatak has joined #pypy
<arigato> mattip: I don't think a 64-bit chroot is different from a 32-bit chroot
nunatak has quit [Quit: Leaving]
<arigato> sorry, I don't remember exactly where /mnt/ubuntu came from
<arigato> (or why it is in /mnt in the first place)
<bbot2> Failure: http://buildbot.pypy.org/builders/own-win-x86-32/builds/1980 [mattip: force build, py3.6]
marky1991 has joined #pypy
marky1991 has quit [Ping timeout: 246 seconds]
<arigato> vstinner: by now I think that if we're seriously trying to help pypy, we need to define a new C API that works differently; just cleaning up the existing C API is nice but doesn't help pypy at all now
<kenaan> arigo cffi/cffi edac34833b64 /doc/source/ref.rst: Issue #402 Fix bug in documentation
marky1991 has joined #pypy
marky1991 has quit [Ping timeout: 245 seconds]
<kenaan> stevie_92 cpyext-gc-cycle 78d356baf93c /pypy/module/cpyext/slotdefs.py: Implemented wrapper for tp_finalize
marky1991 has joined #pypy
marky1991 has quit [Ping timeout: 272 seconds]
marky1991 has joined #pypy
zzarr has joined #pypy
zzarr has left #pypy ["Leaving"]
zzarr has joined #pypy
<zzarr> Hello!
<zzarr> I have build a project using the ABI method of cffi
Rhy0lite has joined #pypy
<zzarr> but, I have problems with callbacks, the callbacks are running in different threads
<zzarr> but the callbacks are not "visible" to the GIL
danieljabailey has joined #pypy
<mattip> arigato: ok, I am mostly set up with the chroot. How did you update mercurial to 3.7? I see you did not install it from apt
forgottenone has quit [Ping timeout: 255 seconds]
<zzarr> I'll be back tomorrow
marky1991 has quit [Remote host closed the connection]
zzarr has quit [Ping timeout: 272 seconds]
marky1991 has joined #pypy
<vstinner> fijal: "it's a typical very-long-discussion with no real consensus?" i'm not sure that you expect an answer to this question...
<vstinner> arigato: "just cleaning up the existing C API is nice but doesn't help pypy at all now" ok, i started to understand that, but well, it should still help PyPy :)
<vstinner> arigato: oops. i would still help CPython :)
<vstinner> (hum. i'm unable to write a correct sentence, let me retry one more time) ok, i started to understand that, but well, it should still help CPython :)
<vstinner> i wanted to share the link since you are using the C API, and you may want to contribute to the discussion to give your feedback
<vstinner> if your point is just to say "it's a typical very-long-discussion with no real consensus?" i prefer leave this room
vstinner has left #pypy ["User left"]
<fijal> We are all for making pythreadstate internal
<fijal> But we can't be bothered to argue the finest points of "what if someone relies on it, despite not being documented"
<fijal> So yes, making pythreadstate and py interpreters tatę and all the structs opaque is a good development I think
Ai9zO5AP has joined #pypy
bogner has quit [Ping timeout: 244 seconds]
bogner has joined #pypy
marky1991 has quit [Ping timeout: 250 seconds]
marky1991 has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
forgottenone has joined #pypy
Zaab1t has joined #pypy
Zaab1t has quit [Quit: bye bye friends]
<lesshaste> what are the current main goals for pypy? Are people still working hard at making numpy faster?
<lesshaste> as in speeding up cpyext
forgottenone has quit [Quit: Konversation terminated!]
<pjenvey> mattip: there's some docs in pypy/lib-python/stdlib-upgrade.txt
jcea has joined #pypy
epelesis has joined #pypy
igitoor has quit [Remote host closed the connection]
marky1991 has quit [Ping timeout: 255 seconds]
marky1991 has joined #pypy
nimaje has quit [Ping timeout: 258 seconds]
nimaje has joined #pypy
igitoor has joined #pypy
marky1991 has quit [Ping timeout: 246 seconds]
Ai9zO5AP has quit [Ping timeout: 245 seconds]
marky1991 has joined #pypy
igitoor has quit [Changing host]
igitoor has joined #pypy
Rhy0lite has quit [Quit: Leaving]
<fijal> lesshaste: pypy is done we can all go home ;-)
<cfbolz> pjenvey: awesome, thanks
<cfbolz> lesshaste: yes, we plan to work on cpyext performance and py 3.6 (and later 3.7) support
<cfbolz> pjenvey: ah, seems we are now doing the first import after their move to git
epelesis has quit [Remote host closed the connection]
marky1991 has quit [Read error: Connection reset by peer]
inhahe has quit [Ping timeout: 255 seconds]
dddddd has quit [Remote host closed the connection]
inhahe has joined #pypy
inhahe_ has joined #pypy
<Hodgestar> lesshaste: Not sure if you have read https://morepypy.blogspot.com/2018/09/inside-cpyext-why-emulating-cpython-c.html or not, but worthwhile reading if you haven't and are interested in cpyext performance (it was written in late 2018, work started in 2017). Summary: The story for improving cpyext performance changed from ??? to lots-of-work-but-there-is-a-plan-without-too-many-question-marks (and maybe particular cases could
<Hodgestar> be made faster without a huge amount of work).
kipras has joined #pypy
lritter has quit [Ping timeout: 246 seconds]
inhahe has quit []