cfbolz 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 | if a pep adds a mere 25-30 [C-API] functions or so, it's a drop in the ocean (cough) - Armin
lritter has quit [Ping timeout: 260 seconds]
lritter has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
lritter has quit [Ping timeout: 240 seconds]
isidentical has joined #pypy
lritter has joined #pypy
marky1991_2 has joined #pypy
marky1991 has quit [Ping timeout: 256 seconds]
otisolsen70 has joined #pypy
dnshane has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
dnshane has joined #pypy
<bbot2> Started: http://buildbot.pypy.org/builders/own-win-x86-32/builds/2599 [mattip: 32-bit windows]
<bbot2> Failure: http://buildbot.pypy.org/builders/own-win-x86-32/builds/2599 [mattip: 32-bit windows]
<bbot2> Started: http://buildbot.pypy.org/builders/own-win-x86-32/builds/2600 [mattip: 32-bit windows, py3.7]
lritter has quit [Remote host closed the connection]
_whitelogger has joined #pypy
tos9 has joined #pypy
marky1991_2 has quit [Ping timeout: 256 seconds]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/5760 [mattip: 32-bit windows, py3.7]
mattip has quit [*.net *.split]
lazka has quit [*.net *.split]
gsnedders has quit [*.net *.split]
luizirber has quit [*.net *.split]
holdsworth_ has quit [*.net *.split]
exarkun has quit [*.net *.split]
exarkun has joined #pypy
mattip has joined #pypy
otisolsen70_ has joined #pypy
lazka has joined #pypy
holdsworth_ has joined #pypy
luizirber has joined #pypy
holdsworth_ has quit [Max SendQ exceeded]
otisolsen70 has quit [Ping timeout: 256 seconds]
holdsworth has joined #pypy
gsnedders has joined #pypy
isidentical has quit [Quit: isidentical]
otisolsen70_ has quit [Quit: Leaving]
otisolsen70 has joined #pypy
<bbot2> Failure: http://buildbot.pypy.org/builders/own-win-x86-32/builds/2600 [mattip: 32-bit windows, py3.7]
Dejan has joined #pypy
<altendky> is there a pleasant way to get openssl available in github actions windows to build cryptography against while testing with pypy? their own ci involves calling out to another workflow to build openssl and choco seems to only provide 64 bit but the github actions python setup provides 32-bit pypy. i'm working on this as i try to expand pymodbus' build matrix to include more than just cpython on linux.
nimaje has quit [Read error: Connection reset by peer]
<simpson> altendky: There's Nix, which IMO is pleasant, but it's an acquired taste which requires effort. https://github.com/cachix/install-nix-action can set up Nix and nixpkgs and then nearly any other software you like; https://github.com/cachix/cachix-action then pushes it to public cache for others to use.
<altendky> simpson: nix for windows?
<simpson> altendky: Oh, sorry, completely missed "windows". Good point, not sure what to use.
<altendky> simpson: i'm looking again at the cryptography workflow and maybe it'll be 'ok'... might make a public action if i get it working.
<altendky> would sure be nice to just get pypy wheels up for cryptography but they decline on the basis of having to create a wheel per release due to (supposed) regular abi changes
<altendky> i don't even personally get to use pypy since i mostly end up working with qt for my applications
<mattip> altendky: you can get the builds the pypy buildbot does by
<mattip> hg up win32_14x
<mattip> or
<mattip> hg up win64_14x
<altendky> mattip: ooh, thanks.
<nulano> note that openssl in pypy/externals is just a copy of the cpython bins from https://github.com/python/cpython-bin-deps/tree/openssl-bin
nimaje has joined #pypy
<mattip> ahh, right. I forgot that is where we got them from
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/5760 [mattip: 32-bit windows, py3.7]
<altendky> mattip: thanks. https://github.com/altendky/pymodbus/blob/d4e5740e3efdb69bd28e0295e5377aca6f700b9f/.github/workflows/ci.yml#L146-L159 it is worki... well, it installed anyways and got to running the tests. do you think there would be interest in a github action repo to make this more easily available on gha?
<altendky> as mentioned i don't get to use pypy, but i still try to get more repos testing against it towards the end of making it an easier thing to use in terms of the community size
otisolsen70_ has joined #pypy
nimaje1 has joined #pypy
nimaje has quit [Killed (cherryh.freenode.net (Nickname regained by services))]
nimaje1 is now known as nimaje
otisolsen70 has quit [Ping timeout: 256 seconds]
marky1991 has joined #pypy
<mattip> altendky: did you see this https://github.com/actions/setup-python ?
<altendky> mattip: i do use that. did i miss a feature that would help with getting openssl?
<altendky> or are you suggesting that the 'get pypy externals' feature ought be added to that existing action?
<mattip> pypy and cpython ship libssl*.dll as part of the package
<mattip> maybe I don't understand the problem
<mattip> on any python you should be able to do "import _ssl"
<mattip> including the pypy installed by that action
<altendky> mattip: i need cryptography and they decline to make wheels for pypy so cryptography must be built which requires the openssl headers etc.
<altendky> really i'm working on pymodbus to add trio support, and on the side trying to get their test matrix expanded from linux/cpython. in so doing i find that in github actions you have to do something extra to get cryptography installed in pypy on windows. cryptography complains about not finding one of the openssl headers.
<LarstiQ> altendky: you might be interested in https://github.com/antocuni/pypy-wheels
<altendky> (while cryptography is being compiled)
<LarstiQ> not sure if adding support for Windows there is a one time affair, or needs to happen per package
<altendky> LarstiQ: not building at all does seem even better. i'll leave that tab around and consider helping out there as well.
<LarstiQ> antocuni: do you have thoughts on Windows wheels?
<antocuni> LarstiQ: not really
<antocuni> even the linux ones are half unmantained by now because I don't have time/feel like to actively working on them
<antocuni> but if anybody wants to volunteer in helping/improving/taking control I'll be glad
<mattip> I am not sure building our own pypi is maintainable long term.
<mattip> conda-forge is very helpful about getting pypy3.7 going
<antocuni> I agree, my wheels were a hacky solution which I needed to cut down building time for some docker images at my previous job
otisolsen70_ has quit [Quit: Leaving]
<mattip> https://conda-forge.org/status/#pypy37 shows 580 done, 442 in pr 1027 awaiting parents
<mattip> conda create -n pypy37 pypy3.7; conda activate pypy37
<mattip> gets me pretty far, including cryptography
<antocuni> maybe we should declare conda as the default/preferred/supported way to use pypy and write nice docs
marky1991 has quit [Ping timeout: 256 seconds]
marky1991 has joined #pypy
oberstet has joined #pypy
<mattip> antocuni: sounds good
marky1991 has quit [Ping timeout: 260 seconds]
marky1991 has joined #pypy
marky1991 has quit [Ping timeout: 256 seconds]
otisolsen70 has joined #pypy
isidentical has joined #pypy
<mattip> pyston was on podcast.__init__ this week
<cfbolz> cool
<cfbolz> did you listen to it?
<cfbolz> going on podcasts is still something I want to try at some point
<mattip> yes. I found it interesting to hear the perspective of another implementation.
<mattip> they are targeting web apps (django, tornado) which I thought we do quite well at
<mattip> I think they now get like 20% faster than CPython (if I understood correctly)
<Gustavo6046> weird, pyenv only has pyston up to 0.6.1
<Gustavo6046> ops, wrong channel
<Gustavo6046> well, not wrong but
<Gustavo6046> eh
<simpson> Gustavo6046: No worries. At least you're being productive, trying out Pyston and seeing how it works.
otisolsen70 has quit [Remote host closed the connection]
<cfbolz> Gustavo6046: "Our plan is to open-source the code in the future, but since compiler projects are expensive and we no longer have benevolent corporate sponsorship, it is currently closed-source while we iron out our business model."
<cfbolz> that might be the reason
otisolsen70 has joined #pypy
<simpson> Looking for nice things to say while listening to this podcast. Their thoughts on GC vs RC when it comes to removing the GIL are good, though.
isidentical has quit [Quit: isidentical]
ctismer has quit [Ping timeout: 264 seconds]
otisolsen70 has quit [Remote host closed the connection]
ctismer has joined #pypy
otisolsen70 has joined #pypy
otisolsen70 has quit [Remote host closed the connection]
otisolsen70 has joined #pypy
otisolsen70 has quit [Remote host closed the connection]
otisolsen70 has joined #pypy
otisolsen70 has quit [Remote host closed the connection]
oberstet has quit [Remote host closed the connection]
<Gustavo6046> Ah
<Gustavo6046> cfbolz: hmm, I see
<Gustavo6046> simpson: thanks!
<Gustavo6046> But yeah, doesn't Pyston already distribute binaries or something?