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
mgedmin has quit [Ping timeout: 245 seconds]
holdsworth_ has joined #pypy
astrojl_matrix has quit [*.net *.split]
papangoo[m]1 has quit [*.net *.split]
holdsworth has quit [*.net *.split]
astrojl_matrix has joined #pypy
papangoo[m]1 has joined #pypy
lritter has quit [Ping timeout: 264 seconds]
lritter has joined #pypy
jcea has quit [Ping timeout: 260 seconds]
forgottenone has joined #pypy
<mattip> it's live: https://www.pypy.org/blog/
<mattip> today's puzzle: why is the win64 run of rpython hitting the pypy directory instead?
<mattip> win32 does not do that
jacob22 has quit [Ping timeout: 264 seconds]
<mattip> fixed that was easier to find than I thought
oberstet has joined #pypy
lritter has quit [Ping timeout: 246 seconds]
lritter has joined #pypy
Taggnostr2 has quit [Ping timeout: 240 seconds]
Taggnostr has joined #pypy
Taggnostr has quit [Quit: Switching to single player mode.]
<cfbolz> mattip: yay, thanks a lot for your work on this!
<cfbolz> We should ask Planet python for replacing the blog (there's still a surprising number of people reading it) https://github.com/python/planet/issues
<mattip> hehe, the blog post is already there https://planetpython.org
<mattip> so we need to make a PR to switch the fed?
<mattip> feed
Taggnostr has joined #pypy
<cfbolz> mattip: or an issue, I am not sure
otisolsen70 has joined #pypy
<antocuni> mattip: congrats for the blog, and thank you!
<mattip> :)
<antocuni> we should do the same for hpy :)
<antocuni> mattip: I tried to write a comment to check that they work
<mattip> grr. I couldn't try it out till it was live
<antocuni> it seems it work, doesn't it?
<mattip> not on firefox
<mattip> I get a "Cross-Origin Request Blocked: The Same Origin Policy disallows reading the remote resource at https://api.github.com ..."
<antocuni> it worked for me with chrome
* antocuni tries with firefox as well
<mattip> ahh, it is ghostery blocking something
<antocuni> I just made another comment from firefox
<mattip> if I disable ghostery it works
<antocuni> makes sense I suppose, it seems that ghostery is explicitly trying to block these kind of inter-website integrations
<mattip> it's OK if I "trust" pypy.org in the ghostery UI
<antocuni> ok
<cfbolz> mattip: I'm not sure what we should do with the heap pull request. We usually try to keep the stdlib diff to CPython small
<mattip> cfbolz: maybe put it off till 3.10, when the change will show up in the stdlib?
<cfbolz> mattip: or we accept it in order to put pressure on cpython to accept it as well 😅
<cfbolz> I'll take a more detailed look later in any case
<mattip> cool
jryans has quit [Quit: Bridge terminating on SIGTERM]
ronny has quit [Quit: Bridge terminating on SIGTERM]
ShadeJonathan[m] has quit [Quit: Bridge terminating on SIGTERM]
astrojl_matrix has quit [Quit: Bridge terminating on SIGTERM]
jevinskie[m] has quit [Quit: Bridge terminating on SIGTERM]
toad_polo has quit [Quit: Bridge terminating on SIGTERM]
papangoo[m]1 has quit [Quit: Bridge terminating on SIGTERM]
papangoo[m] has joined #pypy
toad_polo has joined #pypy
astrojl_matrix has joined #pypy
ShadeJonathan[m] has joined #pypy
ronny has joined #pypy
jryans has joined #pypy
jevinskie[m] has joined #pypy
lritter has quit [Ping timeout: 260 seconds]
Dejan has joined #pypy
jacob22 has joined #pypy
jacob22 has quit [Read error: No route to host]
jcea has joined #pypy
DRMacIver has quit []
DRMacIver has joined #pypy
jacob22 has joined #pypy
cfbolz has quit []
cfbolz has joined #pypy
jacob22_ has joined #pypy
jacob22 has quit [Ping timeout: 272 seconds]
jacob22_ has quit [Ping timeout: 264 seconds]
<fijal> ok, so the pypy nightly comes with no six?
<fijal> which means pip does not work, at least in some version
<mattip> which nightly did you try?
<mattip> ensurepip should install a functioning pip
<mattip> pypy -m ensurepip
<fijal> ok
<fijal> I got a pip from ubuntu and upgrading it helped
<fijal> I mean "would be cool if pypy warned you about stuff" but I really don't know why, it's a very unconstructive problem
<fijal> but maybe we can detect the version of pip and error with "this will not work"
<fijal> ?
<mattip> I don't understand. Can you describe what you did, starting from which nightly you used?
infernix has quit [Remote host closed the connection]
infernix has joined #pypy
<fijal> I downloaded latest 3.7 nightly
<fijal> and downloaded some recent ubuntu, 20.10
<fijal> you create a virtualenv using nightly pypy and boom, nothing works
<larstiq> fijal: yeah, that's a known problem
<larstiq> I've been trying to find a writeup about that I recently read, but having some trouble
<larstiq> in https://groups.google.com/g/dev-python/c/UhcKOVyugcA it's " SC discussed how Debian (and thus Ubuntu) ships Python, in particular how this affects
<larstiq> users and their view of Python. "
<larstiq> but no link to the discussion, hrmf
<cfbolz> larstiq: the discussion was some gist, I think
<cfbolz> pretty weird place to talk
<larstiq> cfbolz: yeah
<larstiq> I remember I found that link from the cryptography rust bug report
<fijal> well, maybe pypy can detect the situation?
<fijal> because I'm pretty sure the answer will be "pypy does not work"
<fijal> anyway, no, I don't know how to fix it from pypy side
<fijal> but I think pretty much I ever try pypy + virtualenv + pip it explodes in bizzare ways that have possibly nothing to do with pypy
jacob22_ has joined #pypy
<mattip> you mean "sudo apt install virtualenv" and then using that, right?
<Dejan> i stopped using virtualenv the moment they added pip to python3
<Dejan> pardon, the "venv" module
<Dejan> <python3 interpreter> -m venv <env name> is how I do it for quite some time
<Dejan> colleagues love conda
<Dejan> ImportError: /opt/jobserve/pypy/site-packages/numpy/core/_multiarray_umath.pypy37-pp73-x86_64-linux-gnu.so: undefined symbol: PyPyUnicode_Check
<Dejan> is this a known issue?
<Dejan> i was basically trying to install numexpr
<mattip> that is meant to be a macro, so maybe you used the wrong headers somehow?
<mattip> where did you get numpy from?
<mattip> fijal: this works for me:
<mattip> virtualenv -p pypy/bin/pypy /tmp/venv
<mattip> /tmp/venv/bin/pypy -mpip install numpy
<mattip> but maybe that is not a clean, untouched pypy/bin/pypy, let me try again
<mattip> this works too
<mattip> mkdir /tmp/clean; cd /tmp/clean
<fijal> this is what pip?
<mattip> tar -xf <latest pypy 3.7>
<mattip> pip 20.0.2 from /tmp/venv/site-packages/pip (python 3.7)
<Dejan> i tried to pip3 install numexpr
<Dejan> in pypy3.7 environment
<mattip> grr. another meeting
<mattip> fijal: can you put up a complete reproducer somewhere?
<fijal> mattip: I had pip from ubuntu?
<fijal> maybe that's the problem?
<fijal> came with no six installed
<fijal> tthat's 20.1.1
<fijal> no, now works
<fijal> I HAVE NO CLUE
<cfbolz> fijal: you are still not very precise in the steps that lead to the problem :-P
<fijal> well, this is the history that I found
<fijal> I *think* this is not a reproducer because I can't reproduce it any more
<cfbolz> I can reproduce
<cfbolz> with the python3-virtualenv ubuntu package
<cfbolz> fijal: ^^
<fijal> oh cool
<fijal> I don't know what I changed that I can no longer reproduce
<fijal> (and I changed quite a bit I guess, trying to frantically get it to work)
dmalcolm__ has joined #pypy
lazka8 has joined #pypy
whitewlf has joined #pypy
runciter_ has joined #pypy
runciter has quit [*.net *.split]
dmalcolm_ has quit [*.net *.split]
lazka has quit [*.net *.split]
whitewolf has quit [*.net *.split]
energizer has quit [*.net *.split]
lazka8 is now known as lazka
energizer has joined #pypy
ulope has quit [Ping timeout: 246 seconds]
oberstet has quit [Quit: Leaving]
ulope has joined #pypy
forgottenone has quit [Ping timeout: 246 seconds]
<mattip> cfbolz: this is without running "pypy -m ensurepip" ?
EWDurbin has quit []
EWDurbin has joined #pypy
michelp has quit []
michelp has joined #pypy
<cfbolz> mattip: I'm not sure, honestly
<cfbolz> let me check
<cfbolz> mattip: doesn't matter whether I run ensurepip or not, always get the failing six import
<mattip> thanks
otisolsen70 has quit [Quit: Leaving]
bbot2_ has quit [Quit: buildmaster reconfigured: bot disconnecting]
pjenvey1 is now known as pjenvey
* mattip restarting buildbot for win64 rpython run typo
bbot2 has joined #pypy
<mattip> cfbolz: mine looks different, it is installing six https://paste.ubuntu.com/p/pW97YMJNmd
<cfbolz> mattip: and virtualenv is from python3-virtualenv?
<mattip> (reformatted for more clarity) https://paste.ubuntu.com/p/T9tq7FbrFr/
<mattip> virtualenv is in /usr/bin/virtualenv, so I assume that is from ubuntu
<cfbolz> mattip: yes, but I think ubuntu has two such packages, a python3 based one and a python2 one
<mattip> but note the line app_data_dir=/home/matti/.local/share/virtualenv/seed-app-data/v1.0.1.debian
<mattip> you have app_data_dir=/home/cfbolz/.local/share/virtualenv
<mattip> in mine there is all this stuff in /home/matti/.local/share/virtualenv/seed-app-data/v1.0.1.debian/3.7/image/CopyPipInstall
<cfbolz> yeah, it's a game of hunt-the-difference now :-(
<cfbolz> maybe it's because I indeed have another virtualenv script installed somehow, that's not from ubuntu
* cfbolz needs to sleep
<mattip> apparently there is a --download option to virtualenv, that will download lots of packages including six
<mattip> maybe that is stored across invocations and once downloaded always used?
<mattip> and a --six option to choose which version of six to install
graingert has quit []
graingert has joined #pypy
* tumbleweed has a look at the virtualenv thing
<tumbleweed> looks like that's Ubuntu 20.10
jaraco has quit []
jaraco has joined #pypy
<tumbleweed> OK, so this is caused by assumptions in the pip devendoring patch in Debian/Ubuntu
<tumbleweed> I bet it breaks any python built from source
<tumbleweed> s/built/running/
jcea has quit [Ping timeout: 260 seconds]
jcea has joined #pypy