00:21
lritter has joined #pypy
00:32
jcea has quit [Quit: jcea]
00:45
<
ronan >
firespeaker: pypy3 is packaged in the same way as cpython, so you can't pip install anything on the system pypy3, which makes it rather useless as you've noticed. OTOH, if you create a virtualenv based on it, you should be able to install what you want
01:52
gracinet has quit [Ping timeout: 246 seconds]
02:41
xcm has quit [Remote host closed the connection]
02:42
andi- has quit [Remote host closed the connection]
02:43
xcm has joined #pypy
02:47
andi- has joined #pypy
02:56
lritter has quit [Ping timeout: 265 seconds]
02:57
lritter has joined #pypy
03:00
xcm has quit [Remote host closed the connection]
03:01
xcm has joined #pypy
03:43
xcm has quit [Remote host closed the connection]
03:46
xcm has joined #pypy
04:18
jvesely has quit [Quit: jvesely]
04:26
dddddd has quit [Ping timeout: 240 seconds]
04:29
dddddd has joined #pypy
05:27
<
mattip >
cfbolz: 5dd2d3cca75b broke translation
05:29
xcm has quit [Remote host closed the connection]
05:31
xcm has joined #pypy
05:36
dddddd has quit [Remote host closed the connection]
05:47
DarkPlutonium has joined #pypy
06:08
xcm has quit [Remote host closed the connection]
06:09
xcm has joined #pypy
06:11
dustinm- has joined #pypy
06:12
dustinm has quit [Ping timeout: 240 seconds]
06:13
oberstet has quit [Ping timeout: 250 seconds]
06:20
zmt01 has quit [Read error: Connection reset by peer]
06:21
zmt01 has joined #pypy
06:38
oberstet has joined #pypy
07:10
<
cfbolz >
mattip: looking
07:31
oberstet has quit [Remote host closed the connection]
07:41
oberstet has joined #pypy
07:56
xcm has quit [Remote host closed the connection]
07:57
xcm has joined #pypy
08:09
antocuni has joined #pypy
08:18
Ai9zO5AP has joined #pypy
08:30
xcm has quit [Remote host closed the connection]
08:32
xcm has joined #pypy
08:45
<
kenaan >
cfbolz default c866b476cfee /rpython/jit/metainterp/quasiimmut.py: fix translation, tweak output
09:20
gracinet has joined #pypy
09:28
jeroud has quit [Read error: Connection reset by peer]
09:29
jeroud has joined #pypy
10:11
antocuni has quit [Ping timeout: 265 seconds]
10:20
rubdos has quit [Ping timeout: 246 seconds]
10:22
rubdos has joined #pypy
10:31
jcea has joined #pypy
10:41
<
kenaan >
mattip default 241777245f33 /pypy/tool/release/repackage.sh: refactor repackage script to generate exe
10:53
lritter has quit [Remote host closed the connection]
11:03
xcm has quit [Ping timeout: 268 seconds]
11:07
xcm has joined #pypy
11:25
jcea has quit [Quit: jcea]
11:30
<
kenaan >
mattip py3.6 f963c5cf8e2e /: merge default into py3.6
11:42
antocuni has joined #pypy
12:15
xcm has quit [Remote host closed the connection]
12:17
xcm has joined #pypy
12:24
ionelmc has quit [Quit: Connection closed for inactivity]
12:53
Rhy0lite has joined #pypy
13:01
jvesely has joined #pypy
13:10
dddddd has joined #pypy
13:15
dmalcolm has quit [Remote host closed the connection]
13:20
DarkPlutonium is now known as DarkUranium
13:20
david has joined #pypy
13:21
david is now known as Guest9758
14:18
BPL has joined #pypy
14:20
astronavt___ is now known as astronavt
14:37
<
firespeaker >
ronan, so the suggested way to numpy with pypy3 on debian unstable is in a virtualenv?
14:37
<
firespeaker >
how do I go about doing that?
14:39
<
firespeaker >
ronan: also, is there any reason pypy3 is packaged using cpython stuff in debian?
14:39
<
firespeaker >
it means pretty much everything fails out of the box
14:39
<
ronan >
it's something like 'virtualenv -p `which pypy3` my-pypy-env', then 'source my-pypy-env/bin/activate'
14:40
<
ronan >
as for why it's like that, ask tumbleweed
14:41
<
mattip >
maybe even "pypy3 -mvenv my-pypy-env", if they didn't remove venv as well as ensurepip
14:41
<
ronan >
yes, that should work too
14:41
<
ronan >
but virtualenv installs pip, which venv doesn't IIRC
14:59
<
firespeaker >
by `virtualenv`, you mean `python3 /usr/lib/python3/dist-packages/virtualenv.py` ?
14:59
<
firespeaker >
or something else?
15:01
<
ronan >
firespeaker: yes, that thing
15:01
<
firespeaker >
cool, seems to be letting me install numpy from pip in that virtual environment
15:01
<
firespeaker >
though mattip said something about it not being optimised when built with pip
15:03
<
ronan >
well, it can be optimised if you install and configure the right things (see numpy docs)
15:04
<
firespeaker >
in the virtual environment?
15:04
<
firespeaker >
what are wheels?
15:06
<
firespeaker >
like, a python install with a bunch of libraries?
15:06
<
ronan >
a binary distribution format for python packages
15:06
<
firespeaker >
ah, okay
15:07
<
ronan >
pip uses them automatically if it finds an appropriate one
15:07
<
firespeaker >
hm, it says those are for ubuntu
15:08
<
ronan >
they're very likely to work on debian as well
15:08
<
firespeaker >
lol, the pypy3 virtualenv I made runs this script at over twice as long as just cpython
15:10
<
firespeaker >
oh man, and I screwed something up on my system last night; now pip3 says "no module named pip"
15:12
<
firespeaker >
okay an apt reinstall fixed that
15:13
<
firespeaker >
ronan: hm, should I uninstall my package manager's version of numpy before I install that wheel?
15:13
<
firespeaker >
otherwise I get this:
15:13
<
firespeaker >
Requirement already satisfied: numpy in /usr/lib/python3/dist-packages (1.16.5)
15:14
<
firespeaker >
or should I do it in my virtualenv?
15:14
<
ronan >
you should do it in your virtualenv
15:14
<
firespeaker >
ahkay
15:15
<
ronan >
the whole point is that the package directories are separate
15:15
<
firespeaker >
when I use that numpy I get errors
15:16
<
ronan >
firespeaker: check the pypy-wheels docs, you need to install some libs
15:17
<
firespeaker >
hm, not finding those docs...
15:18
<
firespeaker >
I see it says three libraries
15:18
<
firespeaker >
ah, I see, misread it at first
15:20
<
firespeaker >
aha, first debian/ubuntu difference found
15:20
<
firespeaker >
libopenblas, not libblas
15:22
<
firespeaker >
okay, still taking over twice as long to run this script with pypy3 and that wheel version of numpy
15:22
<
firespeaker >
anyway, matplotlib is the bottleneck in this script
15:26
oberstet has quit [Quit: Leaving]
15:26
<
firespeaker >
I guess I should try to make a matplotlib wheel for pypy?
15:26
* firespeaker
guesses this isn't trivial
15:49
marky1991 has joined #pypy
15:55
<
ronan >
Yet another oddity in the encode()/decode() protocol: error handlers can completely replace the input
15:56
<
ronan >
I'm wondering if we should implement it or just say "Go home CPython, you're drunk!"
15:58
ekaologik has joined #pypy
16:08
marky1991 has quit [Remote host closed the connection]
16:09
<
ronan >
actually, it's only for decoding
16:10
marky1991 has joined #pypy
16:29
marky1991 has quit [Remote host closed the connection]
16:31
marky1991 has joined #pypy
17:00
antocuni has quit [Ping timeout: 268 seconds]
17:09
marky1991 has quit [Remote host closed the connection]
17:19
<
cfbolz >
ronan: yes, it's completely ridiculous
17:19
<
cfbolz >
I don't think it makes sense to try to implement it
17:25
<
ronan >
well, it doesn't seem that hard to implement, but pfff...
17:43
gracinet has quit [Quit: Leaving.]
18:07
Rhy0lite has quit [Quit: Leaving]
18:11
xcm has quit [Ping timeout: 268 seconds]
18:15
xcm has joined #pypy
18:18
Ai9zO5AP has quit [Quit: WeeChat 2.5]
18:35
xcm is now known as Guest19094
18:35
Guest19094 has quit [Killed (wolfe.freenode.net (Nickname regained by services))]
18:36
xcm has joined #pypy
19:01
antocuni has joined #pypy
19:07
Ai9zO5AP has joined #pypy
19:12
<
cfbolz >
ronan: let's raise a WhatWereYouSmokingError
19:48
antocuni has quit [Ping timeout: 240 seconds]
20:33
<
mattip >
ronan: I started dealing with absurd error handlers on the codec_errorhandler branch
20:35
<
mattip >
feel free to continue it if this is what you meant or to redo it if not
21:53
inhahe has joined #pypy
22:20
_whitelogger has joined #pypy
23:24
jcea has joined #pypy