02:00
jcea has quit [Ping timeout: 260 seconds]
02:50
lritter has quit [Ping timeout: 256 seconds]
02:50
lritter has joined #pypy
03:55
oberstet has quit [Remote host closed the connection]
04:22
_whitelogger has joined #pypy
05:03
forgottenone has joined #pypy
07:06
marvin_ has quit [Remote host closed the connection]
07:06
marvin__ has joined #pypy
07:07
proteusguy has quit [Remote host closed the connection]
08:25
swills has quit [Ping timeout: 240 seconds]
08:38
swills has joined #pypy
09:07
swills has quit [Ping timeout: 240 seconds]
09:08
swills has joined #pypy
09:23
swills has quit [Ping timeout: 246 seconds]
09:29
swills has joined #pypy
09:33
swills has quit [Ping timeout: 246 seconds]
09:37
swills has joined #pypy
09:41
swills has quit [Ping timeout: 240 seconds]
10:24
jcea has joined #pypy
10:42
swills has joined #pypy
10:44
jcea has quit [Quit: jcea]
10:47
jcea has joined #pypy
10:54
proteusguy has joined #pypy
10:56
Dejan has quit [Ping timeout: 246 seconds]
11:32
<
cfbolz >
mattip_, arigo: so with my hack (of sometimes popping two vmprof frames) the crash is fixed
12:18
<
LarstiQ >
cfbolz: how do you determine when to do that?
12:27
<
cfbolz >
LarstiQ: at the place where we had the failing assert, I check whether one frame up has the correct code object and then I pop twice. if not, it still crashes
12:27
<
cfbolz >
it's a hack, but I think it's better than crashing the vm
13:47
oberstet has joined #pypy
14:26
marky1991 has quit [Ping timeout: 256 seconds]
17:08
mattip_ is now known as mattip
17:09
<
mattip >
LarstiQ: I don't think that particular problem is relevant for PyPy, we have a pure-python _ctypes module
18:42
<
fijal >
yay memory safe languages
18:43
<
fijal >
but also, if you control ctype.from_param argument, surely you can crash something?
18:43
<
cfbolz >
weeell, python is only safeish
18:43
<
cfbolz >
too many loopholes
18:43
<
fijal >
like, when not using ctypes for example?
18:44
<
fijal >
I guess you can end up with a sane path to that function call
18:44
<
cfbolz >
fijal: cffi? ;-)
18:45
<
fijal >
cfbolz: we have a bug in sketchup ruby GC (probably) where an object of
*completely random type* lands somewhere
20:16
oberstet has quit [Remote host closed the connection]
20:49
marky1991 has joined #pypy
20:51
marky1991 has quit [Remote host closed the connection]
20:52
marky1991 has joined #pypy
20:57
marky1991 has quit [Ping timeout: 256 seconds]
20:59
amirouche has joined #pypy
20:59
<
amirouche >
hello is there a howto use cffi with poetry package manager?
21:53
<
mattip >
amirouche: I doubt anyone here uses poetry. What have you tried, what did you expect, and what happened instead?
21:53
<
amirouche >
I tried nothing :)
21:54
<
amirouche >
poetry reinvent many things compared to setup.py
21:54
<
amirouche >
I mean does differently.
21:55
<
mattip >
I still don't understand what your goal is
21:57
<
mattip >
cfbolz: sorry, I missed the notification that you were coding
21:57
<
cfbolz >
mattip: poetry is another python packaging thingie. how do you package a cffi library with it?
21:57
<
cfbolz >
mattip: oops, maybe I should ping here consistently?
21:57
<
amirouche >
mattip: I try to create bindings for C library and package it with poetry (which does rely on setup.py)
21:59
<
cfbolz >
I think the docs probably should contain an example. but first somebody needs to figure out how it works :-(
21:59
<
mattip >
"Note that, at the moment, only pure python wheels are supported."
22:01
<
mattip >
as far as I know, the only way to build a wheel for c-extension modules is with setup,py
22:01
<
mattip >
even things like scikit-build use setup.py underneath
22:02
<
amirouche >
tx matt
22:55
forgottenone has quit [Quit: Konversation terminated!]