00:08
swills has quit [Ping timeout: 240 seconds]
00:34
swills has joined #pypy
01:02
swills has quit [Ping timeout: 260 seconds]
01:25
swills has joined #pypy
01:33
lritter has quit [Ping timeout: 240 seconds]
02:52
Techcable has joined #pypy
03:02
jcea has quit [Ping timeout: 240 seconds]
04:40
forgottenone has joined #pypy
05:19
marky1991 has quit [Ping timeout: 260 seconds]
05:20
forgottenone has quit [Remote host closed the connection]
06:01
amirouche1 has left #pypy ["WeeChat 2.8"]
06:23
<
mattip >
using pypy instead of python3 seems to be a mixed bag when running "emerge" on gentoo
06:24
<
mattip >
sometimes faster, sometimes slower
06:55
_whitelogger has joined #pypy
07:10
isidentical has joined #pypy
07:28
dmalcolm has quit [Ping timeout: 265 seconds]
07:34
mgorny has quit [Quit: No Ping reply in 60 seconds.]
07:35
mgorny has joined #pypy
07:50
dmalcolm has joined #pypy
08:33
<
mgorny >
mattip: it's not clear though if it depends on the data or hardware
08:33
<
mgorny >
might be a case of underoptimization
09:35
<
cfbolz >
Bit hard to investigate without having a gentoo installation, I suppose?
09:44
<
fijal >
so apparently we don't have a working distribution for OS X
09:44
<
fijal >
(which is not called OS X but I don't care)
09:45
<
fijal >
got glyph to test our nightly
10:25
glyph has quit [Quit: End of line.]
10:25
glyph has joined #pypy
10:59
forgottenone has joined #pypy
13:05
marky1991 has joined #pypy
13:10
marky1991 has quit [Remote host closed the connection]
13:11
marky1991 has joined #pypy
13:31
marky1991 has quit [Remote host closed the connection]
13:32
marky1991 has joined #pypy
13:37
forgottenone has quit [Ping timeout: 260 seconds]
13:44
oberstet has joined #pypy
13:51
lritter has joined #pypy
14:28
marky1991 has quit [Ping timeout: 256 seconds]
14:42
<
mattip >
So far we have avoided wrapping tp_finalize when creating a PyObject* from a RPython w_obj
14:43
<
mattip >
According to the PEP for tp_finalize, calling it is equivalent to calling obj.__del__
14:45
<
mattip >
is this true in PyPy as well, or should it call a finalizer instead?
14:59
Dejan has joined #pypy
14:59
Dejan has joined #pypy
15:25
<
mattip >
in type instantiation, there is a check if the class has a
__del__ then the typedef gets a hasuserdel flag
15:26
<
mattip >
then in allocate_instance, that causes a call to self.finalizer_queue.register_finalizer(instance)
15:30
<
mattip >
so I guess tp_del should be a NOP: our GC will clean up the instance when it does gc.collect()
15:35
jcea has joined #pypy
15:40
<
cfbolz >
mattip: I don't understand the cpython slots enough to usefully answer, I fear :-(
16:25
marky1991 has joined #pypy
16:25
marky1991 has quit [Remote host closed the connection]
16:25
marky1991 has joined #pypy
16:35
marky1991 has quit [Ping timeout: 240 seconds]
17:28
marky1991 has joined #pypy
18:02
marky1991 has quit [Ping timeout: 240 seconds]
18:32
marky1991 has joined #pypy
19:49
<
cfbolz >
mattip: good find!
20:09
<
mattip >
I was answering a stack overflow question about multiprocessing and looked through the issues
20:09
<
mattip >
it turns out _pickle is important
20:10
<
cfbolz >
mattip: I should find out this week whether the student wants to do it. if not, I might work on it for a bit myself
20:37
<
mattip >
cool. it seems we should do a release soon, it would be nice to get a _pickle in if it is not too fiddly
20:39
<
cfbolz >
mattip: students don't work
*that* quickly ;-)
20:40
<
cfbolz >
but indeed, if he says no we'll see
20:40
marky1991 has quit [Ping timeout: 265 seconds]
21:53
rubdos has quit [Quit: WeeChat 2.8]
22:02
isidentical has quit [Quit: isidentical]
22:28
<
tos9 >
"Unable to copy 'C:\\hostedtoolcache\\windows\\PyPy\\3.7.9\\x86\\venvlauncher.exe'"
22:29
<
tos9 >
But yeah does anyone know if that indicates GHActions' PyPy is missing a file on Windows, or is that our/PyPy's issue
22:29
<
tos9 >
(Or our/Julian's bad CI setup)
22:29
<
tos9 >
I suppose in the normal Yak stack it can also be a tox-on-pypy-on-windows issue
22:39
<
tos9 >
(Never mind)
23:42
Techcable has quit [Remote host closed the connection]
23:42
Techcable has joined #pypy