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
oberstet has quit [Remote host closed the connection]
dddddd__ has quit [Ping timeout: 264 seconds]
_whitelogger has joined #pypy
jcea has quit [Quit: jcea]
<jaraco>
Good news - I’ve gotten pypa/distutils@master to run tests as a standalone implementation of distutils.
<jaraco>
(just run ‘tox’ to run the distutils tests)
Rhy0lite has quit [Quit: This computer has gone to sleep]
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
thrnciar_ has joined #pypy
oberstet has joined #pypy
<mattip>
jaraco: did you see my PR to your branch?
<arigo>
we never did anything related to PyThreadState_SetAsyncExc, right?
<arigo>
also asking for people that use ctypes.pythonapi.PyThreadState_SetAsyncExc, which is the semi-official way in CPython to interrupt a non-main thread
phlebas has joined #pypy
<mattip>
it has been in stubs.py for quite a while
<mattip>
I compared a run of untranslated pypy/module/signal/test between default and py3.6, after skipping the new AppTestPThread tests
<the_drow[m]>
Has any of you checked the pattern matching PEP?
<the_drow[m]>
How hard will it be to implement?
<cfbolz>
that's a long way off
<cfbolz>
we're going to have to implement the new parser first
<njs>
arigo: mattip: pypy's lack of set_wakeup_fd on windows is somewhat annoying btw, in case that's related to your issue there
glyph has quit [Quit: End of line.]
glyph has joined #pypy
<the_drow[m]>
<cfbolz "we're going to have to implement"> Oh I thought that PyPy's parser is more advanced.
<the_drow[m]>
Are we anywhere near a working Python 3.7 release?
<cfbolz>
the_drow[m]: yes, the idea is to do an alpha soon
<cfbolz>
The main missing feature is some re changes
<cfbolz>
Our parser is a pretty exact port of the old cpython one
<the_drow[m]>
Is the upgrade from 3.6 to 3.7 that hard? Or were the dev team lacking the time/funds to do it?
<cfbolz>
every upgrade is hard ;-)
<cfbolz>
but yes, we have been losing contributors a bit
<cfbolz>
corona isn't helping either
<the_drow[m]>
;/
<the_drow[m]>
With what's going on in the world it's completely understandable
<the_drow[m]>
it's hard to know what's missing from the issues alone
<the_drow[m]>
Maybe you should consider using boards or milestones
<cfbolz>
the_drow[m]: we use the failing tests as an indication
<the_drow[m]>
Yes but that's not visible to most users
<the_drow[m]>
Have you guys considered using Gitlab CI instead of buildbot on the same machines you use to test PyPy?
<cfbolz>
that's work though ;-)
<the_drow[m]>
It's work I can do
<cfbolz>
(we are in fact using gitlab CI as a very simple "stuff isn't completely broken" check nowadays already)
<the_drow[m]>
I wonder if implementing the new parser earlier will help with the walrus operator
<the_drow[m]>
You're introducing another operator anyway
<cfbolz>
the_drow[m]: no, the parser is extremely easy to adapt with the old parser: it's automatically generated out of the same grammar that CPython has been using so far
<cfbolz>
this will actually be much harder in the future, the new grammar contains snippets of C code that don't help us at all
<njs>
mattip: hmm, did you also update the signal handling code to handle sockets on windows? the actual implementation is pretty different on unix and windows...
<mattip>
nope
<cfbolz>
mattip: ah, great, thank you
<njs>
mattip: in particular, on windows the "fd" is not an fd, but a socket handle, and you have to use socket functions like 'send' (on windows, you can't call 'write' on a socket handle)
<mattip>
the code in rpython/translator/c/signals.c might need some work then
<mattip>
if I understand you correctly, the fd in pypysig_set_wakeup_fd can be a socket?
* mattip
trying to compare CPython code
jcea has joined #pypy
<mattip>
in CPython the win32 and non-win32 implementations of signal_set_wakeup_fd share very little code
<mattip>
njs: does this work properly on pypy2.7 ?
<njs>
mattip: dunno
<mattip>
in cpython, the wakup object has a use_send boolean used in trip_send
<mattip>
ok, so in short, a mess, and untested in PyPy apparently since tests pass (like most of windows)
<mattip>
I will revert that last commit on py3.6 and add a more descriptive comment
todda7 has joined #pypy
<njs>
mattip: the trio testsuite will definitely exercise this if you need some test
xcm has quit [Ping timeout: 240 seconds]
<mattip>
thanks
xcm has joined #pypy
<mattip>
we should document all of the places things work in linux64 and not in windows
otisolsen70 has joined #pypy
<the_drow[m]>
<cfbolz "the former is implementation-lan"> I see...
<the_drow[m]>
So we need it to use RPython instead?
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
YannickJadoul has joined #pypy
epony has quit [Ping timeout: 258 seconds]
xcm has quit [Remote host closed the connection]
<arigo>
mattip: fwiw, I'm pretty sure none of the wakeup mess code is in CPython 2.7 either (i.e. it was a unix-only feature)
xcm has joined #pypy
Rhy0lite has joined #pypy
the_drow[m] has quit [Quit: killed]
agates[m] has quit [Quit: killed]
salotz[m] has quit [Quit: killed]
toad_polo has quit [Quit: killed]
lritter has joined #pypy
todda7 has quit [Read error: Connection reset by peer]
tsaka__ has joined #pypy
agates[m] has joined #pypy
dddddd has joined #pypy
YannickJadoul has quit [Ping timeout: 240 seconds]
salotz[m] has joined #pypy
the_drow[m] has joined #pypy
toad_polo has joined #pypy
dmalcolm_ has quit [Remote host closed the connection]
dmalcolm_ has joined #pypy
epony has joined #pypy
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
<jaraco>
mattip: I haven’t seen the PR yet, but I’ll look soon. Thank you again.
<mattip>
iterating on it now: I removed cov so now one of the runs finishes with 3 failures, which is what I was seeing
<mattip>
locally
tsaka__ has quit [Ping timeout: 244 seconds]
lastmikoi has quit [Ping timeout: 265 seconds]
lastmikoi has joined #pypy
salotz[m] has quit [Quit: Idle for 30+ days]
lastmikoi has quit [Quit: ...]
petronny has joined #pypy
lastmikoi has joined #pypy
<cfbolz>
the_drow[m]: yes
Dejan has joined #pypy
Dejan has quit [Changing host]
Dejan has joined #pypy
tsaka__ has joined #pypy
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
<mattip>
gracinet has imported all the attachments to the issues which were missing in our original repo import
<mattip>
I deleted two very old issues with spam content as attachments.