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
lritter has quit [Ping timeout: 258 seconds]
lritter has joined #pypy
todda7 has quit [Ping timeout: 264 seconds]
lritter has quit [Quit: Leaving]
todda7 has joined #pypy
<njs> mattip: Python's 'atexit' is implemented by the interpreter, so the semantics are the same across platforms I think
jcea has quit [Quit: jcea]
jvesely has quit [Quit: jvesely]
todda7 has quit [Ping timeout: 246 seconds]
inhahe has quit []
jvesely has joined #pypy
jvesely has quit [Remote host closed the connection]
jvesely has joined #pypy
inhahe has joined #pypy
jvesely has quit [Quit: jvesely]
mgedmin has quit [Quit: ZNC - https://wiki.znc.in/ZNC]
mgedmin has joined #pypy
jvesely has joined #pypy
jvesely has quit [Client Quit]
jvesely has joined #pypy
jvesely has quit [Remote host closed the connection]
jvesely has joined #pypy
Rhy0lite has quit [Quit: This computer has gone to sleep]
mukundan314 has joined #pypy
epsilonKNOT has quit [Ping timeout: 258 seconds]
mukundan314 is now known as mukundan3141
mukundan3141 is now known as mukundan314
<mattip> njs: I mean the _exit() C function call which will immediately terminate the program on POSIX systems, maybe _Exit is an alternative spelling
_whitelogger has joined #pypy
epsilonKNOT has joined #pypy
<mattip> it seems wb_before_move in general gives a ~5% gain, but then there are benchmarks that are slower
<mattip> the permalink on the comparison page doesn't work properly, here is how I got a comparison to the latest linux64:
<mattip> then in executables use only pypy-jit-64 latest in branch 'wb_before_move' and pypy-jit-64 latest
<mattip> then at the top of the page "horizontal"
<mattip> and normalization (all the way at the bottom) pypy-jit-4-latest
<mattip> I think some of the benchmarks that show wb_before_move as slower are jittery, like sympy_sum
oberstet has joined #pypy
mukundan314 has quit [Remote host closed the connection]
dddddd has quit [Ping timeout: 260 seconds]
<arigo> mattip: I'm not up to digging much into these cases, I *think* it's good enough that it doesn't seem to have massive negative effects (and it has massive positive effects in the issue's code)
<mattip> +1
<mattip> Once we sort out the macOS packaging issues I would like to do a 7.3.2
<mattip> so next week I think
<arigo> nice
* mattip still dreaming of a python3.7 alpha
<mattip> I wonder what project's test suite would flush out important 3.7 problems that we should fix before an alpha
<arigo> nothing seems to have been done about 're' so far
todda7 has joined #pypy
<mattip> is 're' broken for normal use cases?
<arigo> I don't know
<arigo> people mentioned here that the re engine changed in 3.7
<arigo> I don't know more though
<mattip> we have that packaging script that runs against the top 1000 pypi packages, but there is no standard way to run test suites
<mattip> so it only checks if the package can be installed
todda7 has quit [Ping timeout: 240 seconds]
<mattip> there is regex, which has some tests https://bitbucket.org/mrabarnett/mrab-regex/src/hg/
todda7 has joined #pypy
mwhudson has quit [Ping timeout: 244 seconds]
<arigo> it's c1aa5a9acd68 I am talking about
todda7 has quit [Ping timeout: 260 seconds]
marvin__ has quit [Remote host closed the connection]
marvin has joined #pypy
mwhudson has joined #pypy
mwhudson has quit [Changing host]
mwhudson has joined #pypy
todda7 has joined #pypy
<mattip> got it, but I wonder how many of the failures of test_re.py are used by packages/end-users
todda7 has quit [Read error: No route to host]
<cfbolz> I started looking into sre but then got distracted
<cfbolz> It's not superhard it's just that most of the bytecode numbers changed for not a lot of reasons
<cfbolz> There are also a couple of new ones but they don't seem hard to implement or are variants of ones that we already have
<arigo> OK
<cfbolz> Ist nobody else get to it, I will probably resume working on the problem from July on
<arigo> OK
otisolsen70 has joined #pypy
Dejan has joined #pypy
Dejan has quit [Changing host]
Dejan has joined #pypy
marky1991 has quit [Ping timeout: 264 seconds]
<mattip> hmm. we fail 14/100 of the regex tests from https://bitbucket.org/mrabarnett/mrab-regex
<mattip> both with py3.6 and py3.7, so I guess 3.7 is no worse than 3.6 ;/
<Dejan> NumPy will drop 3.6 support it seems
<Dejan> actually, mattip you mentioned that here yesterday, did you?
<mattip> numpy won't until December, but other packages in the scientific data stack might
<mattip> that is why I am trying to figure out if we are ready for a py3.7 release
<mattip> alpha release
dddddd has joined #pypy
YannickJadoul has joined #pypy
Rhy0lite has joined #pypy
marky1991 has joined #pypy
BPL has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
jacob22 has quit [Read error: Connection reset by peer]
jacob22 has joined #pypy
lritter has joined #pypy
<Dejan> mattip, 👍
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
YannickJadoul has quit [Quit: Leaving]
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
marky1991 has quit [Ping timeout: 240 seconds]
Dejan has quit [Quit: Leaving]
marky1991 has joined #pypy
oberstet has quit [Quit: Leaving]
otisolsen70 has quit [Ping timeout: 240 seconds]
BPL has quit [Quit: Leaving]
Ninpo has quit [Ping timeout: 264 seconds]
Ninpo has joined #pypy