antocuni changed the topic of #pypy to: PyPy, the flexible snake (IRC logs: https://botbot.me/freenode/pypy/ ) | use cffi for calling C | "PyPy: the Gradual Reduction of Magic (tm)"
__main__ has joined #pypy
yuyichao_ has quit [Ping timeout: 240 seconds]
antocuni has quit [Ping timeout: 260 seconds]
yuyichao_ has joined #pypy
Hotpot33 has quit [Ping timeout: 248 seconds]
Hotpot33 has joined #pypy
dddddd has quit [Remote host closed the connection]
awkwardpenguin has joined #pypy
awkwardpenguin has quit [Ping timeout: 264 seconds]
tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
Hotpot33 has quit [Ping timeout: 240 seconds]
marr has quit [Ping timeout: 260 seconds]
Hotpot33 has joined #pypy
fryguybob has quit [Ping timeout: 256 seconds]
inhahe_ has quit []
pilne has quit [Quit: Quitting!]
_aegis_ has joined #pypy
inhahe_ has joined #pypy
fryguybob has joined #pypy
ArneBab_ has joined #pypy
squeaky_pl has quit [Ping timeout: 252 seconds]
ArneBab has quit [Ping timeout: 264 seconds]
jtl_ has joined #pypy
jcea has quit [Ping timeout: 250 seconds]
forgottenone has joined #pypy
_whitelogger has joined #pypy
Hotpot33 has quit [Ping timeout: 240 seconds]
jtl_ has quit [Remote host closed the connection]
tayfun26 has joined #pypy
Hotpot33 has joined #pypy
dddddd has joined #pypy
<arigato> crash when playing with sys.setrecursionlimit(): ah, I think I get what occurs now
<arigato> or more precisely, I am sure I do
<arigato> the shadow stack overflows
<arigato> random nonsense follows
<kenaan> mattip cffi/cffi[doc-set_source] 889d38233ce8 /doc/source/overview.rst: try to clarify set_source() requires str(<code that implements functions>)
<arigato> compiling the C sources with anything less than "-O3" makes the bug go away---which makes sense:
<arigato> in these modes the C compiler is better at making the C stack compact, and thus the C stack grows more slowly
<arigato> so comparatively, the shadow stack grows more quickly
<arigato> seems it overflows the 1.3MB already when the C stack is only 3.7MB
mattip has joined #pypy
<arigato> mattip: doc-set_source: I see the point quite a lot, but I don't think we should do it by changing the single example to add an indirection
<arigato> because now it looks like you need to write such a dummy indirection for all functions
<mattip> I usually start with something stupid like "double(float x){return x*2.0;}"
<arigato> then we should do the same, I guess
<mattip> s/double/doubler/
<arigato> I'm not against rewriting this overview page more deeply
<arigato> so yes, as a first step, it's likely a good step
<mattip> Is my summary at the end of the stackoverflow question accurate enough? I could add that somewhere too
<arigato> I would prefer if someone (might be you or me) goes over overview.rst and tries to rethink the whole structure
<arigato> based on what we learned people are not getting or the more natural order for teaching
<mattip> agreed
<mattip> cannot promise "when", win32 and osx py3.5 worry me more
<arigato> :-)
<kenaan> mattip extradoc ecb8d7f22159 /planning/sprint-leysin-2018-notes.rst: start thinking about leysin sprint
<mattip> arigato: "the shadow stack overflows" - can we detect that somehow?
<arigato> detecting it is easy: mprotect() the last page
<arigato> what to do is harder
<arigato> for now I'll open an issue
<kenaan> arigo default 0eb832cbb13a /pypy/doc/cpython_differences.rst: Issue #2719: Keyword arguments to built-in functions
<mattip> arigato: can you point me in the direction of fixing winreg.EnumKey ?
<mattip> it is very painful to run own tests on py3.5 win32, they are seriously slow and errors are non-helpful
<arigato> :-/
* arigato looks inside CPython 3.5
<mattip> ahh, ok that already is a hint
<arigato> the problem is all the newtext() in interp_winreg.py
<arigato> they decode utf8
<arigato> which is probably wrong
<mattip> cool. I will try to compare to cpython
<arigato> ah, CPython uses the RegQueryValueW() instead of RegQueryValue() function (and the same for all others I guess)
<arigato> so that gives us an RPython unicode instead of string
<arigato> and then we should use space.newunicode()
<arigato> with no decoding
<arigato> there is an explicit call to decode with argument 'mbcs' too in pypy, needs to be checked if CPython does the same thing in the same cases
<arigato> thanks! :-)
<mattip> :)
jamesaxl has quit [Read error: Connection reset by peer]
jamesaxl has joined #pypy
awkwardpenguin has joined #pypy
marr has joined #pypy
awkwardpenguin has quit [Ping timeout: 248 seconds]
antocuni has joined #pypy
forgottenone has quit [Quit: Konversation terminated!]
exarkun has quit [Ping timeout: 272 seconds]
exarkun has joined #pypy
squeaky_pl has joined #pypy
ronan has joined #pypy
runciter has quit [Ping timeout: 252 seconds]
runciter has joined #pypy
LexLeoGryfon has joined #pypy
<LexLeoGryfon> hello
<LexLeoGryfon> anyone could help me? i'm trying to make embedded binary with cython+pypy https://github.com/cython/cython/wiki/FAQ#id50
<LexLeoGryfon> my question is : what arguments should i specify in this line in order to compile and link with pypy?
<LexLeoGryfon> $(CC) -I /usr/include/python$(PYVERSION) foobar.c -lpython$(PYVERSION) -o foobar
<LexLeoGryfon>
<LexLeoGryfon> the default line is not working
<LexLeoGryfon> for me
<LarstiQ> LexLeoGryfon: I have no idea with cython, but linking against cpython as you do above will for sure not get you pypy
<ronan> LexLeoGryfon: embedding pypy surely requires special support in cython, but I don't know whether it exists. You should ask cython devs
<LarstiQ> LexLeoGryfon: note that you can embed pypy directly via cffi
<LexLeoGryfon> LarstiQ: what I see here http://doc.pypy.org/en/latest/embedding.html is that it really embedding pypy in C code in order to interpret python code.
<LexLeoGryfon> but my goal is - is to build simple single static binary with hello world function
<LexLeoGryfon> build it from python code
<LexLeoGryfon> and in order to achieve that i'm trying to https://github.com/cython/cython/wiki/FAQ#id50
<LarstiQ> LexLeoGryfon: you can do the same with the cffi embedding
<LarstiQ> LexLeoGryfon: as ronan said, cython's --embed probably needs pypy support and we don't know if it has it, ask cython people
<LexLeoGryfon> k, thnx, i'm checking out https://cffi.readthedocs.io/en/latest/embedding.html
<LexLeoGryfon> as for cython devs - I asked them, but you have faster responce time :) They probably havent read my question yet
awkwardpenguin has joined #pypy
antocuni has quit [Ping timeout: 240 seconds]
awkwardpenguin has quit [Ping timeout: 256 seconds]
exarkun has quit [Ping timeout: 264 seconds]
exarkun has joined #pypy
tav has quit [Quit: tav]
tav has joined #pypy
tav has quit [Ping timeout: 240 seconds]
<LarstiQ> you could also start reading how --embed is implemented, see if you figure it out before they respond ;)
oberstet has joined #pypy
<LexLeoGryfon> thank you LarstiQ, for now I managed to compile single binary with cython+cpython according to their guide. however it is dynamically linked. i guess I need to compile static version of interpreter to link statically. As for pypy, I will continue expiriment later
raynold has quit [Quit: Connection closed for inactivity]
<kenaan> mattip py3.5 213db4a13e7c /testrunner/get_info.py: add more info for buildbot virtualenv run with -A
* mattip running ^^ against some non-pushed changes to buildbot to see if it solves win32 virtualenv problems
exarkun has quit [Ping timeout: 248 seconds]
exarkun has joined #pypy
antocuni has joined #pypy
fryguybob has quit [Ping timeout: 264 seconds]
Rhy0lite has joined #pypy
antocuni has quit [Ping timeout: 272 seconds]
<kenaan> mattip py3.5 3e7cb4758b05 /testrunner/get_info.py: fix target_name
<kenaan> mattip buildbot 424f8ecee620 /bot2/pypybuildbot/builds.py: use more Properties from testrunner/get_info, leave old defaults in place
bbot2 has quit [Quit: buildmaster reconfigured: bot disconnecting]
bbot2 has joined #pypy
jcea has joined #pypy
forgottenone has joined #pypy
mattip has left #pypy ["bye"]
realitix has joined #pypy
tayfun26 has quit [Remote host closed the connection]
antocuni has joined #pypy
TheAdversary has joined #pypy
Hasimir has joined #pypy
fryguybob has joined #pypy
oberstet has quit [Ping timeout: 260 seconds]
antocuni has quit [Ping timeout: 240 seconds]
realitix has quit [Quit: realitix]
RemoteFox has joined #pypy
<RemoteFox> hello
<LexLeoGryfon> hai
exarkun has quit [Ping timeout: 264 seconds]
exarkun has joined #pypy
raynold has joined #pypy
forgottenone has quit [Quit: Konversation terminated!]
awkwardpenguin has joined #pypy
awkwardpenguin has quit [Ping timeout: 240 seconds]
<kenaan> rlamy py3.5 49d834e68cbe /pypy/module/posix/interp_posix.py: Fix err.filename when *xattr() functions raise an OSError
<ronan> mattip (logs): ^^^ for 5.10.1
forgottenone has joined #pypy
Rhy0lite has quit [Quit: Leaving]
chelz has joined #pypy
forgottenone has quit [Remote host closed the connection]
oberstet has joined #pypy
forgottenone has joined #pypy
tbodt has joined #pypy
squeaky_pl has quit [Ping timeout: 264 seconds]
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
forgottenone has quit [Remote host closed the connection]
forgottenone has joined #pypy
<glyph> Is the slowness of pip under pypy a known / well understood issue?
<glyph> This is so agonizingly slow (10x, 20x cpython?) I'm wondering if I've misconfigured something with my devpi instance
forgottenone has quit [Remote host closed the connection]
pilne has joined #pypy
forgottenone has joined #pypy
<LexLeoGryfon> glyph: I'm not sure, just using pypy3.5 Nth day myself, but for me pip performsvery well, already installed few packages in venv.
<LexLeoGryfon> maybe your issue relates to specific packages, or your individual environment
<glyph> LexLeoGryfon: it could be the fact that I just upgraded to 5.10 final and now it's building a bunch of new wheels
chelz has quit [Ping timeout: 276 seconds]
exarkun has quit [Ping timeout: 248 seconds]
exarkun has joined #pypy
LexLeoGryfon has quit [Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/]
<tos9> glyph: I don't think I've ever noticed such a thing.
<tos9> glyph: Though I haven't used CPython in a really long time, so maybe I just have normalized to how fast it is on PyPy :)
<glyph> tos9: After caching a bunch of wheels pypy is much faster, so this may just be C compiler invocation time
* tos9 nods
<tos9> glyph: ( c.f. https://github.com/antocuni/pypy-wheels which is nice )
chelz has joined #pypy
exarkun has quit [Ping timeout: 240 seconds]
exarkun has joined #pypy
fryguybob has quit [Ping timeout: 240 seconds]
jcea has quit [Ping timeout: 240 seconds]
fryguybob has joined #pypy
jcea has joined #pypy
jcea has quit [Ping timeout: 265 seconds]
awkwardpenguin has joined #pypy
jcea has joined #pypy
awkwardpenguin has quit [Ping timeout: 248 seconds]
chelz has quit [Ping timeout: 265 seconds]
Greasy-Gappers has joined #pypy