arigato 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 | mac OS and Fedora are not Windows
brent has quit [Read error: Connection reset by peer]
speeder39_ has quit [Quit: Connection closed for inactivity]
antocuni has joined #pypy
antocuni has quit [Ping timeout: 245 seconds]
demonimin has quit [Ping timeout: 245 seconds]
demonimin has joined #pypy
demonimin has joined #pypy
forgottenone has joined #pypy
jcea has quit [Quit: jcea]
tsaka_ has joined #pypy
tsaka__ has quit [Remote host closed the connection]
brent has joined #pypy
brent has quit [Read error: Connection reset by peer]
brent has joined #pypy
dddddd has quit [Remote host closed the connection]
darkman66 has joined #pypy
darkman66 has quit [Ping timeout: 245 seconds]
<mattip> tos9: did you give up?
darkman66 has joined #pypy
<kenaan> andrewjlawrence winoverlapped 79630c7e2718 /: Fixed exception types produced by overlapped.getresult, reverted rwin32.py
<kenaan> andrewjlawrence winoverlapped 60b0c39477d5 /lib_pypy/_winapi.py: Add connection aborted error code.
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4516 [Andy Lawrence: force build, winoverlapped]
xcm has quit [Read error: Connection reset by peer]
xcm has joined #pypy
darkman6_ has joined #pypy
darkman66 has quit [Ping timeout: 250 seconds]
oberstet has joined #pypy
<petronny> mattip: I get this error (https://www.irccloud.com/pastebin/hqhbEe0z/pypy3-armv6h-qemu.log) again I can get `self.graph` now.
<petronny> I run `(Pdb+) p self.graph` and it returns `<FunctionGraph of (pypy.module._sre.interp_sre:109)W_SRE_Pattern.make_ctx at 0x7a4d8fb0>`.
<petronny> What should I do next?
<mattip> ping cfbolz, arigato :)
<mattip> (context: a crash in translation on ARM32 in a qemu)
<mattip> petronny: how did you get a qemu arm32 with more than 256MB ram?
<cfbolz> looking a bit
<petronny> with `qemu-user-static` and `binfmt-qemu-static`
<petronny> I write a script to download the latest archlinuxarm image for rpi and copy `/usr/bin/qemu-arm-static` into it.
<petronny> Then with a little more tweaks I can chroot into it with `arch-chroot` on archlinux.
<cfbolz> petronny: can you give me the output of 'p ARRAY' and 'p FIELD' from the pdb?
<cfbolz> (in the bottommost frame)
<petronny> uh, which command reverses `up`?
<cfbolz> down
<petronny> okay but I have to leave for dinner now ... I'll post the output 1 hour later.
<cfbolz> petronny: thanks a lot for doing this!
<cfbolz> enjoy your dinner
<mattip> arch-chroot seems like a nice script, it handles all the mount/umount needed
<mattip> this site claims archlinux is #1, ubuntu #6. Maybe time to take a look?
<kenaan> cfbolz py3.6 5ea8098b2f2c /pypy/interpreter/astcompiler/: construct huge dicts in constant stack space as well
<mattip> cfbolz: py3.6 currently does not translate
<cfbolz> thanks, will fix
danieljabailey has quit [Ping timeout: 245 seconds]
danieljabailey has joined #pypy
<kenaan> cfbolz py3.6 43bb1c622881 /pypy/interpreter/astcompiler/optimize.py: fix translation
<kenaan> cfbolz default 3929440c517f /rpython/jit/metainterp/test/test_virtualizable.py: the test that I think we should figure out how to pass
demonimin has quit [Ping timeout: 246 seconds]
demonimin has joined #pypy
demonimin has joined #pypy
jacob22_ has joined #pypy
forgottenone has quit [Ping timeout: 246 seconds]
jacob22 has quit [Ping timeout: 245 seconds]
rindolf has joined #pypy
<rindolf> Hi all! In the production phase of https://github.com/shlomif/euler662-pypy3-vs-cpython3 pypy3 is slower than cpython3 - why?
dante has quit [Ping timeout: 246 seconds]
<mattip> rindolf: what do you mean "production phase"? How do you run that code, what does it do, how do you know it is slower?
dante has joined #pypy
<rindolf> mattip: i do some tests first
<rindolf> mattip: the prints are slower
darkman66 has joined #pypy
<cfbolz> rindolf: you are using a C extension module, right?
<cfbolz> those are often a bit slower in pypy
<mattip> (it seems to be cffi?)
<cfbolz> ah, weird, it's both?
<rindolf> cfbolz: yes
<cfbolz> advanced
<rindolf> cfbolz: i can remove the cext stuff
<cfbolz> was just confused
<cfbolz> anyway, numpy is a C extension, so that is often slower than cpython as well
darkman6_ has quit [Ping timeout: 245 seconds]
<mattip> if you were using pypy2 you could sub out np for import _numpypy.multiarray as np
<rindolf> cfbolz: ah
<rindolf> mattip: AttributeError: 'module' object has no attribute 'uint32'
<mattip> np.zeros((10, 10), dtype='uint32') works, but for np.ones you have to do "np.zeros((10, 10), dtype='uint32') + 1"
<rindolf> mattip: ah
<mattip> it is a bit of a rewrite, but using a=ffi.new('uint64[%d]' % (h*w)) and indexing as a 1d vector a[h*y + x] instead of numpy
<mattip> would work on pypy3
<mattip> although no nice `astype` conversions
<mattip> which would save a copy anyway
<rindolf> mattip: pushed a branch - https://github.com/shlomif/euler662-pypy3-vs-cpython3
marky1991 has joined #pypy
<mattip> I made a PR with some tweaks, pypy2 is 2.5 times faster than cpython for 662_v1.py --test
<mattip> rindolf: ^^^
<mattip> printing less frequently might help
<mattip> nope, no big change when printing less
<rindolf> mattip: thanks, let me see
demonimin has quit [Ping timeout: 246 seconds]
demonimin has joined #pypy
demonimin has joined #pypy
forgottenone has joined #pypy
<rindolf> mattip: merged
<kenaan> cfbolz default a8deefd689d8 /pypy/objspace/std/celldict.py: kill identical helper function
alawrence has joined #pypy
alawrence has quit [Ping timeout: 256 seconds]
Rhy0lite has joined #pypy
<tos9> mattip: I gave up yesterday, but going to try again today
<tos9> First by trying it on this completely other mac machine
<mattip> ok. I think you need to set 'MACOSX_DEPLOYMENT_TARGET' to 10.14, like cpython does
<tos9> Yeah I saw that last message -- didn't work with that either unfortunately
<mattip> ahh. do you remember what the error was?
<tos9> mattip: the multiarray one
demonimin has quit [Quit: bye]
<tos9> i.e. the install completed, but then at runtime complained about missing symbols
<mattip> using clang was going to solve that :(
adamholmberg has joined #pypy
antocuni has joined #pypy
danieljabailey has quit [Ping timeout: 268 seconds]
<petronny> cfbolz: I'm back. `p ARRAY` gives `<GcArray of utf8_loc_elem { baseindex: Signed, ofs: FixedSizeArray of 16 Char } >`
<petronny> And `p FIELD` gives `<FixedSizeArray of 16 Char >`
dddddd has joined #pypy
<cfbolz> petronny: great! That helps me, actually. Will try to think of how to fix this
<petronny> waiting for you news then.
<cfbolz> petronny: can you use up until you find a variable called 'op'?
<petronny> `storage_0 = direct_call((<* fn create_utf8_inde...torage>), v3432, v3433)`
<petronny> `File "/build/pypy/src/pypy2.7-v7.1.0-src/rpython/jit/codewriter/call.py", line 317, in getcalldescr`
danieljabailey has joined #pypy
<tos9> mattip: whee so of course on this other mac everything works out of the box
<tos9> no setting MACOSX_DEPLOYMENT_TARGET, no modifying sysconfig_pypy
<tos9> numpy.test is plowing on even
<tos9> bunch of failures, but it runs
<mattip> yes, failures are expected, I need to push fixes to numpy now that 7.1 is out
* tos9 nods
<tos9> ok, I'm going to have to investigate when I get to worko
<tos9> but yeah I did nothing on this machine, which is reasonably similar, in that it's running same version of macos, and it's using my dotfiles, where I pin all the packages I install, so in theory it should be really quite similar to the other one :/
<tos9> but will see what's up
<cfbolz> petronny: thanks
<cfbolz> I am away from my laptop now but will look into it later
* petronny ok
jacob22_ has quit [Ping timeout: 268 seconds]
oberstet has quit [Remote host closed the connection]
<cfbolz> Henri ist cranky
<cfbolz> Hab Sorge dass Ärztin nicht gut geht
<cfbolz> Argh
<cfbolz> Sirry
<cfbolz> Wrong window ;-)
<arigato> :-)
Ai9zO5AP has joined #pypy
jcea has joined #pypy
adamholmberg has quit [Remote host closed the connection]
adamholmberg has joined #pypy
marky1991 has quit [Quit: Saliendo]
marky1991 has joined #pypy
jacob22 has joined #pypy
jacob22 has quit [Client Quit]
andi- has quit [Ping timeout: 258 seconds]
jacob22 has joined #pypy
forgottenone has quit [Ping timeout: 246 seconds]
andi- has joined #pypy
andi- has quit [Excess Flood]
jacob22 has quit [Read error: Connection reset by peer]
jacob22 has joined #pypy
adamholmberg has quit [Remote host closed the connection]
darkman66 has quit [Remote host closed the connection]
<cfbolz> arigato: ping?
<arigato> pong
<arigato> cfbolz: ^
<cfbolz> ah, sorry
<cfbolz> arigato: do you have the slightest clue what is going on with petronny's problem?
<cfbolz> the effectinfo crashes because it sees a utf8 index storage
<cfbolz> but I have no clue why, it uses an interior field, but the JIT should just ignore it
darkman66 has joined #pypy
<arigato> looking, but so far I agree with you
<arigato> do we know if it's pypy2 or pypy3 and at which revision?
<cfbolz> pypy3.6-v7.1.0
<arigato> are you sure? the line number (pypy.module._sre.interp_sre:109)W_SRE_Pattern.make_ctx is correct in default but not in release-pypy3
<cfbolz> hmmmm, no it's just their directory name in the paste
<arigato> right
darkman66 has quit [Remote host closed the connection]
darkman66 has joined #pypy
<arigato> no, we're confused
<arigato> petronny also said " `File "/build/pypy/src/pypy2.7-v7.1.0-src/rpython/jit/codewriter/call.py` above
<cfbolz> hm, but then the two don't agree?
<arigato> yes
jcea has quit [Ping timeout: 264 seconds]
<arigato> maybe it crashes similarly in both
<arigato> looking in the pypy2 branch, it makes no sense because rutf8.create_utf8_index_storage() is only called from a function that is itself only called with jit.conditional_call_elidable
<arigato> same in the py3.6 branch
<cfbolz> yes, but note that we aren't trying to make a jitcode for the function
<cfbolz> it's only about making an effectinfo for it
<arigato> ah
<cfbolz> for a jit-opaque call
<arigato> no, I think not
<arigato> wait a sec, it's getcalldescr(op) and the op is direct_call(create_utf8_index_storage)
andi- has joined #pypy
<arigato> I think this means the op is in a function being jitted
<cfbolz> hmmmm
<cfbolz> that feels impossible to me
<cfbolz> arigato: maybe the codewriter is looking into the "not we_are_jitted()" path?
<cfbolz> shouldn't, I suppose
<arigato> or there is too much inlining before the codewriter pass?
<arigato> ah but it shouldn't see that anyway, unless it looks into the "not we_are_jitted()" path
<arigato> maybe it handles we_are_jitted() correctly but not we_are_translated()?
<cfbolz> ouch
<cfbolz> possible
<cfbolz> arigato: will try to write a unit test
<arigato> no, the "not we_are_translated" is supposed to disappear completely at flowgraph-generation-time
<arigato> I think?
<arigato> yes
<cfbolz> ok, but the unit test fails
<cfbolz> committing
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4516 [Andy Lawrence: force build, winoverlapped]
<arigato> ah, maybe it makes jitcode for the complete function even if there is a we_are_jitted()
<arigato> only with the branch removed
<arigato> and maybe the dead parts are just dead afterwards
* arigato away again
<cfbolz> arigato: pushed the failing test, I am out of time as well. complete mystery to me why it doesn't fail everywhere
<cfbolz> ah, nonsense
darkman66 has quit [Remote host closed the connection]
adamholmberg has joined #pypy
<cfbolz> false alarm, if I follow what sre does more closely, it doesn't fail
jcea has joined #pypy
<kenaan> cfbolz default 439f4942e2bd /rpython/jit/metainterp/test/test_ajit.py: it passes like this
adamholmberg has quit [Remote host closed the connection]
adamholmberg has joined #pypy
adamholmberg has quit [Ping timeout: 245 seconds]
mvantellingen has quit [Quit: ZNC 1.6.3+deb1 - http://znc.in]
mvantellingen has joined #pypy
darkman66 has joined #pypy
andi- has quit [Ping timeout: 255 seconds]
adamholmberg has joined #pypy
xcm has quit [Remote host closed the connection]
adamholmberg has quit [Remote host closed the connection]
adamholmberg has joined #pypy
xcm has joined #pypy
andi- has joined #pypy
adamholmberg has quit [Ping timeout: 272 seconds]
adamholmberg has joined #pypy
adamholmberg has quit [Read error: Connection reset by peer]
antocuni has quit [Ping timeout: 244 seconds]
adamholmberg has joined #pypy
darkman66 has quit [Remote host closed the connection]
adamholmberg has quit [Remote host closed the connection]
darkman66 has joined #pypy
darkman66 has quit [Remote host closed the connection]
darkman66 has joined #pypy
Garen has quit [Read error: Connection reset by peer]
Garen has joined #pypy
darkman66 has quit [Remote host closed the connection]
Rhy0lite has quit [Quit: Leaving]
speeder39_ has joined #pypy
darkman66 has joined #pypy
marky1991 has quit [Read error: Connection reset by peer]
darkman66 has quit [Remote host closed the connection]
marky1991 has joined #pypy
darkman66 has joined #pypy
darkman66 has quit [Remote host closed the connection]
darkman66 has joined #pypy
marky1991 has quit [Read error: Connection reset by peer]
darkman66 has quit [Remote host closed the connection]
darkman66 has joined #pypy
Ai9zO5AP has quit [Quit: WeeChat 2.4]
alawrence has joined #pypy
darkman66 has quit [Remote host closed the connection]
darkman66 has joined #pypy
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
darkman66 has quit [Remote host closed the connection]
antocuni has joined #pypy
moei has quit [Quit: Leaving...]
jacob22 has quit [Read error: Connection reset by peer]
jacob22 has joined #pypy
<kenaan> andrewjlawrence winoverlapped c91d95fe08ff /lib_pypy/_overlapped.py: Correct GetLastError module.
<kenaan> andrewjlawrence winoverlapped 6c5e88b3724e /: Merged py3.6 changes
rindolf has quit [Remote host closed the connection]
alawrence has quit [Ping timeout: 256 seconds]
tsaka_ has quit [Remote host closed the connection]
jcea has quit [Ping timeout: 240 seconds]
darkman66 has joined #pypy
speeder39_ has quit [Quit: Connection closed for inactivity]
darkman66 has quit [Ping timeout: 250 seconds]
hermitdev has joined #pypy
jacob22 has quit [Read error: Connection reset by peer]
jacob22 has joined #pypy
darkman66 has joined #pypy