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
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
lritter has quit [Quit: Leaving]
Rhy0lite has quit [Ping timeout: 246 seconds]
Rhy0lite has joined #pypy
Rhy0lite has quit [Quit: This computer has gone to sleep]
speeder39_ has quit [Quit: Connection closed for inactivity]
xcm has quit [Read error: Connection reset by peer]
xcm has joined #pypy
dddddd has quit [Remote host closed the connection]
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
rindolf has joined #pypy
rindolf has quit [Remote host closed the connection]
<cfbolz> Cool, it seems Firefox is deduplicating strings when copying them out of the nursery
<mattip> it seems they use a reverse debugger extensively
Graypup_ has quit [Quit: ZNC 1.6.1 - http://znc.in]
Graypup_ has joined #pypy
Graypup_ has quit [Remote host closed the connection]
Graypup_ has joined #pypy
<cfbolz> mattip: I think they started rr even
oberstet has joined #pypy
dddddd has joined #pypy
rguillebert has joined #pypy
oberstet_ has joined #pypy
oberstet has quit [Ping timeout: 272 seconds]
oberstet__ has joined #pypy
oberstet_ has quit [Ping timeout: 264 seconds]
<mattip> kiwisolver https://github.com/nucleic/kiwi is a prerequisite for matplotlib, and installing it stumps some users
<mattip> I opened an issue to get PyPy wheels built https://github.com/MacPython/kiwisolver-wheels/issues/9
<mattip> but then go to thinking - why use c++ for PyPy?
<mattip> there is a javascript port https://github.com/IjzerenHein/kiwi.js/ with benchmarks
<mattip> I wonder how hard it would be to convert that to pure python, and pit PyPy's JIT against typescript's JIT
<cfbolz> mattip: that might be a question for plebas
<cfbolz> there is probably a pure python cassowary somewhere
<mattip> yup, https://github.com/brodderickrodriguez/cassowary which says in the readme
<mattip> This project was transferred to Brodderick Rodriguez in February 2020 who will oversee its development from here on out.
<mattip> there was an open issue, I added a comment
<mattip> at least kiwisolver could use cppyy...
BPL has joined #pypy
<mattip> whoops. Looking into mmap failures on win32 and the comment on this numpy issue
<mattip> it seems we never implemented "exports" in rpython.rlib.rmmap, which was part of CPython 2.7
<mattip> so we don't count the number of re-uses of mmap, and mmap.close() succeeds even when on cpython it fails
otisolsen70 has joined #pypy
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7053 [mattip: force build, mmap-exports]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/5354 [mattip: force build, mmap-exports]
rguillebert has quit [Quit: Connection closed for inactivity]
otisolsen70 has quit [Quit: Leaving]
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7053 [mattip: force build, mmap-exports]
YannickJadoul has joined #pypy
Rhy0lite has joined #pypy
<mjacob> mattip: after becoming used to rr, the only reason for me using plain gdb is if rr crashes
BPL has quit [Read error: Connection reset by peer]
<mattip> mjacob: isn't it much slower since it needs to record everything?
<mjacob> mattip: not noticable for the stuff i'm doing
<mjacob> mattip: i think it's around 3x times slower for both recording and replaying, but that quickly pays off
<mattip> thanks
<mjacob> the last time i remember when i found a reverse debugger too slow was when i debugged pyinteractive.py on top of PyPy-RevDB
<mjacob> but that was probably mostly the fault of pyinteractive.py
jcea has joined #pypy
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/5354 [mattip: force build, mmap-exports]
todda7 has joined #pypy
speeder39_ has joined #pypy
BPL has joined #pypy
todda7 has quit [Ping timeout: 272 seconds]
jcea has quit [Quit: jcea]
todda7 has joined #pypy
jcea has joined #pypy
lritter has joined #pypy
todda7 has quit [Ping timeout: 272 seconds]
gsnedders has quit [Ping timeout: 246 seconds]
gsnedders has joined #pypy
speeder39_ has quit [Quit: Connection closed for inactivity]
speeder39_ has joined #pypy
YannickJadoul has quit [Remote host closed the connection]
YannickJadoul has joined #pypy
gsnedders has quit [Ping timeout: 256 seconds]
gsnedders has joined #pypy
YannickJadoul has quit [Quit: Leaving]
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
otisolsen70 has joined #pypy
otisolsen70 has quit [Quit: Leaving]
<mattip> is there a reason we do not expose f_back in the PyFrameObject ? I realize the frame stack may be very different from CPython's,
<mattip> but having it missing causes SIP (needed for PyQT, needed for matplotlib on conda) to fail to build
otisolsen70 has joined #pypy
oberstet__ has quit [Remote host closed the connection]
<mattip> matplotlib on conda split into two packages, the base one works with PyPy (no PyQT)
<cfbolz> mattip: I suspect that it's simply that nobody wanted to think about the ramifications of forcing virtualized frames in the middle of cpyext
<mattip> well, we have averted the problem for now, as matplotlib-base does not need PyQT so we do not need SIP (yet)
<mattip> I think the reasonable solution is to remove frame probing from SIP on PyPy, but that will mean talking to riverbankcomputing
<cfbolz> mattip: I have the vague memory that christian tismer works on that nowadays
<cfbolz> he's one of the pypy founders
<mattip> right, he was doing alot of windows stuff
<mattip> so far I discovered that PyQT wants PyType_GetFlag and PyType_GetSlot, both of which were part of the python 3.4 C-API but we never noticed
<mattip> .. and PyCFunction_Call. With those 3 PyQt builds
BPL has quit [Quit: Leaving]
DanielHolth has joined #pypy
ofir has joined #pypy
<ofir> Hi, I have implemented C python bindings for my libfoo.so via cffi and I'm considering to use cppyy as means to reducing the boilerplate associated with C APIs.
<ofir> I would like to know what are the arguments against using cppyy
speeder39_ has quit [Quit: Connection closed for inactivity]
<ofir> i.e. are there gotcha's / common pitfalls that would justify sticking to a "simple" C API (and ABI)?
<bbot2> Started: http://buildbot.pypy.org/builders/own-win-x86-32/builds/2363 [mattip: force build, winconsoleio]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7054 [mattip: force build, winconsoleio]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/5355 [mattip: force build, winconsoleio]
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7054 [mattip: force build, winconsoleio]
catern has quit [Excess Flood]
epony has quit [Remote host closed the connection]
catern has joined #pypy
DanielHolth has quit [Quit: This computer has gone to sleep]
epony has joined #pypy
DanielHolth has joined #pypy
DanielHolth has quit [Client Quit]
DanielHolth has joined #pypy
DanielHolth has quit [Client Quit]
Ninpo has quit [Ping timeout: 258 seconds]
Ninpo has joined #pypy
ofir has quit [Remote host closed the connection]
ofir has joined #pypy
ofir is now known as Guest35247