cfbolz changed the topic of #pypy to: PyPy, the flexible snake (IRC logs: https://botbot.me/freenode/pypy/ ) | use cffi for calling C | the secret reason for us trying to get PyPy users: to test the JIT well enough that we're somewhat confident about it
<njs>
mattip: fyi, dunno if this is relevant, but numpy is really bad about actually holding on to buffers the way it ought to
<njs>
mattip: it tend sto assume that as long as the original object is kept alive, then the buffer pointer will remain valid, even if we release the buffer
forgottenone has quit [Read error: Connection reset by peer]
forgottenone has joined #pypy
<kenaan>
cfbolz pyparser-improvements-2 fefc40951b6f /pypy/interpreter/test/: fix fallout from position fixes
<kenaan>
cfbolz pyparser-improvements-2 65640d1556ff /lib-python/2.7/test/: fix CPython tests that are affected (in one, we are like CPython now again)
<kenaan>
mattip issue2752 882bbee00812 /pypy/module/cpyext/test/test_bufferobject.py: after PyObject_GetBuffer fails, a second call also fails, PyString_AsString succeeds
inhahe_ has quit [Ping timeout: 256 seconds]
<mattip>
I am tempted to declare issue 2752 "not release critical" since it has existed forever
<cfbolz>
+1
<mattip>
it seems like a failure of get_raw_address_of_string(string) in rffi.py, at some point the interface changed
<mattip>
from (obj,string) to just (string)
<mattip>
the documentation still refers to the old interface, maybe the code does too :/
<fijal>
er
<fijal>
anyone here speaks swedish?
<mattip>
the input, string, is a simple python string, not an rpython string (whatever that is, maybe a CArray?), and there is no obj to create a weak reference
* mattip
taking one last look at win32 and pypy3, to make virtualenv work
<vstinner>
I'm confused by "Donate towards general pypy progress". maybe it should just be removed. the link displays https://pypy.org/don2.html ... which is already the default view
<vstinner>
cfbolz: when i open pypy.org, the Donate button is already there. i'm just complaining then clicking on "Donate towards general pypy progress" text does nothing
<cfbolz>
vstinner: yes, it used to be the button to choose which project to donate to. Now there is only 'general progress'
<vstinner>
cfbolz: i see :) i was just surprised than clicking does nothing :)
<cfbolz>
vstinner: yes, I see how it's confusing, trying to fix
<vstinner>
cfbolz: ok, thanks ;)
illume has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<cfbolz>
vstinner: thank you, it's good feedack
<kenaan>
cfbolz pypy.org[extradoc] 272637cfd909 /: - remove flattr link - make "Donate towards general pypy progress" not a link (since it does nothing wh...
<cfbolz>
vstinner: I always forget how pypy.org gets updated, but hopefully soon
<vstinner>
cfbolz: ok, thanks for the fix ;)
mcyprian has quit [Remote host closed the connection]
vstinner has left #pypy [#pypy]
jaffachief has joined #pypy
<tos9>
I have a .h file that is dynamically generated that I'll need in CFFI
<tos9>
Right now I'm building it as part of build_py, but that's too late for it to be used from within cffi_modules it looks like, which runs like way back during setup requires step
<tos9>
Where should that generation be happening / what command should I overwrite to make sure it's there by the time CFFI is going to build a module for me
<tos9>
Actually now maybe it occurs to me to do the .h generation in the cffi build.py...
<antocuni>
tos9: that's exactly what I was about to suggest
<tos9>
yeah it kind of makes sense I guess
<antocuni>
you could even put it in some tmp directory, to avoid cluttering the checkout