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
<arigato>
bonus points for anyone who has got any idea where the RuntimeError is raised
ajlawrence has joined #pypy
<mattip>
the offending line is self._ext_to_normal(_getfinalpathname(s))
<arigato>
sorry, yes, _getfinalpathname('.') raises RuntimeError for me too
<mattip>
putting a pdb just before it, the _getfinalpathname(s)
<mattip>
ahh
<arigato>
I have no clue at all where inside the source of pypy it comes from
<mattip>
I see it in pypy/module/posix/interp_nt.py
<arigato>
I see no "RuntimeError" in this file
<arigato>
ah wait
<mattip>
what happens when it raises a rwin32.lastSavedWindowsError without turning it into a applevel exception?
<arigato>
aaaah
<arigato>
>>> NotImplementedError.__base__
<arigato>
<class 'RuntimeError'>
<arigato>
and this file declares a class LLNotImplemented
<arigato>
bad idea
<arigato>
so that's where it comes from
<arigato>
I think it's trying to catch it in interp_posix
<arigato>
I'm not surprized it doesn't work in RPython
kenaan has joined #pypy
<kenaan>
arigo py3.6 618365170725 /pypy/module/posix/interp_nt.py: Don't make a subclass of NotImplementedError to avoid using NotImplementedError---it's unlikely to work correctly in ...
<mattip>
if that is true, then check_GetFinalPathNameByHandle is failing, which means GetFinalPathNameByHandleW is not found
<kenaan>
arigo default 2dceead28db5 /rpython/flowspace/: Don't make an RPython subclass of AssertionError or NotImplementedError for the purpose of "catching" it: it would ...
<mattip>
there is a test in posix/test/test_nt.py, but that file seems skipped on windows py3.6 builds in the "app-level (-A) tests", and we don't run own tests
<mattip>
the test passes untranslated
<arigato>
sorry, I'm out of ideas and I don't want to figure out how to translate and debug the source code on windows
<arigato>
the GetModuleHandle/dlsym dance seems to work, because it works in interp_time.py
<mattip>
ok, maaybe if I have some time I will add some print statements and translate
<arigato>
great debugging technique, but yes, thank you
antocuni has joined #pypy
dash_ has joined #pypy
jcea has joined #pypy
dash_ has quit [Quit: leaving]
mdash has joined #pypy
tsaka__ has joined #pypy
Rhy0lite has joined #pypy
xorAxAx has joined #pypy
xorAxAx has quit [Changing host]
xorAxAx has joined #pypy
<xorAxAx>
.
<kenaan>
rlamy py3.7 66823f138171 /pypy/interpreter/app_main.py: Keep app_main sysflags in sync with sys.flags
<simpson>
Turns out that the 99% of Python folks wishing for Python 2 to die are kind of silly. They can't tell the difference between Python and CPython.
inhahe_ has quit []
andi- has joined #pypy
<Dejan>
that is absolutely fine with me, as long as I can grab 3.x builds :)
<Dejan>
question now is - where can i grab a 2.7 aarch64 build?
<Dejan>
using CPython for translation is superslow
<Dejan>
is it possible to run translation concurrently?
<Dejan>
(to use all the cores)
<simpson>
Sadly, no.
<Dejan>
fair enough
xcm has quit [Remote host closed the connection]
jcea has quit [Remote host closed the connection]
xcm has joined #pypy
<Dejan>
i guess i could use rpython instead of python2 on my VPS
<Dejan>
mattip, are you there?
<lesshaste>
ronan, oh wow.. it really is a rolling target
<lesshaste>
ronan, has the numpy work carried on?
<Dejan>
mattip, how do you build aarch64
<Dejan>
mattip, i hope you are not using CPython 2.7 for that
<lesshaste>
it would be great if pypy 3.x builds were available for centos or rhel for example
<lesshaste>
currently I can't use it at work
<lesshaste>
and they don't support python 2
<Dejan>
i use them on fedora
<Dejan>
the x86_64 builds
<Dejan>
i did not try on CentOS, but I see no reason why they would not work there
<Dejan>
i use CentOS on my home workstation, so will check later when I go home
<ronan>
lesshaste: we're not doing much on numpy ATM, it just works
<Dejan>
ah! so the default branch is actually 2.7, right?
jcea has joined #pypy
<ronan>
Dejan: yes
<Dejan>
ok, now it is building faster on my aarch64 vps