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
<nimaje>
seems like the absolute directory hirachy of lib_pypy/_blake2/impl gets recreated under lib_pypy/_blake2/, so I have an empty lib_pypy/_blake2/wrkdirs/usr/ports/lang/pypy3/work/pypy3.6-v7.3.0-src/lib_pypy/_blake2/impl which seems useless and no idea what would create that
<mattip>
hmm. A bit no idea, maybe some interaction with cffi and the package.py script
<mattip>
could you clean that out, do a mkdir lib_pypy/_blake2/wrkdirs, chmod to make it read-only,
<mattip>
then rerun _blake2_build.py and see who tries to write there?
epony has joined #pypy
CrazyPython has joined #pypy
<nimaje>
ok, _blake2_build.py creates the dir, but it does even when I chmod -w
jcea has joined #pypy
<nimaje>
ok, using user nobody helped a traceback https://0x0.st/iBxA.txt (it's usr here as it is a installed version)
CrazyPython has quit [Read error: Connection reset by peer]
<mattip>
it seems tmpdir maybe should be pointing somewhere else, or should be cleaned up by the script
xcm has quit [Ping timeout: 260 seconds]
<mattip>
I am on mobile, will check it deeper when possible
<jneen>
it seems to pass the rtyper, but get stuck on the `ld` step
adamholmberg has quit [Client Quit]
xcm has quit [Remote host closed the connection]
adamholmberg has joined #pypy
xcm has joined #pypy
<cfbolz>
jneen: hey!
<cfbolz>
jneen: which OS are you working on?
<jneen>
hey :]
<jneen>
arch linux, just ran a full system upgrade
<jneen>
kernel 5.5.13-arch1-1
<jneen>
i've made the possibly-slightly-crazy decision to keep the frontend currently written in ruby, and only do the bytecode interpretation in rpython
<jneen>
which means that if i want to eval or load a file i have to run an external program to partially compile it first
<jneen>
(...eventually i'll bootstrap that part away, but...)
<cfbolz>
jneen: that sounds fine
<cfbolz>
(pypy itself had a slightly weird bytecode compiler situation way back when in 2006 or so ;-) )
<jneen>
haha yeah i mean
<jneen>
it makes sense, it's easier to parse and pre-process when you can do more dynamic things than rpython wants to do
<cfbolz>
exactly
<cfbolz>
so we had the parser in full python, and had to do complicated things to freeze the bytecode of the parser into the binary. it was very slow
<cfbolz>
anyway, spawnv: I know nothing about it really, but usually it means that your flavor of linux needs an extra linker option than ubuntu
<jneen>
huh
<cfbolz>
let me dig for a few minutes
<cfbolz>
I might be talking nonsense
<jneen>
thanks a ton
<cfbolz>
jneen: hm, no, seems spawnv also doesn't exist for me
<jneen>
...huh
<cfbolz>
(not even on cpython, fwiw)
<cfbolz>
jneen: fork + exec?
Dejan has quit [Quit: Leaving]
<jneen>
does os.fork work reliably?
<cfbolz>
it should
<cfbolz>
or what do you mean by "reliably"?
dustinm has quit [Quit: Leaving]
<cfbolz>
jneen: I'll be going to bed in a few minutes, should be around more during the weekend daytime (European time)