<mattip>
if you use that, you might want to look at rpython/tool/udir.py, which sets the directory the artifacts are created in,
<mattip>
PYPY_USESSION_DIR sets the top-level directory (by default /tmp)
<tumbleweed>
mattip: so what is the ARM64 box you've been playing on? AWS A1?
<mattip>
debian rule uses a different trick - setting TMPDIR - which does the same thing
<mattip>
tumbleweed: yup
<mattip>
someone at amazon gave me some credits to try it out
<mattip>
it translated in 6200 seconds.
<petronny>
Is parallel translation possible?
<mattip>
no
* tumbleweed
will try his chroot setup on an A1 and see if it's reproduceable
<LarstiQ>
mattip: re docker removing the build attempt, was that because you passed --rm to `docker run`?
<mattip>
tumbleweed: I am trying again with a armhf base image instead of arm32v7
<mattip>
LarstiQ: I am a docker newbie. I have a Dockerfile but it finishes with hg clone
<mattip>
then I "docker run -it <image> bash" and build inside it.
<LarstiQ>
mattip: that should stick around then after you quit
<LarstiQ>
mattip: with `docker ps -a` you can see also the non-running containers, and with `docker start <name of container>` get back to a suspended one
forgottenone has joined #pypy
<mattip>
ahh, so I am not re-entering the container, rather starting a new one each time I call "docker run -it <image>" ?
<LarstiQ>
mattip: yeah, `docker run <image>` will instantiate a new container based on the specified image, you can run multiple in parallel
<LarstiQ>
if you pass --name to run it will use the name you wish, otherwise it gives something random like `sad_lamarr`
<LarstiQ>
mattip: the distinction between containers (docker ps -a) and images (docker images) is a crucial one, and unfortunately a subtle one to beginners
darkman66 has joined #pypy
darkman66 has quit [Remote host closed the connection]
<kenaan>
arigo cffi/cffi[py3.8-interp-dict] 25a0189f4638 /: close branch, ready for merge (python 3.8a3 was released)
<kenaan>
arigo cffi/cffi 5aeca29219db /c/call_python.c: merge py3.8-interp-dict: use the official way to do that
marky1991 has quit [Ping timeout: 245 seconds]
leshaste has quit [Ping timeout: 255 seconds]
xcm has quit [Remote host closed the connection]
<tumbleweed>
mattip: schroot, setup by mk-sbuild, and (in this case) https://github.com/stefanor/ansible-sbuild (although this doesn't know to set the linux32 personality here)
<tumbleweed>
mattip: you want access to an instance like this, before I terminate it?
xcm has joined #pypy
<kenaan>
arigo default 67af7a9e5ca2 /: update to cffi/5aeca29219db
leshaste has joined #pypy
<mattip>
tumbleweed: no, thanks. It has been most helpful.
<tumbleweed>
there's a good chance that lxc/lxd gets all of this stuff right, out of the box, too
<kenaan>
arigo cffi/cffi d5c083291044 /c/: Shut down a warning from recent CPython versions: int() should always return an int, not a bool
darkman6_ has joined #pypy
darkman66 has quit [Ping timeout: 255 seconds]
<kenaan>
arigo cffi/cffi dbc31afe99d0 /c/test_c.py: improve the test on py2.7
<kenaan>
arigo cffi/cffi e43fdc644918 /c/test_c.py: Two more cases
<kenaan>
arigo default afdbea198bb5 /pypy/module/_cffi_backend/: update to cffi/e43fdc644918
leshaste has quit [Quit: Leaving]
fling has quit [Ping timeout: 246 seconds]
fling has joined #pypy
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
darkman6_ has quit [Remote host closed the connection]
darkman66 has joined #pypy
xcm has quit [Remote host closed the connection]
marky1991 has joined #pypy
xcm has joined #pypy
darkman66 has quit [Ping timeout: 246 seconds]
marky1991 has quit [Ping timeout: 250 seconds]
tsaka__ has joined #pypy
Garen_ has joined #pypy
marky1991 has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
Garen has quit [Ping timeout: 255 seconds]
antocuni has quit [Ping timeout: 244 seconds]
<tos9>
So more on the numpy saga...
<mattip>
?
<tos9>
mattip: jupyter notebook I think doesn't run on pypy
<ronan>
tos9: jupyter notebook worked last time I tried
<tos9>
The gc module though is supposed to be interpreter-specific right? So seemingly the right fix for that is for prometheus to not assume anything there?
<tos9>
ronan: Yeah I could have sworn I've done this before too --
<tos9>
The prometheus-looking stuff looks new though I think
mjacob has quit [Ping timeout: 252 seconds]
<tos9>
(And it's interesting to see jupyter expects to interact with prometheus at all...)