!LemuOOvbWqRXodtSsw:nixos.org

NixOS Reproducible Builds

541 Members
Report: https://reproducible.nixos.org Project progress: https://github.com/orgs/NixOS/projects/30122 Servers

Load older messages


SenderMessageTime
25 May 2021
@andi:kack.itandi-
In reply to @baloo_:matrix.org
andi: afaict, python was reproducible on >3.6
Well I changed the compile flags in the linked PR ;)
23:08:15
@baloo_:matrix.orgbaloopretty sweet23:09:14
@baloo_:matrix.orgbaloothat's the thing I had in mind https://www.python.org/dev/peps/pep-0552/23:09:54
@baloo_:matrix.orgbaloowhich is a different thing23:10:06
@baloo_:matrix.orgbaloo(reproducibility of the .pyc files)23:10:15
@andi:kack.itandi-Didn't those only have very limited impact on performance anyway?23:14:53
@baloo_:matrix.orgbaloohuuum23:15:59
@baloo_:matrix.orgbaloodepends on your workflow I'd say23:16:22
@andi:kack.itandi-Let's say I am running synapse :)23:16:41
@baloo_:matrix.orgbalooit affects the boot performance of python23:16:44
@baloo_:matrix.orgbaloostartup performance*23:16:53
@andi:kack.itandi-So probably only relevant for short-lived processes23:19:12
@baloo_:matrix.orgbalooyeah23:19:22
@andi:kack.itandi-That don't have a local dir to write cache files to23:19:23
@baloo_:matrix.orgbaloorunning tests23:19:28
@baloo_:matrix.orgbaloowell, iirc a process will try to write to /nix/store/blabla-python/share/foo.pyc if you open the foo.py23:20:11
@baloo_:matrix.orgbaloonot sure there is a per-user cache of those23:20:30
@andi:kack.itandi-Well the cache dir would be more like while developing on python software23:20:44
@baloo_:matrix.orgbalooif it can't store the pyc, it will work just fine, but needs to rebuild23:20:49
@baloo_:matrix.orgbalooyeah but usually you invoke libraries and stuff23:21:05
@baloo_:matrix.orgbaloo(I stopped using venv since nix)23:21:35
@andi:kack.itandi-Anyway, determinism > speed for now :-)23:22:10
@baloo_:matrix.orgbalooyeah that's nice. I'd like to debug for the gcc build23:30:54
@baloo_:matrix.orgbaloobut I need to step through the different phases and compare the intermediate results23:31:15
@baloo_:matrix.orgbaloothe non-reproducibility there does not make much sense to me23:31:52
@andi:kack.itandi-oh, GCC is still not reproducible? I thought we got that one covered.23:32:32
@baloo_:matrix.orgbalooit is now23:32:41
@baloo_:matrix.orgbaloobut it's 8-12% slower23:32:49
@andi:kack.itandi-would be worse if the build results were slower now23:33:08
@baloo_:matrix.orgbaloothe build results are the same23:33:38

Show newer messages


Back to Room ListRoom Version: 6