Closed turion closed 1 year ago
This seems to actually push objects to the cache.
Post job cleanup.
Cachix: push
/home/runner/work/_actions/cachix/cachix-action/v12/dist/main/push-paths.sh cachix tweag-monad-bayes
compressing using zstd and pushing /nix/store/4w4b4mxfhdc9lf3rsf4qqa9z8xnijbxl-source (26.27 KiB)
compressing using zstd and pushing /nix/store/7rvyhm99h13ddfwblbixhr7iy9mzr4zj-cabal-fmt-0.1.6 (6.79 MiB)
compressing using zstd and pushing /nix/store/5lzm0zkydfl9bs19zs0lbr7y0qsxh613-source (8.03 KiB)
compressing using zstd and pushing /nix/store/9p17l6hmyz32hjma72nw6dfqxfnqk8x8-source (53.46 KiB)
compressing using zstd and pushing /nix/store/gsslxw2mhiygis23y7nrizbpjl1pgbaq-source (162.64 KiB)
compressing using zstd and pushing /nix/store/kz15p1pxm5dw3hnay3bnqhz4mvsg8jyl-source (34.86 MiB)
compressing using zstd and pushing /nix/store/plcgf3d938380z6qaim2fa22gna3c12l-pre-commit-run (112.00 B)
compressing using zstd and pushing /nix/store/fxf5927vw1wl2ap0lhr72fx3jbj1wmwj-source (127.23 KiB)
compressing using zstd and pushing /nix/store/qwk3nqcb5k3y51k1kbrqamgcp83544zl-pre-commit-config.json (1.60 KiB)
compressing using zstd and pushing /nix/store/ymdgrjb15pg6w0dhwx006kyzngrmnv4y-cabal-fmt-0.1.6-bin (3.94 MiB)
All done.
And it can read as well:
Run nix --print-build-logs build .#pre-commit --accept-flake-config
nix --print-build-logs build .#pre-commit --accept-flake-config
shell: /usr/bin/bash -e {0}
env:
NIX_PATH: nixpkgs=channel:nixos-unstable
copying path '/nix/store/4w4b4mxfhdc9lf3rsf4qqa9z8xnijbxl-source' from 'https://tweag-monad-bayes.cachix.org/'...
copying path '/nix/store/fxf5927vw1wl2ap0lhr72fx3jbj1wmwj-source' from 'https://tweag-monad-bayes.cachix.org/'...
copying path '/nix/store/zjpd6dn9kza8zl3l17cwgyifsxbvi3nb-source' from 'https://cache.nixos.org/'...
Now the question is whether the common nix setup already worked in the first place. And if yes, if --accept-flake-config
is needed at all. And if no, why we never saw a benefit of caching before. Maybe we simply never populated the cache. To do so, we could for example run master
after every PR when it's merged.
Now the question is whether the common nix setup already worked in the first place
Yes it does, see https://github.com/tweag/monad-bayes/actions/runs/5004889063/jobs/8968073752?pr=277.
And if yes, if
--accept-flake-config
is needed at all.
warning: ignoring untrusted flake configuration setting 'extra-substituters'.
Pass '--accept-flake-config' to trust it
warning: ignoring untrusted flake configuration setting 'extra-trusted-public-keys'.
Pass '--accept-flake-config' to trust it
copying path '/nix/store/4w4b4mxfhdc9lf3rsf4qqa9z8xnijbxl-source' from 'https://tweag-monad-bayes.cachix.org/'...
copying path '/nix/store/fxf5927vw1wl2ap0lhr72fx3jbj1wmwj-source' from 'https://tweag-monad-bayes.cachix.org/'...
So no, it isn't.
And if no, why we never saw a benefit of caching before. Maybe we simply never populated the cache.
Yes, I forgot that the cachix action part was commented all the time.
@idontgetoutmuch @reubenharry can you review?
I'll defer to @idontgetoutmuch , since I don't know any nix.
I'll defer to @idontgetoutmuch , since I don't know any nix.
In the kingdom of the blind...
@turion can we talk for 10 minutes? It looks fine but I really don't know much about CI.
Fixes #273