Attendees: @regnat @infinisil @edolstra @fricklerhandwerk @tomberek @Ericson2314
Notes: @fricklerhandwerk
Agenda
- check in on fixing the 2.18 regression
- CLI stabilisation
- flakes breakdown
https://github.com/NixOS/nix/issues/9052
- @ericson2314 did not have enough time to reduce the reproducer to be minimal enough for writing a regression test
CLI stabilisation
-
nix store path-from-hash-part- postponed to second batch
-
nix store prefetch-file- @Ericson2314: this should probably its own command to better represent how Nix is compartmentalised internally
- @edolstra: the internal architecture should stay internal
-
@Ericson2314: we have RFC 134 saying we want a store-only Nix
- @regnat: doesn’t mean the full CLI can’t have these commands even if they are not directly part of the store
-
@fricklerhandwerk: would it make sense to make it a top-level command e.g.
nix fetchto do anything related to remote FSOs?- @ericson2314: possibly, for instance it could operate on lock files if we were to stabilise them separately
- @regnat: would have to use noun-verb pattern though
-
@regnat: since it’s not directly related to the store, we could set it aside for now
- @Ericson2314: agreed
- discussion of command line flags
- decision: @Ericson2314 and @fricklerhandwerk will look through the mixins to check if anything needs to be moved around or discussed in the team
- next time:
- take a look at
nix build - otherwise reserve time for reviews of PRs for stabilising the CLI
- take a look at
Flakes breakdown
Since we did not make much material progress the last two times, this time we started preparing a joint statement of what we’re intending to do about flakes. The goal is to establish clarity and agreement on our values and objectives, and communicate that appropriately. It will be published when ready. We will continue to take time each week to narrow down a realistic roadmap that takes into account our available resources, and keep working on and reviewing implementation aspects.