fixed here: https://github.com/NixOS/nixpkgs/pull/297968
Also, hydra-checks pass on all four platforms:
```sh
shivaraj in 🌐 nixos in ~
❯ nix run github:nix-community/hydra-check -- grafana --arch aarch64-darwin
Build Status for grafana.aarch64-darwin on jobset nixpkgs/trunk
✔ grafana-10.4.1 from 2024-03-27 - https://hydra.nixos.org/build/254421303
shivaraj in 🌐 nixos in ~ took 45s
❯ nix run github:nix-community/hydra-check -- grafana --arch x86_64-darwin
Build Status for grafana.x86_64-darwin on jobset nixpkgs/trunk
✔ grafana-10.4.1 from 2024-03-27 - https://hydra.nixos.org/build/254422747
shivaraj in 🌐 nixos in ~
❯ nix run github:nix-community/hydra-check -- grafana --arch x86_64-linux
Build Status for nixpkgs.grafana.x86_64-linux on jobset nixos/trunk-combined
✔ grafana-10.4.0 from 2024-03-27 - https://hydra.nixos.org/build/254385494
shivaraj in 🌐 nixos in ~
❯ nix run github:nix-community/hydra-check -- grafana --arch aarch64-linux
Build Status for nixpkgs.grafana.aarch64-linux on jobset nixos/trunk-combined
✔ grafana-10.4.0 from 2024-03-27 - https://hydra.nixos.org/build/254392946
```
* mysql service would assume that all the `*.sql` files in the directory, provided by `initialDatabases [{ schema = <directory>; … }]`, exists in a folder named [mysql-databases](291e7be83a/nix/mysql.nix (L239-L241)), after this change it would just look for all the `*.sql` files in the top-level schema directory.
* add tests
We don't need per-file package overlays, since we don't envision having too many of them. Whenever possible, get them from nixpkgs, to keep our tests nimble.
* fix: pgadmin-init
in pgadmin-8.2, the command to setup configuration db is `setup-db`
* fix: add grafana overlay
grafana 10.3.3 from upstream nixpkgs doesn’t work in darwin
* chore: free up space in ubuntu-latest
---------
Co-authored-by: shivaraj-bh <sbh69840@gmail.com>
* explicitly define project root in treefmt
* define `projectRootFile` in `treefmt`
* make `treefmt.projectRoot` as `inputs.services-flake`
* use github flake url
* flake.nix as projectRootFile
* add comment
* force override-input in test flake
---------
Co-authored-by: shivaraj-bh <sbh69840@gmail.com>
* clickhouse init script
* clickhouse schema test
* clickhouse folder
* docs: add initialDatabases to tips & tricks
* docs: use heading anchor
* add TODO to find an alternative to start clickhouse-server during init
---------
Co-authored-by: shivaraj-bh <sbh69840@gmail.com>
* Use psql instead of postgres, as a result, use `pg_ctl` to temporarily start the postgres server in `${name}-init` process.
* Avoid creating a custom `postgresPkg` with extensions by using `apply` attribute provided by `mkOption`
* Refactor setup script, use `writeShellApplication` and resolve shell-check errors
---------
Co-authored-by: shivaraj-bh <sbh69840@gmail.com>