Drop the nixpkgs-old flake input and old-pkgs binding (after tauri removal)

[thedavidmeister]

Drop the nixpkgs-old flake input and old-pkgs binding

The nixpkgs-old input pinned at 48975d7f9b9960ed33c4e8561bcce20cc0c2de5b exists solely for tauri-shell (specifically for webkitgtk and libsoup-2.4 / glib-networking / gsettings-desktop-schemas / gtk3 / glib that need to track an older nixpkgs). Every other consumer of pkgs uses the current nixpkgs.

References to old-pkgs in flake.nix:

• tauri-build-inputs (glib, gtk3, libsoup_2_4, glib-networking, optional webkitgtk)
• devShells.tauri-shell shellHook (gsettings-desktop-schemas, gtk3, glib-networking)

Plus old-pkgs is exported as a reusable output (line ~595).

Why drop

• nixpkgs-old is two and a half years stale (rev from 2023). Maintaining a parallel dep tree adds complexity for one consumer.
• The exported old-pkgs lets downstream consumers reach into the pinned-stale nixpkgs, increasing the surface area of "what depends on the old pin."
• After Remove tauri-shell devShell and tauri-build-inputs #129 lands (remove tauri-shell entirely), nothing in rainix needs old-pkgs.

Sequencing

Blocked on #129 (remove tauri-shell). Once that merges, this becomes a 4-line PR:

1. Drop inputs.nixpkgs-old.
2. Drop the nixpkgs-old, arg in outputs = { ... }.
3. Drop old-pkgs = import nixpkgs-old { inherit system; };.
4. Remove old-pkgs from the inherit ... reusable-outputs block.

Migration for downstream consumers

Anyone consuming rainix.outputs.old-pkgs directly would need to either:

• Pin their own nixpkgs-old input, or
• Switch to current nixpkgs (preferred — they should not be on a 2023 pin in 2026).

Worth grepping the rain ecosystem for rainix.old-pkgs references before merging.

Related

• Remove tauri-shell devShell and tauri-build-inputs #129 — remove tauri-shell (blocking dependency).
• feat: add sol-shell devShell, drop sol-prelude and rs-prelude #130 — sol-shell devShell, prelude removal.

[thedavidmeister]

Done in #159.