As a consequence of restrictions imposed by RFC 140 - Simple Package Paths [1]
-, files related to a package should be confined on the package directory.
Certainly this restriction does not apply to packages outside by-name hierarchy.
Nonetheless, this is an interesting organization heuristics: things that affect
Emacs should be confined inside Emacs directory. Besides a future migration, the
"debuggability" of a framework is way more enhanced when we know how to find all
its files.
A similar task was done before, when RFC 140 was not a thging yet - namely, the
migration of emacs-modes to elisp-packages [2].
[1] https://github.com/NixOS/rfcs/pull/140
[2] https://github.com/NixOS/nixpkgs/pull/123859
ada-mode includes extra binaries which are used by it at runtime. They
require gnat, gprbuild, gnatcoll-core and wisi's source checked out in
`../wisi-*`.
Telega uploads packages that are incompatible with stable tdlib
releases to melpa and ones that are compatible to melpa stable.
This makes the melpa packages very unreliable and we should prefer the
one from melpa stable.
With these changes it's possible to override pkgs & lib by:
```
emacs.pkgs.override { pkgs = fooPkgs; }
```
and get them to coherently apply to all emacs packages.
This makes it much easier to create customisations around emacs via
the a new convenience passthru attr:
- `emacs.pkgs`: What used to be emacsPackages is now `emacs.pkgs`
The previous versioned names `emacs*Packages` have been moved to
aliases.nix and are now considered deprecated in favour of `emacs*.pkgs`.
This removes the unnecessary compiler build dependency. We also set
preferLocalBuild = true;
allowSubstitutes = false;
to not farm out the build on a remote builder or bother with trying to
find a binary substitution.
This approach has several differences with emacs2nix:
- the updater uses a downloaded recipes.json and archive.json for commit information, it uses a local checkout only for hashing the recipes
- the generated file is JSON
- the updater is written in emacs lisp
- prefetch errors are put into an error key in the JSON, for review + meta.broken attributes are generated from it
The updater re-uses the existing generated file to memoize prefetched content-sha256s for commits, thus prefetching should normally be quite fast.
