emacs.pkgs.treesit-grammars: init fake package

[DamienCassou]

Description of changes

Adding this package gives Emacs access to tree-sitter grammars.

How to test

Given that there is an emacs29 package, the changes in this PR can be tested with:

$ nix-build -E 'with import ./default.nix {}; pkgs.emacs29.pkgs.withPackages (epkgs: [(epkgs.treesit-grammars.with-grammars (grammars: [grammars.tree-sitter-bash]))])'

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandbox = true set in nix.conf? (See Nix manual)
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 23.05 Release Notes (or backporting 22.11 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
• Fits CONTRIBUTING.md.

[DamienCassou]

Result of nixpkgs-review pr 230751 run on x86_64-linux 1

[DamienCassou]

/cc @clhodapp

[clhodapp]

Cool! I keep managing to falsely convince myself that linkFarm doesn't do subfolders like this. Taking advantage of it makes this much cleaner that what I would have done.

There are definitely some benefits to putting this where you're putting it so that it fits more ergonomically with emacsWithPackges. I guess the emacs package maintainers will need to sign off on it being OK to structure things like this.

One thing I will say: if this is going to go into nixpkgs, you should probably fix it so that it uses the correct shared library extension when building for Darwin. My code didn't do that because I was lazy and all of my machines are Linux at present.

[DamienCassou]

Result of nixpkgs-review pr 230751 run on x86_64-linux 1

[DamienCassou]

/cc @adisbladis @jian-lin I would appreciate your feedback on this one. Thank you.

[DamienCassou]

@clhodapp I took your feedback into account. Thank you very much.

[DamienCassou]

If I don't get more feedback, I will soon merge this PR.

[wentasah]

This works well for me. Thanks!

I've noticed that in addition to libtree-sitter-*.so the resulting derivation contains also:

libmarkdown_inline.so -> /nix/store/pbyd7m44l79jbg9q0mjjnjg7vcp78svi-markdown_inline-grammar-0.20.8/parser
liborg.so -> /nix/store/8wsmj257ifn6n5vj0ir405pphi12bdpa-org-grammar-0.20.8/parser

However, I don't think it can cause big problems other than not being able to load those grammars. And in case of org, emacs users will probably not want to use it.

Also, when one wants to use this with emacsUnstable (or maybe also emacsGit) from emacs-overlay, it is necessary to override the passtthru attribute and set treeSitter to true. Without that, treesit-extra-load-path variable is not set by the emacs wrapper. See this commit in my config.

[DamienCassou]

Thank you very much for the feedback @wentasah.