mu: enable AOT native-comp for emacs lisp

[jian-lin]

Description of changes

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/)
• 22.11 Release Notes (or backporting 22.05 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
  • (Release notes changes) Ran nixos/doc/manual/md-to-db.sh to update generated release notes
• Fits CONTRIBUTING.md.

[peterhoeg]

Make sense, but any chance we can move this out into a separate hook and just call that? I’m sure there are other emacs packages that would benefit from this.

[jian-lin]

any chance we can move this out into a separate hook and just call that?

Good idea, I will wait for #194279 to be merged first.

I’m sure there are other emacs packages that would benefit from this.

Are there? To the best of my knowledge, mu is the only package. Could you list a few?

[jian-lin]

Are there?

Indeed, there are many packages like mu4e.

rg emacs -t nix --glob '!pkgs/applications/editors/emacs' --glob '
!pkgs/build-support/emacs' --glob '!nixos' --glob '!test'

I am working on it.

[peterhoeg]

Even our regular emacs packages (under emacsPackages) require nativecomp so they should simply use the same hook/helper.

[jian-lin]

Even our regular emacs packages (under emacsPackages) require nativecomp

They have infra for native comp since a7cb8e3, not a hook though.

they should simply use the same hook/helper

Yeah, that is also a good idea.

I will mark this as draft to prevent it from being merged.

[jian-lin]

I give up making a setup hook for native-comp because I cannot figure out how to put everything needed by native-comp into a hook, e.g. these lines. I will leave that for future contributors.

I make native-comp conditionally depend on emacs.nativeComp.

I suggest we merge this for now and transfer it to a hook when the hook is implemented.

[jian-lin]

Any updates?

[chvp]

I give up making a setup hook for native-comp because I cannot figure out how to put everything needed by native-comp into a hook, e.g. these lines. I will leave that for future contributors.

Is this not already solved in a general way by the current emacs infrastructure/emacs-overlay. All my emacs packages (except mu4e) are AOT-compiled already, and I can't imagine someone added AOT compilation to all those packages seperately.

[jian-lin]

Is this not already solved in a general way by the current emacs infrastructure/emacs-overlay.

No and emacs-overlay is not involved here.

All my emacs packages (except mu4e) are AOT-compiled already

Let me explain some details.

Currently, only elisp packages, which are built using elpaBuild, melpaBuild or trivialBuild, enable AOT native-comp after my pr. This is because only these three builders run native-comp in the postInstall. As a result, mu4e does not do AOT native-comp.

I can't imagine someone added AOT compilation to all those packages seperately.

Yeah, the infra of emacs needs some improvement to do it the right way. One possible way is to make native-comp a hook, as @peterhoeg says.

Unfortunately, it seems few people care about native-comp in emacs. Almost all recent native-comp related prs are all from me. As I said, I tried to make native-comp a hook, but failed. I would appreciate it if someone could improve the emacs infra. mu4e is the last piece I care for native-comp, so here is this pr.

[jian-lin]

It seems this pr does not reach the bar for maintainers to merge it and there is also nothing more I can do. Closing.

[peterhoeg]

I am truly sorry to leave you hanging like this. I absolutely think there is value in your work - I just haven't had tiem to do anything about it.

[jian-lin]

Oh, I misunderstood. Let's reopen it and give it some more time.

[LiGoldragon]

Unless emacs *absolutely requires' the elisp files, the nativeComp outputs should go in their own derivations. Lean emacs deploys. Easier to think about and modularize. Just a thought. I might implement something and will let you know if I do.

[peterhoeg]

In principle that makes sense, but is there really a usecase for emacs without the natively compiles files?

[bhankas]

Just so it doesn't get forgotten to sands of time, @peterhoeg can we do something about this? Anybody else who might be able to review if you're busy?

[peterhoeg]

My apologies. I will take the time to do what is needed during this coming week.

[LiGoldragon]

In principle that makes sense, but is there really a usecase for emacs without the natively compiles files?

I was thinking more about emacs without the elisp files

[bhankas]

ping @peterhoeg ! Apologies if I'm being annoying, but I'd quite like to see this merged. If Peter is busy, can @adisbladis take a look?

[adisbladis]

While it would indeed be better to solve in a hook or other central location it's OK to have a workaround until then.

[jian-lin]

FYI, I made a follow-up PR to package mu4e as a separate Emacs lisp package.