rustup-init: install proxy binaries

[branchvincent]

• Have you followed the guidelines for contributing?
• Have you ensured that your commits follow the commit style guide?
• Have you checked that there aren't other open pull requests for the same formula update/change?
• Have you built your formula locally with HOMEBREW_NO_INSTALL_FROM_API=1 brew install --build-from-source <formula>, where <formula> is the name of the formula you're submitting?
• Is your test running fine brew test <formula>, where <formula> is the name of the formula you're submitting?
• Does your build pass brew audit --strict <formula> (after doing HOMEBREW_NO_INSTALL_FROM_API=1 brew install --build-from-source <formula>)? If this is a new formula, does it pass brew audit --new <formula>?

---

Currently, this formula only contains a rustup-init binary (an installer) which a user must call once (and then will never need it again). It's not really useful in its current form because users can just do curl https://sh.rustup.rs | sh to achieve the same.

Instead, this actually installs all rustup-managed proxies (which are all just rustup symlinked under a different name). All users need to do is configure a default toolchain on the first run.

The only downside is conflicting with rust, which was noted in the original PR #9617 but I think we should reconsider since it gives users the choice (and most users won't notice now that we have bottles and rust is only a build dep). It is also the same approach other packagers seem to have taken: arch, debian, nix

[branchvincent]

Ah but we need rust to build rustup. Let me rethink this

[branchvincent]

Ok moved proxy binaries to libexec to avoid conflicts

[carlocab]

The only downside is conflicting with rust, which was noted in the original PR #9617 but I think we should reconsider since it gives users the choice (and most users won't notice now that we have bottles and rust is only a build dep). It is also the same approach other packagers seem to have taken: arch, debian, nix

We could probably make one of them keg-only to fix the conflict.

[branchvincent]

I do like that approach more since users could also brew link --force. Looking at analytics:

rust:

install: 38,827 (30 days), 113,069 (90 days), 440,514 (365 days)
install-on-request: 28,106 (30 days), 81,005 (90 days), 333,059 (365 days)

rustup-init:

install: 3,643 (30 days), 16,870 (90 days), 51,411 (365 days)
install-on-request: 3,393 (30 days), 15,613 (90 days), 47,587 (365 days)

So I say we make rustup-init keg-only, rename it to rustup (note we do have an existing alias) and stop installing the rustup-init binary (since it no longer serves a purpose). Only issue is breaking current installs, do we care? We could of course just note it in the caveats or instead deprecate rustup-init and make rustup a new formulae

[carlocab]

We could manually link rustup and rustup-init in post_install since those wouldn't conflict, I think.

[branchvincent]

great idea, done! This looks pretty good to me now, my only other nit is I want to rename this to rustup and have rustup-init be the alias instead, but I can save that for another PR

[github-actions]

🤖 An automated task has requested bottles to be published to this PR.