feat: introduced new downloads page

[ovflowd]

This PR introduces the shiny new Downloads page envisioned by the package maintenance Working Group (#7037).

It also officially unblocks community members from adding package managers and installation methods to the Downloads page!

Warning!: This PR also resets all translations for the downloads page (unfortunately, as a preventive method)

Closes #7037
Closes #7296
Unblocks #6953
Unblocks #6978
Unblocks #6832

CHECK THE PREVIEW HERE

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated (UTC)
nodejs-org	✅ Ready (Inspect)	Visit Preview	Dec 28, 2024 3:03am

[github-actions]

Unit Test Coverage Report

Lines	Statements	Branches	Functions
	84.08% (597/710)	65.12% (183/281)	82.7% (110/133)

Unit Test Report

Tests	Skipped	Failures	Errors	Time
124	0 💤	0 ❌	0 🔥	4.903s ⏱️

[ovflowd]

Out of curiosity, what determines if an installation method is "official"? What makes nvm official that fnm doesn't have -- both are third-party version managers not maintained by the Node.js org?

That is a good question. I will write up the guidelines for Official vs Community on a follow-up PR and how to get your installation method or package manager added to the current Downloads page.

However, regarding official versus community, I'd say there needs to be some criteria for the official status to be attained. (Keep in mind these are criteria I've come up with and were NOT agreed upon yet; these are pending discussion and agreement)

• As soon as a Node.js release gets out, said version manager/installation mode needs to have it out, so either coordinated releases or something that listens for our releases
• The project maintainers must have a close relationship with Node.js, at least a direct communication way
• The installation method must download the official binaries bundled by the Node.js project; if the installation method builds it from the source and/or modifies what gets bundled (SSL version, Corepack or not, NPM or not or removed modules), then it also does not qualify as an official method.
• Maybe other criteria.

fnm may already fit the criteria. Still, at the moment, I only marked (intentionally) "nvm" as the only Official one as it already fits all these criteria above, and it is the most used version manager out there. I'm more than happy adding "fnm" to the list, I don't even know who the fnm maintainers are and if they have any ties with the Node project.

[github-actions]

Lighthouse Results

URL	Performance	Accessibility	Best Practices	SEO	Report
/en	🟢 91	🟢 100	🟢 100	🟢 91	🔗
/en/about	🟢 100	🟢 100	🟢 100	🟢 91	🔗
/en/about/previous-releases	🟢 99	🟢 100	🟢 100	🟢 92	🔗
/en/download	🟠 89	🟢 100	🟢 100	🟢 91	🔗
/en/blog	🟢 100	🟢 100	🟢 96	🟢 92	🔗

[ljharb]

i'm happy with the current state of the PR, thanks!

[ovflowd]

Cane someone from @nodejs/web-infra approve this PR too?

[MattIPv4]

LGTM.

Looking forward to the follow-up docs PR for folks wanting to add other community options, and what constitutes an official option vs. community.

As a future improvement, would love the over-the-wire HTML to have some default selected (perhaps based on stats for the most common OS) and the snippet pre-rendered (could just be plain-text if Shiki pre-rendering is too complex).

[MattIPv4]

You mean when JavaScript is not available

Yah indeed, if JS isn't available and before JS has run on the page but once the HTML has arrived (with my slow:tm: internet, I see a solid few seconds of the empty code block which I'd love to see have a default of some sort loaded instead from the start)

[ovflowd]

You mean when JavaScript is not available

Yah indeed, if JS isn't available and before JS has run on the page but once the HTML has arrived (with my slow:tm: internet, I see a solid few seconds of the empty code block which I'd love to see have a default of some sort loaded instead from the start)

I tested with slow 3g and it took 2-3 seconds, but honestly that shouldn't be a concern. The issue of having an initial codebox regardless of environment, @MattIPv4 is that it will cause a flash effect, since most likely it will show a different one once it renders, hence why the Skeleton.

It would also require Shiki + Eval'ing the snippet to happen twice. My last commit was done as a performance pass meticulously 👀 the performance graphs on DevTools.