nushell / nushell.github.io

Nushell's main website, blog, book, and more
https://www.nushell.sh/book/
MIT License
183 stars 437 forks source link

Consider moving away from VuePress #1126

Open Hofer-Julian opened 1 year ago

Hofer-Julian commented 1 year ago

I've been wondering about the low development activity of VuePress and found this: https://vitepress.dev/guide/what-is-vitepress#what-about-vuepress

I am not sure if I can work on this soon, but I wanted to check if there would be any objections for moving this site to VitePress.

Migration guide: https://vitepress.dev/guide/migration-from-vuepress

hustcer commented 1 year ago

Thanks for the suggestion. Actually I have thought about this for quite a while, there are three choices I have considered:

As we may support multiple doc versions in the future and we have lots of docs already, the build speed is quite important, after reading this I think RSPress is a better choice than VitePress and it has multiple version support: https://rspress.dev/guide/default-theme/multi-version.html Hugo seems doesn't have builtin multiple versions support? I'm not quite sure. What's more I haven't found a hugo theme that's better than current one. So RsPress is the one I prefer currently, however it doesn't support shiki yet, I have already create an issue here: https://github.com/web-infra-dev/rspress/issues/211, we need this plugin to support real Nushell syntax highlighting. So my suggestion is keep on waiting.

Hofer-Julian commented 1 year ago

Thanks for the write-up @hustcer.

Never heard of RSPress, but sounds like an interesting tool. I wouldn't advise using Hugo. It's a great tool, but reproducing the features we use by VuePress sounds like a big maintenance burden.

Waiting a bit sounds good to me. VuePress still works quite well after all.

Hofer-Julian commented 1 year ago

I've changed the issue title accordingly

NotTheDr01ds commented 3 months ago

Copying over some relevant information from the duplicate #1231:

From @fdncred, some requirements:

... it would have to:

  1. be as easy or easier to manage than what we have now
  2. look as good or better than what we have now
  3. translations would need to work as good or better than what we have now
  4. @hustcer would need to approve of it since he manages a lot of the site for us now (we'd have to have other core-team members approve too in order to move to it, but not really needed for a preview)
  5. we'd also need a way to preview it somehow

From @hustcer:

There are two things I most care about:

  1. Build speed, we have lots of docs now, the build speed is not fast enough, in the future we also intend to support multiple versions and the build speed requirements are higher. We do not support multiple versions also due to the slow build speed.
  2. The new approach need to support multiple versions, which is especially important after version v1.0. We may have docs for v1 and v1.1, etc. at the same time. You can try to solve these two points with Starlight.

And a Starlight PoC by @jacobdalamb in #1276