I am looking to create a website that crawlers index properly, what is the best way to go about creating a site like that? Is it possible to do with shadow-cljs and these libraries to interact with Gatsby or Next.js? I would like to use Reagent I think, but I do not know if I can.
How would you create a site that renders to crawlers well?
I see that both gatsby-cljs and next-cljs have been archived. I would be interested in hearing Thomas’ perspective on where his thinking has gone since his interest in Gatsby and Next.js.
Or is the answer obvious: is shadow-grove your (Thomas?) new focus that more-or-less subsumes interest in Gatsby and Next.js?
I archived those since they were meant to demonstrate on how you could potentially use CLJS with those platforms. They were written a long time ago and I haven’t looked at either platform since. Given these are JS platforms I doubt those examples even still work. I didn’t want to delete the repos, but if you intent to build something using these platforms you are probably better off starting from scratch.
shadow-grove is in no way comparable to those platforms. Server-side rendering is not supported and as of now is a non-goal, so I do not intend to work on anything in that area anytime soon.
If you ask me how to create a static site I would tell you to create a normal dynamic website rendering HTML via CLJ (Clojure, not ClojureScript). I would use hiccup but any other CLJ lib is fine. Anything goes. Use any CLJ server lib you want. ring+hiccup is fine. To generate the static part you just
curl every URL you need, store the
.html where needed and publish it to a server. Much more manual than the JS platforms, but also inifinetly less complex.
Rum might be worth a look. It’s a minimal React wrapper that supports server-side rendering - so will tick the box for crawling purposes.
From the docs:
Rum is a client/server library for HTML UI. In ClojureScript, it works as React wrapper, in Clojure, it is a static HTML generator.
You can use nbb to render Reagent forms on the server side. In Sitefox I use the following pattern:
["fs" :refer [readFileSync writeFileSync]]
[reagent.dom.server :refer [render-to-static-markup]]
["node-html-parser$default" :refer [parse]]))
(let [template (readFileSync "index.html")
html (render-to-static-markup [:div "hello"])
document (parse template)
body (.querySelector document "body")]
(aset body "innerHTML" html)
(writeFileSync "page.html" (.toString document)))
You will need to
npm i nbb node-html-parser react react-dom in order to run this nbb script.
What this does is load up the template
index.html and replace the body tag with the Reagent rendered div tag. You can scale this to write many files with different content based on one or more templates. You can also do things like loading Markdown using e.g. the
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.