Markofile

[Jack5079]

Svelte is turning to shit so we will need to find a new framework. I propose Marko.

Advantages

• It is trivial to combine a Marko Run and Hono server: npm run build src/server.ts, and in there something like

  import * as Run from "@marko/run/router" // Run uses Request and Response
  import { Hono } from "hono" // So does Hono
  app.use("/", (c, next) => Run.fetch(c.req.raw) ?? next()) // i think? I had something before that worked I need to test more

• Marko lets you stream pages. This means, for example, we can send the CSS of a preview page before we have the metadata
• Marko's Tags API has reactivity like Svelte
• Marko 5 (current) won't hydrate components that don't have JS
• Marko 6 (future) is resumable - no hydration at all

Disadvantages

• Data loading not clear when combining it with Hono (if we want to use idiomatic Marko Run) although we could simply do whatever we do in Svelte since we aren't using SvelteKit anyway
• Less Marko developers
• No transitions
• No scoped CSS (although we aren't using that anyway)

[nbitzz]

Thank god marko starts with an M every new improvement we make regarding frameworks must have a related pun on monofile or fyle

[nbitzz]

For Markofile, since we'll be using Marko Run (probably), let's take care of account management (and maybe the file list display) on separate routes I think this should make #28 easier (or slightly less of a weird problem) to solve

[Jack5079]

Also @nbitzz I reccomend you read https://github.com/marko-js/run#path-structure

[Jack5079]

actually the whole readme

[nbitzz]

actually the whole readme

was planning on it, dw

[Jack5079]

Marko 5 now has Tags API built-in as of 5.33. Install @marko/translator-tags & @marko/runtime-tags. Tags API is still very much in alpha

[Jack5079]

I'm rewriting monofile in SvelteKit instead, when Marko Run improves I will revisit it