svelteland/svelte-kit-blog-demo

A demo of a markdown based blog with SvelteKit, deployed to Github Pages.

svelte-kit-blog-demo's Issues

Installing svelte kit demo

Hi,
First of all I am a web dev beginner, I have some issue trying to install your blog. Can you provide a little document wich explain how to install your blog. thanks

Can't make dynamic routes

Hello, I've followed your tutorial to make a blog based in markdown posts.
I have same code as your github repo, and I have this error when accessing https://myblog/blog/someSlug

500
Unexpected token < in JSON at position 0
SyntaxError: Unexpected token < in JSON at position 0

I've looked around for a few hours, and I can't figure out what's wrong! Everything works fine in my app except that. Have you encountered this error before? Thanks a lot

Update to work with latest changes

Thanks for this starter repo. Any chance it can be updated to reflect the recent changes?

https://svelte.dev/blog/whats-new-in-svelte-june-2021

Namely,

svelte.config.js config files are now loaded in ESM format (.js instead of .cjs).

Confusing errors when frontmatter is missing

Just a note that plain .md files error out. To avoid very confusing error messages here's a modification to [slug].svelte

        if (!post.metadata || !post.metadata.title || !post.metadata.date) {
            post.metadata = {
                title: "Error!",
                date: "?"
            };
            post.content = "<p>Missing Frontmatter! Expected at least a title and a date!</p><p>Example:</p><p>---<br>\
title: Some title<br>\
date: 2021-04-03<br>\
excerpt: A short bit of the article<br>\
---</p>";
        }

Blog performance limited by process()

Playing around with my own local build I found that:

svelte-kit-blog-demo/src/lib/markdown.js

Line 30 in f8b7ac2

let content = runner.stringify(runner.runSync(tree));

Takes an incredible amount of time, this negatively impacts the rendering performance when listing out articles as all files under the target folder are processed each time. This is a somewhat bizarre browsing experience as the article list page with less content is considerably slower than visiting an individual article. It's in part dependent on how many nodes are generated as part of the ast from the markdown file. I'm finding pages with heavy usage of code blocks generate an absurd amount of nodes to handle syntax highlighting. This in turn means the runner.runSync(tree) can easily take seconds long. Not sure if you have an approach in mind, my most immediate thought was to have the backend use a bit of caching.

The counter to this is to avoid or be very specific about code blocks and whether any highlighting is needed. I found
"```txt" to work.

This assumes the website is run off the node adapter.