Frustration Leads to Iteration

I've spent a very long time trying to migrate this blog over to Next.js (opens in a new tab). I forked a lot of projects, from Lee Robinson's excellent personal website (opens in a new tab)¹ to perhaps my favorite dev on the internet, Andrew Healey's (opens in a new tab) wonderfully simple (opens in a new tab) Next.js blog to finally settling on a fork of Reilly O'Donnel's Next.js Blog (opens in a new tab). I've done some heavy moderation to some of the components of the fork and plan on continuing. However the issue I ran into with Next.js was MDX (opens in a new tab)²

What's wrong with MDX?

Well, what was wrong, to me at least, was the pain of moving hundreds of plain markdown files to it. To be fair, I've been writing at this blog for 8 years now. I started off with WordPress.com, then WordPress self-hosted, then to Jekyll, then to WordPress again, then to Jekyll again (opens in a new tab), then to Ghost and now to Next.js. Gatsby, too (opens in a new tab). Round and round I went.

Moving all those posts around, from WordPress when I wasn't using markdown, to exporting from WordPress, etc, meant that there were errors in the frontmatter, extra tags in the frontmatter, weird date formatting, and more.

MDX is unforgiving with these issues as it is JavaScript in your markdown. I have 200+ files that I needed to get into the new system with as little a hiccup as possible.

The other problem was my lack of understanding of how Next.js works.

Importing all posts

I moved all 200+ posts into each of the above repos to build on Vercel (opens in a new tab) which is my preferred deployment platform after years of using Netlify. Each push resulted in multiple build errors.

I changed each .md extension to .mdx however there was raw html in several of the posts with weird formatting that kept Next.js from building properly. I wasn't willing to give up, but I knew I needed to figure something out because I wanted something simple, Ghost was expensive and too much for what I need my personal website to be.

Remembering Max Stoiber's archive blog

Somehow I remembered that Max Stoiber of styled-components fame had deprecated his old blog (opens in a new tab) and made a new one (opens in a new tab). I figured instead of fighing all 200 of these posts, that I would keep the ones from 2020 to current on the new site, and archive the rest, which you can find on the archives (opens in a new tab).

Frustration

I spent almost three months on this site and I feel like I've gained tons of insight and knowledge on how a new Next.js site works, so much so that my next project is using Next and Supabase.

Was it worth it? Technically yes. Emotionally? Probably not. But it is finished. So have at it.