Migrating from Jekyll to Hugo
I originally chose to use Jekyll because it’s supported by GitHub Pages. At the time, I also wasn’t aware of other static site generators, and Jekyll seemed to be widely used. But when I recently decided to put more effort into my website, I looked into other options to make sure that I have the best possible foundation to work with. I figured migrating when my site is tiny would be easier than trying to migrate later when my site has more content and features.
I looked into this list of static site generators and decided to investigate Hugo more, mainly because of its popularity and because I had read a fair amount of praise for it beforehand. I don’t anticipate adding much dynamic content to my website, or else I would have given more consideration to Gatsby and Next, both of which use React.
I found quite a few posts from people who have migrated from Jekyll to Hugo:
- Docs are Migrating from Jekyll to Hugo
- Bye Bye, Jekyll. Hello, Hugo.
- Switching to Hugo
- Hugo vs. Jekyll: a showdown of static site generator (sic)
- Migrating from Jekyll+Github Pages to Hugo+Netlify
Hugo’s key differentiators: 1. Ease of install 2. Speed (critical for large sites) 3. Integrated live-reload while editing in near realtime 4. Multilingual capabilities 5. Flexible 6. Very strong community 7. Very good & comprehensive documentation (but not perfect…yet)
Hugo struggles with: 1. Not integrated with Github like Jekyll (though webhooks solve this to a large degree) 2. No plugin support 3. Media & Asset processing not tightly integrated
The cons are not important to me. I don’t need GitHub integration because I switched my website hosting from GitHub Pages to Netlify. I don’t think I will miss plugins because Hugo seems to have many built-in features (like RSS feed generation) that might be plugins for other generators. And I don’t need a sophisticated asset pipeline because I want to keep my website relatively minimal and static anyway.
Hugo builds my site in tens of milliseconds, and incremental builds are also measured in milliseconds. Jekyll, in contrast, took a few seconds for a full build, and even incremental builds took over a second. While Jekyll’s times might not sound so bad right now, my site doesn’t have much content. According to what I have read, Jekyll does not scale well, and it can take several minutes to build larger sites.
While Jekyll’s speed was not a pain point for me yet, I did want live reloading. I followed this pull request for adding live reloading to Jekyll without having to use a plugin. The feature was released with version 3.7.0 on January 2, 2018. However, I experienced an issue when I tried to use it. I would get
404 errors when pages were regenerated, presumably because the browser requests the page before it is ready. There is an open pull request to fix it.
This was the problem that really pushed me to switch to Hugo. It has natively supported live reloading for much longer, so I was confident that the kinks would be worked out already. That has turned out to be true in practice. Live reloading with Hugo works quickly and reliably.
This is the commit where I did the migration. Hugo has a command to import from Jekyll, so I ran that first. Most of the work after that was just converting the templates to Hugo’s syntax, which took a bit of time since I was not familiar with Go’s templating libraries.
I also had to convert my Sass files to raw CSS, but this wasn’t a big deal for me because I was barely using any Sass specific features. There’s an open issue for adding native Sass support, and this tweet confirms that it’s coming:
I did find the layout system to be a little confusing in terms of figuring out how Hugo determines which layouts to use for which pages. I had to re-read the documentation a couple times before I got it working.
Later on, there were multiple times when I discovered that Hugo has a convenient method for doing something, like adding Google Analytics, embedding YouTube videos, or adding Disqus comments. It wouldn’t be difficult to do these things with Jekyll, but they would involve adding boilerplate that isn’t necessary with Hugo. Small but nice touches like these give me confidence that I’ll want to stick with Hugo for a long time.
Jekyll is still a solid choice for creating a static website. It has a large community and its own advantages (like a plugin system). However, I’m happy that I switched to Hugo and that it’s in my development toolbox now. I encourage anyone who maintains or needs a static website to check it out.
Originally published at www.dannyguo.com.