There are several static site generators that I have tried, and Hugo is now up for review.

It claimms to be the fastest, with pagebuilds of 1 ms per page. That is certainly impressive, but there are other aspects of Hugo that appeal to me.

I have looked at Jekyll and [Pelican]. I like Jekyll because it is tightly integrated with Github pages, making it easy to launch a site. What I don’t like about Jekyl is that it is written in Ruby, mainly because I don’t know Ruby well enough.

For that reason, I have also looked at [Pelican] because it is written in [Python], which is a language I know and like. Even though you’d think it is an advantage to use a site generator in a language you know, the truth is that I never get to the point that I really need or want to look at the underlying code: if the static site generator does what it is supposed to do, I already spend enough time on writing content and trying to make it look good.

That made me circle back to Jekyll: easy to launch a site on Github pages, what more do you need?

Well, there are some things left to desire if you look at Jekyll. And that brought me to Hugo.

First of all, the community around it seems very active, it has 53.9K stars on Github. It also seems to be aiming at providing a solid solution for building a static site, and not just a blog (like Jekyll seems to be doing). It’s written in Go and is supposedly blazingly fast.

Whatever, I am trying to justify that I am looking into Hugo, let’s just have a go and tell you what I’m coming across.

  • install

  • restart shell so hugo command can be found

  • create new hugo site: hugo new site name_of_new_site

  • git init

  • git add .

  • add a theme:

    $ git submodule add https://github.com/luizdepra/hugo-coder.git themes/hugo-coder

  • fill out _config.toml

  • Build and launch site:

    $ hugo server -D
    Info
    You are reading this, so it works!
Site been build and deployed.
Very enthusiastic about Hugo, it is blazing fast

Getting the site fully working on AWS S3 was not completely trouble free, main problem turned out to be the fact that S3 isn’t a webserver, so you can’t tell it to look for index.html if it gets to a directory. Hugo has a solution for this, called uglyURLs, so setting that to true fixed the issue partially. New problem I then encountered was the very good caching that Cloudfront uses: the changes to uglyURLs was in effect, but I was still seeing the old (wrong) URLs, which gave errors.

Luckily there is a way of clearing the cache, which is called Invalidation on Cloudfront. Not a vbery intuitive name, I would say, but once you know that, you ca make it work.