Haven’t posted here in a while, but thought it was time to do a little cleaning.
Almost 3 years since the last post. Lots of water under the bridge and lots of changes in technology.
Getting this website back up and running took a few updates.
Yea, lots of updates in the last 3 years, but this really wasn’t a big issues. Run
npm update and I was done.
Again, lots of changes, but I actuall haven’t had to change anything to website source in order to get it “compile”. Kudos to the team.
This was pain. I had to update from v1.something to v3. A lot has changed in tailwind and I had to spend a fair amount of time getting the css nice and the website looking okay.
Also, the recommended way to integrate tailwind with hugo shifted. Originally, tailwind and hugo were both being run via npm scripts. But now, tailwind is being run via hugo iteslf and the postcss integration. It does make things a bit nicer.
During the deploy builds, I still kick off hugo via an
npm script, but that is
just to make sure that everything is installed for tailwind and postcss.
There is still work to be done on this front. The website isn’t fully onboard with tailwind. I still have some custom css. But I’ll save fixing that when I do a total redesign the website.
I originally bought the domain via Google Domains. It was the easiest thing to do given that the site runs on Firebase. This also let me do some stuff around automating the deploys so I could schedule posts.
Buuut - Google Domains is shutting down and the assets (e.g. my domain) are being sold to SquareSpace. Didn’t really want any part of that to be honest. So, I decided to move my domain to Cloudflare.
This turned out to be a bit of a problem. Actually moving the registration and getting all the records set up was supereasy. But when it was all over, the website went into a redirect loop.
Turns out that Cloudflare, by default “proxies” the DNS record ( See here) which basically puts Cloudflare in the middle of every transaction. This was breaking Firebase with a continual redirect loop. A few trips through the docs and I finally found the page I just linked and how to fix the issue.
No problems since.
Firebase/ GCP/ Github
I had set up a series of Github workflow actions to automate Firebase deploy and gather the information needed to do schedule posts. Of course, everything about this was now wrong. Mostly it was how to go about authentication the two sides (github and firebase/gcp) to each other and was solved fairly quickly.
All-in-all, not too terrible. The tailwind/ccs stuff was the most anger inducing.
More to come.