I really deviated from the topic today. That is what happens when I start with front-end stuff - I drift. Drifting is otherwise a talent, but most often it just wastes time.
This Sunday, I set out to explore AWS SES to set up some email notification features for one of my upcoming projects. Being a backend guy, I thought it would be a good idea to prepare the backend first before we put something on the front end. Then one thing led to another and I ended up exploring various options to implement an admin panel for Hugo generated static website.
Long story short, I am still sticking with my good old IDE - VSCode. However, I still want to document my experience of exploring a few options and conclude this topic by answering why I am sticking to my IDE-based automated workflow which I will share in a later post.
What are the options?
A quick Google search landed me here on the Hugo website's frontend section that presents us with links to various options. Out of all the options, I spent the most time on Netlify CMS and Hokus CMS.
My experience with Netlify was good, but not good enough. I mean I like Netlify, but here I am trying to do something solo, and depending on Netlify did not make a lot of sense. Providing access to my GitHub private repositories, setting up the authentication flow, etc. seemed too much in terms of dependency.
It eventually led to the DNS setting of my domain which was required for the auth to the admin panel to work properly. This is where I gave up. I am trying to manage all of this on the least number of platforms, and Netlify CMS service was just not part of it. I had to apply brakes here as this just leads to the proliferation of tools in my ecosystem and thus, a huge web of who accesses what is created.
Plus I did not want to manage any more DNSes. It is just too much to remember.
I also checked Hokus CMS, which is a desktop app, which made a bit more sense as I thought it would just connect to my Git repository or something like S3, where I could edit and create posts. For some reason, the setup did not work on my M1 Mac.
I also checked the repository, the project is no more active. The last update was made 2 years ago and Juliano has clearly mentioned that it is a work-in-progress project. The interface screenshot looks promising and I really hope this project picks up.
I did not click on Lipi as it mentioned Java in its one-liner description. 10 years ago I was in love with Java. But I have moved on now and similar to Kubernetes, it can be a world of its own. I am just trying to resist the temptation here.
Enwrite mentioned it is evernote-powered. That introduces another tool. And by this time I was done. I mean I never intended to explore admin CMS panels today, still, I ended up spending almost 2 hours on this.
Why do I reject all?
Short answer - priorities.
For now, I just need to prove that automating static content delivery is possible. Which I have already proved using GitHub Actions, S3, etc. I have other priorities like email features, user session management, etc. to worry about on an immediate basis.
Maybe, in the future, I will take this up again. At least Hugo is finalized for the static part of the front end.