Mark Hernandez

Updating generator-lionbyte

On March 9, 2018, I updated my package generator-lionbyte to v1.0.1!

Today, I want to share all the changes I made and my thoughts as I was working on it.

If you want to follow along or review the changes at your leisure, here are links to the version tags:

You can check the diffs on this page:

Why I Decided to Update

I was using VS Code's debugger for Java for homework in my network communications class, and I thought it was a pretty great tool. I wanted to do a better job of debugging my JavaScript projects and rely less on console.log().

I did a quick search for any how-to articles and found one by Matt Rozema. Essentially, there were 3 options: debug the transpiled version, installing another package for debugging, or setting the IDE's configuration to use babel-node. All of these options seemed tedious to me, and I wanted to avoid it if I could.

Then I thought, "Maybe I don't need to use babel-node?" The feature I absolutely wanted to keep was async/await; they make it easier to work with Promises. According to, that particular ES2017 feature is available all the way back to Node v7.10.1! I use the latest available Node version on my personal machines, so I'm content.

As I was looking through the page, I saw that ES6 import/export was not natively supported. I read Vance Lucas's article Don’t Transpile JavaScript for Node.js afterward. He basically reaffirmed that I came to the right conclusions. He mentioned that babel converts import/export down to require() and module.exports. It's still useful for client-side code, since webpack can parse and transpile it.

Thus, I decided to start updating my package.


Although prettier doesn't necessarily affect the functionality, this basically helped save me from my past self. The last time I really worked on this project was around mid-December, so I should make everything easier to read and resume. I saw that prettier paired well with standard.

Removing babel-node Usage

This was tedious, but not difficult. I re-wrote the modules to use require and module.exports as needed.

Then I saw that the build scripts I wrote using gulp would be useless for Node.js specific projects. It was using gulp-babel to transpile the code, but I don't intend to write any code to be used for Node versions < 8.

Therefore, I decided that I don't need gulp anymore.

Removing gulp Usage

There were a lot of tasks, and I was using gulp-hub to let the tasks be split into separate files. The problem was, that took a long time to load all the separate files before running a single task.

The task files were distributed among the subgenerators. The server and default project types no longer needed them, so that was easy. In a way, it was pretty satisfying to delete all those files.

When I went to the static-site project type, I saw that it was heavily relying on gulp. It was... not the best, to say the least. I'm starting to realize that I was using way too much RAM during development with that configuration.

Using webpack@4 (Legato)

Since I was removing all gulp tasks, I thought it would be better to fully utilize webpack. Although it recently updated to v4, codenamed "Legato," it's a good chance for me upgrade my other projects using v3.

I previously only used webpack to package the JavaScript portion of a project. Now I'm letting it build and bundle everything. Most plugins have already updated to support this version, but some have to be installed with either @latest or @next. As of March 11, 2018, it is the following:

  • extract-text-webpack-plugin@next
  • html-webpack-plugin@latest

You can look at the webpack configuration files that I defined here.

Lessons Learned

  • Node.js versions > 8 already support all JavaScript features I want and more
  • prettier and standard code formatters are a nice pair
  • gulp is still pretty good, but has better use cases for it
  • webpack does so much for web development