typedoc | Documentation generator for TypeScript projects | Generator Utils library

I solved this issue using a typedoc plugin lib made specifically for this purpose, typedoc-plugin-not-exported. This flawlessly addresses this exact problem. Exposing types on the global scope isn't technically exporting them, and so the updated version of typedoc (I'm thinking >=0.20.0) won't include them. This plugin library fixes that.

Okay let's make a simpler test case:

Found where the issue comes from.

It looks like oat++ is a bit finnicky when it comes about declaring the list object.

Firstly, rimraf will not do anything to the tsconfig.json file. It’s important to understand that the docs script in the package.json file, i.e. this part;

From one of the Typedoc Collaborators:

You see that error because the eslint and estree (parser that eslint relies on) have mismatched versions. I'd recommend making the versions match if possible, or turning on skipLibCheck.

I ended up adding skipLibCheck to the tsconfig.json compiler options, and that fixed it.

You may consider using Nodemon.

Okay, I was able to get this working but I am not sure why I had to do it this way after it's been working for over a year with the current system. I even have other devs using it with no issue. Very strange.

To get this working, I had to make a simple change.

I modified cdk.json to be:

This is due to Terser being unable to reason about the side-effects in your colors.ts enum, so Terser keeps all three definitions even though it only exports one of them.

If these weren't transpiled TypeScript enums, I'd recommend to simplify the declarations, ideally by marking each function /*#__PURE__*/ and making it return its intended value. However, since they are TypeScript enums, you might need to convert them to object literals as const, are certainly easier for Terser to reason about and are likely sufficient for your needs.

If I'm reading your output right, the arrays you're trying to remove are present in both development and runtime builds; you've omitted them with "..." but they're there.

According to your package.json you are using both sideEffects and usedExports of Webpack's tree-shaking feature set. sideEffects correctly asserts that you aren't changing anything aside from your exports, so Webpack can safely skip your whole module if your project consumes none of its exports. However, usedExports might not be as smart as you would hope:

usedExports relies on terser to detect side effects in statements. It is a difficult task in JavaScript and not as effective as straightforward sideEffects flag. It also can't skip subtree/dependencies since the spec says that side effects need to be evaluated.

It seems that for both development and production Webpack is smart enough to detect that your HueColors is the only export you consume, but Terser is not smart enough to determine that each self-initializing IIFE is free of side-effects that would affect the others. Technically, as a human, I can't reason about it either: Some other piece of code might have changed the Object or Array prototype in a bizarre way, even if your functions didn't use inline assignment or modify same-named shadowed variables of an enclosing scope in your IIFEs.

With an in-browser copy of terser I've been able to reproduce your problem.

First of all, switching to const object literals would be completely effective:

You want to document the props interface, and not the component itself. Which means this is the same as documenting fields of an interface.