recursive-proxy-mock VS nosx

Let's take a specific example. One of my libraries recursive-proxy-mock is bundled with Rollup. If you take a look at dist/cjs/index.js you can see there is one big file with everything in it, all bundled together. There are named exports for each thing. This can easily have unused code eliminated either via tree shaking or dead code elimination. I've never made a distinction as to which applies in which scenario since both should always be used together. But I guess my point is, a bundled library doesn't need to have any special format which would require a bundlers help to drop dead code.

If you want to check out my build config for my library that sounds very similar in architecture to yours, feel free: https://github.com/CreativeTechGuy/recursive-proxy-mock

Take a look at the recursive-proxy-mock library. The ability to create an object where every property or function "exists" with infinite depth is a really powerful primitive. Not just for providing mock implementations to complex objects for testing, but also for the ability to store and replay operations on another object in the future.

If you are just looking for a TS library config (not related to react) then here's a good example I can share: https://github.com/CreativeTechGuy/recursive-proxy-mock/blob/main/rollup.config.js

I just did this prototype library yesterday that uses proxies internally: https://github.com/jehna/nosx