Parcel vs Waku

[klis87]

I am aware that Waku required tremendous amount of work to learn how to implement RSC. I wonder though, now, as Parcel bundler implemented RSC in its core, how much Waku could be simplified, if switched to a bundler which supports RSC natively?

This is also question for the future, as Vite which Waku is built upon plans to add RSC as well.

[dai-shi]

Waku consists of various features, such as:

• bundler (including transformer)
• http server (for both RSC and SSR)
• dev server
• router
• cloud adapters
• cli
• (examples)
• (docs)

I guess developing a Waku-like framework bases on Parcel would still be a non-trivial task.

The biggest part I would like to offload from Waku is the bundler including the transformer, which should be done in the near future. We also extend the transformer for Waku-specific semantics, so that's also a required capability.

That said, if someone were to start developing a new framework from now, Parcel should provide good building blocks.

[klis87]

Thank you for the detailed answer, indeed it is much more than just a bundler, and I am really crossing fingers for Waku! I really admire your work, I know how hard it had to be to implement RSC without any documentation how to do it!

[dai-shi]

Thanks.

As a reflection, my original goal was to create a bundler and a very minimal server. That was a project that took a couple of months, but it wasn't a "framework," and the DX wasn't great. It was good as an RSC playground, but it wasn't a tool for production. Then, we pivoted to a real framework.