Update nodejs bindings to use runtime field count

[jtraglia]

This PR updates the Node.js bindings to work with runtime field counts.

Note: the JSON trusted setup was outdated. It contained 4096 G2 points. Pulled in the latest version.

[asn-d6]

@matthewkeil would be great if you could take a look here. Thanks!

[matthewkeil]

You already got most of the items @jtraglia. Just the Error vs TypeError and the question about removing the Blob type. I tend to think that was a cleaner way of handling the array of bytes so its fixed length. What do you think? What about you @asn-d6?

[matthewkeil]

LGTM!! 🚀

Would love to actually run the tests to make sure all is well once #352 is approved and merged.

[jtraglia]

Great. Yes, we'll be sure that all of the tests pass before merging!

[asn-d6]

I'm slightly confused on why the bindings here use Blob in the typescript code, but uint8_t *blob in the C++ code.
From reading the comments above I see this is because "To my knowledge, like in C, it's not very practical in C++ to have a custom Blob type.".

Is this mixed usage of Blob preferable to just using uint8_t everywhere?

[jtraglia]

I'm slightly confused on why the bindings here use Blob in the typescript code, but uint8_t *blob in the C++ code.

For the API, I think Blob is a more human friendly name. Keep in mind Blob is a Uint8Array under the hood:

c-kzg-4844/bindings/node.js/lib/kzg.d.ts

Line 9 in bdcc3b9

export type Blob = Uint8Array; // Dynamic size

Is this mixed usage of Blob preferable to just using uint8_t everywhere?

For me, yeah. As long as it's Blob everywhere in the TS code & uint8_t * in the C++ code I'm good.

I suppose I don't know what the alternative is in the C++ code. Would you prefer we typedef Blob uint8_t? That would work, but I chose not to do that in the C code. We could also use std::vector<uint8_t> if you think that's a good idea.

[asn-d6]

LGTM! Thanks all!