Documenting the expected behavior when one of the inputs is zero for batchInvert about noble-curves HOT 6 CLOSED

It also seems that the non batched version of invert will panic on zero, if I'm understanding correctly:

  if (number === _0n || modulo <= _0n) {
    throw new Error(`invert: expected positive integers, got n=${number} mod=${modulo}`);
  }

from noble-curves.

Document where?

from noble-curves.

It's reasonable. It should be undefined or error since inverse doesn't exist.

from noble-curves.

Right, I'm not commenting on whether it's reasonable, just that it would be good to have it documented, preferably on top of the function?

Though as commented above, batchInvert and Invert seem to have different behaviour on zero, what was the rationale behind this?

from noble-curves.

Though as commented above, batchInvert and Invert seem to have different behaviour on zero, what was the rationale behind this?

Performance reasons and easier testing

from noble-curves.

done

from noble-curves.