Improve comments on high/low-S handling

[webmaster128]

of secp256k1_verify/secp256k1_recover_pubkey

[maurolacy]

Suggested change

	// Since the recovery param is mossing in the test vectors, we try both 0 and 1
	// Since the recovery param is missing in the test vectors, we try both 0 and 1

I understand that in a real scenario you need to know the recovery param. Otherwise, you can get a recovered pubkey that is not the right one.

I also understand that this is not related to the high/low-S comments, but another issue / improvement.

Will read more about all this and comment back if needed.

[maurolacy]

I understand that in a real scenario you need to know the recovery param. Otherwise, you can get a recovered pubkey that is not the right one.

Will read more about all this and comment back if needed.

Here are some references about the recovery param:

• https://ethereum.stackexchange.com/questions/42455/during-ecdsa-signing-how-do-i-generate-the-recovery-id
• https://bitcoin.stackexchange.com/questions/38351/ecdsa-v-r-s-what-is-v

So, it seems that there's a small probability of getting two valid pub keys for different recovery params.

[webmaster128]

I also understand that this is not related to the high/low-S comments, but another issue / improvement.

Exactly. The only reason they are mixed together here is because I pull in the COSMOS_SECP256K1_TESTS_JSON tests which are missing the recovery param.

[webmaster128]

Thanks a lot for the thorough review, @maurolacy!