Add ECPair.toUncheckedHex() and ECPair.fromUncheckedHex(hex, network, compressed)

[woodser]

This PR adds the ability to get the private key as unchecked hex (without network or compression information) and initialize a new ECPair from unchecked hex, network, and compression.

This is helpful to minimize the encoded data for encryption and splitting with Shamir's Secret Sharing which is why other libraries usually offer this in their API.

I'm using this API in the CryptoStorage tool here and here.

[dcousens]

NACK, already in #1070

[woodser]

The motivation of this PR is to encapsulate initialization from hex. It looks like #1070 does not permit a caller to initialize an ECPair from a hex string without using BitcoinJS's internal Buffer class. Please correct me if I'm wrong. Otherwise I think initializing from unchecked hex should be added and I can make the PR off of #1070 if you agree?

[dcousens]

@woodser Buffer isn't internal to bitcoinjs, it is required for many things to interact with this library.

[woodser]

I've been reading through #1070. If I understand correctly, ECPair is going to accept a buffer as its first constructor argument, and the way to get the hexadecimal private key will be keypair.privateKey.toString('hex'), correct?

The hex will not be padded so it may contain fewer than 64 characters for some private keys, such as this LTC private key: T34MuTEaRxt3kj6dn2MMHbvSPQHrsBjKrdz6WtwrLe1KzCzYkMup.

Shouldn't it be padded to conform to 64 characters?

Also, is d.toBuffer(32).toString('hex') a valid way of creating a padded hex from a BigInteger?

Thanks.

[dcousens]

@woodser

bitcoinjs-lib/src/ecpair.js

Lines 51 to 57 in d7eb6c8

	function fromPrivateKey (buffer, options) {
	typeforce(types.Buffer256bit, buffer)
	if (!ecc.isPrivate(buffer)) throw new TypeError('Private key not in range [1, n)')
	typeforce(isOptions, options)
	return new ECPair(buffer, null, options)
	}

Note specifically

bitcoinjs-lib/src/ecpair.js

Line 52 in d7eb6c8

typeforce(types.Buffer256bit, buffer)

[dcousens]

Also, is d.toBuffer(32).toString('hex') a valid way of creating a padded hex from a BigInteger?

That is what we did previously 👍

[woodser]

Ok, thanks!