forked from LBRYCommunity/lbry-sdk
41 lines
No EOL
1.6 KiB
Python
41 lines
No EOL
1.6 KiB
Python
class SignatureSerializationFlag:
|
|
UNSIGNED = 0
|
|
'''
|
|
Format:
|
|
<FLAG><CLAIM BINARY>
|
|
or (legacy)
|
|
<CLAIM BINARY>
|
|
'''
|
|
ECDSA_LEGACY = 1
|
|
'''
|
|
Old claim format, which carried the signature inside the protobuf. Requires serializing back the claim with
|
|
signature stripped out for validation. This process requires knowledge on how a claim is serialized, thus requires
|
|
old fixed protobuf schema to work.
|
|
|
|
Format:
|
|
<CLAIM PROTOBUF SERIALIZED>
|
|
Curves: NIST256p, NIST384p, SECP256k1
|
|
Signature content: `r` and `s` in each half of the 64 or 96 bytes (depends on curve)
|
|
Signed payload:
|
|
1. Claim transaction output address (raw, decoded using base58)
|
|
2. Stripped out claim protobuf serialization (without the signature)
|
|
3. Certificate claim id (binary, not in network byte order)
|
|
'''
|
|
ECDSA_SECP256K1 = 2
|
|
'''
|
|
Format:
|
|
<FLAG><CERTIFICATE ID><SIGNATURE><BINARY PAYLOAD>
|
|
Curve: SECP256K1
|
|
Signature content: 64 bytes total, each half represents `r` and `s`
|
|
Signed payload:
|
|
1. raw claim name as bytes
|
|
2. Claim transaction output address (raw, decoded using base58)
|
|
3. Binary payload, independent of serialization (everything after the signature last byte)
|
|
4. Certificate claim id, not in network byte order.
|
|
|
|
A certificate can be signed as well, but this serialization model is unaware of content type or protobuf format.
|
|
'''
|
|
@classmethod
|
|
def is_flag_valid(cls, flag):
|
|
# todo: use python 3 enum when fully ported, but not worth now as its an extra dependency for py2
|
|
return 0 <= flag <= 2 |