Add golden tests for signing of transactions #203
Conversation
| @@ -396,7 +396,7 @@ newtype SignedTx = SignedTx { signedTx :: ByteString } | |||
| deriving (Show, Eq, Generic) | |||
|
|
|||
| data TxWitness | |||
| = PublicKeyWitness ByteString | |||
| = PublicKeyWitness ByteString (Hash "signature") | |||
There was a problem hiding this comment.
Key 'AddressK XPub would feel better instead of ByteString, will do in separate pr because of import cycles
There was a problem hiding this comment.
XPub would probably be sufficient here.
| addr `shouldBe` bin | ||
|
|
||
|
|
||
| it "decodeSignedTx slBinary == walletTx" $ do |
There was a problem hiding this comment.
I opted to test both using decodeSignedTx and encodeSignedTx. Should add a roundtrip property regardless (I think it doesn't exist. We only test the roundtrip on txs from the BinarySpec blocks.)
| bs <- CBOR.decodeBytes | ||
| case CBOR.deserialiseFromBytes decodePKWitness (BL.fromStrict bs) of | ||
| Left err -> fail $ show err | ||
| Right (_, input) -> return input | ||
| 1 -> CBOR.decodeTag *> (ScriptWitness <$> CBOR.decodeBytes) | ||
| 2 -> CBOR.decodeTag *> (RedeemWitness <$> CBOR.decodeBytes) | ||
| _ -> fail |
There was a problem hiding this comment.
As I had problems with it previously, I should check that this fail works, and doesn't produce ~infinite output.
|
|
||
| fromR = either undefined Prelude.id | ||
| in | ||
| fromR $ mkStdTx m (error "no root key", pass) [ownedIns] outs |
There was a problem hiding this comment.
Similarly for what we did for Yoroi golden tests in the AddressDerivationSpec module, it would be better to have a goldenSignedTx utility to generate and compare a transaction binary format from a list of (Address, Key 'AddressK XPub)
| Mainnet -> Mainnet | ||
| Staging -> Mainnet | ||
| Testnet -> Testnet | ||
| Local -> Testnet |
There was a problem hiding this comment.
If I recall correctly, the witness includes the protocol magic too, no ? So all tx should be slightly different even between staging and mainnet. Also, for consistency, we're probably better off with having the same file for local and testnet (not sure there's actually any value checking this for the local network...)
KtorZ
force-pushed
the
anviking/95/SigningSpec
branch
from
2af4b0b
to
ddf4939
Compare
Basically cherry pick of cardano-foundation/cardano-wallet#203
Issue Number
#95
Overview
Comments