Extract Http-Bridge specifics into a dedicated package #212
Conversation
| @@ -64,7 +64,7 @@ jobs: | |||
| name: "Compiling" | |||
| script: | |||
| - mkdir -p ~/.local/bin | |||
| - travis_retry curl -L -o cardano-node-simple.tar.gz https://raw.githubusercontent.com/input-output-hk/cardano-wallet/master/lib/core/test/data/cardano-node-simple/cardano-node-simple-3.0.1.tar.gz | |||
| - travis_retry curl -L -o cardano-node-simple.tar.gz https://raw.githubusercontent.com/input-output-hk/cardano-wallet/KtorZ/extract-http-bridge/lib/http-bridge/test/data/cardano-node-simple/cardano-node-simple-3.0.1.tar.gz | |||
There was a problem hiding this comment.
Again, temporary, will change after merged.
| @@ -93,14 +93,14 @@ jobs: | |||
| script: | |||
| - export NETWORK=mainnet | |||
| - tar xzf $STACK_WORK_CACHE | |||
| - stack --no-terminal test cardano-wallet-core:unit | |||
| - stack --no-terminal test cardano-wallet-http-bridge:unit | |||
There was a problem hiding this comment.
So in CI, we have two jobs running tests. One for testnet which run everything including integration tests. And another one for mainnet, to run a couple of unit tests with a different protocol magic. Yet, we only pattern-match on the protocol magic (and actually, have access to it) in the bridge code so there's no need to re-run the core tests on mainnet.
| return $ ApiT . SignedTx $ bs | ||
| bs <- (base64 <$> (p .: "signedTx")) | ||
| >>= either fail return | ||
| tx <- pure (CBOR.deserialiseFromBytes decodeSignedTx (BL.fromStrict bs)) |
There was a problem hiding this comment.
is this change compatible with "old data" - if tx witness is stored on a blockchain
There was a problem hiding this comment.
aha - in fact I see it didn't change at all. You have just moved some code bits here
There was a problem hiding this comment.
Indeed. The code is the same, but now done within the http-bridge package (rather than in the core), because it uses some cbor-specific encoding. So, before that, the core was sending a 'SignedTx ByteString, where the ByteStringwas supposedly the serialized tx, and it now sends a(Tx, [TxWitness])` and the network layer takes care of serializing it. This keeps the network layer agnostic to the binary format.
| chgs'' = drop 1 chgs | ||
| inps' = if length inps > 1 then drop 1 inps else inps | ||
| outs' = if length outs > 1 then drop 1 outs else outs | ||
| chgs' = drop 1 chgs | ||
| in | ||
| filter (\s -> s /= sel && isValidSelection s) | ||
| [ CoinSelection inps' outs' chgs' |
There was a problem hiding this comment.
I see that we have reduced generator scope of CoinSelection - hope coverage is still big enough to test relevant bits
There was a problem hiding this comment.
I've only adjusted the shrinker here to be less aggressive and therefore a bit more readable. The more aggressive version is especially needed for the property that has been relocated.
There was a problem hiding this comment.
yes I see it here 088c8cc#diff-80e66ce4975ed1ec822ac0aa910b4553R775
Issue Number
Overview
Comments