Wire up outstanding new IPC queries to the cli #2290
Conversation
Jimbo4350
force-pushed
the
jordan/wire-up-remaining-api-queries-final
branch
4 times, most recently
from
4f63c04
to
67db64f
Compare
Jimbo4350
force-pushed
the
jordan/wire-up-remaining-api-queries-final
branch
2 times, most recently
from
be36ba2
to
ec658d4
Compare
Jimbo4350
requested review from
DavidEichmann,
dcoutts,
erikd,
intricate and
newhoggy
as code owners
Jimbo4350
force-pushed
the
jordan/wire-up-remaining-api-queries-final
branch
4 times, most recently
from
56c7a9b
to
da28b27
Compare
Jimbo4350
force-pushed
the
jordan/wire-up-remaining-api-queries-final
branch
from
da28b27
to
5f7ff36
Compare
Jimbo4350
force-pushed
the
jordan/wire-up-remaining-api-queries-final
branch
from
25329bb
to
e3f8b18
Compare
There was a problem hiding this comment.
Definitely going in the right direction.
Main thing is to look more closely at the ToJSON instances you're providing. Where we do provide ToJSON instances they should be sane machine-readable choices, and reflect the surface structure of the types, not revealing lots of internals (so not using Generic or Show).
Jimbo4350
force-pushed
the
jordan/wire-up-remaining-api-queries-final
branch
3 times, most recently
from
ef64b98
to
250cf11
Compare
| instance ToJSON TxIn where | ||
| toJSON (TxIn txId (TxIx ix)) = | ||
| Aeson.String ( Text.decodeUtf8 (serialiseToRawBytesHex txId) | ||
| <> "#" | ||
| <> Text.pack (show ix) | ||
| ) |
There was a problem hiding this comment.
Oh what happened to the ToJSONKey instance? Perhaps I misread and you only have the ToJSONKey for the underlying Shelley type.
We should provide it for the surface type because it's directly useful and so we can use the simple instance for UTxO above.
The ToJSONKey instance should use this rendering as a single string with the #, but the ToJSON should still give us an object with two fieldstxid and txix.
| -- Orphan instances involved in the JSON output of the API queries. | ||
| -- We will remove/replace these as we provide more API wrapper types |
There was a problem hiding this comment.
Mmm. This does make me a little nervous. Where do we still need these? Is this for the ToJSON instance for the ledger state? Otherwise I think we should be able to do everything with JSON instances only for the API surface types.
There was a problem hiding this comment.
Yes these are for the LedgerState and ProtocolState which are just type wrappers around the underlying ledger-specs types. We could do something similar for them (e.g the api's UTxO and having conversion function from the underlying ledger types) but I figured implementing JSON instances would be less time consuming and OK for now as these are used in rendering the output for protocol-state and ledger-state cli commands.
| :: Query.SerialisedLedgerState era | ||
| -> Either LBS.ByteString (Query.LedgerState era) | ||
| decodeLedgerState (Query.SerialisedLedgerState (Serialised ls)) = | ||
| first (const ls) (decodeFull ls) |
There was a problem hiding this comment.
We should probably provide something like thisin the API itself to get from the SerialisedLedgerState to a LedgerState (or error), rather than having to know the underlying details of how it works here.
There was a problem hiding this comment.
I'll add this to my TODO list or I can delegate this as a small task. This might be a good task for John or Yuri to get their feet wet in the API.
Jimbo4350
force-pushed
the
jordan/wire-up-remaining-api-queries-final
branch
from
250cf11
to
03783fe
Compare
Jimbo4350
force-pushed
the
jordan/wire-up-remaining-api-queries-final
branch
from
5ddada9
to
f3088d7
Compare
|
bors r+ |
|
Build succeeded: |
Depends on #2275
The following cli commands were updated with the new IPC.
tiputxoledger-stateprotocol-state