More memory efficient query ledger-state command

[newhoggy]

This implements toEncoding of various ToJSON instances because it allows for more memory efficient lazy encoding.

As part of the change, the output of the query command will no longer be pretty printed.

Before uses about 16.4G:

$ time CARDANO_NODE_SOCKET_PATH=/Users/jky/wrk/iohk/mainnet/cardano-node.socket cardano-cli query ledger-state --mainnet > /tmp/ledger-state.bin

CARDANO_NODE_SOCKET_PATH=/Users/jky/wrk/iohk/mainnet/cardano-node.socket       302.63s user 14.17s system 91% cpu 5:47.89 total

After uses about 11.5G:

$ time CARDANO_NODE_SOCKET_PATH=/Users/jky/wrk/iohk/mainnet/cardano-node.socket $(./scripts/bin-path.sh cardano-cli) query ledger-state --mainnet > ledger-state.bin

CARDANO_NODE_SOCKET_PATH=/Users/jky/wrk/iohk/mainnet/cardano-node.socket       114.72s user 5.11s system 80% cpu 2:27.96 total

[Jimbo4350]

Nice work 👍 . You have introduce alignment where there wasn't any in the JSON instances. Lets stick to non-alignment. I think we should adopt an approach similar to ledger where we run a script with ormolu so there is no more ambiguity of the formatting.

[lehins]

We should really try avoiding duplication. If at any point we need to refactor this code we'll have to do double amount of work. And even worse we have more room for mistakes, say if one of the duplicate implementations will get out of sync (eg. change one field name and then forget the other)

comments addressed

[lehins]

Looks great, thank you! Very minor syntactic adjustments suggested

[Jimbo4350]

Good work! Please squash the commits.

[newhoggy]

Squashed

[newhoggy]

bors r+

[iohk-bors]

Build succeeded:

• buildkite/cardano-node
• ci/hydra-build:required