feat(cli,world): register system ABI onchain

[holic]

• fix ABI race condition in world (world uses build script to generate ABIs but that same build script needs ABIs now)
  • maybe easiest to fix my moving to parsing AST instead of ABI
• register ABIs in metadata table during deploy
• add .mud to .gitignore in templates, etc.
• add namespaceLabel to config's system output

[changeset-bot]

🦋 Changeset detected

Latest commit: 5ebb331

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 26 packages

Name	Type
@latticexyz/cli	Patch
@latticexyz/world-modules	Patch
mock-game-contracts	Patch
@latticexyz/abi-ts	Patch
@latticexyz/block-logs-stream	Patch
@latticexyz/common	Patch
@latticexyz/config	Patch
create-mud	Patch
@latticexyz/dev-tools	Patch
@latticexyz/explorer	Patch
@latticexyz/faucet	Patch
@latticexyz/gas-report	Patch
@latticexyz/protocol-parser	Patch
@latticexyz/query	Patch
@latticexyz/react	Patch
@latticexyz/recs	Patch
@latticexyz/schema-type	Patch
solhint-config-mud	Patch
solhint-plugin-mud	Patch
@latticexyz/store-indexer	Patch
@latticexyz/store-sync	Patch
@latticexyz/store	Patch
@latticexyz/utils	Patch
@latticexyz/world-module-metadata	Patch
@latticexyz/world	Patch
ts-benchmarks	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[holic]

okay so after two attempts, I understand this a bit better

we currently have a constraint/order of operations

• parse solidity AST to codegen system interfaces + world interface
• import/use those codegen interfaces
• compile
• register ABI from compile artifact

I tried to simplify by just

• compile
• register ABI from compile artifact
• (eventually codegen interface from ABI)

but we lean on just ABI for a number of reasons:

• source won't compile if it references codegen interfaces that don't exist yet (chicken and egg problem)
• ABI is a subset of interface and is missing things like param modifiers (string memory vs string calldata, address vs address payable)

so then I went the other way

• parse solidity AST to codegen ABI
• register ABI
• (eventually codegen interface from ABI)

but that is also fraught because just parsing solidity skips parts of the compiler we need, like resolving imports with remappings, resolving underlying types of enums/user types, converting structs to tuples

so we can't actually create pure/true representation of the ABI and reimplementing all that in the "parse solidity" step seems like a bad idea

all this to figure out where the right "entry point" is for generating the local system manifest and I think it's in mud build after forge build

to recap:

• we need to keep our order-of-operations (parse source, codegen interfaces, compile, extract ABI)
• generate system manifest from ABI, after forge build but before abi-ts (in the future we may want these JSON manifests to be strongly typed)

there's still the problem of the ABI not including enough data for a 1:1 interface (e.g. memory vs calldata) but maybe we can just make some assumptions for now and fill that in later when we have some time to combine the ABI with AST (both available in forge output, which is nice, but still a pain to unwind)

[alvrs]

🔥🔥🔥