test(precompiles): add golden tests for asset V1 (BOP-423)

[stephancill]

Summary

Adds a golden/snapshot test suite pinning current Asset V1 behavior of the B-20 precompile (BOP-423), before/while it is frozen behind the versioning seam from #3969 (BOP-417). Sibling of the stablecoin golden suite (#3971).

Every op is driven through the version-resolver-gated path (inner(..., BaseUpgrade::Beryl) → AssetVersions::from_base_upgrade → V1) against the real EVM-backed B20AssetStorage over HashMapStorageProvider, with InMemoryPolicy for deterministic allow/block. Each mutation case asserts:

1. exact returned ABI bytes (or the typed revert),
2. resulting state (balances / supply / roles / allowances / multiplier / metadata / storage),
3. emitted event signature(s), and
4. a pinned keccak storage hash over the sorted storage triples (hash_state) — a reference baseline for the CI frozen-manifest check (BOP-421).

Plus a dedicated golden_gas_footprints test pinning each mutating op's storage-access footprint — the (SLOAD, SSTORE, KECCAK256) op counts, the deterministic gas-schedule-independent signal that drives real gas — so an extra storage access in V1 is caught even when bytes/state/events match.

Op coverage: the shared B-20 surface (transfer / transferFrom / approve + memo variants, mint / burn / burnBlocked, pause / unpause, updateSupplyCap|Name|Symbol|ContractURI, grant/revoke/renounce/renounceLastAdmin/setRoleAdmin, updatePolicy, permit, all reads + role/policy constants), plus the asset-specific ops: updateMultiplier (operator role, zero-reject), updateExtraMetadata (metadata role, empty-key reject), batchMint (mint role, length-mismatch / empty-batch), announce (operator role, internal-call loop, reused-id, nested-announce, malformed / failing internal call), and the scaled-balance reads (multiplier, toScaledBalance, toRawBalance, scaledBalanceOf, configurable decimals). Privileged vs unprivileged via inner_with_privilege vs inner; the guard envelope (nonpayable / uninitialized / pre-Beryl) via dispatch_with_observer; plus the factory grant_role_unchecked bootstrap.

Edge cases: zero-address (sender / receiver / approver), insufficient balance/allowance, supply-cap, paused, policy-blocked sender + executor policy, empty-feature-set, last-admin protection (revoke/renounce), expired permit, and every unprivileged authorization revert (caller lacking the required role) — including the asset-specific batchMint / updateExtraMetadata / announce / updateMultiplier role checks.

Meta coverage guard: a compile-time v1_op_coverage_checklist — an exhaustive match over the generated IB20Calls / IB20AssetCalls enums whose arms reference the covering golden #[test] fns. Adding an ABI op fails the build (non-exhaustive match) until a golden is added; renaming/removing a golden fails the build (missing fn reference). This intentionally couples to the shared ABI surface: when a future version extends the ABI, this suite is expected to be updated to assert the new op reverts for V1.

No source/logic changes — frozen V1 is untouched; only a new integration test file and one [[test]] stanza are added.

Coverage

Region/line coverage of the frozen logic from this suite alone (via cargo llvm-cov, this test target only):

• b20_asset/logic/v1.rs: 100% line, 100% function, 91.1% region
• b20_asset/dispatch.rs: 97.2% line, 92.7% region

Remaining uncovered regions are defensive checked_*().ok_or_else(under_overflow) closures (incl. the scaled-balance overflow guard) and &&/|| short-circuit halves — unreachable without contriving arithmetic overflow.

Testing

• cargo test -p base-common-precompiles --features test-utils --test b20_asset_v1_golden → 98 passed
• full crate (cargo test -p base-common-precompiles --features test-utils) green; clippy (--all-targets) + fmt clean

State hashes are pinned; regenerate with BLESS_GOLDEN=1 cargo test ... -- --nocapture (documented in the file header).

[linear]

BOP-423

[cb-heimdall]

🟡 Heimdall Review Status

Requirement	Status	More Info
Reviews	🟡 0/1	Denominator calculation Show calculation 1 if user is bot 0 1 if user is external 0 2 if repo is sensitive 0 From .codeflow.yml 1 Additional review requirements Show calculation Max 0 0 From CODEOWNERS 0 Global minimum 0 Max 1 1 1 if commit is unverified 0 Sum 1

[github-actions]

Three unused imports: SolError (line 24), Asset and TokenAccounting (lines 27, 29). Clippy should catch these — remove them to keep CI clean.

Suggested change

	use alloy_sol_types::{SolCall, SolError, SolEvent, SolInterface, SolValue};
	use base_common_genesis::BaseUpgrade;
	use base_common_precompiles::{
	Asset, AssetAccounting, AssetV1, AssetVersion, AssetVersions, B20_MAX_SUPPLY_CAP, B20AssetInit,
	B20AssetStorage, B20AssetToken, B20PolicyType, B20TokenRole, IB20, IB20Asset, InMemoryPolicy,
	NoopPrecompileCallObserver, PermitArgs, TokenAccounting,
	use alloy_sol_types::{SolCall, SolEvent, SolInterface, SolValue};
	use base_common_genesis::BaseUpgrade;
	use base_common_precompiles::{
	AssetAccounting, AssetV1, AssetVersion, AssetVersions, B20_MAX_SUPPLY_CAP, B20AssetInit,
	B20AssetStorage, B20AssetToken, B20PolicyType, B20TokenRole, IB20, IB20Asset, InMemoryPolicy,
	NoopPrecompileCallObserver, PermitArgs,

[github-actions]

Review Summary

Test-only PR adding 2711 lines of golden/snapshot tests for Asset V1. No production code changes.

Finding

Unused imports (already flagged inline): SolError, Asset, and TokenAccounting are imported but never used. The existing inline comment covers this — note that its suggestion to add SolInterface is also unused and should be omitted from the fix.

General observations

• The compile-time v1_op_coverage_checklist (exhaustive match over IB20Calls/IB20AssetCalls without a wildcard arm, with covered(&[...]) referencing actual #[test] fns) is a solid pattern for ensuring golden coverage stays in sync with the ABI surface.
• hash_state capacity hint (events.len() * 64) underestimates for events with >2 topics or large data payloads, but this only affects pre-allocation efficiency in tests — not correctness.
• All .unwrap() / .expect() usage is within test code, which is appropriate.
• Cargo.toml change is clean: [[test]] stanza placed correctly before [features], [lints] workspace = true already present.

[github-actions]

❌ base-std fork tests did not run

The build or setup step failed before any tests could execute. Check the workflow logs for details.

Dependency	Ref	Commit
base-std	main	4658f1b7
base-anvil	f4370bc97756557d0ef1a6c325ad7d2532e71112	f4370bc9