test: remove dual-denom test paths and add invariants coverage

[aljo242]

Summary

• remove remaining dual-denom (6/12/2 decimal) support paths from test constants and test suites now that runtime enforces 18-decimal-only configuration
• update non-18 rejection tests to use inline custom coin-info fixtures instead of shared multi-decimal constants
• add randomized ERC20 convert round-trip invariant coverage and a Go fuzz target for scaling/denom conversion idempotence

Test plan

• go test -tags=test ./x/vm/types ./utils
• go test -tags=test ./tests/integration/testutil
• go test -tags=test ./tests/integration -run 'TestERC20KeeperTestSuite|TestWERC20PrecompileUnitTestSuite' (from evmd/)

Made with Cursor

[greptile-apps]

Greptile Summary

This PR removes all non-18-decimal test fixtures and constants (SixDecimalsChainID, TwelveDecimalsChainID, TwoDecimalsChainID) now that the runtime enforces 18-decimal-only chains, replacing them with inline EvmCoinInfo fixtures in the rejection tests. It also adds a new randomized ERC20 convert round-trip invariant test and a Go fuzz target for the denom-relabeling conversion, including a regression unit test for the previously reported uint64 overflow guard.

• Constants cleanup: Three non-18-decimal ChainID entries and their associated EvmCoinInfo maps are removed from testutil/constants/constants.go; rejection tests in utils_test.go and configurator_test.go now use self-contained inline fixtures.
• New randomized invariant test (test_conversion_invariants_randomized.go): Runs 50 random ConvertERC20 ↔ ConvertCoin operations across five deterministic seeds, asserting strict value conservation (erc20Bal + coinBal == 5000) at every step.
• New fuzz target (scaling_fuzz_test.go): Verifies idempotency and amount preservation of ConvertCoinsDenomToExtendedDenomWithEvmParams, with a correctly guarded uint64 overflow path and a dedicated regression unit test.

Confidence Score: 5/5

Safe to merge — all changes are test-only, removing dead dual-decimal paths and adding well-guarded new coverage.

All modifications are confined to test files and documentation. The non-18-decimal constants and chain IDs are cleanly removed, the rejection tests are correctly rewritten with inline fixtures, the randomized invariant test properly guards against zero-balance panics, and the fuzz target uint64 overflow is addressed with both a runtime guard and a dedicated regression unit test.

No files require special attention.

Important Files Changed

Filename	Overview
tests/integration/x/erc20/test_conversion_invariants_randomized.go	New randomized round-trip invariant test; covers deterministic edge cases and 50 random steps per seed with correct balance-conservation assertions. No issues found.
x/vm/types/scaling_fuzz_test.go	New fuzz target for denom-relabeling conversion with idempotency and amount-preservation checks; the previously reported uint64 overflow guard is correctly implemented and backed by TestBuildFuzzInputCoinsMaxUint64Guard.
testutil/constants/constants.go	Removes SixDecimalsChainID, TwelveDecimalsChainID, TwoDecimalsChainID and their EvmCoinInfo entries; only 18-decimal entries remain.
utils/utils_test.go	TestCalcBaseFeeRejectsUnsupportedDecimals now uses inline EvmCoinInfo fixtures instead of the removed shared constants; test intent and coverage are preserved.
x/vm/types/configurator_test.go	TestEVMConfiguratorRejectsNon18Decimals migrated to inline fixtures for six- and twelve-decimal rejection cases; no behavioral change.
tests/integration/precompiles/werc20/test_events.go	Removes six-decimal test cases from TestEmitDepositEvent and TestEmitWithdrawalEvent; straightforward cleanup.
tests/integration/testutil/test_config.go	Removes the six-decimal test case from TestWithChainID; consistent with the constant cleanup.
README.md	Adds fuzz target invocation examples with 30s and 5m fuzz windows; documentation only.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[Mint 5000 ERC20 tokens] --> B[assertInvariants: initial]
    B --> C[ConvertERC20 amount=1]
    C --> D[assertInvariants: after convertERC20-1]
    D --> E[ConvertCoin amount=1]
    E --> F[assertInvariants: after convertCoin-1]
    F --> G{allERC20 > 0?}
    G -- yes --> H[ConvertERC20 full balance]
    H --> I[assertInvariants: after full convertERC20]
    I --> J{allCoin > 0?}
    J -- yes --> K[ConvertCoin full balance]
    K --> L[assertInvariants: after full convertCoin]
    L --> M[Random loop: 50 steps per seed]
    M --> N{erc20Bal == 0?}
    N -- yes --> O[convertFromERC20 = false]
    N -- no --> P{coinBal == 0?}
    P -- yes --> Q[convertFromERC20 = true]
    P -- no --> R[random choice]
    O --> S[ConvertCoin rand amount]
    Q --> T[ConvertERC20 rand amount]
    R --> S
    R --> T
    S --> U[assertInvariants: erc20Bal + coinBal == 5000]
    T --> U
    U --> M
    M --> V[assertInvariants: final]

Loading

_{Reviews (2): Last reviewed commit: "test: strengthen conversion invariants a..." | Re-trigger Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 66.92%. Comparing base (7d11196) to head (100f0b8).
⚠️ Report is 9 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1185      +/-   ##
==========================================
+ Coverage   66.32%   66.92%   +0.60%     
==========================================
  Files         336      320      -16     
  Lines       23959    23409     -550     
==========================================
- Hits        15891    15667     -224     
+ Misses       6891     6583     -308     
+ Partials     1177     1159      -18     

see 2 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[aljo242]

@greptile re-review

[github-actions]

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days-before-close if no further activity occurs.