safe-global / safe-client-gateway / 23241548536

90%

push

github

fix(tests): resolve flaky tests from chain ID collisions and missing mock resets (#2977)

* fix(tests): prevent chain ID collisions in recipient-analysis tests

faker.string.numeric(3) for both source and target chain IDs has a ~0.1%
collision chance per test, causing mock routing to return the source Safe
for both chains — making bridge analysis return empty results.

Add uniqueTargetChainId() helper that retries until it generates a chain
ID different from mockChainId. Also remove redundant dead-code branch in
the INCOMPATIBLE_SAFE test.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix(tests): add missing mockReset() calls in CoW Swap TWAP tests

The TWAPs describe block inconsistently called mockReset() before
mockImplementation(), unlike the Swaps block. Add networkService.get
and .post mockReset() calls before every mockImplementation() in the
TWAPs tests to prevent mock state leaking between tests.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* refactor(tests): simplify uniqueTargetChainId using faker exclude

Use faker.string.numeric({ length: 3, exclude: [mockChainId] }) instead
of a manual do/while loop, as suggested in PR review.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(tests): scope fake timers to individual TWAPs tests

jest.useFakeTimers() in beforeAll caused the NestJS app to be
initialized under fake timers in beforeEach, potentially preventing
timer-based operations during module creation from completing.

Move jest.useFakeTimers()/useRealTimers() into the two tests that
actually need time control, so the app always initializes with real
timers. Also remove redundant mockReset() calls — jest.resetAllMocks()
in beforeEach already resets the module-level singleton networkService.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

3074 of 3879 branches covered (79.25%)

Branch coverage included in aggregate %.

14757 of 16037 relevant lines covered (92.02%)

557.54 hits per line