Snapshot Auto-Update

GitAuto automatically updates snapshots when generating or modifying tests that use snapshot assertions. When a snapshot mismatch is detected during test execution, GitAuto reruns the tests with the update flag (e.g., Jest's --updateSnapshot) to create or update the snapshot files.

New Snapshot Tests Always Fail on First Run

New or modified tests that use toMatchSnapshot() or toMatchInlineSnapshot() always fail on their first run because the snapshot file does not exist yet. Snapshots are created by running the test and recording the output. But in GitAuto's pipeline, this means every snapshot test would report as a failure, triggering unnecessary correction iterations.

Auto-updating snapshots prevents these false failures. When GitAuto generates a new test that uses snapshots, it detects the "snapshot mismatch" error pattern and automatically creates the baseline snapshot, allowing the test to pass on subsequent runs.

Why Models Can't Handle Snapshots

Snapshot testing is a chicken-and-egg problem: the snapshot file is created by running the test, but the test can't pass without the snapshot file. Models can write expect(component).toMatchSnapshot() but they can't create the actual snapshot file - that requires executing the test. Without application-layer support to run the test with --updateSnapshot, every new snapshot test would fail on first run. Snapshot testing requires runtime execution to create the baseline, and no training or benchmark simulates this workflow, so models have no learned strategy for handling the create-then-compare lifecycle.

How It Works

During test execution, if the test runner reports a snapshot mismatch or missing snapshot error, GitAuto detects this specific failure pattern in the test output. It then reruns the failing test with the update flag (e.g., Jest's --updateSnapshot or -u), which tells the runner to create or overwrite the snapshot file with the current output.

The updated snapshot file is then included in the PR commit alongside the test file. This ensures the PR is self-contained - reviewers can see both the test and its expected output in a single diff.

Related Features

• Test Execution - the test runner that detects snapshot mismatches
• Coverage Enforcement - measures coverage after snapshots are updated and tests pass