Release 0.21

Wednesday, June 18, 2025

🔧 New features

Benchmark your tests

You can now measure the runtime of any test using the new bench feature. If a test exceeds your expected threshold, use the max_ms option to mark it as failed.

Example

# @bench max_ms=10
function test_should_be_fast() {
  sleep 0.01
}

Output

✗ Failed: test_should_be_fast (Expected < 10ms, got 11ms)

Placeholder support in snapshots

Ignore dynamic values in your snapshots using placeholders like . This makes tests stable even when parts of the output vary between runs.

Example

function test_snapshot_with_placeholder() {
  echo "ID: 12345" | assert_match_snapshot --ignore "{{ignore}}"
}

Snapshot

ID: {{ignore}}

Count data providers accurately

find_total_tests now includes cases provided by @data_provider, giving a more accurate count of total executed tests.

🐛 Bugfix

Parallel mode stability

Fixes were applied to ensure parallel mode works consistently across platforms, including improved detection on Windows and consistent execution on arch64.

Example

# Works across systems and architectures
BASHUNIT_PARALLEL_RUN=true ./bashunit

Validate exit codes

A new test ensures the bashunit_exit_code logic is correctly handled under different execution paths.

🌾 Miscellaneous

• Improved performance of internal clock utility
• Flexible argument handling in the installer
• Project overview added to the docs site
• CI jobs now split between macOS and Ubuntu
• Fixed minor typo in spy assertion message

---

See the full changelog on GitHub