run-tests

name	run-tests
description	Execute AL test codeunits. Use al-runner for fast pure-logic tests (no BC required). Use bc-test for full integration tests against a running BC instance.

/run-tests — Execute AL Tests

Two Test Runners

Tool	When to use	BC required?	Speed
al-runner	Pure-logic unit tests (in-memory tables, no pages/HTTP/events)	No	Milliseconds
bc-test	Full integration tests against a running BC instance	Yes	Seconds–minutes

al-runner (preferred for unit tests)

Full AI agent guide:

al-runner --guide

Basic usage:

# Run all tests
al-runner ./src ./test

# With symbol packages
al-runner --packages .alpackages ./src ./test

# With stubs for unsupported dependencies
al-runner --packages .alpackages --stubs ./stubs ./src ./test

# Run a single test procedure
al-runner --run TestMyProcedure ./src ./test

# With coverage report
al-runner --coverage ./src ./test

# Machine-readable JSON output
al-runner --output-json ./src ./test

# JUnit XML for CI
al-runner --output-junit results.xml ./src ./test

Exit codes:

Code	Meaning
0	All tests passed
1	Real assertion failures
2	Runner limitations only (compilation gaps, missing mocks — not real failures)
3	AL compilation error

Use exit code 2 to distinguish runner gaps from real failures in CI.

What al-runner supports: in-memory tables, CRUD, filters, cross-codeunit calls, interfaces, asserterror, JSON types, BLOB/InStream/OutStream, RecordRef/FieldRef, IsolatedStorage, TextBuilder, and more. See al-runner --guide for the full list.

What al-runner does NOT support: Pages, Reports, XMLports, HTTP calls, event subscribers. Inject these via AL interfaces or use --stubs.

bc-test (full BC integration)

Expects bc-test on PATH (from bc-linux or bc-tools).

bc-test --help

Use when tests require BC's service tier, events, UI flows, or features not supported by al-runner.

Behavior

1. Determine which runner is appropriate for the test type
2. Parse $ARGUMENTS and pass through to the selected tool
3. If there's an active task folder (.dev/<task-slug>/), capture output to .dev/<task-slug>/test-results.txt

Handling Results

• All passing (al-runner exit 0): Report pass count. Note "Tests green."
• Exit 2 (al-runner): Runner limitations only. Not a real failure — add stubs or use bc-test for the blocked tests.
• Real failures: List each failed test with codeunit, method name, and error message. Offer to analyze and suggest fixes.
• No tests found: Check that test codeunits use Subtype = Test and mark procedures with [Test].