| name | dal-unit-test-skill |
| description | Run and inspect this repository's build and Google Test workflow. Use when the user asks to run all tests, verify the whole codebase, execute the Linux or Windows build/test scripts, inspect `test_output.txt`, or confirm whether the test suite passes. |
| user-invocable | true |
Unit Test Skill
Instructions
- Before starting, delete any existing
test_output.txt in the repository root so it cannot be confused with fresh output.
Windows
- In a Windows terminal, run the following command to execute the full build-and-test workflow:
$ ./build_windows.bat > 'test_output.txt' 2>&1
- Review the output captured in the root-level
test_output.txt. The build
scripts run CTest, so a representative passing tail looks like:
100% tests passed, 0 tests failed out of <N>
Label Time Summary:
...
Total Test time (real) = <seconds> sec
Linux
- In a Linux bash shell, run the following command to execute the full build-and-test workflow:
$ bash ./build_linux.sh > test_output.txt 2>&1
- Review the output captured in the root-level
test_output.txt. The build
script excludes the benchmark CTest label from its normal test pass, so a
representative passing tail looks like:
100% tests passed, 0 tests failed out of <N>
Label Time Summary:
...
Total Test time (real) = <seconds> sec
Requirements
- Make sure the full test suite passes. The authoritative acceptance line is:
100% tests passed, 0 tests failed out of <N>
Here <N> is the number of Google Test cases discovered by CTest for the active
configuration. Do not pin a historical test count: the discovered count changes
as tests and platform-specific targets change. If any test fails, CTest names
the failing case and --output-on-failure prints its Google Test diagnostics.