Skip to main content

Runs and tests

Symptoms​

  • the agent cannot run tests or the build;
  • a run uses the wrong environment;
  • the agent cannot write suitable unit tests;
  • coverage stays low;
  • a test passes and fails intermittently;
  • you need to connect your work to a TMS.

What to check​

  1. Whether there is a working run configuration in the IDE.
  2. Whether the correct SDK, working directory, and environment variables are selected.
  3. Whether the feature is supported for your IDE and language in the feature matrix.
  4. Whether the original test passes manually through the same configuration.
  5. Whether the run or the required files are blocked in Tools, Edit Scope, or the ignore files.

Fix​

For tests and builds, use IDE run configurations rather than an arbitrary terminal command. Start with a single test method or class, then widen the check.

If ordinary unit test generation does not fit a complex scenario, create a test from a real execution. For uncovered branches, use coverage increase. If a test is unstable, run flaky test analysis, and repeat the analysis after the fix.

To work with issues and test cases, connect a TMS.