Test Smell Detector for VS Code

This repository contains a minimum VS Code extension prototype for integrating the Java-based Test Smell Detector (tsDetect).

Current prototype

The extension contributes one command:

• Test Smell Detector: Analyze Current File

The command validates the active editor, asks for the corresponding production Java file, runs the bundled tsDetect engine, and publishes detected test smells to both the Test Smell Detector Output panel and VS Code's Problems panel.

The extension also performs incremental analysis when a Java test file is saved. Only the saved test/production pair is analyzed, and only that test file's diagnostics are refreshed. Automatic analysis can be disabled with the testSmellDetector.analyzeOnSave setting.

Each detected smell is represented as a warning diagnostic. Because the current tsDetect output reports smells at file level rather than exact source locations, the warning is anchored to the Java test class name. VS Code also shows a warning squiggle and its standard diagnostic hover at that location.

tsDetect uses a test/production file pair. The extension first remembers a production file selected during a manual analysis. It can also match common names such as CalculatorTest.java to Calculator.java and prefers files under src/main. If an automatic run cannot find a match, run Analyze Current File once and select the production file manually.

Requirements

• Java must be installed and available as java on the system path.
• The test and production Java files must be saved locally.

If Java is installed under a different executable path, set testSmellDetector.javaPath in VS Code settings.

Run locally

1. Run npm install.
2. Open this folder in VS Code.
3. Press F5 to open an Extension Development Host.
4. In the new window, open a saved Java test file and keep it as the active editor.
5. Open the Command Palette and run Test Smell Detector: Analyze Current File.
6. Select the corresponding production Java file when prompted.
7. Review the detected smells in the Problems panel and the detailed run summary in the Test Smell Detector Output panel.
8. Modify and save the Java test file. The extension reruns tsDetect after a short delay and replaces that file's diagnostics automatically.

Incremental analysis

• Runs only for saved local Java test files.
• Recognizes common test suffixes such as Test, Tests, IT, and Spec, as well as Java files under src/test.
• Waits briefly after a save to avoid redundant work.
• Serializes detector runs for the same file and ignores outdated results.
• Allows different test files to be analyzed independently.
• Removes stale diagnostics as soon as a Java file changes.
• Can be disabled through the testSmellDetector.analyzeOnSave setting.

Automated checks

The GitHub Actions workflow runs for pull requests targeting main, pushes to main, and manual workflow runs. It installs dependencies with npm ci, type-checks and compiles the TypeScript extension, compares all regression fixtures with their expected tsDetect results, and builds an installable VSIX.

The same checks can be run locally with Node.js 22 or newer and Java 11 or newer:

npm ci
npm run check
npm run compile
npm run test:corpus
npm run package:vsix

npm run test:corpus expects the compiled adapter in out/, so run npm run compile first. The packaging command creates test-smell-detector.vsix. In GitHub Actions, that file is retained as the test-smell-detector-vsix workflow artifact.

Prototype limitations

• Production-file matching follows common project and filename conventions; unusual project layouts may require one manual selection.
• Results are reported at file level, so diagnostics mark the test class rather than the exact expression that caused a smell.