Ryan Kegel
a8d7427a95
Some checks failed
Chore App Build, Test, and Push Docker Images / build-and-push (push) Failing after 1m44s
- Added E2E test setup in `auth_api.py` with `/e2e-seed` endpoint for database reset and test user creation. - Integrated Playwright for end-to-end testing in the frontend with necessary dependencies in `package.json` and `package-lock.json`. - Created Playwright configuration in `playwright.config.ts` to manage test execution and server setup. - Developed new skills for Playwright best practices, visual regression, smoke test generation, and self-healing tests. - Implemented new test cases for chore creation in `chores-create.smoke.spec.ts` and `chores-create.spec.ts`. - Added page object models for `ChildEditPage` and `LandingPage` to streamline test interactions. - Updated `.gitignore` to exclude Playwright reports and test results. - Enhanced documentation in `copilot-instructions.md` for testing and E2E setup.
14 lines
798 B
Markdown
14 lines
798 B
Markdown
---
|
|
name: playwright-healer
|
|
description: Analyzes failing Playwright tests and suggests automated fixes based on UI changes.
|
|
---
|
|
# Playwright Self-Healing Instructions
|
|
When a user asks to "fix" or "heal" a test:
|
|
|
|
1. **Analyze the Trace**: Use the Playwright MCP or CLI to open the latest trace file in `.playwright-cli/traces/`.
|
|
2. **Compare Snapshots**: If a locator failed, take a fresh `snapshot` of the page. Identify if the element moved, changed its ARIA role, or had its text updated.
|
|
3. **Propose the Patch**:
|
|
- If the UI changed, suggest the updated locator.
|
|
- If it's a timing issue, suggest adding an `expect(locator).toBeVisible()` wait.
|
|
- If it's a data issue, check the mock definitions.
|
|
4. **Verify**: Run the patched test once before presenting the final code to the user. |