Feat/playwright runner - #2
Open
Octagon-simon wants to merge 15 commits into
Open
Conversation
Recognise and run Playwright as a test runner alongside the unit runners. Adds @playwright/test detection, a parser for the JSON reporter (resolving spec paths against the report's rootDir, not cwd), best-effort config parsing for baseURL/testDir/webServer, and per-file run commands. Playwright is opt-in only (never auto-detected over Vitest/Jest); detectEnvironment gains type overloads so the no-config path never yields it. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add `lacuna generate --e2e` and `lacuna fix --e2e`, a browser-layer counterpart to the unit/integration agent. generate --e2e: - Route discovery for Next.js (app + pages router) and React Router, dependency-gated so a Vite app with src/pages/ isn't misread as Next. - Per-route DOM snapshot via the project's own Playwright (the webServer starts the app once): aria tree + data-testids (with tag/role), redirect-aware. - DOM-aware spec generation with strict selector discipline (role/label/ testid only, no CSS/XPath, no arbitrary waits, no forced interactions, post-action validation, auth/isolation rules), priority-ordered truncation, parallel workers with the live panel, and flake confirmation. A spec that never goes green is deleted, never left broken. fix --e2e: - Repairs failing specs with a failure-analysis prompt (selector drift, strict-mode violations, sync, redirect) driven by a FRESH snapshot of the failing route, with an explicit repair-priority order that never weakens assertions to force a pass. --inject-testids (opt-in, the only path that edits app source): - Adds data-testid to page sources, then VERIFIES against a re-snapshot: a testid is kept only if it lands on an actual interactive element, else the source is reverted. So a non-forwarding or custom component leaves source untouched regardless of prompt guidance. - Library-aware (lib/flows/ui-libraries.ts + resolve-libraries.ts): detects the UI library by following the import chain, including barrel re-exports, so a custom component behind a barrel isn't mistaken for a library one; falls back to installed deps only when the chain can't be resolved. The agent loop, provider, retry/oscillation machinery, and result parser are shared with the unit path; target selection, context, and verification differ. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
README End-to-end testing section, landing-page feature card, and a full project-map pass: the runner, src/lib/flows/, the e2e-loop, the prompt suite, the data flows, and the E2E invariants. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Resolved conflicts: - src/agent/fix-loop.ts: combined main's keep-best (write bestCode on exhaustion/oscillation, report bestPassCount) with the branch's E2E helperBackups restore — both run; helperBackups is a no-op for unit fixes. - project-map.md: merged the fix-loop entry (keep-best + targeted --file hint + scoped-per-tsconfig type-check from main; E2E mode paragraphs from branch) and took main's loop.ts entry (cause-aware retry wording). - README.md: auto-merged (keep-best + governing-tsconfig wording alongside the E2E docs); also updated the fix-side diagram to "keep the best attempt". main brings: patch-anchor robustness, fix-loop keep-best, monorepo type-check via governing tsconfig, and the v0.2.4 release bump.
E2E runs needed @playwright/test but nothing helped the user get it — `generate --e2e` just printed "none was found" and stopped, and `init` never mentioned Playwright at all. - init: for React/Next.js projects, add an opt-in step after unit setup — "Set up end-to-end (Playwright) testing?" (default no) — that installs @playwright/test + browser binaries. This is the consistent home for a heavy, explicit dependency, matching how init handles the unit-test deps. - playwright.ts: add installPlaywright() and ensurePlaywrightForRun(), the shared guard for the e2e/fix loops. When @playwright/test is missing it prints the exact install command (+ an `lacuna init` pointer) and, only when interactive (TTY, not CI) and not --dry-run, offers a confirm+install inline. The guard runs before the worker pool spawns, so the prompt never races workers; in CI/non-TTY it prints the hint and returns without hanging. - e2e-loop.ts / fix-loop.ts: replace the dead-end "none was found" error with ensurePlaywrightForRun (offerInstall: !dryRun). - Docs: README E2E section + project-map (init, playwright, e2e/fix loops).
… auto-refresh E2E generation now recovers app workflows and asserts real outcomes: - FlowMap (src/lib/flows/flowmap.ts): static AST control→outcome map (toast/ redirect/modal), v2 one-hop cross-file resolution; fed per-control into the prompt so each action asserts its own result. Open-sourced: https://github.com/Octagon-simon/flowmap - Explorer (--deep): walks multi-step flows; drives ARIA comboboxes (widget-driver: https://github.com/Octagon-simon/widget-driver), captures the real success/validation toast, P6 locator typing, section/nav feature boundaries, live per-probe progress. - Auth: auto-refresh stale/missing sessions (refreshAuthState), IndexedDB capture, signed-in readiness wait. - Coverage guards: never go green by deleting tests; keep-best by passing-test count across generate AND fix (countTestFunctions). - Assertion quality: ban persistent-landmark/taut - Fixes: shared-test 30s timeout truncation (scaled test.setTimeout), networkidle on realtime apps, toast-junk capture, e2e fix no longer reads the unit mock / suggests --fix-polluters. - Docs: wiki/, lessons post, README + project-map updates.
Bring the patch-coverage (@diff) + class-aware skeleton work from main into the Playwright/E2E branch. Both feature sets are kept side by side: - generate.ts: unified flags — E2E (--e2e/--route/--max-routes/--inject-testids /--deep) alongside patch-coverage (@diff positional, --improve). E2E path branches first (route-driven, no unit runner); the @diff/dir/file paths follow. - fix-loop.ts: import block reconciled to include BOTH E2E (playwright, snapshot, spec-helpers, e2e prompts) and the scope/patch/format symbols; the merged body keeps the E2E repair path and the --types/scope repair path. - detector.ts: scoped/related coverage commands + Playwright runner coexist. - project-map.md / README.md: rows document both features. Verified: tsc clean, npm run build succeeds, `generate @diff` and `generate --e2e` both execute, patch-coverage unit checks pass. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The E2E generate path called showStarNudge() directly, so a run that generated some specs AND failed others still showed the "thanks, star us" nudge instead of the issue-report nudge — the exact double/contradictory outcome showOutcomeNudge exists to prevent. Route it through showOutcomeNudge(specsGenerated, specsFailed, 'generate') like the unit path, which shows the issue nudge when failed > succeeded, else the star. showStarNudge is no longer imported directly here. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
| GitGuardian id | GitGuardian status | Secret | Commit | Filename | |
|---|---|---|---|---|---|
| 34577921 | Triggered | Generic Password | 6e87a1b | src/agent/e2e-loop.ts | View secret |
🛠 Guidelines to remediate hardcoded secrets
- Understand the implications of revoking this secret by investigating where it is used in your code.
- Replace and store your secret safely. Learn here the best practices.
- Revoke and rotate this secret.
- If possible, rewrite git history. Rewriting git history is not a trivial act. You might completely break other contributing developers' workflow and you risk accidentally deleting legitimate data.
To avoid such incidents in the future consider
- following these best practices for managing and storing secrets including API keys and other credentials
- install secret detection on pre-commit to catch secret before it leaves your machine and ease remediation.
🦉 GitGuardian detects secrets in your source code to help developers and security teams secure the modern development process. You are seeing this because you or someone else with access to this repository has authorized GitGuardian to scan your pull request.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.