mirror of
https://github.com/supabase/supabase.git
synced 2026-10-06 01:45:10 +03:00
codex/fix-tanstack-e2e
2
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Miranda LimonczenkoandClaude Opus 5
|
f10f00ae69 |
fix(e2e): install e2e-shared when CI filters to a single suite (#48960)
## I have read the [CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md) file. YES ## What kind of change does this PR introduce? Bug fix. Unblocks the WWW E2E check on `master`. ## What is the current behavior? The WWW E2E job fails before running any test: ``` Error: Cannot find package '@axe-core/playwright' imported from /home/runner/_work/supabase/supabase/e2e/shared/axe.ts Error: No tests found ``` Both E2E workflows install with a filter: ``` pnpm install --frozen-lockfile --filter=e2e-www... ``` The `...` suffix pulls in a package's declared dependencies. Neither `e2e-www` nor `e2e-docs` declared `e2e-shared`; both reach it through relative imports such as `../../shared/axe.ts`, which pnpm's dependency graph cannot see. So the filter selected one project, `e2e/shared/node_modules` was never created, and Node resolving `@axe-core/playwright` from `e2e/shared/axe.ts` walked up to a root that does not carry it under pnpm's isolated layout. `e2e-docs` is broken the same way. It had not run against the shared module yet, so it has not gone red. ## What is the new behavior? `e2e-shared` is declared as a workspace dependency of both suites, so the filter installs it. | | Filter scope | Importing `e2e/shared/axe.ts` | | --- | --- | --- | | Before | 1 of 28 projects | `Cannot find package '@axe-core/playwright'` | | After | 2 of 28 projects | Imports cleanly | The lockfile gains two `link:../shared` entries and no new downloads. ## Manual Testing 1. Check out this branch and delete the shared package's modules: `rm -rf e2e/shared/node_modules` 2. Run the command CI runs: `pnpm install --frozen-lockfile --filter=e2e-www...` 3. Confirm the output reports `Scope: 2 of 28 workspace projects` and that `e2e/shared/node_modules` exists again. 4. Repeat steps 1 - 3 with `--filter=e2e-docs...`. ## Additional context Fixing only the workflow lines, by adding a second `--filter=e2e-shared`, would work as well. Declaring the dependency was chosen instead because the dependency is real and every consumer of the filter gets it, not just the two workflow files. The imports stay relative. Declaring the workspace dependency is enough to get the package installed, so no import paths change in this PR. <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **Chores** * Updated end-to-end test packages to use shared testing utilities at runtime. * Improved consistency between documentation and website test suites. <!-- end of auto-generated comment: release notes by coderabbit.ai --> Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |
||
|
Miranda LimonczenkoandClaude Opus 5
|
6d3a4bcc48 |
feat(www) Add scaffolding for WWW E2E tests and CI check (#48861)
Closes DOCS-1278 ## I have read the [CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md) file. YES ## What kind of change does this PR introduce? Feature. Adds E2E test scaffolding and a CI check for the marketing site. ## What is the current behavior? Closes [FE-4047](https://linear.app/supabase/issue/FE-4047). The marketing site has no E2E coverage. Docs has a suite in `e2e/docs`, but its runner, git helpers and axe reporting are private to that package, so a second site cannot reuse them. ## What is the new behavior? * **A www suite scoped to changed content.** Changed `.mdx` files in `_blog`, `_events`, `_customers` and `_alternatives` map to the URLs they render. Pages with `disable_page_build: true` are skipped because they 404 by design. Capped at 20 pages. Enforces `heading-order` and `page-has-heading-one`, matching docs. * **`e2e/shared` The docs site is also static with similar needs. This folder shares the docs logic with www. * **A CI check that is safe to mark required.** Path scoping lives in a `Detect changed paths` step rather than a `paths:` trigger, so the check reports on every pull request instead of being skipped. `waitForVercelDocsPreview.js` becomes `waitForVercelPreview.js`, shared by both workflows. ## How the check behaves The job always reports a check run, so it is safe to mark required. Path scoping happens in a step rather than a `paths:` trigger, which would leave non-www pull requests waiting on a check that never reports. | Case | Behavior | | --- | --- | | Fork pull request adds new pages | Passes without testing. The Vercel wait is gated on `head.repo.full_name == github.repository`, so forks resolve no preview URL. The job emits a `::warning` and a job summary containing a ready-to-run `gh workflow run www-e2e.yml` command with the resolved page paths, so a maintainer can run it against the preview. | | Vercel preview times out or fails | Passes without testing. The wait step is `continue-on-error: true`, so a 900s timeout or a failed deployment leaves the URL unset and the suite skips. Vercel's own `Vercel – zone-www-dot-com` check already reports the failure. | | Draft pull request | Job does not run at all, gated at the job level on `pull_request.draft == false`. `ready_for_review` is in the trigger's `types`, so marking it ready runs the check. | | Another app changed, www untouched | Job runs and every step skips. The `www` filter matches only the four content directories, `e2e/www`, `e2e/shared`, the lockfile, and this workflow. | | Only the harness changed | Passes without testing. Scope resolves to zero pages, and the Vercel wait is additionally gated on `www_app`, so it does not wait for a preview Vercel skipped. | | No preview resolves, any reason | Skips rather than falling back to production. Production does not serve pages the pull request adds, so testing it would fail a valid change. | ### Not covered Changes to `apps/www` components and routes do not trigger this check — only the four content directories do. A follow-up can check global components such as the navigation and the footer. ## Manual testing 1. Start the site: `pnpm dev:www` 2. Run `pnpm e2e:www` with no www content changed. It should resolve zero pages and skip Playwright, not fail. 3. Touch a post, then run `pnpm e2e:www` again: `echo "" >> apps/www/_blog/2024-01-01-some-post.mdx`. The resolved `/blog/...` path should be listed before Playwright starts. 4. Run against production with no local server: `PLAYWRIGHT_BASE_URL=https://supabase.com WWW_E2E_PAGE_PATHS=/blog/postgres-language-server pnpm e2e:www` 5. Point step 4 at a page with a known heading problem. The failure should name the rule, the CSS selector and the markup. 6. Confirm docs still passes on the shared runner: `pnpm dev:docs`, then `pnpm e2e:docs` <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **New Features** * Added WWW end-to-end testing for affected content pages, including accessibility checks. * Added standard and full-site test commands, configurable preview testing, and failure reports. * Added shared utilities for page discovery, accessibility scanning, and test execution. * **Documentation** * Documented WWW test setup, coverage, debugging, CI behavior, and running checks against production or preview environments. * **Improvements** * Updated documentation test workflows to better identify affected changes and handle preview environments. <!-- end of auto-generated comment: release notes by coderabbit.ai --> --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com> |