v1.3.5 - Dependency Updates + E2E Test Repair
Highlights
- Security: vite 8.0.3 → 8.0.12 closes 2 HIGH CVEs (dev server WebSocket arbitrary file read,
server.fs.denybypass). - E2E suite repaired: previously broken on local Directus instances; now 13/14 passing (1 skipped when fixture collection is absent).
- No breaking changes: all bumps are patch/minor within the same major.
Changed
Patch/minor dependency updates across the toolchain:
- vite 8.0.3 → 8.0.12
- @directus/extensions-sdk 17.1.1 → 17.1.4, @directus/types 15.0.1 → 15.0.3
- vue 3.5.32 → 3.5.34, @vitejs/plugin-vue 6.0.5 → 6.0.6, @vue/test-utils 2.4.6 → 2.4.10
- vitest, @vitest/coverage-v8, @vitest/ui 4.1.2 → 4.1.6
- @playwright/test 1.59.1 → 1.60.0, happy-dom 20.8.9 → 20.9.0
- typescript 6.0.2 → 6.0.3, eslint 10.1.0 → 10.3.0, eslint-plugin-vue 10.8.0 → 10.9.1
- @typescript-eslint/eslint-plugin, @typescript-eslint/parser 8.58.0 → 8.59.3
- dotenv 17.4.0 → 17.4.2
Fixed
- E2E configuration:
DIRECTUS_URLis now scheme-aware — accepts full URLs (http://localhost:8058) while bare hostnames still default tohttps://(backwards-compatible). Playwright configs load.envfrom project root. - E2E UI tests: Replaced Bearer-header auth (incompatible with the Directus Vue admin SPA) with a new
loginAdminUIhelper that performs API login in cookie mode and attaches thedirectus_session_tokento the browser context. - E2E
/server/infoassertion: Modernized for Directus 11 (data.project.*instead ofdata.directus.version). - E2E selector: Fixed a latent CSS+text-engine mix in the content navigation locator.
- E2E fixture handling: Split the ExpandableBlocks-interface test — admin-shell check always runs, the fixture-dependent part skips with a clear reason when the test collection is absent.
Notes
Remaining pnpm audit findings (16 after this release, down from 30) stem from transitive dependencies of @directus/extensions-sdk (axios, lodash-es subdep, fast-xml-parser, unhead, postcss, follow-redirects) and require an upstream Directus SDK release to resolve.
Full diff: v1.3.4...v1.3.5