[lucasbraum]

Pack

vitest, >=4 <5 → >=5 <6. Status: candidate (1 false positive among breaking findings; see below). Queue entry set to candidate. No pack-request issue existed for it.

Sources

• Migrating to Vitest 5.0: https://vitest.dev/guide/migration (vitest.dev is not reachable from this sandbox; read as docs/guide/migration/index.md of vitest-dev/vitest at 69e6369)
• Vitest 5.0.0 release notes: https://github.com/vitest-dev/vitest/releases/tag/v5.0.0
• Benchmarking guide: https://vitest.dev/guide/benchmarking
• Extending matchers: https://vitest.dev/guide/extending-matchers

No official codemod exists for this major.

Rules and behavior notes

Rules:

• matcher-types (breaking, detect + compiler claims, agent): matchers on the global jest.Matchers, or on a one-parameter Assertion<T> / Matchers<R>. Site = the declare global block or the interface.
• jest-dom-matchers (breaking, detect, agent): import '@testing-library/jest-dom[/vitest]', whose types no longer reach expect.
• bench-api (breaking, type diff, agent): bench and the Bench* types removed.
• removed-entrypoints (breaking, detect, rewrites): vitest/coverage and vitest/reporters → vitest/node, vitest/environments and vitest/snapshot → vitest/runtime.
• removed-entrypoints-manual (breaking, detect, agent): vitest/runners, vitest/suite, vitest/mocker, vitest/internal/module-runner.
• sequential-removed (breaking, type diff, agent): test.sequential, describe.sequential, sequential option.

Behavior notes (all "Decisions for you", some also test follow-ups): coverage-patterns (also a finding, found from source: bare directory in coverage.include/exclude), clear-mocks-default, test-name-pattern, inline-projects-extend, hoisted-calls-top-level, mock-behavior, assertion-timing, pretty-format-output, fake-timers-temporal, artifact-locations, config-resolution, browser-mode, environment-requirements, node-api.

Ground-truth repositories

• votingworks/vxsuite @ 5cb3c7d9c1abc891e7b12de57c0278790dd6160a, upgrade commit bb04c5fb01aa9bcd285170d2277df04ad5d7e09c (vitest 4.1.11 → 5.0.2, with @vitest/coverage-istanbul 5.0.2): 30 expected findings (3 matcher-types, 11 jest-dom-matchers, 3 bench-api, 13 coverage-patterns).
• vitorvasc/opentelemetry-ecosystem-explorer @ 46d5b37d00bbf27f394e33cee64a0e206c31315f (directory ecosystem-explorer), upgrade PR head 2b0497b35d0e49dbeeb0d6174ad09aed03d46697 (4.1.11 → 5.0.0): 2 expected findings (jest-dom-matchers).

Expected findings come from pnpm packs:truth (their compiler plus the upgrade commit's changed lines), read by hand; the 169 compiler diagnostics under vxsuite's libs/image-utils/src/vitest_setup.ts are evidence of that one declaration and are recorded once at it. Exclusions are listed in the $comment of ground-truth.json.

Considered and dropped: siemens/element (a real migration, but its @simpl/* and @siemens-ux/* packages come from code.siemens.com, so it cannot install from a public registry); google/budoux, ckeditor/ckeditor5-vue, gopherium/gophenberg, justinlocsei/larkspur and abraham/nutmeg (version bumps with no source change, so no site).

uptide pack test output

New pack (pnpm uptide pack test vitest --write; the ~18,900 generic (unverified) lines are omitted):

uptide pack test vitest · candidate
fixtures      5 cases, 17 sites · pass
ground truth
  votingworks/vxsuite@5cb3c7d9c1ab             vitest 4.1.11 → 5.0.2  predicted 15334, expected 30, false positives 15320, false negatives 16
    with @vitest/coverage-istanbul 4.1.11 → 5.0.2
  vitorvasc/opentelemetry-ecosystem-explorer@46d5b37d00bb vitest 4.1.11 → 5.0.0  predicted 3621, expected 2, false positives 3619, false negatives 0

  rule                         precision  recall    tp   fp   fn
  bench-api                          15%     67%     2   11    1
  coverage-patterns                 100%      0%     0    0   13
  generic                             0%    100%     0 18927    0
  jest-dom-matchers                 100%     85%    11    0    2
  matcher-types                      75%    100%     3    1    0
  overall (sites)                     0%     50%    16 18939   16
  breaking (sites)                   94%     50%    16    1   16

false positives among breaking:
  votingworks/vxsuite  libs/image-utils/src/jest_pdf_snapshot.ts:42  matcher-types (breaking)

false negatives (16)
  13 x coverage-patterns (vitest.config.ts files that import a local `defineConfig` wrapper, not vitest)
  jest-dom-matchers: apps/print/frontend/src/setupTests.ts:2, libs/dev-dock/frontend/src/setupTests.ts:1 (files that do not import vitest)
  bench-api: libs/hmpb/src/render_ballot.bench.ts:16

Existing packs, rerun in this sandbox, all unchanged:

zod     verified  overall 100% / 84%, breaking 100% / 71%, 0 false positives  (PASS)
stripe  verified  PASS
ai      verified  PASS
react   verified  excalidraw + tldraw, breaking 100% precision (89 sites, 0 false), recall 69%  (PASS, same numbers as the recorded verification)

react note: the first run could not install tldraw (its packageManager is yarn@4.7.0 and the yarn on PATH here is 1.22.22; the installer runs it directly). That is an environment issue, not a regression. With a corepack shim on PATH (corepack enable --install-directory <dir>), the rerun matched the recorded numbers.

pnpm lint, pnpm typecheck and pnpm test pass.

What is not covered

• Why it is a candidate: one false positive among breaking findings. jest_pdf_snapshot.ts:42 is a real new TS2551 (toMatchImageSnapshot missing on Assertion), but it is a symptom of the declaration at vitest_setup.ts:10, which the pack reports. check did not trace it to that cause (a global jest.Matchers augmentation), so it stands as a second breaking site. The ground truth records one site at the cause, as docs/packs.md says, and I did not add the symptom to pass.
• Pack detect runs only on files that import the package, so these expected sites are missed: 13 bare-directory coverage entries (configs that import a shared defineConfig wrapper), and jest-dom imports in two setup files that import nothing from vitest. Recall on breaking sites is 50%.
• check could not resolve imports in 69 of 71 vxsuite files (workspace packages are not built), so only a few of the 169 compiler diagnostics are visible there.
• bench-api claims 11 inferred unverified sites on Bench* types in vxsuite that the upgrade did not touch (overall tally only, not breaking), and the describe wrapper at render_ballot.bench.ts:16 is missed.
• Rules and notes with no evidence in either repo, resting on fixtures and the guide only: removed-entrypoints*, sequential-removed, nested vi.mock, and all behavior notes except coverage-patterns.
• A second repo that really migrated and installs here is needed for verified to mean more than one monorepo: the other candidates found are version-only bumps.

Tooling friction

• pnpm uptide check from source fails inside its worker (Cannot find module .../packages/core/src/errors.js); pnpm build and the dist CLI work. pack test is unaffected.
• pnpm packs:truth requires full 40-character SHAs, and for a file the upgrade moved it prints the new path with the old line number; the paths in the ground truth were read from git diff -M.
• The README packs table includes candidates and the README test caps it at 150 lines; adding the row went one line over, so this PR joins two lines of the documentation list (no content change).
• vitest.dev and its guide are not reachable from the sandbox; the guide was read from the vitest repo over git.

🤖 Generated with Claude Code

https://claude.ai/code/session_01YNCujWu9MXTm48xoYVKHJC

---

Generated by Claude Code

[lucasbraum]

Superseded by #63

---

Generated by Claude Code