ContextQMD
Back to Plate

Coverage Priority Map Testing Review Non React Refresh

docs/plans/2026-03-24-coverage-priority-map-testing-review-non-react-refresh.md

53.0.57.2 KB
Original Source

Coverage Priority Map Testing Review Non React Refresh

Inputs

  • Coverage source: lcov.info
  • Constraints:
    • temporary exclude /react and obvious React surfaces outside /react
    • no browser or e2e coverage work
    • no coverage vanity
    • file-first ranking, not package cosplay

Coverage Run

  • Command: bun test --coverage --coverage-reporter=lcov --coverage-dir=.coverage-repo-2026-03-24i --reporter=dots
  • Result: Fresh coverage: 2898 pass, 0 fail, 582 files, 2.47s.

Suite Health

  • pnpm test:profile -- --top 25: clean. No fast-lane threshold breach.
  • pnpm test:slowest -- --top 25: clean. Same story.
  • skipped-test debt scan: none worth fixing
  • commented-out spec scan: none worth fixing
  • cross-spec import scan: none

Scoring Rules

  • Scope is packages/**/src/**.
  • Exclude test files, barrels, declaration files, obvious type-only files, test-support helpers, test-infra packages, zero-value crumbs, and the temporary non-React cut.
  • Reward deterministic transforms, queries, merge helpers, parser or serializer seams, plugin resolution, plugin contracts, overrides, and bounded runtime utilities.
  • Penalize DOM-heavy leftovers, thin wrappers, giant files, and tiny crumb gaps.
  • Missing-from-lcov runtime files are treated as fully uncovered, because pretending they have zero uncovered lines is clown math.
  • package_score is the sum of the top 5 remaining file scores in that package.

Strong Take

Do not do another package sweep.

Raw score says core first, but I would split that into two moves: do the plugin and HTML seam files first, then decide if the DOM-ish static helper is worth it.

Strict batch, sorted by value:

  1. core: getSelectedDomFragment.tsx — score 8
  2. core: resolvePlugin.ts — score 8
  3. core: resolvePlugins.ts — score 7
  4. autoformat: AutoformatPlugin.ts — score 6
  5. callout: BaseCalloutPlugin.ts — score 6
  6. table: withNormalizeTable.ts — score 6
  7. list-classic: withNormalizeList.ts — score 6
  8. core: deserializeHtmlNode.ts — score 6

Threshold Counts

  • score >= 7: 3 files
  • score >= 6: 11 files
  • score >= 5: 52 files
  • score >= 4: 87 files
  • score >= 3: 131 files
  • score >= 2: 174 files
  • score >= 1: 196 files

Packages By Value

  1. core — package score 35, top files getSelectedDomFragment.tsx:8; resolvePlugin.ts:8; resolvePlugins.ts:7; deserializeHtmlNode.ts:6; pipeNormalizeInitialValue.ts:6
  2. markdown — package score 27, top files splitIncompleteMdx.ts:6; deserializeInlineMd.ts:6; mdast.ts:5; convertNodesSerialize.ts:5; customMdxDeserialize.ts:5
  3. list-classic — package score 26, top files withNormalizeList.ts:6; withInsertFragmentList.ts:5; withList.ts:5; insertTodoListItem.ts:5; unwrapList.ts:5
  4. table — package score 26, top files withNormalizeTable.ts:6; withApplyTable.ts:5; getSelectedCellsBorders.ts:5; setBorderSize.ts:5; moveSelectionFromCell.ts:5
  5. docx-io — package score 25, top files importDocx.ts:5; html-to-docx.ts:5; font-table.ts:5; content-types.ts:5; settings.ts:5
  6. slate — package score 24, top files node-entry.ts:5; legacy-editor.ts:5; editor-type.ts:5; deleteText.ts:5; mergeNodes.ts:4
  7. suggestion — package score 21, top files deleteSuggestion.ts:5; withSuggestion.ts:4; removeMarkSuggestion.ts:4; acceptSuggestion.ts:4; insertFragmentSuggestion.ts:4
  8. code-block — package score 17, top files BaseCodeBlockPlugin.ts:5; insertCodeBlock.ts:5; withCodeBlock.ts:3; htmlDeserializerCodeBlock.ts:3; formatter.ts:1
  9. basic-nodes — package score 13, top files BaseHeadingPlugin.ts:4; BaseCodePlugin.ts:3; BaseItalicPlugin.ts:2; BaseUnderlinePlugin.ts:2; BaseStrikethroughPlugin.ts:2
  10. media — package score 13, top files insertImageFromFiles.ts:4; withImageUpload.ts:3; BaseImagePlugin.ts:3; withImageEmbed.ts:3

Best Files By Value

  1. core — getSelectedDomFragment.tsx — score 8, coverage 13.9%, uncovered 31
  2. core — resolvePlugin.ts — score 8, coverage 88.1%, uncovered 8
  3. core — resolvePlugins.ts — score 7, coverage 95.8%, uncovered 15
  4. autoformat — AutoformatPlugin.ts — score 6, coverage 37.0%, uncovered 46
  5. callout — BaseCalloutPlugin.ts — score 6, coverage 0.0%, uncovered 20
  6. table — withNormalizeTable.ts — score 6, coverage 92.7%, uncovered 8
  7. list-classic — withNormalizeList.ts — score 6, coverage 92.5%, uncovered 7
  8. core — deserializeHtmlNode.ts — score 6, coverage 94.4%, uncovered 4
  9. markdown — splitIncompleteMdx.ts — score 6, coverage 95.7%, uncovered 3
  10. markdown — deserializeInlineMd.ts — score 6, coverage 93.3%, uncovered 2
  11. core — pipeNormalizeInitialValue.ts — score 6, coverage 93.8%, uncovered 1
  12. core — SlatePlugin.ts — score 5, coverage 0.0%, uncovered 641
  13. core — BasePlugin.ts — score 5, coverage 0.0%, uncovered 541
  14. utils — plate-types.ts — score 5, coverage 0.0%, uncovered 251
  15. core — SlateEditor.ts — score 5, coverage 0.0%, uncovered 197

Stop Condition

Stop when the remaining misses are mostly wrappers, DOM-only seams, giant sludge, or crumbs that would only move the percentage and not your regression confidence.

Caveats

  • Package totals are noisier than file totals. Trust the file ranking more.
  • The temporary non-React cut hides React-heavy hotspots on purpose. That is a sequencing choice, not proof they are solved.
  • Some DOM-ish non-React files still rank. That is fine, but they come after parser, plugin, and structural seams if you want the smarter order.

Full Data