Back to Plate

Slate v2 Triage Close Queue

docs/slate-issues/triage-close-queue.md

53.0.67.4 KB
Original Source

Slate v2 Triage Close Queue

Purpose

This is the first safe bulk-triage queue for open Slate issues.

It is now also a historical execution record.

It is intentionally conservative.

It does not include:

  • stale-only closures
  • medium or low confidence closures
  • direct-v2 issues, even if they smell duplicate or invalid

Execution Status

Batch A was executed by Dylan after this queue was prepared.

Verified post-run status:

  • Batch A queue size: 54
  • Batch A issues still open: 0
  • live repo open-issue count after Batch A: 628

That means the 682-issue research snapshot is still the analysis baseline, but this queue is no longer a pending first-run artifact.

Do not rerun Batch A blindly.

Confidence Review

Action Matrix

ActionCountShape
close-invalid6337 invalid, 26 likely-invalid
close-stale5756 stale-candidate, 1 valid
mark-duplicate106 valid, 3 duplicate-candidate, 1 likely-valid
close-duplicate88 duplicate-candidate

Safe-Now Rule

Safe now means:

  • close-invalid
    • confidence = high
    • v2 relevance != direct
  • mark-duplicate or close-duplicate
    • confidence = high
    • explicit duplicate target exists
    • v2 relevance != direct

That yields:

  • safe now: 54
  • manual review: 27
  • stale bucket deferred entirely: 57

Spot Check

Representative dossier spot-checks passed:

  • #6007: duplicate to #6002
  • #5958: upstream Chromium issue, likely not Slate-owned
  • #5748: already answered in-thread, API expectation mismatch
  • #5655: consumer misuse, maintainer already gave working integration shape
  • #5457: pure CSS/support confusion
  • #4561: old duplicate already pointed at canonical thread

Batch A: Executed

A1. Mark Duplicate

IssueTargetWhy
#6007#6002release-process duplicate already identified in-thread
#4929#4882deserialize-example stack overflow duplicate
#4895#4914readonly copy duplicate
#4580#4573paste HTML crash duplicate
#4561#4081hot refresh duplicate

A2. Close Duplicate

IssueTargetWhy
#3710#4268paste-html example extra-newlines duplicate

A3. Close Invalid

IssueWhy
#5958thread points to upstream Chromium bug
#5895consumer/editor-behavior expectation, not a clean Slate bug
#5838third-party slate-vue3 announcement noise
#5824iOS style-preserving paste expectation, not current Slate-owned bug
#5820invalid path/leaf expectation mismatch
#5748already answered in-thread, API usage mismatch
#5735animation request, not Slate bug
#5732noise
#5704popup textbox selection expectation, not Slate-owned
#5692legacy browser support expectation
#5671generic support complaint
#5667invalid state wiring in consumer array mapping
#5655consumer misuse; maintainer already showed working dnd-kit shape
#5548docs/support mismatch around root path semantics
#5498consumer styling/padding issue
#5478old non-hooks support ask
#5457CSS line-height/support confusion
#5432invalid link-focus expectation
#5413focus prerequisite/support confusion
#5379styling flag expectation, not engine bug
#5333“markdown editor out of the box” expectation
#5280broad paste-format complaint without clean Slate bug seam
#5250invalid current ticket as framed; should not stay open as bug
#5174iOS autofocus limitation/support issue
#5132invalid value / stale-read consumer misuse
#5101invalid decoration/search-highlight framing
#5051old slate-html-serializer ecosystem issue
#5032broad DOM-point complaint without stable current repro
#4827Chrome extension interference
#4800import/setup misuse
#4692product request, not issue
#4648punctuation semantics expectation, weak bug claim
#4513collapsed-selection onCopy expectation in old stack
#4438LSP feature request, not bug
#4322vague ecosystem issue, not clean core bug
#4281Firefox shortcut expectation
#4239chrome extension integration noise
#4146example/integration misuse
#4145support question
#4029community chat noise
#3954invalid current framing around empty non-void copy
#3870invalid old void-boundary report as filed
#3826vague mark-removal issue, not credible current bug
#3765old serializer ecosystem issue
#3763support/API usage question
#3753old invalid framing around empty-editor insertion
#3453support request for IDs in React component
#3935example limitation, not core bug

Batch B: Manual Review Before Any Bulk Action

These are not safe for an automatic first sweep.

Reasons:

  • direct-v2 relevance
  • medium or low confidence
  • duplicate target exists but the thread still smells like a real canonical bug family
IssueProposed actionWhy it stays out of batch A
#6016close-invaliddirect-v2, identity/runtime signal
#5912close-invalidonly medium confidence
#5698mark-duplicate to #5711medium confidence and direct-v2
#5686close-invaliddirect-v2 and touches real API semantics
#5647close-invalidonly medium confidence
#5634close-invalidonly medium confidence
#5605mark-duplicate to #6013direct-v2 API ergonomics
#5550close-invalidonly medium confidence
#5472close-invalidlow confidence
#5417close-invalidonly medium confidence
#3798close-invalidonly medium confidence
#3723close-duplicate to #3421only medium confidence
#3705close-duplicate to #3921medium confidence and direct-v2 history signal
#3621close-invalidonly medium confidence
#3466close-duplicate to #3339medium confidence and direct-v2
#3433close-duplicate to #3412only medium confidence
#3369close-duplicate to #3824direct-v2 mobile/input signal
#3313close-duplicate to #3112direct-v2 mobile/input signal
#5066mark-duplicate to #5762direct-v2 mobile/input signal
#5026mark-duplicate to #4994direct-v2 mobile/runtime signal
#5019mark-duplicate to #4988direct-v2 mobile/input signal
#4971close-invalidonly medium confidence
#4936close-invalidmedium confidence and indirect-v2
#4680close-invalidmedium confidence and indirect-v2
#4377close-invalidonly medium confidence
#3977close-invalidmedium confidence and indirect-v2
#2597close-duplicate to #1971direct-v2 sentinel/runtime issue

Explicitly Deferred

Stale Queue

57 issues are currently marked close-stale.

Do not bulk-close them yet.

That bucket needs one more pass because stale is where fake confidence goes to hide.

  1. Treat Batch A as executed historical record, not an open queue.
  2. Start from Batch B manual review if more triage work is needed.
  3. Review stale after that, not before.
  4. If another bulk script is generated later, refresh GitHub state first instead of reusing Batch A assumptions.