Back to Plate

Slate Issues Gitcrawl Rebuild

docs/plans/2026-05-04-slate-issues-gitcrawl-rebuild.md

53.0.63.2 KB
Original Source

Slate Issues Gitcrawl Rebuild

Goal

Rebuild docs/slate-issues from a live gitcrawl mirror instead of the old pre-v2 frozen snapshot. The output should be better triage, not a prettier copy of the April ledger.

Hard Rules

  • gitcrawl is the archive-first source for candidate discovery, neighbors, clusters, and stale/duplicate pressure.
  • Live GitHub remains final truth for issue state when a claim depends on current status or comments.
  • Do not keep the 682-issue snapshot as the active truth if live data disagrees.
  • Keep exact Fixes #... claims limited to issue-level proof in docs/slate-v2/ledgers/issue-coverage-matrix.md.
  • Document gaps brutally: missing mirror coverage, weak clusters, duplicate-uncertain families, stale rows, no-repro rows, and v2 proof holes.
  • No GitHub comments, label writes, closes, commits, pushes, or PRs.

Source Files To Rebuild Or Reconcile

  • docs/slate-issues/open-issues-ledger.md
  • docs/slate-issues/open-issues-dossiers.md
  • docs/slate-issues/test-candidate-map.md
  • docs/slate-issues/benchmark-candidate-map.md
  • docs/slate-issues/issue-clusters.md
  • docs/slate-issues/package-impact-matrix.md
  • docs/slate-issues/requirements-from-issues.md
  • docs/slate-issues/roadmap-from-issues.md
  • docs/slate-issues/triage-close-queue.md

Execution Phases

PhaseStatusOutput
1. Mirror refreshcompleteFresh gitcrawl sync for ianstormtaylor/slate open issues and comments.
2. Cluster buildcompletegitcrawl cluster output plus weak-cluster and singleton gap notes.
3. Gap auditcompleteCurrent docs vs live mirror deltas: closed, new, missing, stale, weak duplicate candidates.
4. Corpus rebuildcompleteUpdated docs/slate-issues summary docs and generated/detailed ledgers.
5. V2 claim reconciliationcompleteIssue coverage matrix claims left unchanged; new live rows stay untriaged until issue-level proof.
6. VerificationcompleteFormatting, row/count scripts, bun run completion-check.

Current Findings

  • Existing corpus is a 2026-04-02 frozen 682-issue snapshot.
  • Gitcrawl live mirror now reports 630 open issues and 29 open PRs: 659 open threads total.
  • Gitcrawl hydrated 1856 comments/reviews, embedded 659 rows, and produced 617 clusters.
  • Only 28 clusters are multi-member; 589 are singletons. Raw gitcrawl clustering is useful for candidate discovery but too conservative for the Slate v2 architecture taxonomy.
  • New live issues absent from the frozen ledger: #6051, #6053.
  • Frozen rows no longer live-open: 54; GitHub confirms all are closed.
  • V2 exact issue claims are unchanged. The new live ledger preserves old sync status where available and marks new live rows as untriaged-live-new.