ContextQMD
Back to Plate

Slate Tranche 3 Execution

docs/slate-v2/slate-tranche-3-execution.md

53.0.68.0 KB
Original Source

Slate Tranche 3 Execution

Goal

Do an honest tranche-3 review for packages/slate against:

  • live plate-2 tranche/ledger docs
  • current .tmp/slate-v2/packages/slate/**
  • missing draft-backed proof owners under .tmp/slate-v2-draft/packages/slate/**
  • missing perf gate ownership

This file is prior context only now.

The mutable execution-state owner was moved to:

Execution Model

  • control/docs repo: plate-2
  • execution repo: .tmp/slate-v2
  • package scope: .tmp/slate-v2/packages/slate/**

Current Tactic

  • burn down the missing tranche-3 kept owners from smallest honest slice to largest
  • prefer owner files that either:
    • prove an already-existing seam, or
    • expose one narrow missing public seam without dragging in a broader lane

Pivot History

  • rejected the false “package-local suite green means tranche complete” read
  • rejected starting with bookmarks first because that would blob together API, refs, durability semantics, and later decoration pressure
  • chose text-units before interfaces-contract because it was the smallest real missing public seam plus missing proof owner
  • chose transaction-contract before bookmarks because it could expose core-owned transaction drift without reopening broader API/design lanes
  • chose headless-contract before bookmarks because it was small enough to falsify package-local ownership and exposed a narrow static-surface gap
  • pivoted out of the cheap owner-file phase once the driver gate returned to a real core normalization row
  • replan triggered after full package-local gates went green again while the remaining-owner ledger stayed non-empty
  • pivoted away from range-ref-contract.ts as the next slice because the first probe exposed a deeper public-vs-internal ref publication design seam

Questions

  1. What is actually still missing for packages/slate?
  2. Which missing items are package-local proof/contract gaps vs broader migration-program gaps?
  3. Which live plate-2 docs drifted into a false “slate is done” story?
  4. What should the roadmap say now?

Method

  1. Read live tranche-3 owner docs in plate-2.
  2. Inventory current slate-v2/packages/slate/test/** and slate-v2/packages/slate/src/**.
  3. Compare against the draft-backed proof-owner list already named in slate-legacy-draft-contract-corpus.md.
  4. Classify missing items:
    • package-local must-have now
    • package-local maybe later / post RC
    • broader non-slate program work
  5. Rewrite live docs to match the honest classification.

Findings

Draft-backed proof owners already named by the corpus

The corpus doc says tranche 3 keeps these proof-owner lanes alive:

  • query-contract.ts
  • legacy-editor-nodes-fixtures.ts
  • interfaces-contract.ts
  • legacy-interfaces-fixtures.ts
  • transforms-contract.ts
  • operations-contract.ts
  • legacy-transforms-fixtures.ts
  • surface-contract.ts
  • transaction-contract.ts
  • snapshot-contract.ts
  • bookmark-contract.ts
  • range-ref-contract.ts
  • clipboard-contract.ts
  • normalization-contract.ts
  • headless-contract.ts
  • text-units-contract.ts

Current read

  • package-local owner files are landed
  • package-local correctness/build/type/lint gates are green
  • the #6038 benchmark lane exists and runs
  • the broad standalone oracle file .tmp/slate-v2/packages/slate/test/snapshot-contract.ts is not part of the package-closeout proof by default and is currently red when run directly

Remaining Items

Package-local must-have now

  • none

Perf Owner Status

  • landed:
    • #6038 benchmark lane for transaction execution and mixed structural updates
  • command:
    • bun run bench:slate:6038:local

Current Gates

  • targeted correctness:
    • cd .tmp/slate-v2 && bun test ./packages/slate/test/text-units-contract.ts
  • package closeout gates:
    • cd .tmp/slate-v2 && bun test ./packages/slate/test
    • cd .tmp/slate-v2 && bunx turbo build --filter=./packages/slate
    • cd .tmp/slate-v2 && bunx turbo typecheck --filter=./packages/slate
    • cd .tmp/slate-v2 && bun run lint:fix
    • cd .tmp/slate-v2 && bun run lint

Latest Slice

  • landed packages/slate/src/text-units.ts
  • exported the seam from packages/slate/src/index.ts
  • landed packages/slate/test/text-units-contract.ts
  • landed packages/slate/test/interfaces-contract.ts
  • landed packages/slate/test/transaction-contract.ts
  • landed packages/slate/test/headless-contract.ts
  • landed packages/slate/test/bookmark-contract.ts
  • landed packages/slate/test/normalization-contract.ts
  • landed packages/slate/test/transforms-contract.ts
  • landed packages/slate/test/extension-contract.ts
  • landed packages/slate/test/surface-contract.ts
  • landed packages/slate/test/clipboard-contract.ts
  • landed packages/slate/test/range-ref-contract.ts
  • landed scripts/benchmarks/slate/6038-transaction-execution.mjs
  • landed package.json#bench:slate:6038:local
  • fixed transaction rollback in packages/slate/src/core/public-state.ts so restored draft state keeps the pre-transaction runtime-id index
  • recovered missing static Editor.getFragment(editor) in packages/slate/src/interfaces/editor.ts
  • recovered bookmark surface:
    • packages/slate/src/interfaces/bookmark.ts
    • packages/slate/src/editor/bookmark.ts
    • Editor.bookmark(...)
    • root Bookmark* exports through interfaces/index.ts
  • verification:
    • bun test ./packages/slate/test/text-units-contract.ts
    • bun test ./packages/slate/test/interfaces-contract.ts
    • bun test ./packages/slate/test/transaction-contract.ts
    • bun test ./packages/slate/test/headless-contract.ts
    • bun test ./packages/slate/test/bookmark-contract.ts
    • bun test ./packages/slate/test/normalization-contract.ts
    • bun test ./packages/slate/test/transforms-contract.ts
    • bun test ./packages/slate/test/extension-contract.ts
    • bun test ./packages/slate/test/surface-contract.ts
    • bun test ./packages/slate/test/clipboard-contract.ts
    • bun test ./packages/slate/test/range-ref-contract.ts
    • bun test ./packages/slate/test
    • bunx turbo build --filter=./packages/slate
    • bunx turbo typecheck --filter=./packages/slate
    • bun run lint:fix
    • bun run lint
    • bun run bench:slate:6038:local
  • current driver-gate return:
    • green again on the current tree after the normalization split repair

Current Status

  • package-local correctness/build/type/lint gates are green
  • package-local benchmark owner exists and runs
  • this closeout read is stale
  • packages/slate is reopened under the absolute-api replan
  • standalone broad-oracle proof from snapshot-contract.ts is green again
  • the active remaining owner is API direction + explicit-cut classification, not missing package-local proof files
  • package-local perf coverage is now tracked separately in:
    • docs/plans/2026-04-19-slate-core-perf-coverage-plan.md
  • do not treat the old “move to tranche 4 next” read as the current execution order when that perf plan is active
  • do not treat this file as permission to move to tranche 4

Broader program work, not packages/slate-local

  • support-package work remains blocked until the slate core redesign is honestly settled:
    • slate-history
    • slate-hyperscript
  • tranche 5:
    • slate-dom
  • tranche 6:
    • slate-react
  • tranche 7+:
    • example parity
    • root proof closure
    • overlay / huge-document / locality benchmark lanes

Live Doc Sync Applied

  • synced live control docs to the completed packages/slate tranche-3 read:
    • docs/slate-v2/master-roadmap.md
    • docs/slate-v2/replacement-gates-scoreboard.md
    • docs/slate-v2/release-readiness-decision.md
    • docs/slate-v2/ledgers/slate-transforms-api.md
    • docs/slate-v2/ledgers/slate-editor-api.md
    • docs/slate-v2/ledgers/slate-interfaces-api.md

Next Move