slate-side-rich-text-benchmark-artifact

Objective: Add a real Slate-side artifact to the rich-text benchmark for a previously Slate-v2-only detailed table. The chosen layer is the huge-document Chromium browser trace because Slate has the same route surface and the result can be measured directly instead of represented as an adapter gap.

Goal plan: docs/plans/2026-05-28-slate-side-rich-text-benchmark-artifact.md

Completion threshold: The goal is complete when a local Slate browser trace artifact is generated from /Users/zbeyens/git/slate/site/out/examples/huge-document, ingested into benchmarks/editor, rendered in rich-text.html as aligned Slate v2 and Slate columns, verified by package checks, verified through the served data route, and this plan passes the autogoal completion check.

Verification surface:

• Slate artifact generator: /Users/zbeyens/git/plate-2/.tmp/slate-v2/scripts/benchmarks/browser/react/huge-document-slate-browser-trace.mjs
• Generated Slate artifact: /Users/zbeyens/git/plate-2/.tmp/slate-v2/tmp/slate-react-huge-document-slate-browser-trace-benchmark-surfaces-legacyChunkOn-legacyChunkOff-blocks-5000-iters-3-ops-10.json
• Evidence Kit ingestion: /Users/zbeyens/git/plate-2/benchmarks/editor/src/index.mjs
• Viewer data: /Users/zbeyens/git/plate-2/benchmarks/editor/docs/perf/rich-text-data.json
• Served route: http://127.0.0.1:8765/rich-text.html

Constraints:

• This step adds one real Slate artifact first; it does not pretend every Slate-v2-only category has a legacy equivalent.
• The measured route is legacy Slate huge-document browser behavior, not a synthetic placeholder row.
• Browser trace rows compare equivalent operation/metric fixtures across surface columns.
• The Slate site was served from existing site/out; the generator can rebuild it with SLATE_LEGACY_BROWSER_TRACE_BUILD=1.

Boundaries:

• Source of truth: .tmp/slate-v2 benchmark scripts and /Users/zbeyens/git/slate/site/examples/ts/huge-document.tsx.
• Allowed edit scope: .tmp/slate-v2 benchmark script/package script, benchmarks/editor ingestion/viewer/generated data, and this plan.
• External sources: N/A; local Slate and Slate v2 checkouts are enough.
• Browser surface: legacy Slate static huge-document route served to Playwright, plus rich-text.html served on port 8765.
• Tracker sync: N/A; no issue or PR requested.
• Non-goals: Slate artifacts for v2-only collaboration, transaction-kernel, ref-projection, and internal store categories.

Blocked condition: Work would stop only if legacy Slate could not serve the huge-document route, Playwright could not select/type into the legacy editor, or Evidence Kit could not normalize the resulting trace rows. All three paths worked.

Major source:

• type: local benchmark/browser artifact implementation
• id / link: .tmp/slate-v2 plus /Users/zbeyens/git/slate
• title: Slate-side huge-document browser trace artifact
• decision to make: which hidden-Slate detailed table deserves the first real Slate artifact
• decision criteria: comparable legacy route exists, Playwright can measure it, table rows align by operation/metric, package check stays green, served viewer exposes Slate columns

Major lane:

• lane: benchmark/performance
• output type: generated browser trace artifact plus Evidence Kit rows
• implementation expected: yes
• affected packages / surfaces: .tmp/slate-v2 benchmark tooling and benchmarks/editor
• dominant risk: adding a column without comparable row alignment

Phase / pass table:

Phase	Status	Evidence	Next
Intake and source read	complete	Read current v2-only tables, Slate huge-document route, and existing browser trace harness.	Artifact implementation
Current-state map	complete	Identified slate-react-huge-document-browser-trace as comparable and high value.	Generator
Options and recommendation	complete	Chose real Slate browser trace over fake missing rows or v2-only internal artifacts.	Ingestion
Review / pressure pass	complete	Corrected first ingestion so rows align by operation, not by surface-specific fixture names.	Verification
Implementation or plan artifact	complete	Added generator, package script, Evidence Kit spec, browser-trace normalizer, and generated docs/data.	Closeout
Verification	complete	Benchmark artifact, package check, docs check, and served data proof all passed.	Final response
Closeout	complete	This plan records evidence and passes check-complete.	Done

Start Gates:

Gate	Applies	Evidence
major-task loaded	yes	Used because this is benchmark comparison work.
Active goal checked or created	yes	Active autogoal created for the Slate-side artifact objective.
Source of truth read before analysis	yes	Read Slate and Slate v2 huge-document routes and benchmark ingestion code.
Major lane selected	yes	Benchmark/performance lane selected.
Decision criteria stated	yes	Completion threshold names artifact, ingestion, row alignment, checks, and served proof.
Existing repo patterns / prior decisions checked	yes	Reused .tmp/slate-v2 benchmark scripts and benchmarks/editor Evidence Kit ingestion.
Helper stack selected	yes	Used local script, Playwright, Evidence Kit, and static HTTP proof only.
External research decision recorded	yes	N/A because local clone behavior settled the task.
Implementation expectation recorded	yes	Implementation expected and completed.
Workspace authority selected	yes	plate-2 owns benchmark harness; .tmp/slate-v2 owns benchmark script; /Users/zbeyens/git/slate owns legacy route.
Branch / PR expectation decided	yes	No commit, push, or PR requested.
Browser pack selected	yes	Browser trace generation and served viewer proof required browser coverage.
Browser route / app surface identified	yes	Legacy Slate /examples/huge-document plus http://127.0.0.1:8765/rich-text.html.
Browser tool decision recorded	yes	Browser MCP was not exposed; Playwright generated the benchmark artifact and curl verified the served static viewer.
Console/network caveat policy recorded	yes	Artifact route success, completed DOM surface counts, and served JSON proof were recorded; separate console audit was not needed for static viewer data.

Work Checklist:

• Objective includes outcome, completion threshold, verification surface, constraints, boundaries, and blocked condition.
• Major source records source type, id/link, title, decision type, expected outcome, decision criteria, likely files/packages/surfaces, browser surface, and highest-leverage owner.
• Current state is mapped before proposing a new architecture, migration, benchmark, or plan.
• Existing repo patterns, prior decisions, and nearby implementation constraints are recorded before external research.
• External docs or source are used only where repo evidence does not settle the question, or N/A reason is recorded.
• Options, recommendation, tradeoffs, blast radius, and rejection reasons are recorded.
• Facts, inference, and recommendation are separated.
• Review or pressure lenses are selected and completed, or marked N/A with reason.
• If implementation happens, touched-surface packs cover docs, browser, package/API, or agent-native surfaces as needed.
• Workspace authority recorded: every proof command names the cwd/tool that owns the analyzed or changed behavior.
• Accepted/actionable review findings are fixed or explicitly rejected with evidence.
• Browser pack: route, interaction path, and expected visible outcome are recorded before proof.
• Browser pack: browser proof uses the repo-approved browser tool or records a blocker/waiver.
• Browser pack: console and network errors are checked or explicitly out of scope.
• Browser pack: screenshot, trace, or exact verification caveat is ready for final handoff.

Completion Gates:

Gate	Applies	Required action	Evidence
Named verification threshold	yes	Generate Slate artifact, ingest it, regenerate viewer, and prove served data	Slate artifact generated; served data has 904 rows and aligned browser trace table.
Current-state source audit	yes	Map current owner, boundaries, constraints, and affected surfaces	Read Slate route, v2 browser trace, Evidence Kit specs, and viewer grouping.
Decision criteria closure	yes	Mark each criterion satisfied, narrowed, rejected, or blocked with evidence	Comparable route, artifact rows, aligned cells, package check, and served proof all satisfied.
Options / tradeoffs / rejection record	yes	Record viable options and chosen recommendation	Chose browser trace first; rejected fake cells and v2-only internal categories for this step.
Review / pressure pass	yes	Validate the table is meaningful	Changed normalization from surface-in-fixture to surface-as-library so every row has all comparison cells.
Review findings closure	yes	Fix accepted finding and record proof	Aligned table proof showed 26 fully aligned browser-trace rows.
External-source audit	no	Cite official/local clone/external sources when used, or record N/A	N/A: local clone source and measured browser artifact were enough.
Implementation gates	yes	Close primary-template and browser gates	Generator, ingestion, docs, full package check, and served proof completed.
Final handoff contract	yes	Record recommendation, evidence, caveats, residual risk, and next owner	Final handoff below records the artifact and remaining non-goals.
Final lint	yes	Run scoped equivalent when files changed	Biome checks passed for touched .tmp/slate-v2 and benchmarks/editor files.
Goal plan complete	yes	Run node .agents/rules/autogoal/scripts/check-complete.mjs docs/plans/2026-05-28-slate-side-rich-text-benchmark-artifact.md	Completion check passed after this update.
Browser interaction proof	yes	Exercise the target route/interaction with approved browser tool or record blocker	Playwright typed into legacy Slate huge-document route and produced timing rows.
Browser console/network check	no	Record console/network state or why it is not applicable	N/A: static viewer proof is JSON/HTTP; benchmark route proof is the generated trace artifact.
Browser final proof artifact	yes	Record screenshot/trace/route proof or exact caveat	Artifact JSON plus served rich-text-data.json smoke proof recorded.

Findings:

• slate-react-huge-document-browser-trace was the right first missing Slate artifact because legacy Slate exposes the same huge-document editor route.
• A single slate:browser-trace library column was weaker than needed because it produced surface-specific fixtures. Surface-as-library normalization gives directly aligned operation/metric rows.
• Collaboration, transaction execution, ref projection, and editor-store rows remain intentionally v2-only until a real Slate-equivalent workload is defined.

Decisions and tradeoffs:

• Chosen: add Slate huge-document browser trace for legacyChunkOn and legacyChunkOff.
• Rejected: fill detailed tables with synthetic adapter-missing rows. That would improve cosmetics and make the benchmark less honest.
• Rejected for this slice: implement Slate equivalents for v2-only internal runtime categories. Those need separate workload design.
• Tradeoff: the generator uses existing site/out by default to avoid rewriting the sibling Slate checkout; set SLATE_LEGACY_BROWSER_TRACE_BUILD=1 when a fresh legacy site build is required.

Implementation notes:

• Added bench:react:huge-document:slate-browser-trace:local in .tmp/slate-v2/package.json.
• Added huge-document-slate-browser-trace.mjs to serve legacy Slate static output, select DOM text nodes, type through Playwright, and summarize DOM, heap, long task, selection, and typing timing metrics.
• Added browser-trace artifact normalization to benchmarks/editor/src/index.mjs.
• Normalized browser trace surfaces as library columns: slate-v2:default-render-auto, slate-v2:dom-present, slate:chunk-on, and slate:chunk-off.

Review fixes:

• Fixed initial non-aligned browser trace rows by moving surface identity from fixture path to library label.

Error attempts:

Error / failed attempt	Count	Next different move	Resolution
Browser-side helper was stringified inside page.evaluate and broke syntax	1	Inline the helper inside the browser function	node --check and Biome passed.
First ingestion shape produced a Slate column without shared fixtures	1	Normalize surfaces as libraries and metrics as shared fixtures	Served data proved 26 fully aligned rows.

Verification evidence:

• node --check .tmp/slate-v2/scripts/benchmarks/browser/react/huge-document-slate-browser-trace.mjs passed.
• cd .tmp/slate-v2 && bunx biome check package.json scripts/benchmarks/browser/react/huge-document-slate-browser-trace.mjs --fix passed.
• cd .tmp/slate-v2 && bun run bench:react:huge-document:slate-browser-trace:local passed.
• Slate artifact evidence: legacyChunkOn and legacyChunkOff completed 3 measured iterations each, observed 5,000 blocks, and had 0 native-surface timeouts.
• cd benchmarks/editor && npx biome check src/index.mjs benchmarks/render-rich-text-viewer.mjs --fix passed.
• cd benchmarks/editor && npm run bench:rich-text:check generated 904 rows.
• Browser trace category evidence: 104 rows, 26 table fixtures, four libraries: slate-v2:default-render-auto, slate-v2:dom-present, slate:chunk-on, slate:chunk-off.
• cd benchmarks/editor && npm run check passed.
• curl -I --max-time 2 http://127.0.0.1:8765/rich-text.html returned HTTP 200.
• Served JSON smoke proof returned rowCount: 904, browserTraceRows: 104, browserTraceTableRows: 26, fullyAlignedRows: 26, workloadSlateStatus: ok, and status counts { adapter-missing: 54, coverage-gap: 130, ok: 716, optional-missing-artifact: 2, over-budget: 2 }.

Final handoff contract:

• Recommendation: use browser-trace as the pattern for future Slate artifacts: real measured artifact first, surface-as-library normalization second.
• Confidence: high for huge-document browser trace coverage.
• Evidence: artifact generation, row counts, aligned served table rows, and full package check.
• Tests / commands: listed in verification evidence.
• Browser proof: Playwright artifact plus served rich-text.html/JSON proof.
• PR / tracker: N/A.
• Caveats: static legacy Slate site/out was used; rebuild with SLATE_LEGACY_BROWSER_TRACE_BUILD=1 for a fresh Slate site artifact.
• Next owner: add Slate artifacts only where a real equivalent workload exists.

Timeline:

• 2026-05-28: Created major-task autogoal plan.
• 2026-05-28: Added and ran Slate huge-document browser trace artifact.
• 2026-05-28: Regenerated Evidence Kit rich-text benchmark and static viewer.
• 2026-05-28: Verified package checks and served route data.

Reboot status:

Question	Answer
Where am I?	Complete.
Where am I going?	Final handoff.
What is the goal?	Add the first real Slate-side artifact for a v2-only rich-text detailed benchmark table.
What have I learned?	Browser trace is comparable; some other v2-only lanes are not comparable without new workload design.
What have I done?	Generated and ingested Slate huge-document browser trace rows with aligned Slate v2/Slate columns.

Open risks: The new artifact is real browser evidence, but it is limited to the legacy Slate static site output that already existed locally. For publication-grade numbers, rerun with SLATE_LEGACY_BROWSER_TRACE_BUILD=1 and treat machine/browser variance like any other performance benchmark.