perf(web): format minimap previews only when opened

[juliusmarminge]

Streaming updates eagerly compacted every loaded prompt and response for the timeline minimap, including when its preview was closed. Keep the source text in the minimap index and compact only the selected preview. Message targets and preview content stay the same.

Production client replay	Before	After
Median update	23.0 ms	20.9 ms (−9.1%)
Median run p95	36.9 ms	32.4 ms (−12.2%)

Five rotated-order runs, 120 measured updates per run, 12 stable TypeScript fences and a growing prose tail. Same browser, origin, viewport and synthetic data; no concurrent builds or local checks. The measurements compare ab67795 with this PR's unchanged runtime implementation. Later commits adjust test fixtures only. CI validates integration with newer main. These are client rendering measurements, not provider/network throughput or production user percentiles.

A separate 200-response × 20KB stage benchmark reduced minimap item construction from 21.25 ms to below 0.1 ms. This is an isolated stage saving, not a whole-app multiplier. An open preview still formats its selected response in full.

Exploration and rejected experiments · Production samples · Minimap stage samples · Benchmark source and reproduction notes

Impacts the web timeline and the web renderer used by desktop. Mobile, server, providers, wire contracts and file/diff implementations are unchanged. No new dependency, truncation, observer or state-library migration. Conditional HTML parsing, toolbar measurement observers, and code-block memoization were discarded after compatibility or timing checks; this PR contains only lazy preview formatting. Further measured improvements are separate layers in #11193, #11196 and #11198.

Validation: 103 focused timeline tests, web typecheck, targeted lint with pre-existing timeline warnings, and current-head CI. Browser checks cover minimap keyboard navigation, long-history navigation, selected text/DOM identity through 30 streaming updates, Markdown disclosures, a 10,000-line file to EOF, and split/collapsed working-tree diffs. No new flicker was observed in these checks. Native mobile and Electron shell behavior were not exercised.

Before	After

Before navigation and streaming:

https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/56590efe867cdef9/before-demo.webm

After navigation and streaming:

https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/789dea2b5f537b59/after-demo.webm

Recordings show visual states and compress dropped-frame intervals; timing comes from separate unrecorded runs. 0.8 seconds of recorder-induced startup was trimmed. No private transcript or authentication data is included. The source tree contains no probe or evidence files.

Model: GPT-6. Harness: Codex.

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.5 KiB	13.5 KiB	+2 B (+0.0%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.1 KiB	7.1 KiB	+1 B (+0.0%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.5 KiB	+1 B (+0.0%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	57.0 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	8	8	0 (0.0%)	21	✅
Claude	Total thread wire	13.6 KiB	13.6 KiB	+7 B (+0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.1 KiB	7.1 KiB	+3 B (+0.0%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.5 KiB	6.5 KiB	+4 B (+0.1%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	9	9	0 (0.0%)	21	✅

Baseline: 57aee3e · PR result: 3b9cf6e · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 113.9 KiB
• Claude decoded thread snapshot: 114.6 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[macroscopeapp]

Approvability

Verdict: Approved at 8c1b016

Macroscope's review found this PR approvable — This localized web optimization defers minimap text formatting until a preview is opened while preserving navigation targets and displayed content. The supporting tests cover the extracted logic, and no schema, deployment, security, billing, default, or static-analysis behavior is changed.

No code changes detected at 3b9cf6e. Prior analysis still applies.

^{You can add or adjust custom eligibility rules. Learn more.}

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Team

Run ID: fafd6ea9-0c3c-4609-b968-935203ec5dd1

📥 Commits

Reviewing files that changed from the base of the PR and between 57aee3e and 3b9cf6e.

📒 Files selected for processing (3)

• apps/web/src/components/chat/MessagesTimeline.tsx
• apps/web/src/components/chat/timelineMinimapItems.test.ts
• apps/web/src/components/chat/timelineMinimapItems.ts

Included review availability: Your plan provides up to 10 included reviews per hour; 9 remain after this review.

---

📝 Walkthrough

Walkthrough

Changes

Timeline minimap

Layer / File(s)	Summary
Minimap item derivation and preview resolution apps/web/src/components/chat/timelineMinimapItems.ts, apps/web/src/components/chat/timelineMinimapItems.test.ts	Adds shared minimap item types and helpers. User turns retain their source text and receive the final assistant response before the next user turn. Tests cover grouping, normalization, null previews, jump targets, and streaming text.
MessagesTimeline integration apps/web/src/components/chat/MessagesTimeline.tsx	Uses the shared derivation and preview resolver. Removes local minimap helpers and memoizes the active preview.

Priority: ⬇️ Low

Estimated code review effort: 3 (Moderate) | ~20 minutes

Change: Refactor

Sequence Diagram(s)

sequenceDiagram
  participant MessagesTimeline
  participant deriveTimelineMinimapItems
  participant resolveTimelineMinimapPreview
  MessagesTimeline->>deriveTimelineMinimapItems: pass timeline rows
  deriveTimelineMinimapItems-->>MessagesTimeline: return minimap items
  MessagesTimeline->>resolveTimelineMinimapPreview: resolve active item
  resolveTimelineMinimapPreview-->>MessagesTimeline: return normalized preview or null

Loading

Merge Risk: ⚪ Minimal · up to 3b9cf

The minimap change preserves source text for indexing while resolving formatting only for the active preview, with no identified merge-blocking behavior regression.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 12.50% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 8 functions across 3 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Title check	✅ Passed	The title clearly and concisely describes the main change: minimap previews are formatted only when opened.
Description check	✅ Passed	The description explains what changed, why it changed, scope, performance results, UI evidence, validation, and limitations. It does not use the template headings or include the checklist explicitly, …

• Fix all pre-merge checks with AI

✨ Finishing Touches 💡 1

📝 Generate docstrings 💡

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch legend-perf/timeline-updates

---

_{Comment @coderabbitai help to get the list of available commands.}