perf(client): reduce repeated sorting and date formatting

[Bil0000]

What Changed

Thread ordering reparses the same dates during each sort comparison, project selection sorts every matching thread to return one, and hourly usage labels repeatedly construct identical Intl.DateTimeFormat instances. This PR removes that repeated work in three production files.

• Parse sort timestamps once per thread. Keep the existing order keys, ID/environment tie breaks, stable ordering, and original thread references. Find the latest project thread in one scan.
• Filter plan updates before sorting them, so unrelated tool activity does not enter the plan sort.
• Reuse date formatters for explicit time zones, with a cache capped at 16 entries. Leave implicit local-time formatting uncached so system time-zone changes still take effect.

No dependencies, settings, wire contracts, or visual changes.

Benchmarks

Measured actual before/after exports from base 8d8189e67d and this branch. These are function-level wall times, not whole-app CPU, memory, GPU, or battery measurements.

Workload	Chromium before	Chromium after	Less time
Sort 1,000 threads by recent activity	1.401 ms	0.262 ms	81.3%
Sort 1,000 active threads without order keys	1.205 ms	0.306 ms	74.6%
Sort 1,000 pinned threads without order keys	1.317 ms	0.357 ms	72.9%
Select latest project thread from 1,000 threads	1.461 ms	0.208 ms	85.8%
Derive a plan from 500 activities / 5 plan updates	0.01537 ms	0.00495 ms	67.8%
Format 24 hourly usage labels and tooltips	12.115 ms	0.375 ms	96.9%

The usage batch saves about 11.7 ms of main-thread work in this fixture. The plan change is a small absolute saving. Benefits depend on history size and how often the caller runs.

Method: 200 warm-up calls per version, nine batches alternating before/after order, median time per operation. Each batch ran 100 list operations, 10,000 plan operations, or 20 usage batches. Fixtures match the committed benchmark: a fixed timestamp permutation for the threads, sequence-ordered activities, and America/New_York hourly formatting. Timings include allocation and garbage collection and varied between runs. The warm cache is intentional for repeated chart labels.

Separate instrumentation, outside timed runs, counted 5,794 → 1,000 Date.parse calls per list operation and 120 → 0 formatter constructions per warm hourly batch. The formatter cache retains at most 16 instances; total process memory was not measured.

Reproduce with the committed Node benchmark

apps/web/src/performance.bench.ts runs the same six workloads with the existing test runner. Copy that file to a checkout of base 8d8189e67d for the before run. From apps/web in each checkout:

# Base checkout, with the benchmark file copied in
vp test bench src/performance.bench.ts --run --outputJson /tmp/t3-perf-before.json

# PR checkout
vp test bench src/performance.bench.ts --run --compare /tmp/t3-perf-before.json --outputJson /tmp/t3-perf-after.json

Node median results, in the same row order as the browser table:

Workload	Node before	Node after
Recent threads	1.635046 ms	0.254633 ms
Active threads	1.403014 ms	0.363550 ms
Keyless pinned threads	4.716217 ms	0.391587 ms
Latest project thread	1.488331 ms	0.204262 ms
Plan state	0.014992 ms	0.009198 ms
Hourly labels and tooltips	10.591893 ms	0.587797 ms

Node and browser numbers are separate runs, not interchangeable. The browser measurements use the actual bundled modules in Chromium; the command above reproduces the Node suite.

Verification

• 407 focused tests passed across shared formatting, shared/web thread ordering, plan state, sidebar, command palette, and timeline logic.
• Added coverage for invalid dates, both sort modes, ties, archived/project filtering, unchanged input/references, mixed plan activity, explicit time zones, and changes to the system time zone. Existing DST tests pass.
• 14,400 additional deterministic comparisons against the base exports returned identical outputs and thread references, including mixed order keys, bad dates, plan clears, DST transitions, and cache eviction.
• Scoped type checks for client-runtime, shared, and web passed. Targeted lint/format checks, web file-discovery check, and production web build passed.
• Verified with 12 synthetic threads and 500 activities: pinned/recent order, thread navigation, command palette, expanded work log, composer edit/clear, usage chart hover, daily/hourly switching, and wide/narrow layouts.

Shared logic serves web/desktop and mobile callers. Native desktop/mobile runtime tests and power measurements were not performed. No provider adapter or connection behavior changes.

Checklist

• Small changes limited to repeated client-side work
• Explained the changes and measured their effect
• Added regression coverage and a repeatable benchmark
• No visual or animation change requiring before/after media

Model: GPT-6. Harness: Codex.

Summary by CodeRabbit

• Improvements

  • Improved date and time formatting performance across usage and activity views.
  • Increased consistency when displaying hourly usage across time zones, including system-default time zone handling.
  • Improved thread ordering and selection reliability, including consistent tie-breaking for equally dated threads.
  • Improved active plan state handling when activity history includes unrelated events.

• Tests & Performance

  • Added broader automated coverage and performance benchmarks for thread sorting, plan selection, and usage formatting.

[macroscopeapp]

Approvability

Verdict: Approved at fdc44a7

Macroscope's review found this PR approvable — This is a contained client-side performance refactor that reduces repeated sorting, date parsing, and formatter construction without adding capabilities, changing defaults, or altering interfaces. The benchmark and regression coverage are development-only, while the production edits have bounded performance and memory effects.

No code changes detected at 3183fbc. 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: Advanced

Run ID: 677c3e61-c7ba-42cd-98d7-81deed44a647

📥 Commits

Reviewing files that changed from the base of the PR and between fdc44a7 and ec055a9.

📒 Files selected for processing (1)

• apps/web/src/session-logic.ts

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

---

📝 Walkthrough

Walkthrough

The changes optimize thread ordering, plan activity selection, and date-time formatter creation. New tests verify ordering, time-zone behavior, input preservation, and edge cases. A benchmark suite measures the affected operations with deterministic datasets.

Changes

Runtime performance updates

Layer / File(s)	Summary
Thread ordering and selection packages/client-runtime/src/state/threadSort.ts, packages/client-runtime/src/state/threadSort.test.ts	Thread sorting and project-thread selection use native sorting, direct scanning, timestamp precomputation, and short-array fast paths. Tests cover ordering, ties, invalid timestamps, filtering, and reference preservation.
Plan activity ordering apps/web/src/session-logic.ts, apps/web/src/session-logic.test.ts	deriveActivePlanState filters plan activities before sorting them. Tests cover sequence ordering, ignored activities, duration calculation, input preservation, and null results.
Date-time formatter caching packages/shared/src/usageFormat.ts, packages/shared/src/usageFormat.test.ts	Date-time formatting functions use a bounded formatter cache. Tests cover multiple time zones, invalid inputs, unknown zones, and system-zone behavior.
Performance benchmark coverage apps/web/src/performance.bench.ts	Deterministic benchmarks cover thread sorting, project-thread selection, plan derivation, and time-zone-aware usage formatting.

Priority: ⬇️ Low

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

Change: Refactor

Suggested reviewers: juliusmarminge, t3dotgg

Merge Risk: ⚪ Minimal · up to 3183f

The performance changes preserve the covered behavior and no actionable merge-blocking risk remains.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 54.55% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 11 functions across 7 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and concisely describes the primary change: reducing repeated sorting and date-formatting work.
Description check	✅ Passed	The description is detailed and relevant. It explains the changes, motivation, benchmarks, verification, testing, and checklist status. The template's separate "Why" heading is omitted, but the ration…
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.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[Bil0000]

Reviewed the advisory docstring-coverage warning. This patch adds no public API, preserves the existing behavior, and includes focused tests for the changed paths. Extra docstrings solely to meet a coverage percentage would add noise to this small performance change, so I am leaving that advisory warning unchanged. The CodeRabbit check passed and reported no actionable code findings.