perf(web): pull request chains parse each link's URL once

[juliusmarminge]

The chat view and every sidebar row resolve a thread's current pull request on each render. resolveThreadPullRequestChains normalized each link's key several times per call (stack grouping, branchKey for head and base, threadPullRequestKeyOf for placement), and every normalization ran parseChangeRequestUrl, which builds a new URL() and runs a few regexes. A thread with 157 linked PRs (a real stack-heavy thread from the seed data) cost about 1.3–1.6 ms per resolve, several times per ChatView render. Over a streamed turn in that thread this added up to 150–370 ms of URL parsing inside long tasks (CDP traces of three full turns on main).

Normalized keys are now cached per link object in a WeakMap. Links are immutable decoded values, and the shell reducer keeps unchanged links by reference, so the cache hits across renders. A changed link is a new object and gets a fresh key.

Measured
Microbenchmark on the real 157-link array (node, 100 iterations after warmup):

Call	main	this PR
resolveThreadCurrentPullRequestLink	1.57 ms	0.68 ms
resolveThreadPullRequestChains	1.30 ms	0.21 ms

In the app: production build (vp build --minify false), headless Chromium, same seeded snapshot, CDP trace per window, median of 3 runs with the build order alternated:

Window	Metric	main	this PR
Switch 3 threads x3	changeRequestUrl.ts ms inside long tasks (≥50 ms)	40	0
Switch 3 threads x3	threadPullRequests.ts ms inside long tasks	25	16
Switch 3 threads x3	ScriptDuration ms	1,561	1,507
Open large thread (warm, n=6)	URL parse + key ms inside long tasks	22	5
Open large thread (warm, n=6)	ScriptDuration ms	302	258

Long-task counts and total long-task ms for open and switch are within noise (1,363 vs 1,352 ms on switch). This takes a few percent off each frame and doesn't remove any. Streaming, where ChatView re-renders on every delta, is where it adds up (see the numbers above).

Tests: packages/shared/src/threadPullRequests.test.ts passes unchanged. Lint and typecheck pass for packages/shared.

Claude Opus 5.5 via Claude Code.

🤖 Generated with Claude Code

[macroscopeapp]

Approvability

Verdict: Approved at c5e2bb4

Macroscope's review found this PR approvable — This is a contained performance optimization that caches normalized pull-request keys without changing the existing chain-resolution behavior for immutable links. It introduces no new capability, schema change, default change, or sensitive-path modification.

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

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

ℹ️ The exact PR base did not have a successful artifact. Baseline uses the latest successful main measurement shown below.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	5.0 KiB	4.9 KiB	−30 B (−0.6%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.2 KiB	1.2 KiB	−30 B (−2.5%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	20.9 KiB	20.8 KiB	−41 B (−0.2%)	29.3 KiB	✅
Codex	Live turn messages	2	1	−1 (−50.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.2 KiB	21.2 KiB	0 B (0.0%)	29.3 KiB	✅
Claude	Live turn messages	2	2	0 (0.0%)	8	✅

Baseline: 9ba6143 · PR result: c5e2bb4 · 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: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

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

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Path: .coderabbit.config.ts
• Review profile: CHILL
• Plan: Team
• Run ID: ed9638a7-b0b8-4640-b2be-2d1607204a75

📥 Commits

Reviewing files that changed from the base of the PR and between eb1d80c and c5e2bb4.

📒 Files selected for processing: 1

• packages/shared/src/threadPullRequests.ts

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 4 remain after this review.

📝 Walkthrough

threadPullRequestKeyOf now caches normalized pull-request keys by object identity. Native-stack and derived-branch grouping use the cached lookup.

Changes: Pull-request key normalization

Layer / File(s)	Summary
Cache normalized keys and use them in grouping packages/shared/src/threadPullRequests.ts	threadPullRequestKeyOf caches normalized keys in a WeakMap. Native-stack and derived-branch grouping use this helper.

Priority: ⬇️ Low

Estimated code review effort: 2 (Simple) | ~8 minutes

Change: Refactor

Suggested reviewers: maria-rcks

Merge Risk: ⚪ Minimal · up to c5e2b

Changed link data receives a new object, so the cache does not leave supported callers with stale keys. No actionable merge-blocking risk remains.

Pre-merge checks: 4

Check name	Status	Explanation
Title check		The title clearly and concisely describes the main performance change: parsing each pull-request link URL once through caching.
Description check		The description clearly explains the performance problem, implementation, measured results, verification steps, and agent usage. It does not explicitly provide issue or maintainer-approval details, or…
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Description check

Explanation

The description clearly explains the performance problem, implementation, measured results, verification steps, and agent usage. It does not explicitly provide issue or maintainer-approval details, or explain why this focused change qualifies without them.

✨ Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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