[fireairforce]

Summary

Add UTOO_ROOT_PATH so worktrees can access shared node_modules and linked dependencies outside the project. It overrides explicit rootPath and automatic detection; relative values resolve from the process working directory and empty values are ignored.

Use one resolver for build, serve, and direct project creation. Preserve the public ProjectOptions.rootPath: string type while allowing creation callers to omit it. Config-only updates retain the current root.

Keep findRootDir() independent of the override so framework plugin lookup and generated file locations stay stable, including Smallfish Less plugins/.sml and Umi's explicit utoopack.root.

UTOO_ROOT_PATH=/path/to/common-parent smallfish dev
UTOO_ROOT_PATH=.. utx up dev

Test Plan

• 30 targeted Vitest tests passed, including seven root creation/update regressions; strict TypeScript compatibility fixture passed.
• Built @utoo/pack-shared, pack CJS/ESM, and @utoo/pack-cli; formatting and spelling checks passed.
• Umi 4.7.26: real build/dev with env unset/set, explicit root, symlink dependency, HTTP 200, and Chrome rendering.
• Smallfish 2.96.20 CLI with React 18: production build and dev/HTTP/Chrome rendering passed with env unset/set, including default PostCSS processing.
• Smallfish 2.96.19 adapter: real build/dev with env unset/set, multiple entries, Less plugin, HTTP 200, and Chrome rendering; plugin and .sml locations stayed unchanged.
• An external symlink fails at the default filesystem boundary and builds successfully with a common parent override.

Framework smoke tests use this PR's JavaScript and existing native binaries. Full Smallfish CLI validation uses the packed candidate with its declared runtime dependencies installed in an isolated directory under the consuming project.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-27T03:43:35.698702Z	8a10d87	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[github-actions]

📊 Performance Benchmark Report (with-antd)

Utoopack Performance Report

Report ID: utoopack_performance_report_20261027_044150
Generated: 2026-10-27 04:41:50
Trace File: trace_antd.json (0.3GB, 0.84M spans)
Test Project: examples/with-antd

---

Executive Summary

Metric	Value	Assessment
Total Wall Time	5,475.3 ms	Baseline
Total Thread Work (de-duped)	16,738.0 ms	Non-overlapping busy time
Effective Parallelism	3.1x	thread_work / wall_time
Working Threads	10	Threads with actual spans
Thread Utilization	30.6%	⚠️ Suboptimal
Total Spans	844,436	All B/E + X events
Meaningful Spans (>= 10us)	184,510	(21.9% of total)
Tracing Noise (< 10us)	659,926	(78.1% of total)

---

Build Phase Timeline

Shows when each build phase is active and how much CPU it consumes.
Self-Time is the time spent exclusively in that phase (excluding children).

Phase	Spans	Inclusive (ms)	Self-Time (ms)	Wall Range (ms)
Resolve	38,350	4,865.3	1,309.2	2,448.2
Parse	8,128	1,030.4	766.2	4,629.6
Analyze	122,295	32,418.6	6,827.7	4,568.2
Chunk	4,773	4,628.4	763.1	1,929.1
Codegen	8,902	1,747.0	1,120.6	1,658.2
Emit	29	37.1	18.6	8.2
Other	2,033	7,208.0	4,407.1	5,475.3

---

Workload Distribution by Diagnostic Tier

Category	Spans	Inclusive (ms)	% Work	Self-Time (ms)	% Self
P0: Scheduling & Resolution	160,807	37,701.5	225.2%	8,325.7	49.7%
P1: I/O & Heavy Tasks	2,745	91.8	0.5%	73.3	0.4%
P2: Architecture (Locks/Memory)	0	0.0	0.0%	0.0	0.0%
P3: Asset Pipeline	20,216	7,407.7	44.3%	2,642.2	15.8%
P4: Bridge/Interop	0	0.0	0.0%	0.0	0.0%
Other	742	6,733.8	40.2%	4,171.2	24.9%

---

Top 20 Tasks by Self-Time

Self-time is the exclusive duration: time spent in the task itself, not in sub-tasks.
This is the most accurate indicator of where CPU cycles are actually spent.

Self (ms)	Inclusive (ms)	Count	Avg Self (us)	P95 Self (ms)	Max Self (ms)	% Work	Task Name	Top Caller
3,402.0	19,381.5	82,566	41.2	0.1	8.9	20.3%	module	module (60%)
3,066.7	3,333.6	16	191666.7	381.1	636.6	18.3%	save snapshot	persist (6%)
1,654.1	1,736.4	2,225	743.4	2.2	161.4	9.9%	analyze ecmascript module	module (79%)
1,131.2	10,461.9	29,783	38.0	0.0	5.5	6.8%	process module	process module (80%)
881.0	2,358.8	20,769	42.4	0.1	7.3	5.3%	internal resolving	internal resolving (71%)
717.2	981.4	5,959	120.4	0.5	28.4	4.3%	parse ecmascript	parse ecmascript (66%)
621.8	4,401.3	3,820	162.8	0.1	133.2	3.7%	chunking	chunking (61%)
618.4	670.5	6,837	90.5	0.4	7.0	3.7%	precompute code generation	generate merged code (53%)
563.4	680.9	5,672	99.3	0.3	115.1	3.4%	compute async module info	compute async module info (51%)
519.3	1,669.3	510	1018.3	2.3	213.8	3.1%	generate merged code	chunking (81%)
422.4	2,500.7	17,034	24.8	0.0	4.7	2.5%	resolving	module (64%)
349.5	349.5	329	1062.3	0.8	213.3	2.1%	generate source map	code generation (83%)
293.5	621.8	132	2223.8	3.7	179.3	1.8%	emit code	generate merged code (40%)
269.9	269.9	9	29992.3	138.8	170.1	1.6%	blocking	save snapshot (56%)
227.7	411.2	930	244.8	0.1	135.7	1.4%	write all entrypoints to disk	write all entrypoints to disk (8%)
152.6	727.0	1,736	87.9	0.2	34.6	0.9%	code generation	code generation (82%)
139.1	224.4	919	151.3	0.4	22.6	0.8%	compute async chunks	compute async chunks (52%)
49.0	49.0	2,167	22.6	0.0	1.7	0.3%	read file	parse ecmascript (91%)
41.3	50.8	582	70.9	0.0	24.4	0.2%	compute binding usage info	compute binding usage info (59%)
25.3	49.2	1,450	17.4	0.0	9.5	0.2%	collect mergeable modules	collect mergeable modules (100%)

---

Critical Path Analysis

The longest sequential dependency chains that determine wall-clock time.
Focus on reducing the depth of these chains to improve parallelism.

Rank	Self-Time (ms)	Depth	Path
1	806.6	3	persist → save snapshot → blocking
2	395.8	6	chunking → generate merged code → emit code → emit code → emit code → read file
3	260.4	2	save snapshot → blocking
4	247.5	4	chunking → generate merged code → emit code → generate source map
5	161.5	3	process module → process module → analyze ecmascript module

---

Batching Candidates

High-volume tasks dominated by a single parent. If the parent can batch them,
it drastically reduces scheduler overhead.

Task Name	Count	Top Caller (Attribution)	Avg Self	P95 Self	Total Self
process module	29,783	process module (80%)	38.0 us	0.05 ms	1,131.2 ms
internal resolving	20,769	internal resolving (71%)	42.4 us	0.06 ms	881.0 ms

---

Duration Distribution

Range	Count	Percentage
<10us	659,926	78.1%
10us-100us	116,789	13.8%
100us-1ms	59,799	7.1%
1ms-10ms	7,744	0.9%
10ms-100ms	135	0.0%
>100ms	43	0.0%

---

Action Items

1. [P0] Focus on tasks with the highest Self-Time — these are where CPU cycles are actually spent.
2. [P0] Use Batching Candidates to identify callers that should use try_join or reduce #[turbo_tasks::function] granularity.
3. [P1] Check Build Phase Timeline for phases with disproportionate wall range vs. self-time (= serialization).
4. [P1] Inspect P95 Self (ms) for heavy monolith tasks. Focus on long-tail outliers, not averages.
5. [P1] Review Critical Paths — reducing the longest chain depth directly improves wall-clock time.
6. [P2] If Thread Utilization < 60%, investigate scheduling gaps (lock contention or deep dependency chains).

Report generated by Utoopack Performance Analysis Agent