feat(pack): add tracing option to control default tracing output

[fireairforce]

Summary

Add option for control @utoo/pack's tracing output.

issue: #2792

PR: umijs/umi#13297

Test Plan

[gemini-code-assist]

Code Review

This pull request introduces a "tracing" configuration option across the CLI, shared configuration, and core build logic, allowing users to toggle default tracing logs. The option is integrated into the NAPI project structures, CLI command arguments for build and dev, and the internal bundle options where it defaults to true. I have no feedback to provide.

[xusd320]

支持完全关掉 tracing ? 那 dev 的启动过程怎么透出呢？

[fireairforce]

支持完全关掉 tracing ? 那 dev 的启动过程怎么透出呢？

默认没关，透传一个配置，umi 那边可以拿到 dev 完成的状态

[github-actions]

📊 Performance Benchmark Report (with-antd)

Utoopack Performance Report

Report ID: utoopack_performance_report_20260419_113154
Generated: 2026-04-19 11:31:54
Trace File: trace_antd.json (0.6GB, 1.61M spans)
Test Project: examples/with-antd

---

Executive Summary

Metric	Value	Assessment
Total Wall Time	14,090.0 ms	Baseline
Total Thread Work (de-duped)	35,984.1 ms	Non-overlapping busy time
Effective Parallelism	2.6x	thread_work / wall_time
Working Threads	5	Threads with actual spans
Thread Utilization	51.1%	⚠️ Suboptimal
Total Spans	1,608,406	All B/E + X events
Meaningful Spans (>= 10us)	523,256	(32.5% of total)
Tracing Noise (< 10us)	1,085,150	(67.5% 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	124,951	5,017.2	4,075.8	8,015.0
Parse	12,135	2,244.6	1,905.4	13,972.5
Analyze	301,235	24,017.8	17,745.9	13,485.9
Chunk	30,179	2,790.4	2,557.5	11,135.8
Codegen	42,153	5,070.1	3,822.5	10,700.9
Emit	75	62.3	29.7	8,986.7
Other	12,528	1,930.8	1,670.0	14,090.0

---

Workload Distribution by Diagnostic Tier

Category	Spans	Inclusive (ms)	% Work	Self-Time (ms)	% Self
P0: Scheduling & Resolution	435,065	30,321.8	84.3%	22,866.5	63.5%
P1: I/O & Heavy Tasks	3,262	162.5	0.5%	129.9	0.4%
P2: Architecture (Locks/Memory)	0	0.0	0.0%	0.0	0.0%
P3: Asset Pipeline	83,143	10,118.8	28.1%	8,299.1	23.1%
P4: Bridge/Interop	0	0.0	0.0%	0.0	0.0%
Other	1,786	530.0	1.5%	511.2	1.4%

---

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
10,385.7	11,996.8	170,098	61.1	0.1	60.3	28.9%	module	write all entrypoints to disk (1%)
3,643.0	4,991.7	39,382	92.5	0.2	263.5	10.1%	analyze ecmascript module	process module (76%)
2,248.1	5,522.6	72,296	31.1	0.1	15.4	6.2%	process module	module (14%)
2,188.1	2,375.6	62,603	35.0	0.0	20.1	6.1%	internal resolving	resolving (30%)
1,877.6	2,631.5	61,677	30.4	0.0	16.8	5.2%	resolving	module (35%)
1,791.3	3,038.9	23,287	76.9	0.3	44.4	5.0%	code generation	chunking (8%)
1,716.0	1,716.0	16,784	102.2	0.4	36.5	4.8%	precompute code generation	code generation (44%)
1,701.0	1,898.2	16,489	103.2	0.3	107.4	4.7%	chunking	write all entrypoints to disk (0%)
1,559.4	1,640.0	8,783	177.6	0.5	68.6	4.3%	parse ecmascript	analyze ecmascript module (27%)
1,282.4	1,282.4	16,223	79.0	0.3	133.7	3.6%	compute async module info	chunking (0%)
1,099.8	1,336.1	9,929	110.8	0.0	287.8	3.1%	write all entrypoints to disk	None (0%)
826.8	827.5	13,520	61.2	0.1	73.6	2.3%	compute async chunks	compute async chunks (0%)
480.5	491.7	1,560	308.0	1.0	18.3	1.3%	webpack loader	parse css (10%)
315.2	315.2	2,082	151.4	0.4	17.2	0.9%	generate source map	code generation (96%)
255.8	514.4	836	306.0	1.5	22.6	0.7%	parse css	module (5%)
104.0	104.0	1,192	87.2	0.0	28.3	0.3%	compute binding usage info	write all entrypoints to disk (0%)
87.4	87.4	2,504	34.9	0.0	7.8	0.2%	read file	parse ecmascript (85%)
64.5	64.5	2,017	32.0	0.0	20.9	0.2%	collect mergeable modules	compute merged modules (0%)
59.0	64.8	813	72.6	0.1	6.4	0.2%	async reference	write all entrypoints to disk (1%)
29.6	64.7	170	174.1	0.4	16.2	0.1%	make production chunks	chunking (2%)

---

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	263.6	2	process module → analyze ecmascript module
2	129.6	2	process module → analyze ecmascript module
3	68.7	2	analyze ecmascript module → parse ecmascript
4	61.6	2	code generation → generate source map
5	47.1	2	analyze ecmascript module → parse ecmascript

---

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
analyze ecmascript module	39,382	process module (76%)	92.5 us	0.16 ms	3,643.0 ms

---

Duration Distribution

Range	Count	Percentage
<10us	1,085,150	67.5%
10us-100us	491,159	30.5%
100us-1ms	25,476	1.6%
1ms-10ms	6,341	0.4%
10ms-100ms	271	0.0%
>100ms	9	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