feat(pack): support mdx option

[fireairforce]

Summary

Align with umi's: https://github.com/umijs/umi/blob/master/packages/bundler-webpack/src/config/javaScriptRules.ts#L200-L207

And mako's: https://makojs.dev/docs/config#mdx

Default is false, not enable.

Test Plan

add snapshot test case

[gemini-code-assist]

Code Review

This pull request introduces support for Rust-based MDX transformations across the codebase. Key changes include updating the core configuration structures, integrating MDX options into both client and server module contexts, and expanding the public schema in Rust, TypeScript, and JSON formats to support MDX settings. Additionally, a new snapshot test has been added to verify basic MDX compilation. Feedback focuses on critical serialization issues in the Config struct; specifically, the mdx field requires serde_self_describing for bincode compatibility, and the MdxOptions enum needs to derive Serialize instead of Encode/Decode to prevent compilation failures.

[github-actions]

📊 Performance Benchmark Report (with-antd)

Utoopack Performance Report

Report ID: utoopack_performance_report_20260428_103442
Generated: 2026-04-28 10:34:42
Trace File: trace_antd.json (0.5GB, 1.57M spans)
Test Project: examples/with-antd

---

Executive Summary

Metric	Value	Assessment
Total Wall Time	11,274.6 ms	Baseline
Total Thread Work (de-duped)	33,560.2 ms	Non-overlapping busy time
Effective Parallelism	3.0x	thread_work / wall_time
Working Threads	5	Threads with actual spans
Thread Utilization	59.5%	⚠️ Suboptimal
Total Spans	1,570,063	All B/E + X events
Meaningful Spans (>= 10us)	375,306	(23.9% of total)
Tracing Noise (< 10us)	1,194,757	(76.1% 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	82,833	5,377.8	4,248.6	7,757.5
Parse	10,367	3,696.2	2,756.9	11,129.6
Analyze	221,462	22,991.0	16,512.1	10,790.5
Chunk	22,136	1,817.2	1,696.7	8,563.7
Codegen	29,598	3,921.5	2,911.2	8,739.8
Emit	38	30.6	15.2	8,650.1
Other	8,872	1,486.5	1,200.7	11,274.5

---

Workload Distribution by Diagnostic Tier

Category	Spans	Inclusive (ms)	% Work	Self-Time (ms)	% Self
P0: Scheduling & Resolution	310,668	29,338.2	87.4%	21,458.8	63.9%
P1: I/O & Heavy Tasks	2,936	799.1	2.4%	783.7	2.3%
P2: Architecture (Locks/Memory)	0	0.0	0.0%	0.0	0.0%
P3: Asset Pipeline	60,583	8,748.3	26.1%	6,678.1	19.9%
P4: Bridge/Interop	0	0.0	0.0%	0.0	0.0%
Other	1,119	435.3	1.3%	420.7	1.3%

---

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
9,768.2	11,451.4	118,829	82.2	0.1	28.0	29.1%	module	write all entrypoints to disk (2%)
3,765.1	5,295.5	35,341	106.5	0.2	186.4	11.2%	analyze ecmascript module	process module (79%)
2,298.7	2,493.7	45,489	50.5	0.0	20.8	6.8%	internal resolving	resolving (30%)
2,134.9	5,376.7	57,379	37.2	0.0	20.8	6.4%	process module	module (10%)
1,927.8	2,862.1	36,787	52.4	0.0	20.9	5.7%	resolving	module (16%)
1,762.6	2,491.5	7,297	241.5	0.7	50.6	5.3%	parse ecmascript	analyze ecmascript module (31%)
1,688.7	2,699.0	17,512	96.4	0.4	61.3	5.0%	code generation	chunking (2%)
1,075.6	1,173.8	12,834	83.8	0.2	22.6	3.2%	chunking	write all entrypoints to disk (0%)
912.8	912.8	10,004	91.2	0.3	9.1	2.7%	precompute code generation	code generation (58%)
746.3	746.3	2,335	319.6	2.3	19.1	2.2%	read file	parse ecmascript (91%)
744.2	1,006.8	7,139	104.2	0.0	136.8	2.2%	write all entrypoints to disk	None (0%)
740.7	740.7	7,825	94.7	0.4	67.3	2.2%	compute async module info	chunking (0%)
603.8	604.3	9,199	65.6	0.1	42.3	1.8%	compute async chunks	webpack loader (0%)
378.3	386.2	1,007	375.7	1.8	18.3	1.1%	webpack loader	parse css (15%)
309.7	309.7	2,082	148.8	0.4	15.9	0.9%	generate source map	code generation (96%)
247.9	458.4	729	340.0	1.4	24.1	0.7%	parse css	module (6%)
59.8	59.8	823	72.6	0.0	13.6	0.2%	compute binding usage info	write all entrypoints to disk (0%)
35.8	44.4	614	58.4	0.1	6.0	0.1%	async reference	write all entrypoints to disk (1%)
29.8	29.8	1,243	23.9	0.0	7.1	0.1%	collect mergeable modules	compute merged modules (0%)
22.1	22.1	557	39.7	0.0	12.0	0.1%	read directory	internal resolving (99%)

---

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	186.5	2	process module → analyze ecmascript module
2	77.3	2	code generation → generate source map
3	69.1	2	process module → analyze ecmascript module
4	68.8	2	code generation → generate source map
5	50.7	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	35,341	process module (79%)	106.5 us	0.16 ms	3,765.1 ms

---

Duration Distribution

Range	Count	Percentage
<10us	1,194,757	76.1%
10us-100us	348,950	22.2%
100us-1ms	20,159	1.3%
1ms-10ms	5,790	0.4%
10ms-100ms	403	0.0%
>100ms	4	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