fix(pack): inlineCss with autoCssModules cause panic

[fireairforce]

Summary

styles.inlineCss + autoCssModules will cause panic, a case from smallfish widget build.

Test Plan

Update test case for both library and client

[gemini-code-assist]

Code Review

This pull request introduces performance optimizations for one-shot build sessions by adding an is_short_session hint to the turbo-tasks engine. This allows the system to use ReadWriteOnShutdown storage mode and conditionally disable dependency tracking when persistent caching is not required, reducing bookkeeping overhead. Additionally, the changes include a fix for CSS module facade analysis to prevent class extraction failures when Analyze references are inlined into JavaScript. New snapshot tests for Less CSS modules have also been added to ensure correct behavior. I have no feedback to provide as there were no review comments to evaluate.

[github-actions]

📊 Performance Benchmark Report (with-antd)

Utoopack Performance Report

Report ID: utoopack_performance_report_20260420_074426
Generated: 2026-04-20 07:44:26
Trace File: trace_antd.json (0.6GB, 1.61M spans)
Test Project: examples/with-antd

---

Executive Summary

Metric	Value	Assessment
Total Wall Time	11,286.5 ms	Baseline
Total Thread Work (de-duped)	24,430.4 ms	Non-overlapping busy time
Effective Parallelism	2.2x	thread_work / wall_time
Working Threads	5	Threads with actual spans
Thread Utilization	43.3%	⚠️ Suboptimal
Total Spans	1,605,835	All B/E + X events
Meaningful Spans (>= 10us)	371,054	(23.1% of total)
Tracing Noise (< 10us)	1,234,781	(76.9% 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	82,013	2,585.3	2,047.4	8,706.3
Parse	10,015	1,936.3	1,689.2	11,162.7
Analyze	220,928	15,817.2	10,850.2	10,922.6
Chunk	20,020	1,923.8	1,784.9	9,307.1
Codegen	28,674	3,914.2	2,892.1	9,395.2
Emit	76	69.6	36.0	7,631.9
Other	9,328	1,149.3	972.2	11,286.5

---

Workload Distribution by Diagnostic Tier

Category	Spans	Inclusive (ms)	% Work	Self-Time (ms)	% Self
P0: Scheduling & Resolution	309,751	19,144.6	78.4%	13,475.8	55.2%
P1: I/O & Heavy Tasks	3,078	147.1	0.6%	113.5	0.5%
P2: Architecture (Locks/Memory)	0	0.0	0.0%	0.0	0.0%
P3: Asset Pipeline	57,153	7,765.5	31.8%	6,357.3	26.0%
P4: Bridge/Interop	0	0.0	0.0%	0.0	0.0%
Other	1,072	338.7	1.4%	325.4	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
5,628.0	6,643.0	119,221	47.2	0.1	18.2	23.0%	module	write all entrypoints to disk (1%)
3,022.5	4,179.9	36,260	83.4	0.2	194.0	12.4%	analyze ecmascript module	process module (80%)
1,677.6	2,699.8	17,157	97.8	0.4	46.2	6.9%	code generation	chunking (1%)
1,392.8	4,159.3	56,575	24.6	0.0	20.3	5.7%	process module	module (9%)
1,331.9	1,444.3	12,929	103.0	0.2	187.5	5.5%	chunking	write all entrypoints to disk (0%)
1,324.5	1,386.5	7,078	187.1	0.6	35.9	5.4%	parse ecmascript	analyze ecmascript module (32%)
1,162.2	1,277.8	45,211	25.7	0.0	9.0	4.8%	internal resolving	resolving (29%)
887.0	887.0	9,434	94.0	0.3	5.9	3.6%	precompute code generation	code generation (59%)
870.5	1,292.8	36,158	24.1	0.0	15.6	3.6%	resolving	module (14%)
715.8	715.8	7,004	102.2	0.4	59.2	2.9%	compute async module info	chunking (0%)
622.2	783.5	7,722	80.6	0.0	129.6	2.5%	write all entrypoints to disk	None (0%)
432.8	433.2	6,984	62.0	0.0	41.4	1.8%	compute async chunks	webpack loader (0%)
327.4	327.4	2,083	157.2	0.4	17.5	1.3%	generate source map	code generation (96%)
301.8	487.0	579	521.3	2.3	47.3	1.2%	parse css	module (8%)
297.4	304.4	964	308.5	0.8	20.2	1.2%	webpack loader	parse css (12%)
62.6	62.6	2,340	26.7	0.0	3.5	0.3%	read file	parse ecmascript (91%)
53.9	53.9	802	67.2	0.0	13.2	0.2%	compute binding usage info	write all entrypoints to disk (0%)
32.8	32.8	37	885.3	5.9	10.9	0.1%	write file	apply effects (100%)
24.6	27.1	534	46.1	0.1	2.4	0.1%	async reference	write all entrypoints to disk (1%)
21.8	21.8	1,034	21.1	0.0	9.8	0.1%	collect mergeable modules	compute merged modules (0%)

---

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	194.1	2	process module → analyze ecmascript module
2	63.7	2	code generation → generate source map
3	59.7	2	module → parse css
4	47.3	2	code generation → generate source map
5	45.8	2	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
analyze ecmascript module	36,260	process module (80%)	83.4 us	0.15 ms	3,022.5 ms

---

Duration Distribution

Range	Count	Percentage
<10us	1,234,781	76.9%
10us-100us	347,785	21.7%
100us-1ms	18,827	1.2%
1ms-10ms	4,248	0.3%
10ms-100ms	190	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