fix(pack): cover merged re-exports from shared modules

[fireairforce]

Summary

Add a concatenate_modules regression fixture for a shared import re-exported through multiple merged barrels. The fixture executes the emitted module factories and reads the re-export getters. Before the Turbopack fix, it throws the unbound __TURBOPACK__imported__module__* ReferenceError seen in imall-main.

Update the next.js submodule to a787299fbf, the merged fix from utooland/next.js#196. This PR is rebased onto next after #3391 merged.

Verification

• cargo test -p pack-tests --test snapshot concatenate_modules -- --nocapture (6 passed on the rebased commit)
• cargo fmt --check in the next.js submodule
• cargo clippy --all-targets -- -D warnings --no-deps in the next.js submodule on the equivalent pre-merge patch
• Full imall-main production build with the patched release-local native binding and affected dependency versions succeeded. A Babel scope scan of all 932 JS output files found zero unbound generated import identifiers; the same scan found two in the original build.

[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 2026-09-26T09:53:28.020822Z	22bdcbc	Draft marked ready

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.

[PeachScript]

+1

[github-actions]

📊 Performance Benchmark Report (with-antd)

Utoopack Performance Report

Report ID: utoopack_performance_report_20260926_100618
Generated: 2026-09-26 10:06:18
Trace File: trace_antd.json (0.3GB, 0.83M spans)
Test Project: examples/with-antd

---

Executive Summary

Metric	Value	Assessment
Total Wall Time	5,966.1 ms	Baseline
Total Thread Work (de-duped)	17,910.8 ms	Non-overlapping busy time
Effective Parallelism	3.0x	thread_work / wall_time
Working Threads	10	Threads with actual spans
Thread Utilization	30.0%	⚠️ Suboptimal
Total Spans	826,031	All B/E + X events
Meaningful Spans (>= 10us)	196,974	(23.8% of total)
Tracing Noise (< 10us)	629,057	(76.2% 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	41,449	5,429.9	1,462.5	2,668.1
Parse	8,154	1,253.7	933.9	5,164.0
Analyze	127,060	35,660.9	7,500.4	5,079.3
Chunk	5,281	5,274.6	843.0	2,174.2
Codegen	11,812	2,047.2	1,321.3	1,875.5
Emit	34	31.3	15.7	6.4
Other	3,184	7,263.2	4,265.4	5,966.0

---

Workload Distribution by Diagnostic Tier

Category	Spans	Inclusive (ms)	% Work	Self-Time (ms)	% Self
P0: Scheduling & Resolution	169,149	41,567.7	232.1%	9,145.3	51.1%
P1: I/O & Heavy Tasks	2,700	90.3	0.5%	74.7	0.4%
P2: Architecture (Locks/Memory)	0	0.0	0.0%	0.0	0.0%
P3: Asset Pipeline	23,829	8,624.2	48.2%	3,132.9	17.5%
P4: Bridge/Interop	0	0.0	0.0%	0.0	0.0%
Other	1,296	6,678.7	37.3%	3,989.3	22.3%

---

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
3,630.9	20,612.3	85,152	42.6	0.1	11.9	20.3%	module	module (60%)
2,866.6	3,054.1	18	159254.2	319.2	594.7	16.0%	save snapshot	persist (6%)
1,806.8	1,926.6	2,532	713.6	2.2	242.7	10.1%	analyze ecmascript module	module (69%)
1,246.9	12,126.1	30,829	40.4	0.1	4.8	7.0%	process module	process module (81%)
969.7	2,603.5	22,454	43.2	0.1	4.0	5.4%	internal resolving	internal resolving (73%)
880.5	1,200.3	5,994	146.9	0.6	40.0	4.9%	parse ecmascript	parse ecmascript (66%)
731.1	4,986.6	3,966	184.3	0.2	125.2	4.1%	chunking	chunking (45%)
724.5	796.5	9,703	74.7	0.3	6.0	4.0%	precompute code generation	generate merged code (42%)
684.8	770.6	6,039	113.4	0.4	109.8	3.8%	compute async module info	compute async module info (51%)
584.5	1,906.4	1,032	566.4	1.3	197.2	3.3%	generate merged code	chunking (47%)
487.2	2,820.9	18,489	26.4	0.0	4.9	2.7%	resolving	module (57%)
417.6	417.6	329	1269.2	0.8	243.3	2.3%	generate source map	code generation (83%)
321.2	722.7	130	2470.6	3.8	194.7	1.8%	emit code	generate merged code (42%)
261.8	491.9	1,495	175.1	0.1	135.4	1.5%	write all entrypoints to disk	write all entrypoints to disk (14%)
190.8	190.8	9	21194.8	95.4	144.8	1.1%	blocking	save snapshot (56%)
179.2	833.1	1,780	100.7	0.2	40.8	1.0%	code generation	code generation (83%)
109.4	284.9	1,282	85.3	0.1	11.0	0.6%	compute async chunks	compute async chunks (38%)
88.1	102.1	742	118.7	0.0	35.1	0.5%	compute binding usage info	compute binding usage info (49%)
53.4	53.4	2,157	24.8	0.0	3.0	0.3%	read file	parse ecmascript (92%)
31.9	58.0	1,744	18.3	0.0	12.1	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	739.6	3	persist → save snapshot → blocking
2	394.2	6	chunking → generate merged code → emit code → emit code → emit code → read file
3	280.9	4	chunking → generate merged code → emit code → generate source map
4	268.3	2	save snapshot → blocking
5	242.8	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	30,829	process module (81%)	40.4 us	0.06 ms	1,246.9 ms
internal resolving	22,454	internal resolving (73%)	43.2 us	0.08 ms	969.7 ms

---

Duration Distribution

Range	Count	Percentage
<10us	629,057	76.2%
10us-100us	126,259	15.3%
100us-1ms	61,407	7.4%
1ms-10ms	9,111	1.1%
10ms-100ms	156	0.0%
>100ms	41	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