perf: Optimize ModuleGraph API implementation

[hardfist]

Summary

• Refine the ModuleGraph API with performance-focused improvements to better support downstream work on the codex/modulegraph branch.
• Keep behavior consistent while targeting efficiency gains in the module graph operations.

Testing

• Not run (not requested)

[Copilot]

Pull request overview

Refines ModuleGraph batch update APIs to avoid parallel cloning/reinsertion and instead mutate connections/modules in-place, aiming to reduce overhead in module graph operations.

Changes:

• Remove Rayon-based parallel mapping in several ModuleGraph batch update methods.
• Update connections and module graph modules in-place via *_mut accessors to avoid cloning and re-inserting.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

expect("should have connection") will panic without identifying which DependencyId was missing, making debugging difficult. Consider using unwrap_or_else/expect with a message that includes dep_id so failures in batch updates are actionable.

[Copilot]

Same as above: this expect("should have connection") panic message lacks the DependencyId context. Including dep_id in the panic/expect message would make failures easier to diagnose (similar to do_update_module nearby).

[github-actions]

📦 Binary Size-limit

Comparing 0807a01 to refactor: wrap artifact in StealCell (#12979) by hardfist

🎉 Size decreased by 48.00KB from 48.65MB to 48.61MB (⬇️0.10%)

[github-actions]

Rsdoctor Bundle Diff Analysis

Found 5 projects in monorepo, 0 projects with changes.

📊 Quick Summary

Project	Total Size	Change
react-10k	5.7 MB	0
react-1k	825.4 KB	0
react-5k	2.7 MB	0
rome	984.2 KB	0
ui-components	2.2 MB	0

Generated by Rsdoctor GitHub Action

[codspeed-hq]

CodSpeed Performance Report

Merging this PR will not alter performance

_{Comparing codex/modulegraph (0807a01) with main (9f4fc6e)}

Summary

✅ 16 untouched benchmarks
⏩ 1 skipped benchmark¹

Footnotes

1. 1 benchmark was skipped, so the baseline result was used instead. If it was deleted from the codebase, click here and archive it to remove it from the performance reports. ↩