chore: release 0.9.0

[stormslowly]

crate: rspack-resolver@0.9.0
npm: @rspack/resolver@0.5.0

[Copilot]

Pull request overview

Release bump updating the Rust crate and published Node package versions, plus lockfile changes to reflect updated workspace state for the npm release process.

Changes:

• Bump crate rspack_resolver version to 0.9.0 (and update Cargo.lock accordingly).
• Bump npm package @rspack/resolver version to 0.5.0.
• Update pnpm-lock.yaml, including new workspace importer entries under bindings/*.

Reviewed changes

Copilot reviewed 2 out of 4 changed files in this pull request and generated 1 comment.

File	Description
pnpm-lock.yaml	Adds workspace importer entries for per-platform binding packages (and a wasm dependency entry).
npm/package.json	Bumps published npm package version to 0.5.0.
Cargo.toml	Bumps Rust crate version to 0.9.0.
Cargo.lock	Updates locked package version to 0.9.0.

Files not reviewed (1)

• pnpm-lock.yaml: Language not supported

---

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

[codspeed-hq]

Merging this PR will degrade performance by 4.2%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

❌ 2 regressed benchmarks
✅ 10 untouched benchmarks

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Memory	resolver[pnp resolve]	8.4 KB	8.8 KB	-4.07%
❌	Memory	resolver[multi-thread]	11.1 MB	11.6 MB	-4.33%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing release/0.9.0 (115fa0b) with main (6381639)}

[stormslowly]

@codspeedbot explain the memory regression

[chatgpt-codex-connector]

You’re seeing a small reported memory regression (~4.2%), but the CodSpeed report itself strongly suggests this is likely measurement/environment noise rather than a real code-caused regression.

What the report is actually saying

From the PR comments:

• Regressions are only on Memory mode for:
  • resolver[pnp resolve]: 8.4 KB → 8.8 KB (~+0.4 KB)
  • resolver[multi-thread]: 11.1 MB → 11.6 MB (~+0.5 MB)
• CodSpeed includes a warning: “Different runtime environments detected.”

That warning matters a lot: CodSpeed is explicitly telling you the base and head benchmark runs may have happened on different runtime environments (machine/OS/libc/etc), which can shift allocator behavior and memory accounting enough to look like a few-percent regression.

Why this is probably not a real regression in this PR

This PR appears to be a release/version bump PR (Cargo/npm version + lockfile updates), with no substantive resolver algorithm change described in the comments. Given that context, a memory shift can come from:

1. Environment drift (most likely, because CodSpeed flagged it),
2. Lockfile/workspace shape changes affecting build/runtime graph slightly,
3. Natural variance around small absolute deltas (e.g., +0.4 KB is tiny).

Practical interpretation

• Treat this as “needs validation”, not “confirmed regression.”
• The key signal is the CodSpeed warning about runtime mismatch.
• If rerun in a pinned/stable environment and it reproduces consistently, then it’s real; otherwise it’s noise.

What to do next (quick path)

1. Re-run CodSpeed on the same commit.
2. Ensure base/head run on the same runner image/hardware class.
3. If available, run CodSpeed memory locally/CI with fixed toolchain and compare twice.
4. If stable regressions remain, inspect allocation hotspots for:
   • resolver[pnp resolve]
   • resolver[multi-thread]

---

Commands/Checks I ran

• ✅ rg --files -g 'AGENTS.md' (to check repo-specific agent instructions; none found in this workspace)
• ✅ Web lookup of CodSpeed documentation on memory mode and environment variance to validate interpretation.

No code changes were made, so no commit or PR creation was required.

View task →

[Copilot]

Pull request overview

Copilot reviewed 3 out of 5 changed files in this pull request and generated 2 comments.

Files not reviewed (1)

• pnpm-lock.yaml: Language not supported