build: Update swc_core to v66.0.3

[kdy1]

No description provided.

[changeset-bot]

⚠️ No Changeset found

Latest commit: 5235d1b

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[claude]

PR Review: build: Update swc_core to v66.0.3

This is a clean dependency-bump PR — pure version updates across Cargo.toml/Cargo.lock/package.json/CHANGELOG.md, with no Rust source changes. The fact that no source code adapted to swc_core 65→66 means the relevant plugin APIs are presumably source-compatible at the call sites in this repo (the actual signal is the green Integration Tests once they finish — they were still pending at review time).

What I'd flag

1. Empty PR description. The PR body is empty. A one-line summary plus a link to the swc_core 66.0.3 release notes / changelog would help reviewers and future archaeologists understand why this bump landed (any notable behavior changes? bug fix users were waiting on?). Each downstream CHANGELOG.md entry repeats only "Update swc_core to v66.0.3" — the user-visible impact is opaque.

2. Internal transform crates skip a major version. Multiple internal transform crates jump from 3.0.0 → 5.0.0 (e.g. react_remove_properties, remove_console, modularize_imports, styled_components, styled_jsx) and the plugin wrapper crates jump 4.0.0 → 6.0.0 (e.g. swc_emotion, swc_confidential, swc_magic, swc_sdk, swc_formatjs_transform 27 → 29, etc.). If 5.0.0 / 4.0.0 were never published, this is harmless; if they were published from a different branch and yanked, fine; otherwise consumers tracking these crates on crates.io will see a confusing gap. Worth a quick sanity check that cargo publish won't choke.

3. Major-version bumps in transitive swc_* deps. Beyond swc_core itself, the workspace pulls in major-version bumps for preset_env_base (7→8), swc_css_prefixer (25→26), swc_ecma_minifier (52→53), swc_ecma_transforms (52→53), and swc_ecma_transforms_react (46→47). These are re-exported by swc_core so the plugin surface should be fine, but for any package that takes a direct dep on one of these (rather than going through swc_core), confirm features still match.

4. Duplicate dependency versions in the lockfile. Cargo.lock now contains both dashmap 5.5.3 (pulled in by lightningcss) and dashmap 6.2.1 (pulled in by preset_env_base/swc_config), and similarly multiple hashbrown major versions. Not a correctness issue, but it inflates compile time and WASM binary size — worth noting if plugin size budgets matter. Probably nothing to do here until lightningcss catches up.

Looks good

• Changeset / CHANGELOG entries are present and consistent across all affected packages.
• package.json version bumps match Cargo.toml version bumps.
• Workspace versions in the root Cargo.toml line up with what Cargo.lock resolves to.
• Scope is narrow: no behavioral changes mixed in with the version bump.

Test coverage / verification

There's nothing to add here in terms of test coverage — this is a dep bump and the existing integration tests are the right gate. The merge should wait for Integration Test (macos-latest) / Integration Test (ubuntu-latest) / Test - windows-latest / Test - ubuntu-latest to come back green, since those are the only signal that the swc_core API changes between 65 and 66 didn't break anything at runtime.

Security

No new direct dependencies introduced beyond what swc_core 66 transitively brings in. rkyv 0.8.10 → 0.8.16 is a patch-level bump of an unsafe-heavy crate (zero-copy deserialization); worth a glance at their changelog for anything CVE-shaped, but nothing in the diff itself is concerning.