Remove dead code from bun_format and napi.cpp

[robobun]

Problem

• Two trait impls in bun_format have no user: Spanned for FormatExpr (src/format/js/format.rs) and Spanned for FormatSpecifier (src/format/js/print/import_declaration.rs). rustc's dead_code lint exempts trait impls, so the deny lints do not report them.
• Bun__napi_get_version (src/jsc/bindings/napi.cpp) has no caller.

Fix

• Delete the two impls and the function: 3 files, 18 lines.
• Correct because nothing can reach them. No crate of the debug build instantiates either span method. A relink with --gc-sections drops Bun__napi_get_version, and rg finds only its definition. It is not in src/symbols.txt or src/symbols.def.
• Verified: bun run rust:check-all (12 targets), cargo check --workspace with --all-targets and --release, bun bd, test/cli/format/format.test.ts, and the napi_get_version tests of test/napi/napi.test.ts.

Background

• Spanned (bun_lint::span) gives a value its source range. FormatExpr and FormatSpecifier are private wrapper types of the formatter. The formatter asks the wrapped node for its range, not the wrapper.
• To rustc a trait impl is always reachable, because generic code could call it. Only a whole-program view shows that nothing does.
• Remove dead code from bun_format, bun_sema, bun_install, and bun_core #45402 removes six Spanned impls of the same kind. These two have no symbol in any object file, because no crate instantiates them, so a scan for dropped sections does not list them.
• No other design applies: the change only deletes unreferenced code.

Downsides

• None found. All 12 targets, the test targets and the release profile type-check without the impls. The lines that the 5,526 open PRs add use neither the impls nor Bun__napi_get_version.

Behaviour change: none

Notes

How the candidates were found

• Linker: the debug link command plus -Wl,--gc-sections -Wl,--print-gc-sections. It drops 13,148 Rust function sections and 380 functions in handwritten C++ objects.
• Compiler: 10,149 pub items got a temporary #[deprecated] mark (not committed). rustc then reports each use. Passes: the 12 CI targets, --all-targets on three of them, --release, and --cfg bun_asan --cfg bun_debug --cfg socket_fault_injection.
• Trait impls: delete every candidate impl, run cargo check --workspace, put back each impl that an error names, repeat.
• cargo mordant at the pinned rev with an empty baseline, on main and on the candidates. No file on main is over its unused_pub baseline.

Checks in detail

• test/cli/format/format.test.ts with the debug ASAN build: 1015 pass, 0 fail with --timeout 120000. With the default 5 s timeout, 15 tests that start a JavaScript config loader time out on this machine.
• test/napi/napi.test.ts -t napi_get_version: 2 pass.
• cargo clippy -p bun_format --no-deps: clean.

What the scan found and where it went

• 184 pub items have no use on any target. Open dead-code PRs already delete 138 of them.
• The other 46 are not in this diff:
  • 4 are used only under a cfg that no check builds: bun_zstd::inflate_embedded and inflate_embedded_nul (bun_codegen_embed), two functions in src/bun_core/code_generation.rs (bun_disallow_code_generation_from_strings).
  • 2 are named by src/codegen/generate-classes.ts: host_fn_this_value, host_fn_setter.
  • 26 are in tables that mirror OS headers: 22 constants in src/errno/windows_errno.rs, Tag::fchownat, the macOS O::EVTONLY, fault_inject::SHUTDOWN, posix::mode_t.
  • 4 are ArrayListAlignedIn::{init, init_capacity, from_unmanaged, append_assume_capacity}. Their only user is src/jsc/bindgen.rs, which an open PR deletes.
  • 9 are used by open PRs, so a removal breaks main when they merge: O::NOATIME (Remove libuv on Windows #42819), O::NOFOLLOW_ANY (archive: refuse symlinks in the destination with openat2(RESOLVE_NO_SYMLINKS) and O_NOFOLLOW_ANY #41750), posix::s_isdir (fs.watch: watch symlink entries of a recursive tree on Linux #43090, bun patch: keep the copy of a global store package in the project #43556), the Windows umask (Remove libuv on Windows #42819, ini: reject a source of 2 GiB or more in Parser::parse instead of aborting on an int cast #45206), the Windows UTIME_NOW (Remove libuv on Windows #42819, shell(touch): let the OS pick the time so that write access is enough #44937), SSL_set0_verify_cert_store (Add ForeignRef<T> and make FetchHeaders the owned handle #33820, Fold CppWebSocketRef onto ForeignRef with a release marker #33889).
  • 1 is StoredTrace::from, which Remove dead code from bun_format, bun_sema, bun_install, and bun_core #45402 removes.
• Trait impls: 77 impl blocks (not counting Debug) that were added since the base of Remove dead code: 49 trait impls nothing uses, across 12 crates #43664 have no kept function in the link.
  • 41 implement project traits. The workspace does not compile without 24. Remove dead code from bun_format, bun_sema, bun_install, and bun_core #45402 removes 6. This PR removes 2. 9 stay by choice: the three Located impls that the trait's doc comment lists, three StaticallyNamed impls, Emit for f64 and Spanned for Target in bun_lint (the API that lint rules are written against), and AsAstNodes for Class in bun_format (one of a family with one impl per handle kind).
  • 36 implement std traits (Hash, Clone, PartialEq, Default, From, Deref, Display). This PR removes none of them, as in Remove dead code: 49 trait impls nothing uses, across 12 crates #43664.
• Handwritten C++: the linker drops 121 extern "C" functions. Open PRs delete 86. The rest have a caller on macOS or Windows, are --wrap shims, are vendored llhttp, or are in the follow-up list.
• Also clean or already claimed: .rs files outside every mod tree, unused Cargo dependencies (cargo's unused_dependencies lint on every target), src/js (tsc --noUnusedLocals reports 5 locals), scripts/, src/codegen/, packages/.

Follow-up candidates, not in this diff

• hawk.toml: 12 [[override]] entries name enum variants that no longer exist (11 NapiStatus::*, and serialized_failure::ErrorKind::JsError). They touch the hunk of Remove dead code from libtcc1.c, four C++ preprocessor conditions, JSPropertyIterator, and hawk.toml #44126.
• Resource Timing: ResourceTiming has a private constructor and no caller, so PerformanceResourceTiming::create, its constructor, ResourceTiming::populateServerTiming, PerformanceServerTiming::create and NetworkLoadMetrics.cpp are unreachable (about 159 lines). Remove dead code from bun_core, the JSC bindings, the debugger, and 24 other crates #41385 and Remove dead code from the build scripts, the vscode and debug-adapter packages, the bake overlay, and two WebCore files #40122 edit the same functions.
• Two C++ functions have a Rust caller that is never instantiated: Bun__allocUint8ArrayForCopy (every call of ArrayBuffer::alloc passes JSType::ArrayBuffer) and Bun__JSBigIntStatsObjectConstructor (only StatsSmall::get_constructor has a caller). A removal edits live generic code.
• mordant-baseline.toml: once Remove dead code from bun_format, bun_sema, bun_install, and bun_core #45402 lands, "unused_pub:src/bun_core/Global.rs" = 1 is stale. mordant reports 0 findings in that file without StoredTrace::from.
• 27 hand-written Debug impls that nothing formats, mostly in bun_lint and bun_format. They are a debugging aid, so they stay.

[robobun]

^{Updated 12:05 AM PT - Oct 17th, 2026}

❌ @robobun, your commit 83386fc has 2 failures in Build #127070 (All Failures):

• step failed outside runner - exit 1 on 🍎 any aarch64 - test-bun
• ❌ CPU instruction violation on Windows x64 — 1 check(s) failed

The baseline build contains instructions not available on Nehalem (SSE4.2, no AVX/AVX2/AVX512).

• Static instruction scan

Static scan violations

_powf_special  [AVX]  (1 insns)
If these are runtime-dispatched behind a CPUID gate: add each symbol to
scripts/verify-baseline-static/allowlist-x64-windows.txt with a comment pointing at the gate.
If there's no gate: this is a real bug — a -march leaked into a subbuild.

---

🧪   To try this PR locally:

bunx bun-pr 45420

That installs a local version of the PR into your bun-45420 executable, so you can run:

bun-45420 --bun

[robobun]

Status: ready for review. The diff is green. CI is red on two lanes that this change does not touch.

CI on the head commit 83386fc (build 127070): 177 jobs pass, 2 fail.

• windows x64 - verify-baseline: the static scan reports _powf_special [AVX] Vphsubw at 0x014519c364. The build of main at the same base commit d7562f2 fails with the same symbol and the same address (build 126987). Later builds of main pass that job.
• darwin aarch64 - test-bun: the artifact download timed out after 120 s. No test ran.

How the dead code was confirmed:

• Spanned for FormatExpr and Spanned for FormatSpecifier: no object file of the debug build has a symbol for either span method. cargo check --workspace passes without the impls on all 12 targets (bun run rust:check-all), with --all-targets, and with --release. bun_format has no cfg gate and both types are private to the crate.
• Bun__napi_get_version: a relink of the debug build with -Wl,--gc-sections -Wl,--print-gc-sections drops it. rg Bun__napi_get_version finds only the definition.

PR: #45420

[coderabbitai]

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 2 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used all 10 included reviews currently available.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

• Configuration used: Repository: oven-sh/bun/.coderabbit.yaml
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: c4830f1e-4258-470e-aa16-e293883921f4

📥 Commits

Reviewing files that changed from the base of the PR and between 1cf3835 and 83386fc.

📒 Files selected for processing: 3

• src/format/js/format.rs
• src/format/js/print/import_declaration.rs
• src/jsc/bindings/napi.cpp

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Repository: oven-sh/bun/.coderabbit.yaml
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: e3f372e9-fa11-4e40-812c-ea389817c2c2

📥 Commits

Reviewing files that changed from the base of the PR and between 116dd64 and 1cf3835.

📒 Files selected for processing: 3

• src/format/js/format.rs
• src/format/js/print/import_declaration.rs
• src/jsc/bindings/napi.cpp

💤 Files with no reviewable changes: 3

• src/format/js/print/import_declaration.rs
• src/jsc/bindings/napi.cpp
• src/format/js/format.rs

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 0 remain after this review.

Walkthrough

The change removes two Spanned implementations from JavaScript formatting types and removes the exported Bun__napi_get_version function from the N-API bindings.

Changes

Formatter span implementations

Layer / File(s)	Summary
Remove formatter span delegation src/format/js/format.rs, src/format/js/print/import_declaration.rs	FormatExpr and FormatSpecifier no longer implement Spanned by delegating to their wrapped values.

N-API version function

Layer / File(s)	Summary
Remove N-API version function src/jsc/bindings/napi.cpp	The exported Bun__napi_get_version function, which returned the N-API module version, was removed.

Priority: ⬇️ Low

Merge Risk: ⚪ Minimal · up to 1cf38

No actionable merge-blocking risk is established for these removals; merge after normal checks.

Pre-merge checks: 4

Check name	Status	Explanation
Title check		The title clearly identifies the main change: removing unused code from bun_format and napi.cpp.
Description check		The description explains the problem, lists the deleted code, states that behavior does not change, and provides detailed verification results. It does not use the exact template headings, but it cove…
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches: 💡 1

🛠️ Fix failing CI checks: 💡

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[claude]

Code review found no issues

No high-confidence issues detected in this change.