[BUG] Bun Bus error crash regression on macOS Silicon (0xBAD4007) — persists in CC 2.1.126 / Bun 1.3.14

Description

This is a regression / re-report of #41603 (closed). The same Bun Bus error crash (0xBAD4007) is still occurring on macOS Silicon, now on Claude Code 2.1.126 with Bun 1.3.14. The previous report was on CC 2.1.87 / Bun 1.3.11 — the bug was not fixed, it regressed or was never fully resolved.

This time the crash happened after a long session (~1h45m elapsed, 21m active) rather than early in the session (90s in the original report), suggesting this may be a broader memory corruption issue that manifests under different conditions.

Environment

• Claude Code version: 2.1.126
• Bun version (bundled): 1.3.14 (0a466a11)
• OS: macOS 26.4.1 (Darwin 25.4.0)
• Platform: macOS Silicon (Apple M4 Pro)
• RAM: 48GB

Crash Details

Bun v1.3.14 (0a466a11) macOS Silicon
Args: "claude" "--dangerously-skip-permissions" "--thinking-display" "summarized"
Features: Bun.stderr(2) Bun.stdin(2) Bun.stdout(2) WebSocket abort_signal(15) fetch(330) jsc spawn(7417) standalone_executable process_dlopen yaml_parse(330) claude_code
Builtins: "bun:main" "node:assert" "node:async_hooks" "node:buffer" "node:child_process" "node:constants" "node:crypto" "node:dns" "node:events" "node:fs" "node:fs/promises" "node:http" "node:https" "node:module" "node:net" "node:os" "node:path" "node:path/posix" "node:path/win32" "node:perf_hooks" "node:process" "node:readline" "node:stream" "node:stream/consumers" "node:string_decoder" "node:timers/promises" "node:tls" "node:tty" "node:url" "node:util" "node:vm" "node:zlib" "ws" "node:http2"
Elapsed: 6332756ms | User: 209638ms | Sys: 31838ms
RSS: 0.51GB | Peak: 0.80GB | Commit: 0.02ZB | Faults: 254 | Machine: 51.54GB

panic(main thread): Bus error at address 0xBAD4007
oh no: Bun has crashed. This indicates a bug in Bun, not your code.

Bun Crash Report Link

https://bun.report/1.3.14/M_10a466a1mgkgkE+hogC__+mp9H_myptpDurxjpDm/pv5C+phs5C2gsr8C2rrr8C+g1j+Cm34gN_A4Augg11L

Key Differences from Original Report (#41603)

| | Original (#41603) | This report |
|---|---|---|
| CC version | 2.1.87 | 2.1.126 |
| Bun version | 1.3.11 | 1.3.14 |
| macOS | 26.3 | 26.4.1 |
| Session elapsed | ~90s | ~1h45m (6332s) |
| RSS at crash | 0.89GB | 0.51GB |
| Peak RSS | 0.89GB | 0.80GB |
| Trigger | Image paste (chat:imagePaste) | Long session (unclear specific trigger) |
| Spawned processes | 28 | 7417 |
| Fetch calls | 14 | 330 |

Notes

• The crash address 0xBAD4007 is identical to the original report, strongly suggesting the same underlying bug
• This session was much more active: 7,417 spawned processes vs 28, and 330 fetch calls vs 14
• RSS was actually lower at crash (0.51GB vs 0.89GB), so this isn't a simple OOM situation
• The Commit: 0.02ZB value looks like a memory accounting overflow/corruption itself
• The claude_code feature flag is now listed in Features (wasn't in the original), suggesting Bun's claude_code-specific code paths may be involved
• The WebSocket feature is also new compared to the original crash

Expected Behavior

Claude Code should not crash with a Bun Bus error. Long sessions should either remain stable or gracefully handle memory pressure.

Previous Report

#41603 — closed without resolution. This is the same crash signature recurring across multiple CC and Bun versions, indicating a persistent upstream issue that needs to be tracked.