Intel Mac (x64) app hangs on launch, stuck in libffi in cua_driver_node_runtime.node

Open
#27 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
3/5
Estimated time
1-2 days
Newbie friendliness
73/100
Issue type
Bug
Clarity
Clearly specified
Activity status
Active
Tech stack
macos, typescript
Domain
desktop

Research direction

Start by reading build/entitlements.mac.plist and the macOS packaging and signing configuration that applies its entitlements. Reproduce the launch hang with the Intel x64 build and check whether the proposed entitlement fixes it; the report includes steps, a sample stack, and a successful ad-hoc re-signing test. Done when the Intel app launches and its window appears, with the chosen fix verified.

Written by the indexing model from the issue text.

Description

What happened

I downloaded the x64 build on my Intel Mac and opened it, expecting the app window to show up. It never does. The app just beachballs with the main process sitting at around 60% CPU. The server child does start and listens on 8799, but no window ever appears.

I sampled the main process to see what it was doing (stack is in the logs below). The main thread is stuck in libffi inside cua_driver_node_runtime.node, which gets loaded at startup.

My guess at what's going on: on x86_64, libffi needs writable + executable memory for its closures. The app has the hardened runtime on but only has com.apple.security.cs.allow-jit, and that only covers MAP_JIT mappings, which libffi doesn't use. So the mmap fails, the temp file fallback fails too, and it just keeps looping. That would also explain why arm64 is fine, since libffi uses static trampolines there.

Adding this to build/entitlements.mac.plist fixed it for me:

<key>com.apple.security.cs.allow-unsigned-executable-memory</key>
<true/>

To test it, I took the released x64 app, added that entitlement to its existing ones and ad-hoc re-signed it with the same hardened runtime options. After that it launched fine: the server came up, the window showed after about 17s, and ffi_closure_alloc no longer appeared when I sampled the main thread. If you'd rather not widen the entitlements, building libffi in the CUA runtime with static trampolines or MAP_JIT on x86_64 should also work. Happy to open a PR with the entitlement change if that's the approach you want.

Steps to reproduce
  1. Download DuskBot-0.2.15-x64.zip (or DuskBot-intel.dmg) on an Intel Mac
  2. Open it
  3. No window, app stops responding. sample <pid> on the main process shows the stack below
Version

0.2.15 (x64)

Platform

macOS

Engine

No response

Logs or screenshots
macOS 15.7.9 (24G830), Intel MacBookPro16,1

Main thread, from `sample`:

ModuleWrap::Evaluate
  napi trampoline                      (cua_driver_node_runtime.node)
    uniffi_runtime_napi::call::call_ffi_function
      uniffi_runtime_napi::callback::vtable::build_vtable_struct
        uniffi_runtime_core::callback::Module::make_callback_trampoline
          libffi::middle::Closure::try_new
            ffi_closure_alloc
              dlmalloc
                dlmmap
                  mmap                 <- fails
                  open_temp_exec_file_dir
                    mkstemp / openat / fstatat   <- keeps retrying forever

server.log (startup gets as far as forking the server, then nothing):
move to applications: moved, relaunching
startup HTTP cache cleared (no-persisted-version: unknown -> 0.2.15)
fork /Applications/Dusk Bot.app/Contents/Resources/server/desktop-entry.mjs port=8799
Dominant language
TypeScript
Stars
277
Forks
27
Avg merge
2h 47m
Merged PRs (30d)
1

Getting set up

This project ships no dev container, Dockerfile or contributing guide, so setting up is up to you: start from its README, and see our first-contribution guide for the general steps.

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from DuskBotAI/duskbot

All issues in DuskBotAI/duskbot

Similar issues

More TypeScript issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.