SIGABRT (SI_TKILL) in pacquet-main thread on pnpm 12.14.0 during 'pnpm exec tauri build' #17204

Activity

1. github-actions commented on Oct 21, 2026

   

   github-actionsBot

   on Oct 21, 2026 – with GitHub Actions

   Contributor

   More actions

   @marktantongco, are you planning to open a pull request for this? If so, reply here and a maintainer will assign the issue to you, so nobody else starts on the same fix.
2. zkochan commented on Oct 21, 2026

   

   zkochan

   on Oct 21, 2026

   Member

   More actions

   Opened #17214 to fix this completely.

   What happened in your stack trace

   Disassembling the Arch Linux pnpm 12.14.0-1 binary (Build ID: f72e42f09888d483fcb097eaf32c4e7e95622f72) at your PIE base (0x55eb0ae7e000) resolves the 8 /usr/bin/pnpm frames on pacquet-main to:

   Frame	Address	Offset	Function
   #4	0x55eb0cc940c8	0x1e160c8	pnpm_fs::pending_temp::die_from_signal(6) (raise(SIGABRT))
   #5	0x55eb0cc81275	0x1e03275	pnpm_executor::interrupt::exit_like
   #6	0x55eb0bf01535	0x1083535	pnpm::commands::exec::ExecArgs::run
   #7	0x55eb0c435a86	0x15b7a86	pnpm::commands::run::execute_script_command
   #8–#11	0x55eb0b8eaa8a…	0xa6ca8a…	Tokio block_on / pacquet-main thread entry

   pnpm itself did not crash during the build. Instead, the child tauri process aborted with SIGABRT (signal 6), and pnpm exec called exit_like -> die_from_signal(SIGABRT) to re-raise the child's termination signal on pnpm so the parent shell receives the same exit status. Because die_from_signal did not suppress core dumps before calling raise(SIGABRT) (SI_TKILL), Linux's do_coredump invoked systemd-coredump for /usr/bin/pnpm as well.

   Fix in #17214

   pnpm_fs::die_from_signal now sets RLIMIT_CORE to 0 and (on Linux/Android, where do_coredump ignores RLIMIT_CORE when kernel.core_pattern pipes to systemd-coredump) clears PR_SET_DUMPABLE via prctl(PR_SET_DUMPABLE, 0, ...) before re-raising the signal. pnpm still exits with the same signal (WTERMSIG(status) == SIGABRT), without generating a bogus core dump for /usr/bin/pnpm.

   Note on the underlying tauri build abort in 68proxy

   In 68proxy (src-tauri/Cargo.toml pulls in tauri with tray-icon -> libappindicator), cargo build --release succeeds without libayatana-appindicator installed because libappindicator-sys loads the .so via dlopen rather than linking it at compile time, and then @tauri-apps/cli-linux-x64-gnu panics on a worker thread during the post-compile bundling step (crates/tauri-cli/src/interface/rust.rs:1722:17: Can't detect any appindicator library), aborting the node (tauri) process with SIGABRT. Installing libayatana-appindicator on Arch/Omarchy (sudo pacman -S libayatana-appindicator) resolves the tauri abort.

   Written by an agent (Antigravity, claude-opus-4-7).
3. zkochan

   closed this as completedin #17214on Oct 21, 2026
4. opensource-fork-sync

   added a commit that references this issue on Oct 21, 2026

   fix(exec): do not dump core for pnpm when re-raising a child signal (p…

   ...

   This issue will close once commit e2cc154 is merged into the 'main' branch.

   e2cc154
5. markharding

   added a commit that references this issue on Nov 3, 2026

   Update pnpm to v12.16.1 (#2164)

   ...

   fd46826