[rush] Global custom commands: SIGTERM to rush is not forwarded and the shell command keeps running #6125

🤖 Filed by GitHub Copilot acting for Naman Kanakiya (@namankanakiya) (light review only).

Package: @microsoft/rush-lib 5.178.1 and 5.181.0. Linux.

Repro: in common/config/rush/command-line.json:

{
  "commandKind": "global",
  "name": "loop",
  "summary": "loop",
  "shellCommand": "node -e \"setInterval(() => console.log('alive'), 1000)\""
}

Run rush loop, then from another shell kill -TERM <pid of the rush node process>.

Observed: rush exits, but sh -c node -e ... keeps running with ppid 1 and keeps printing. Ctrl+C works only because the terminal signals the whole foreground process group. A supervisor, IDE task runner or script that stops rush with SIGTERM leaves the custom command running headless (for a long-running dev-server command, that is the whole dev server and its children).

Expected: terminating rush terminates its shell command.

Cause: GlobalScriptAction runs the command via Utilities.executeLifecycleCommand, which uses spawnSync, so no JavaScript signal handler can run while it blocks.

Suggested fix: use executeLifecycleCommandAsync and forward SIGINT/SIGTERM/SIGHUP to the child (or its process group), then exit with the child's code.

Happy to test a candidate fix.