startupscript

Documentation

Overview

Package startupscript runs an admin-defined shell script when the server boots. Use case: spawn a tunnel (ngrok / cloudflared) so the local HTTP port stays unexposed to the LAN — the admin configures the command via /admin/variables (startup_script + startup_script_enabled) and the server fires it once Bootstrap finishes.

Lifetime is tied to the caller's context: when the context cancels (server stop / tray quit) the subprocess and every child it spawned receive a kill signal via the platform's process-group mechanism (Setpgid + kill -pgid on Unix, Job Object kill-on-close on Windows). Without this, a backgrounded `ngrok &` survives wick shutdown as an orphan re-parented to init — leaking tunnels and accumulating zombies on every restart.

Output (stdout + stderr) is appended to ~/.<appName>/logs/startup-script-YYYY-MM-DD.log.

Multi-line scripts run sequentially in one shell process, exactly like a `.sh` / `.ps1` file. Long-running foreground commands block subsequent lines — use `&` (Unix) or `Start-Process` (Windows) for parallel daemons. Edits to the script row only take effect on next server boot.

Index

• func Run(ctx context.Context, appName, script string) error

Functions

func Run

func Run(ctx context.Context, appName, script string) error

Run spawns the script in a fresh shell, waits for context cancel, then kills the subprocess plus every descendant via the platform's process-group mechanism. Blocking call — callers spawn it in a goroutine. Returns immediately with nil when script is empty or only whitespace.

appName scopes the log file location (~/.<appName>/logs/...). Pass the same string used elsewhere (APP_NAME env / userconfig.Dir input) so logs land next to server/worker/app logs.