Bug: Job cancellation sends SIGTERM only — hung nmap processes never get SIGKILL

[techmore]

Type

bug

Severity

medium

Area

nmapui/jobs.py — job cancellation logic

Description

When a scan job is cancelled, the code sends process.terminate() (SIGTERM) but never escalates to process.kill() (SIGKILL) if the process does not exit. Nmap can sometimes ignore SIGTERM (particularly during certain NSE script phases), leaving zombie processes consuming resources.

if process and process.poll() is None:
    try:
        process.terminate()
    except Exception:
        logger.exception("...")

Proposed Fix

Add a timeout-and-kill escalation:

if process and process.poll() is None:
    try:
        process.terminate()
        try:
            process.wait(timeout=5)
        except subprocess.TimeoutExpired:
            logger.warning("Process %s did not exit after SIGTERM, sending SIGKILL", process.pid)
            process.kill()
            process.wait(timeout=3)
    except Exception:
        logger.exception("Failed to terminate process")

Related Issues

#195 (Menu bar close must kill all app instances)