adding windows support

README.md

@@ -21,7 +21,7 @@ The simplicity of this code also helps promote its support, usage, and adaptatio
 
 # Usage (Installation)
 
-To get started simply install the requirements with `pip install pyyaml` or `pip install -r requirements.txt` and then `python3 run.py` or `./run.py`.  It should immediately start up and run on Linux/Mac.  On Windows you will need a different config file (todo, add an example for windows).
+To get started simply install the requirements with `pip install pyyaml` or `pip install -r requirements.txt` and then `python3 run.py` or `./run.py`.  It should immediately start up and run on Linux/Mac.  On Windows you will need an slightly different config file, basically just use Windows-specific paths for chdir.  See `tasks-windows.yml`.  Also on Windows CTRL-C doesn't always kill the entire engine and all children.  You may need to force-kill the process from the task manager, or simply close the window if you ran it in an Windows Command Prompt/shell.
 
 To begin configuring the tasks you wish to run, please edit the `tasks.yml` file and add the commands you wish to run.  The file is well documented and self explanatory.  Put in the tasks you wish to add, and re-run `run.py` and watch it work.
 
@@ -84,7 +84,6 @@ To this above Docker image you will install the threaded engine, then add your c
 * Make init script(s) for some popular Linux distributions
 * Package this and submit it to package repositories (Ubuntu/Debian/CentOS)
 * Alternatively, package and distribute this via pip
-* Do some/more Windows testing, add documentation and examples for Windows.  It _should_ work but the author is mainly an OS-X/Linux guy.
 * Write some tests for each of the various features to be sure it works and continues to work as intended
 * Make all logs pre-thread have a forced prefix?  Would need to buffer output to accomplish this, and it might break some output data formats especially multi-line ones (eg: JSON, or a stacktrace)
 * Others / Profit ???
@@ -97,6 +96,7 @@ Note: With any of these above, consider the goal of this codebase is to stay sim
 * Documented and cleaned up the codebase, added this awesome readme
 * Integrated the max_runtime feature directly into the codebase instead of being an (annoying) helper script
 * Making a few variables configurable via env vars making it easier to Dockerize in the future
+* Do some/more Windows testing, add documentation and examples for Windows.
 
 # History
 

run.py

@@ -1,7 +1,10 @@
 #!/usr/bin/env python3
 from __future__ import print_function
 from subprocess import Popen, TimeoutExpired
+from pathlib import Path
 import threading
+import platform
+import tempfile
 import datetime
 import signal
 import time
@@ -20,11 +23,12 @@ def readEnvOrDefault(envname, default=False):
         return default
 
 # Tmpdir, this MUST end in /
-tmpdir = readEnvOrDefault('TMPDIR', '/tmp/')
+system_tempdir = Path("/tmp" if platform.system() == "Darwin" else tempfile.gettempdir())
+tmpdir = readEnvOrDefault('TMPDIR', system_tempdir)
 # To prevent flooding, sleep this amount between launching threads
 thread_launch_interval = int(readEnvOrDefault('THREADLAUNCHINTERVAL', '1'))
 # Config file path, defining our tasks to run, default to cwd tasks.yml file
-configpath = readEnvOrDefault('CONFIGPATH', './tasks.yml')
+configpath = readEnvOrDefault('CONFIGPATH', 'tasks.yml')
 # Open our configuration file to load our job configuration
 try:
     with open(configpath, 'r') as stream:
@@ -34,7 +38,7 @@ def readEnvOrDefault(envname, default=False):
             jobs = yaml.load(stream) # This is a fallback to using an older pyyaml without the security patch
 except:
     print("Fatal error: Unable to read or parse {}".format(configpath))
-    raise 
+    raise
 
 
 ###################
@@ -70,7 +74,7 @@ def worker(label, command, interval, start=None, end=None, chdir=None, max_runti
     """thread worker function"""
     print('{}: Worker started for in interval {}'.format(label, interval))
     tmpfile_path = "{}{}.lastrun".format(tmpdir,label)
-    
+
     if chdir is not None:
         print('{}: Worker chdir {}'.format(label, chdir))
         os.chdir(chdir)
@@ -162,6 +166,6 @@ def worker(label, command, interval, start=None, end=None, chdir=None, max_runti
     )
     threads.append(t)
     t.start()
-    
+
     # To prevent flooding of our system, sleep 1 second between new thread creations, just incase
     time.sleep(thread_launch_interval)

tasks-windows.yml

@@ -0,0 +1,11 @@
+# Simple windows-specific examples
+test-1:
+  command: echo "test-1 run"
+  interval: 6    # 10 seconds
+  max_runtime: 3  # 5 seconds
+
+# This one has no max runtime, runs forever (until child script exits), simplest example
+chdir-test:
+  chdir: C:\Users
+  command: echo "chdir test ok"
+  interval: 3600  # 1 hour