fix: add TASKGRAPH_SERIAL and set multiprocessing context for multiprocess kind generation

This was supposed to be done in #765, but clearly I didn't push it before merging.

Author: bhearsum

src/taskgraph/generator.py

@@ -4,6 +4,7 @@
 
 import copy
 import logging
+import multiprocessing
 import os
 import platform
 from concurrent.futures import (
@@ -302,7 +303,7 @@ def _load_tasks_parallel(self, kinds, kind_graph, parameters):
         futures = set()
         edges = set(kind_graph.edges)
 
-        with ProcessPoolExecutor() as executor:
+        with ProcessPoolExecutor(mp_context=multiprocessing.get_context("fork")) as executor:
 
             def submit_ready_kinds():
                 """Create the next batch of tasks for kinds without dependencies."""
@@ -420,7 +421,7 @@ def _run(self):
         # redone in the new processes. Ideally this would be fixed, or we
         # would take another approach to parallel kind generation. In the
         # meantime, it's not supported outside of Linux.
-        if platform.system() != "Linux":
+        if platform.system() != "Linux" or os.environ.get("TASKGRAPH_SERIAL"):
             all_tasks = self._load_tasks_serial(kinds, kind_graph, parameters)
         else:
             all_tasks = self._load_tasks_parallel(kinds, kind_graph, parameters)