Use the default multiprocessing pool configuration.

Previously we were forcing use of fork for all OSes to try to juice
performance at the cost of running into threading issues at some point
when things got more complex. They may have reached that point with
CI lockups happening on macOS. Revert to the default pool setup to
attempt to fix macOS hangs.

pex/jobs.py

@@ -17,7 +17,7 @@
 from pex.common import pluralize
 from pex.compatibility import Queue, cpu_count
 from pex.tracer import TRACER
-from pex.typing import TYPE_CHECKING, Generic, cast
+from pex.typing import Generic, TYPE_CHECKING
 
 if TYPE_CHECKING:
     from typing import (
@@ -679,12 +679,8 @@ def join(self):
 @contextmanager
 def _mp_pool(size):
     # type: (int) -> Iterator[Pool]
-    try:
-        context = multiprocessing.get_context("fork")  # type: ignore[attr-defined]
-        pool = cast("Pool", context.Pool(processes=size))
-    except (AttributeError, ValueError):
-        pool = multiprocessing.Pool(processes=size)
 
+    pool = multiprocessing.Pool(processes=size)
     try:
         yield pool
     finally: