avoid multiprocessing in tests (#46439)

- silences a few pytest warnings related to forking in xdist
- speeds up a few tests / avoids possible oversubscription in xdist

5 files changed, 23 insertions, 15 deletions

diff --git a/lib/spack/spack/detection/path.py b/lib/spack/spack/detection/path.py
index 3588d96115..7316d9124a 100644
--- a/lib/spack/spack/detection/path.py
+++ b/lib/spack/spack/detection/path.py
@@ -23,6 +23,7 @@ import spack.util.elf as elf_utils
 import spack.util.environment
 import spack.util.environment as environment
 import spack.util.ld_so_conf
+import spack.util.parallel
 
 from .common import (
     WindowsCompilerExternalPaths,
@@ -407,7 +408,7 @@ def by_path(
 
     result = collections.defaultdict(list)
     repository = spack.repo.PATH.ensure_unwrapped()
-    with concurrent.futures.ProcessPoolExecutor(max_workers=max_workers) as executor:
+    with spack.util.parallel.make_concurrent_executor(max_workers, require_fork=False) as executor:
         for pkg in packages_to_search:
             executable_future = executor.submit(
                 executables_finder.find,
diff --git a/lib/spack/spack/stage.py b/lib/spack/spack/stage.py
index 776b3ce6b9..ee3b4e95e5 100644
--- a/lib/spack/spack/stage.py
+++ b/lib/spack/spack/stage.py
@@ -2,7 +2,6 @@
 # Spack Project Developers. See the top-level COPYRIGHT file for details.
 #
 # SPDX-License-Identifier: (Apache-2.0 OR MIT)
-import concurrent.futures
 import errno
 import getpass
 import glob
@@ -40,6 +39,7 @@ import spack.resource
 import spack.spec
 import spack.util.crypto
 import spack.util.lock
+import spack.util.parallel
 import spack.util.path as sup
 import spack.util.pattern as pattern
 import spack.util.url as url_util
@@ -1132,7 +1132,7 @@ def get_checksums_for_versions(
         if checksum is not None:
             version_hashes[version] = checksum
 
-    with concurrent.futures.ProcessPoolExecutor(max_workers=concurrency) as executor:
+    with spack.util.parallel.make_concurrent_executor(concurrency, require_fork=False) as executor:
         results = []
         for url, version in search_arguments:
             future = executor.submit(_fetch_and_checksum, url, fetch_options, keep_stage)
diff --git a/lib/spack/spack/test/conftest.py b/lib/spack/spack/test/conftest.py
index 607844265f..2726061cb3 100644
--- a/lib/spack/spack/test/conftest.py
+++ b/lib/spack/spack/test/conftest.py
@@ -1980,12 +1980,14 @@ def pytest_runtest_setup(item):
         pytest.skip(*not_on_windows_marker.args)
 
 
+def _sequential_executor(*args, **kwargs):
+    return spack.util.parallel.SequentialExecutor()
+
+
 @pytest.fixture(autouse=True)
 def disable_parallel_buildcache_push(monkeypatch):
     """Disable process pools in tests."""
-    monkeypatch.setattr(
-        spack.util.parallel, "make_concurrent_executor", spack.util.parallel.SequentialExecutor
-    )
+    monkeypatch.setattr(spack.util.parallel, "make_concurrent_executor", _sequential_executor)
 
 
 def _root_path(x, y, *, path):
diff --git a/lib/spack/spack/util/parallel.py b/lib/spack/spack/util/parallel.py
index 9bbdf5dd7a..b16adb8858 100644
--- a/lib/spack/spack/util/parallel.py
+++ b/lib/spack/spack/util/parallel.py
@@ -9,7 +9,7 @@ import sys
 import traceback
 from typing import Optional
 
-from spack.config import determine_number_of_jobs
+import spack.config
 
 
 class ErrorFromWorker:
@@ -98,9 +98,14 @@ class SequentialExecutor(concurrent.futures.Executor):
         return future
 
 
-def make_concurrent_executor() -> concurrent.futures.Executor:
-    """Can't use threading because it's unsafe, and can't use spawned processes because of globals.
-    That leaves only forking."""
-    if multiprocessing.get_start_method() == "fork":
-        return concurrent.futures.ProcessPoolExecutor(determine_number_of_jobs(parallel=True))
-    return SequentialExecutor()
+def make_concurrent_executor(
+    jobs: Optional[int] = None, *, require_fork: bool = True
+) -> concurrent.futures.Executor:
+    """Create a concurrent executor. If require_fork is True, then the executor is sequential
+    if the platform does not enable forking as the default start method. Effectively
+    require_fork=True makes the executor sequential in the current process on Windows, macOS, and
+    Linux from Python 3.14+ (which changes defaults)"""
+    if require_fork and multiprocessing.get_start_method() != "fork":
+        return SequentialExecutor()
+    jobs = jobs or spack.config.determine_number_of_jobs(parallel=True)
+    return concurrent.futures.ProcessPoolExecutor(jobs)
diff --git a/lib/spack/spack/util/web.py b/lib/spack/spack/util/web.py
index 9a0f1d6e4b..892b64d333 100644
--- a/lib/spack/spack/util/web.py
+++ b/lib/spack/spack/util/web.py
@@ -4,7 +4,6 @@
 # SPDX-License-Identifier: (Apache-2.0 OR MIT)
 
 import codecs
-import concurrent.futures
 import email.message
 import errno
 import json
@@ -30,6 +29,7 @@ from llnl.util.filesystem import mkdirp, rename, working_dir
 import spack.config
 import spack.error
 import spack.util.executable
+import spack.util.parallel
 import spack.util.path
 import spack.util.url as url_util
 
@@ -641,7 +641,7 @@ def spider(
         root = urllib.parse.urlparse(root_str)
         spider_args.append((root, go_deeper, _visited))
 
-    with concurrent.futures.ProcessPoolExecutor(max_workers=concurrency) as tp:
+    with spack.util.parallel.make_concurrent_executor(concurrency, require_fork=False) as tp:
         while current_depth <= depth:
             tty.debug(
                 f"SPIDER: [depth={current_depth}, max_depth={depth}, urls={len(spider_args)}]"