bpo-42955: Add sys.modules_names (GH-24238)

Add sys.module_names, containing the list of the standard library
module names.

1 files changed, 26 insertions, 76 deletions

diff --git a/Tools/scripts/generate_module_names.py b/Tools/scripts/generate_module_names.py
index 985a1a5e5a2..9d363aa04b3 100644
--- a/Tools/scripts/generate_module_names.py
+++ b/Tools/scripts/generate_module_names.py
@@ -17,27 +17,6 @@ IGNORE = {
     '__pycache__',
     'site-packages',
 
-    # Helper modules of public modules.
-    # For example, sysconfig uses _osx_support.
-    '_aix_support',
-    '_collections_abc',
-    '_compat_pickle',
-    '_compression',
-    '_markupbase',
-    '_osx_support',
-    '_sitebuiltins',
-    '_strptime',
-    '_threading_local',
-    '_weakrefset',
-
-    # Used to bootstrap setup.py
-    '_bootsubprocess',
-
-    # pure Python implementation
-    '_py_abc',
-    '_pydecimal',
-    '_pyio',
-
     # test modules
     '__phello__.foo',
     '_ctypes_test',
@@ -69,40 +48,20 @@ WINDOWS_MODULES = (
 )
 
 
-def write_comment(fp, comment):
-    print(f"// {comment}", file=fp)
-
-
-def write_modules(fp, names):
-    for name in sorted(names):
-        if name in IGNORE:
-            continue
-        print(f'"{name}",', file=fp)
-    print(file=fp)
-
-
-def list_builtin_modules(fp):
-    write_comment(fp, "Built-in modules")
-    write_modules(fp, sys.builtin_module_names)
-
-
 # Pure Python modules (Lib/*.py)
-def list_python_modules(fp):
-    write_comment(fp, "Pure Python modules (Lib/*.py)")
-    names = []
+def list_python_modules(names):
     for filename in os.listdir(STDLIB_PATH):
         if not filename.endswith(".py"):
             continue
         name = filename.removesuffix(".py")
-        names.append(name)
-    write_modules(fp, names)
+        names.add(name)
 
 
 def _list_sub_packages(path, names, parent=None):
     for name in os.listdir(path):
-        package_path = os.path.join(path, name)
         if name in IGNORE:
             continue
+        package_path = os.path.join(path, name)
         if not os.path.isdir(package_path):
             continue
         if not any(package_file.endswith(".py")
@@ -114,40 +73,28 @@ def _list_sub_packages(path, names, parent=None):
             qualname = name
         if qualname in IGNORE:
             continue
-        names.append(qualname)
+        names.add(qualname)
         _list_sub_packages(package_path, names, qualname)
 
 
 # Packages and sub-packages
-def list_packages(fp):
-    write_comment(fp, "Packages and sub-packages")
-    names = []
+def list_packages(names):
     _list_sub_packages(STDLIB_PATH, names)
-    write_modules(fp, names)
-
-
-# Windows extensions
-def list_windows_extensions(fp):
-    write_comment(fp, "Windows extension modules")
-    write_modules(fp, WINDOWS_MODULES)
 
 
 # Extension modules built by setup.py
-def list_setup(fp):
+def list_setup_extensions(names):
     cmd = [sys.executable, SETUP_PY, "-q", "build", "--list-module-names"]
     output = subprocess.check_output(cmd)
     output = output.decode("utf8")
-    names = output.splitlines()
-
-    write_comment(fp, "Extension modules built by setup.py")
-    write_modules(fp, names)
+    extensions = output.splitlines()
+    names |= set(extensions)
 
 
 # Built-in and extension modules built by Modules/Setup
-def list_modules_setup(fp):
+def list_modules_setup_extensions(names):
     assign_var = re.compile("^[A-Z]+=")
 
-    names = []
     with open(MODULES_SETUP, encoding="utf-8") as modules_fp:
         for line in modules_fp:
             # Strip comment
@@ -165,25 +112,26 @@ def list_modules_setup(fp):
                 continue
             # "errno errnomodule.c" => write "errno"
             name = parts[0]
-            names.append(name)
+            names.add(name)
 
-    write_comment(fp, "Built-in and extension modules built by Modules/Setup")
-    write_modules(fp, names)
+
+def list_modules():
+    names = set(sys.builtin_module_names) | set(WINDOWS_MODULES)
+    list_modules_setup_extensions(names)
+    list_setup_extensions(names)
+    list_packages(names)
+    list_python_modules(names)
+    names -= set(IGNORE)
+    return names
 
 
-def list_modules(fp):
+def write_modules(fp, names):
     print("// Auto-generated by Tools/scripts/generate_module_names.py.", file=fp)
+    print("// List used to create sys.module_names.", file=fp)
     print(file=fp)
     print("static const char* _Py_module_names[] = {", file=fp)
-    print(file=fp)
-
-    list_builtin_modules(fp)
-    list_python_modules(fp)
-    list_packages(fp)
-    list_setup(fp)
-    list_modules_setup(fp)
-    list_windows_extensions(fp)
-
+    for name in sorted(names):
+        print(f'"{name}",', file=fp)
     print("};", file=fp)
 
 
@@ -193,7 +141,9 @@ def main():
               file=sys.stderr)
         sys.exit(1)
 
-    list_modules(sys.stdout)
+    fp = sys.stdout
+    names = list_modules()
+    write_modules(fp, names)
 
 
 if __name__ == "__main__":