Filter discovered modules to match the given %pyproject_save_files globs

When certain modules are deliberately not included into the built RPM,
they shouldn't be listed in the list of qualified module names which are
used by %pyproject_check_import to test importability of the
distribution.

Related: rhbz#2168193

README.md

@@ -269,6 +269,7 @@ If `%pyproject_save_files` is not used, calling `%pyproject_check_import` will f
 When `%pyproject_save_files` is invoked,
 it creates a list of all valid and public (i.e. not starting with `_`)
 importable module names found in the package.
+Each top-level module name matches at least one of the globs provided as an argument to `%pyproject_save_files`.
 This list is then usable by `%pyproject_check_import` which performs an import check for each listed module.
 When a module fails to import, the build fails.
 

pyproject-rpm-macros.spec

@@ -129,8 +129,10 @@ export HOSTNAME="rpmbuild"  # to speedup tox in network-less mock, see rhbz#1856
 %license LICENSE
 
 %changelog
-* Mon Sep 19 2022 Miro Hrončok <mhroncok@redhat.com> - 1.4.0-1
+* Mon Sep 19 2022 Python Maint <python-maint@redhat.com> - 1.4.0-1
 - %%pyproject_save_files: Support License-Files installed into the *Root License Directory* from PEP 369
+- %%pyproject_check_import: Import only the modules whose top-level names
+  match any of the globs provided to %%pyproject_save_files
 
 * Tue Aug 30 2022 Otto Liljalaakso <otto.liljalaakso@iki.fi> - 1.3.4-1
 - Fix typo in internal function name

pyproject_save_files.py

@@ -536,6 +536,50 @@ def generate_file_list(paths_dict, module_globs, include_others=False):
     return sorted(files)
 
 
+def generate_module_list(paths_dict, module_globs):
+    """
+    This function takes the paths_dict created by the classify_paths() function and
+    reads the modules names from it.
+    It filters those whose top-level module names match any of the provided module_globs.
+
+    Returns list with matching qualified module names.
+
+    Examples:
+
+        >>> generate_module_list({'module_names': {'foo', 'foo.bar', 'baz'}}, {'foo'})
+        ['foo', 'foo.bar']
+
+        >>> generate_module_list({'module_names': {'foo', 'foo.bar', 'baz'}}, {'*foo'})
+        ['foo', 'foo.bar']
+
+        >>> generate_module_list({'module_names': {'foo', 'foo.bar', 'baz'}}, {'foo', 'baz'})
+        ['baz', 'foo', 'foo.bar']
+
+        >>> generate_module_list({'module_names': {'foo', 'foo.bar', 'baz'}}, {'*'})
+        ['baz', 'foo', 'foo.bar']
+
+        >>> generate_module_list({'module_names': {'foo', 'foo.bar', 'baz'}}, {'bar'})
+        []
+
+    Submodules aren't discovered:
+
+        >>> generate_module_list({'module_names': {'foo', 'foo.bar', 'baz'}}, {'*bar'})
+        []
+    """
+
+    module_names = paths_dict['module_names']
+    filtered_module_names = set()
+
+    for glob in module_globs:
+        for name in module_names:
+            # Match the top-level part of the qualified name, eg. 'foo.bar.baz' -> 'foo'
+            top_level_name = name.split('.')[0]
+            if fnmatch.fnmatchcase(top_level_name, glob):
+                filtered_module_names.add(name)
+
+    return sorted(filtered_module_names)
+
+
 def parse_varargs(varargs):
     """
     Parse varargs from the %pyproject_save_files macro
@@ -664,7 +708,7 @@ def pyproject_save_files_and_modules(buildroot, sitelib, sitearch, python_versio
     parsed_records = load_parsed_record(pyproject_record)
 
     final_file_list = []
-    all_module_names = set()
+    final_module_list = []
 
     for record_path, files in parsed_records.items():
         metadata = dist_metadata(buildroot, record_path)
@@ -675,12 +719,11 @@ def pyproject_save_files_and_modules(buildroot, sitelib, sitearch, python_versio
         final_file_list.extend(
             generate_file_list(paths_dict, globs, include_auto)
         )
-        all_module_names.update(paths_dict["module_names"])
+        final_module_list.extend(
+            generate_module_list(paths_dict, globs)
+        )
 
-    # Sort values, so they are always checked in the same order
-    all_module_names = sorted(all_module_names)
-
-    return final_file_list, all_module_names
+    return final_file_list, final_module_list
 
 
 def main(cli_args):

pyproject_save_files_test_data.yaml

@@ -7755,7 +7755,6 @@ dumped:
     - ipykernel.tests.utils
     - ipykernel.trio_runner
     - ipykernel.zmqshell
-    - ipykernel_launcher
 - - zope
   - zope
   - - '%dir /usr/lib/python3.7/site-packages/zope'

tests/python-ldap.spec

@@ -86,6 +86,12 @@ grep -F %{python3_sitearch}/__pycache__/ldif.cpython-%{python3_version_nodots}.p
 grep -F %{python3_sitearch}/__pycache__/ldif.cpython-%{python3_version_nodots}.opt-1.pyc %{pyproject_files} && exit 1 || true
 grep -F %{python3_sitearch}/slapdtest %{pyproject_files} && exit 1 || true
 
+# Internal check: Unmatched modules are not supposed to be listed in %%{_pyproject_modules}
+grep -F slapdtest %{_pyproject_modules} && exit 1 || true
+grep -F ldif %{_pyproject_modules} && exit 1 || true
+# Let's check that at least one module is listed in %%{_pyproject_modules}
+grep -F ldapurl %{_pyproject_modules}
+
 # Internal check: Top level __pycache__ is never owned
 grep -E '/site-packages/__pycache__$' %{pyproject_files} && exit 1 || true
 grep -E '/site-packages/__pycache__/$' %{pyproject_files} && exit 1 || true