fix: skip hash merge backups in every loader

scripts/common.py

@@ -1172,6 +1172,18 @@ def select_emulator_profiles(
     return selected
 
 
+PROFILE_BACKUP_SUFFIX = ".old.yml"
+
+
+def source_yaml_paths(directory: str | Path) -> list[Path]:
+    """YAML sources in a directory, without the copies the hash merge keeps."""
+    return [
+        path
+        for path in sorted(Path(directory).glob("*.yml"))
+        if not path.name.endswith(PROFILE_BACKUP_SUFFIX)
+    ]
+
+
 def load_emulator_profiles(
     emulators_dir: str,
     skip_aliases: bool = True,
@@ -1199,9 +1211,7 @@ def load_emulator_profiles(
             file=sys.stderr,
         )
         return profiles
-    for f in sorted(emu_path.glob("*.yml")):
-        if f.name.endswith(".old.yml"):
-            continue
+    for f in source_yaml_paths(emu_path):
         with open(f) as fh:
             profile = yaml_load(fh) or {}
         if "emulator" not in profile:

scripts/generate_db.py

@@ -27,6 +27,7 @@ from common import (
     list_registered_platforms,
     load_provenance_snapshots,
     require_yaml,
+    source_yaml_paths,
     write_if_changed,
     yaml_load,
 )
@@ -507,9 +508,7 @@ def _collect_all_aliases(files: dict) -> dict:
     # Collect aliases from emulator YAMLs (aliases field on file entries)
     emulators_dir = Path("emulators")
     if emulators_dir.is_dir():
-        for emu_file in emulators_dir.glob("*.yml"):
-            if emu_file.name.endswith(".old.yml"):
-                continue
+        for emu_file in source_yaml_paths(emulators_dir):
             try:
                 with open(emu_file) as f:
                     emu_config = yaml_load(f) or {}

scripts/scraper/_hash_merge.py

@@ -14,7 +14,7 @@ from pathlib import Path
 from typing import Any
 
 
-from common import write_text_atomic, yaml_load
+from common import PROFILE_BACKUP_SUFFIX, write_text_atomic, yaml_load
 
 
 _MAME_RELEASE_RE = re.compile(r"^0\.\d+")
@@ -381,7 +381,7 @@ def _backup_and_write(path: str, data: dict) -> None:
     (core_version, contents, source_ref), and appends new entries.
     """
     p = Path(path)
-    backup = p.with_suffix(".old.yml")
+    backup = p.with_suffix(PROFILE_BACKUP_SUFFIX)
     shutil.copy2(p, backup)
 
     original = p.read_text(encoding="utf-8")
@@ -572,7 +572,7 @@ def _backup_and_write_fbneo(path: str, data: dict, hashes: dict) -> None:
     Existing entries are left untouched (CRC32 changes are rare).
     """
     p = Path(path)
-    backup = p.with_suffix(".old.yml")
+    backup = p.with_suffix(PROFILE_BACKUP_SUFFIX)
     shutil.copy2(p, backup)
 
     original = p.read_text(encoding="utf-8")

scripts/scraper/fbneo_hash_scraper.py

@@ -20,7 +20,7 @@ from typing import Any
 
 import yaml
 
-from common import write_text_atomic, yaml_load
+from common import source_yaml_paths, write_text_atomic, yaml_load
 
 from scripts.scraper._hash_merge import compute_diff, merge_fbneo_profile
 from scripts.scraper.fbneo_parser import parse_fbneo_source_tree
@@ -200,9 +200,7 @@ def fetch_and_cache(force: bool = False) -> dict[str, Any]:
 def _find_fbneo_profiles() -> list[Path]:
     """Find emulator profiles whose upstream references finalburnneo/FBNeo."""
     profiles: list[Path] = []
-    for path in sorted(EMULATORS_DIR.glob("*.yml")):
-        if path.name.endswith(".old.yml"):
-            continue
+    for path in source_yaml_paths(EMULATORS_DIR):
         try:
             data = yaml_load(path.read_text(encoding="utf-8"))
         except (yaml.YAMLError, OSError):

scripts/scraper/mame_hash_scraper.py

@@ -22,7 +22,7 @@ from typing import Any
 
 import yaml
 
-from common import write_text_atomic, yaml_load
+from common import source_yaml_paths, write_text_atomic, yaml_load
 
 from ._hash_merge import compute_diff, merge_mame_profile
 from .mame_parser import parse_mame_source_tree
@@ -181,9 +181,7 @@ _MAME_TREES = frozenset({
 
 def _find_mame_profiles() -> list[Path]:
     profiles: list[Path] = []
-    for path in sorted(_EMULATORS_DIR.glob("*.yml")):
-        if path.name.endswith(".old.yml"):
-            continue
+    for path in source_yaml_paths(_EMULATORS_DIR):
         try:
             with open(path, encoding="utf-8") as f:
                 data = yaml_load(f)

scripts/validate_schemas.py

@@ -12,7 +12,7 @@ from pathlib import Path, PurePosixPath
 
 import yaml
 
-from common import yaml_load
+from common import source_yaml_paths, yaml_load
 from jsonschema import Draft202012Validator, FormatChecker
 
 ROOT = Path(__file__).resolve().parent.parent
@@ -46,7 +46,7 @@ def _validate_yaml_directory(
 ) -> list[str]:
     validator = _validator(schema_name)
     out: list[str] = []
-    for path in sorted(directory.glob("*.yml")):
+    for path in source_yaml_paths(directory):
         if skip_private and path.name.startswith("_"):
             continue
         try:
@@ -214,7 +214,7 @@ def _unknown_data_dir_refs() -> list[str]:
         registered = set((yaml_load(handle) or {}).get("data_directories", {}))
     out: list[str] = []
     for directory in (ROOT / "emulators", ROOT / "platforms"):
-        for path in sorted(directory.glob("*.yml")):
+        for path in source_yaml_paths(directory):
             if path.name.startswith("_"):
                 continue
             with path.open(encoding="utf-8") as handle:

tests/test_audit_regressions.py

@@ -2126,5 +2126,31 @@ class ModuleConstantsDeclaredOnce(unittest.TestCase):
         self.assertEqual(offenders, {})
 
 
+class HashMergeBackupsAreNotSources(unittest.TestCase):
+    """The hash merge copies a profile to <name>.old.yml before rewriting it.
+
+    Every loader skipped the copy except validate_schemas, which then failed
+    on a stale MAME profile nobody would ever ship.
+    """
+
+    def test_the_validator_skips_a_backup(self):
+        sys.path.insert(0, str(ROOT / "scripts"))
+        from validate_schemas import _validate_yaml_directory
+
+        with tempfile.TemporaryDirectory(dir=TMP_ROOT) as directory:
+            (Path(directory) / "mame.old.yml").write_text("emulator: [\n")
+            self.assertEqual(
+                _validate_yaml_directory(Path(directory), "emulator.schema.json"), []
+            )
+
+    def test_the_backup_suffix_is_written_once(self):
+        offenders = [
+            str(path.relative_to(ROOT))
+            for path in sorted(ROOT.glob("scripts/**/*.py"))
+            if path.name != "common.py" and '".old.yml"' in path.read_text()
+        ]
+        self.assertEqual(offenders, [])
+
+
 if __name__ == "__main__":
     unittest.main()