fix: order same-named files by their owner's copy

scripts/common.py

@@ -468,6 +468,36 @@ def _resolve_agnostic(file_entry: dict, files_db: dict, has_strong_hash: bool):
     return None
 
 
+def _affinity_tokens(text: str) -> set[str]:
+    return {
+        token
+        for token in (re.sub(r"[^0-9a-z]", "", part.casefold()) for part in text.split("/"))
+        if token
+    }
+
+
+def _by_affinity(paths: list[str], file_entry: dict, dest_hint: str) -> list[str]:
+    """Same-named candidates, the ones stored with the file's owner first.
+
+    A name step without a tie-break took whichever path sorted first: the
+    TheXTech coin.ogg for Syobon Action, a Hurrican sound for ARMSX2. The
+    destination's directories and the profile that asks name where its own
+    copy lives; the order is otherwise kept.
+    """
+    hint = dest_hint or file_entry.get("path") or file_entry.get("destination") or ""
+    wanted = _affinity_tokens(hint.rsplit("/", 1)[0] if "/" in hint else "")
+    for owner in (file_entry.get("source_profile"), file_entry.get("source_emulator")):
+        if owner:
+            wanted |= _affinity_tokens(str(owner).replace(" ", ""))
+    if not wanted or len(paths) < 2:
+        return paths
+
+    def score(path: str) -> int:
+        return len(wanted & _affinity_tokens(path.rsplit("/", 1)[0]))
+
+    return sorted(paths, key=score, reverse=True)
+
+
 def resolve_local_file(
     file_entry: dict,
     db: dict,
@@ -737,6 +767,7 @@ def resolve_local_file(
         if candidates:
             if zipped_file:
                 candidates = [p for p in candidates if ".zip" in os.path.basename(p)]
+            candidates = _by_affinity(candidates, file_entry, dest_hint)
             primary = [p for p in candidates if "/.variants/" not in p]
             if primary or candidates:
                 return (primary[0] if primary else candidates[0]), "name_exact"
@@ -795,8 +826,9 @@ def resolve_local_file(
                 return (primary[0] if primary else valid[0][0]), "hash_mismatch"
             # No candidate contains the zipped_file -fall through to step 5
         elif not unsourceable:
-            primary = [c[0] for c in candidates if "/.variants/" not in c[0]]
-            return (primary[0] if primary else candidates[0][0]), "hash_mismatch"
+            ordered = _by_affinity([c[0] for c in candidates], file_entry, dest_hint)
+            primary = [c for c in ordered if "/.variants/" not in c]
+            return (primary[0] if primary else ordered[0]), "hash_mismatch"
 
     # 5. zipped_file content match via pre-built index (last resort:
     # matches inner ROM MD5 across ALL ZIPs in the repo, so only use

scripts/generate_pack.py

@@ -1437,7 +1437,7 @@ def generate_emulator_pack(
 
                 dest_hint = fe.get("path", "")
                 local_path, status = resolve_file(
-                    fe,
+                    {**fe, "source_profile": emu_name},
                     db,
                     bios_dir,
                     zip_contents,

scripts/verify.py

@@ -1382,7 +1382,7 @@ def verify_emulator(
 
             dest_hint = file_entry.get("path", "")
             local_path, resolve_status = resolve_local_file(
-                file_entry,
+                {**file_entry, "source_profile": emu_name},
                 db,
                 zip_contents,
                 dest_hint=dest_hint,

tests/test_game_data_homonyms.py

@@ -194,6 +194,43 @@ class ReleaseAssetIsNotANameMatch(unittest.TestCase):
             fetch.assert_not_called()
 
 
+class NameStepPrefersTheOwnersCopy(unittest.TestCase):
+    """Same-named files are ordered by where the asking profile keeps its own.
+
+    The name step took the first path in index order: Syobon Action's
+    coin.ogg came from TheXTech's sound folder while its own copy sat in
+    bios/Other/syobonaction/.
+    """
+
+    def test_the_copy_stored_with_the_profile_wins(self):
+        import tempfile
+
+        from common import resolve_local_file
+
+        with tempfile.TemporaryDirectory() as tmp:
+            previous = os.getcwd()
+            os.chdir(tmp)
+            self.addCleanup(os.chdir, previous)
+            other = Path("bios/Game Engines/TheXTech/sound/coin.ogg")
+            own = Path("bios/Other/syobonaction/coin.ogg")
+            for path, data in ((other, b"a" * 10), (own, b"b" * 12)):
+                path.parent.mkdir(parents=True)
+                path.write_bytes(data)
+            files = {
+                "1" * 40: {"path": str(other), "name": "coin.ogg", "size": 10},
+                "2" * 40: {"path": str(own), "name": "coin.ogg", "size": 12},
+            }
+            db = {"files": files, "indexes": {
+                "by_name": {"coin.ogg": ["1" * 40, "2" * 40]},
+                "by_md5": {}, "by_crc32": {}, "by_path_suffix": {}}}
+            entry = {"name": "coin.ogg", "path": "SyobonAction/SE/coin.ogg"}
+            path, status = resolve_local_file(entry, db, dest_hint=entry["path"])
+            self.assertEqual((path, status), (str(own), "name_exact"))
+            entry = {"name": "coin.ogg", "path": "SE/coin.ogg", "source_profile": "syobonaction"}
+            path, _status = resolve_local_file(entry, db, dest_hint=entry["path"])
+            self.assertEqual(path, str(own))
+
+
 class CollectionCarriesNoGameDataHomonym(unittest.TestCase):
     def test_the_profiles_resolve_no_game_data_to_another_game(self):
         database = REPO_ROOT / "database.json"