fix: group a system once whatever its spelling

The platform's files and the core extras for the Master System sat
in two region groups, sega-master-system and sega-mastersystem,
reduced apart. A US pack kept the declared Japanese BIOS. A core
extra's system group now takes the platform's spelling, in the pack,
the report and the installer's manifest fields alike.

scripts/manifestselect.py

@@ -19,7 +19,7 @@ from dataclasses import dataclass
 
 import region as region_mod
 from common import _norm_system_id, sanitize_pack_path
-from packextras import _extra_system_ids, _kept, extra_region_groups
+from packextras import _extra_system_ids, _kept, extra_region_groups, system_spellings
 
 
 def _dest(entry: dict) -> str:
@@ -58,11 +58,14 @@ class SelectionIndex:
 
         owned_by: dict[str, set[str]] = {}
         extra_groups: dict[str, set[str]] = {}
+        spellings = system_spellings(pack_systems)
         for extra in extras:
             dest = _dest(extra)
             if not dest:
                 continue
-            extra_groups.setdefault(dest, set()).update(extra_region_groups(extra))
+            extra_groups.setdefault(dest, set()).update(
+                extra_region_groups(extra, spellings)
+            )
             # A profile system the platform does not have is no system the
             # pack's --system accepts: it would offer what the pack refuses.
             for sys_id in _extra_system_ids(extra):

scripts/packextras.py

@@ -6,7 +6,9 @@ this module builds so the report and the pack withdraw the same files."""
 
 from __future__ import annotations
 
-from common import PROFILE_IDENTITY_FIELDS
+from collections.abc import Iterable
+
+from common import PROFILE_IDENTITY_FIELDS, _norm_system_id
 from common import archive_members
 from common import declared_root, expand_platform_declared_names, platform_declarations
 from common import load_emulator_profiles
@@ -948,6 +950,7 @@ def platform_region_groups(
         absent=absent,
     )
     held = {id(extra) for extra in extras}
+    spellings = system_spellings(systems)
     for extra in _kept(extras + absent, required_only):
         dest = sanitize_pack_path(extra.get("destination", extra.get("name", "")))
         if not dest:
@@ -965,7 +968,7 @@ def platform_region_groups(
         # looked up by name its scph50006 took DuckStation's [japan], which
         # describes DuckStation's own PS1-on-PS2 choice.
         competes = not extra.get("agnostic_scan") or extra.get("region")
-        for group_id in extra_region_groups(extra) if competes else ():
+        for group_id in extra_region_groups(extra, spellings) if competes else ():
             groups.setdefault(group_id, []).append((dest, name))
         if extras_out is not None and id(extra) in held:
             extras_out.append(extra)
@@ -974,8 +977,31 @@ def platform_region_groups(
     return groups, extra_dests
 
 
-def extra_region_groups(extra: dict) -> list[str]:
-    """The region groups a core extra competes in."""
+def system_spellings(systems: Iterable[str]) -> dict[str, str]:
+    """A platform's own spelling of each system, keyed the way ids compare.
+
+    RetroDECK files the Master System as sega-master-system, the profiles as
+    sega-mastersystem. Grouped under both, the declared bios/bios_J.sms and
+    the core's bios_J.sms were reduced apart and a --region us pack kept the
+    Japanese BIOS.
+    """
+    spellings: dict[str, str] = {}
+    for sys_id in sorted(systems):
+        spellings.setdefault(_norm_system_id(sys_id), sys_id)
+    return spellings
+
+
+def extra_region_groups(
+    extra: dict, spellings: dict[str, str] | None = None
+) -> list[str]:
+    """The region groups a core extra competes in.
+
+    A group named after a system takes the platform's spelling of it, so the
+    extra competes beside the files the platform declares for that system.
+    """
+    def system_group(sys_id: str) -> str:
+        return (spellings or {}).get(_norm_system_id(sys_id), sys_id)
+
     emulator = extra.get("source_emulator", "")
     variant = extra.get("variant_group")
     if variant:
@@ -986,10 +1012,10 @@ def extra_region_groups(extra: dict) -> list[str]:
         return [f"{emulator}:variant:{variant}"]
     explicit = extra.get("source_system")
     if explicit:
-        return [str(explicit)]
+        return [system_group(str(explicit))]
     systems = [str(s) for s in extra.get("source_systems", []) if s]
     if len(systems) == 1:
-        return systems
+        return [system_group(systems[0])]
     # No system names the machine this file is chosen for. Filed under every
     # system of a 30-system profile, CLK's American MSX ROM made the
     # Enterprise group drop its German ROM. Alone, it drops nothing and is

tests/test_e2e.py

@@ -6251,6 +6251,44 @@ struct BurnDriver BurnDrvneogeo = {
              ("declared.bin", "declared.bin")],
         )
 
+    def test_a_system_spelled_two_ways_is_one_group(self):
+        """RetroDECK files the Master System as sega-master-system, the
+        profiles as sega-mastersystem. Two groups reduced apart, and a
+        --region us pack kept the declared Japanese BIOS."""
+        import hashlib  # noqa: PLC0415
+
+        from packextras import platform_region_groups  # noqa: PLC0415
+
+        files, by_name = {}, {}
+        for name in ("bios_U.sms", "bios_J.sms"):
+            payload = name.encode()
+            path = os.path.join(self.bios_dir, name)
+            with open(path, "wb") as handle:
+                handle.write(payload)
+            sha1 = hashlib.sha1(payload).hexdigest()
+            files[sha1] = {"path": path, "name": name, "size": len(payload),
+                           "md5": hashlib.md5(payload).hexdigest()}
+            by_name[name] = [sha1]
+        db = {"files": files, "indexes": {"by_name": by_name, "by_md5": {},
+                                          "by_crc32": {}, "by_path_suffix": {}}}
+        profiles = {"smscore": {
+            "emulator": "SMS Core", "type": "libretro",
+            "systems": ["sega-mastersystem"],
+            "files": [{"name": "bios_U.sms", "region": ["north-america"]},
+                      {"name": "bios_J.sms", "region": ["japan"]}],
+        }}
+        config = {"platform": "P", "cores": ["smscore"], "base_destination": "system",
+                  "systems": {"sega-master-system": {"files": [
+                      {"name": "bios_J.sms", "destination": "bios/bios_J.sms"}]}}}
+        groups, _ = platform_region_groups(
+            config, config["systems"], self.emulators_dir, db, "system", profiles,
+        )
+        self.assertNotIn("sega-mastersystem", groups)
+        self.assertEqual(
+            sorted(dest for dest, _ in groups["sega-master-system"]),
+            ["bios/bios_J.sms", "bios_J.sms", "bios_U.sms"],
+        )
+
     def test_slot_narrowed_pack_gets_its_own_name(self):
         names = {
             os.path.basename(self._region_pack()),

tests/test_region_groups.py

@@ -14,7 +14,7 @@ from pathlib import Path
 REPO_ROOT = Path(__file__).resolve().parent.parent
 sys.path.insert(0, str(REPO_ROOT / "scripts"))
 
-from packextras import extra_region_groups  # noqa: E402
+from packextras import extra_region_groups, system_spellings  # noqa: E402
 
 
 class TheGroupFollowsWhatTheProfileSays(unittest.TestCase):
@@ -63,5 +63,29 @@ class TheEnterpriseKeepsItsRom(unittest.TestCase):
         self.assertNotIn("bios/brd.rom", drops)
 
 
+class TheGroupTakesThePlatformSpelling(unittest.TestCase):
+    def test_a_profile_spelling_maps_to_the_platform_one(self):
+        spellings = system_spellings(["sega-master-system", "sony-playstation"])
+        self.assertEqual(
+            extra_region_groups({"source_system": "sega-mastersystem"}, spellings),
+            ["sega-master-system"],
+        )
+
+    def test_a_system_the_platform_lacks_keeps_its_name(self):
+        spellings = system_spellings(["sony-playstation"])
+        self.assertEqual(
+            extra_region_groups({"source_systems": ["sega-saturn"]}, spellings),
+            ["sega-saturn"],
+        )
+
+    def test_file_and_variant_groups_are_untouched(self):
+        spellings = system_spellings(["sega-megacd"])
+        self.assertEqual(
+            extra_region_groups({"source_emulator": "picodrive", "variant_group": "mcd",
+                                 "source_system": "sega-megacd"}, spellings),
+            ["picodrive:variant:mcd"],
+        )
+
+
 if __name__ == "__main__":
     unittest.main()