feat: hash-based matching for cross-reference

expand_platform_declared_names resolves platform file MD5s
through the database to recover canonical names and aliases,
eliminating false positive undeclared files when a platform
renames a file (e.g. Batocera ROM1 vs gsplus ROM).

scripts/common.py

@@ -915,6 +915,39 @@ def filter_systems_by_target(
     return filtered
 
 
+def expand_platform_declared_names(config: dict, db: dict) -> set[str]:
+    """Build set of file names declared by a platform config.
+
+    Enriches the set with canonical names and aliases from the database
+    by resolving each platform file's MD5 through by_md5.  This handles
+    cases where a platform declares a file under a different name than
+    the emulator profile (e.g. Batocera ROM1 vs gsplus ROM).
+    """
+    declared: set[str] = set()
+    by_md5 = db.get("indexes", {}).get("by_md5", {})
+    files_db = db.get("files", {})
+    for system in config.get("systems", {}).values():
+        for fe in system.get("files", []):
+            name = fe.get("name", "")
+            if name:
+                declared.add(name)
+            md5 = fe.get("md5", "")
+            if not md5:
+                continue
+            # Skip multi-hash and zippedFile entries (inner ROM MD5, not file MD5)
+            if "," in md5 or fe.get("zippedFile"):
+                continue
+            sha1 = by_md5.get(md5.lower())
+            if not sha1:
+                continue
+            entry = files_db.get(sha1, {})
+            db_name = entry.get("name", "")
+            if db_name:
+                declared.add(db_name)
+            for alias in entry.get("aliases", []):
+                declared.add(alias)
+    return declared
+
 
 # Validation and mode filtering -extracted to validation.py for SoC.
 # Re-exported below for backward compatibility.

scripts/generate_pack.py

@@ -28,7 +28,7 @@ sys.path.insert(0, os.path.dirname(__file__))
 from common import (
     MANUFACTURER_PREFIXES,
     build_target_cores_cache, build_zip_contents_index, check_inside_zip,
-    compute_hashes, fetch_large_file, group_identical_platforms,
+    compute_hashes, expand_platform_declared_names, fetch_large_file, group_identical_platforms,
     list_emulator_profiles, list_platform_system_ids, list_registered_platforms,
     filter_systems_by_target, list_system_ids, load_database,
     load_data_dir_registry, load_emulator_profiles, load_platform_config,
@@ -371,12 +371,8 @@ def _collect_emulator_extras(
     by_path_suffix = db.get("indexes", {}).get("by_path_suffix", {})
 
     # Build set of filenames already covered (platform baseline + first pass extras)
-    covered_names: set[str] = set()
+    # Enriched with canonical names from DB via MD5 (handles platform renaming)
-    for sys_id, system in config.get("systems", {}).items():
+    covered_names = expand_platform_declared_names(config, db)
-        for fe in system.get("files", []):
-            n = fe.get("name", "")
-            if n:
-                covered_names.add(n)
     for e in extras:
         covered_names.add(e["name"])
 

scripts/verify.py

@@ -31,10 +31,11 @@ from pathlib import Path
 sys.path.insert(0, os.path.dirname(__file__))
 from common import (
     build_target_cores_cache, build_zip_contents_index, check_inside_zip,
-    compute_hashes, filter_systems_by_target, group_identical_platforms,
+    compute_hashes, expand_platform_declared_names, filter_systems_by_target,
-    list_emulator_profiles, list_system_ids, load_data_dir_registry,
+    group_identical_platforms, list_emulator_profiles, list_system_ids,
-    load_emulator_profiles, load_platform_config, md5sum, md5_composite,
+    load_data_dir_registry, load_emulator_profiles, load_platform_config,
-    require_yaml, resolve_local_file, resolve_platform_cores,
+    md5sum, md5_composite, require_yaml, resolve_local_file,
+    resolve_platform_cores,
 )
 
 yaml = require_yaml()
@@ -261,13 +262,9 @@ def find_undeclared_files(
     data_names: set[str] | None = None,
 ) -> list[dict]:
     """Find files needed by cores but not declared in platform config."""
-    # Collect all filenames declared by this platform
+    # Collect all filenames declared by this platform, enriched with
-    declared_names: set[str] = set()
+    # canonical names from DB via MD5 (handles platform renaming)
-    for sys_id, system in config.get("systems", {}).items():
+    declared_names = expand_platform_declared_names(config, db)
-        for fe in system.get("files", []):
-            name = fe.get("name", "")
-            if name:
-                declared_names.add(name)
 
     # Collect data_directory refs
     declared_dd: set[str] = set()

tests/test_e2e.py

@@ -31,6 +31,7 @@ sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", "scripts"))
 import yaml
 from common import (
     build_zip_contents_index, check_inside_zip, compute_hashes,
+    expand_platform_declared_names,
     group_identical_platforms, load_emulator_profiles, load_platform_config,
     md5_composite, md5sum, parse_md5_list, resolve_local_file,
     resolve_platform_cores, safe_extract_zip,
@@ -269,6 +270,12 @@ class TestE2E(unittest.TestCase):
                         {"ref": "test-data-dir", "destination": "TestData"},
                     ],
                 },
+                "sys-renamed": {
+                    "files": [
+                        {"name": "renamed_file.bin", "destination": "renamed_file.bin",
+                         "md5": f["correct_hash.bin"]["md5"], "required": True},
+                    ],
+                },
             },
         }
         with open(os.path.join(self.platforms_dir, "test_md5.yml"), "w") as fh:
@@ -490,6 +497,19 @@ class TestE2E(unittest.TestCase):
         with open(os.path.join(self.emulators_dir, "test_subdir_core.yml"), "w") as fh:
             yaml.dump(emu_subdir, fh)
 
+        # Emulator whose file is declared by platform under a different name
+        # (e.g. gsplus ROM vs Batocera ROM1) — hash-based matching should resolve
+        emu_renamed = {
+            "emulator": "TestRenamed",
+            "type": "standalone",
+            "systems": ["sys-renamed"],
+            "files": [
+                {"name": "correct_hash.bin", "required": True},
+            ],
+        }
+        with open(os.path.join(self.emulators_dir, "test_renamed.yml"), "w") as fh:
+            yaml.dump(emu_renamed, fh)
+
     # ---------------------------------------------------------------
     # THE TEST -one method per feature area, all using same fixtures
     # ---------------------------------------------------------------
@@ -3251,6 +3271,25 @@ class TestE2E(unittest.TestCase):
         self.assertEqual(div["extra_unprofiled"][0]["name"], "phantom.bin")
         self.assertNotIn("extra_phantom", div)
 
+    def test_173_cross_ref_hash_matching(self):
+        """Platform file under different name matched by MD5 is not undeclared."""
+        config = load_platform_config("test_md5", self.platforms_dir)
+        profiles = load_emulator_profiles(self.emulators_dir)
+        undeclared = find_undeclared_files(config, self.emulators_dir, self.db, profiles)
+        names = {u["name"] for u in undeclared}
+        # correct_hash.bin is declared by platform as renamed_file.bin with same MD5
+        # hash-based matching should suppress it from undeclared
+        self.assertNotIn("correct_hash.bin", names)
+
+    def test_174_expand_platform_declared_names(self):
+        """expand_platform_declared_names enriches with DB canonical names."""
+        config = load_platform_config("test_md5", self.platforms_dir)
+        result = expand_platform_declared_names(config, self.db)
+        # renamed_file.bin is declared directly
+        self.assertIn("renamed_file.bin", result)
+        # correct_hash.bin is the DB canonical name for the same MD5
+        self.assertIn("correct_hash.bin", result)
+
 
 if __name__ == "__main__":
     unittest.main()