From 587aadde0898d1f0f1abf08a47826cfc1573cd2a Mon Sep 17 00:00:00 2001
From: Abdessamad Derraz <3028866+Abdess@users.noreply.github.com>
Date: Fri, 7 Aug 2026 15:30:37 +0200
Subject: [PATCH] fix: coerce numeric hashes to strings

---
 emulators/vice.yml                     |  2 +-
 scripts/common.py                      | 24 ++++++++++++++
 scripts/exporter/systemdat_exporter.py |  2 +-
 tests/test_e2e.py                      | 44 ++++++++++++++++++++++++++
 4 files changed, 70 insertions(+), 2 deletions(-)

diff --git a/emulators/vice.yml b/emulators/vice.yml
index 3f008602..2906c041 100644
--- a/emulators/vice.yml
+++ b/emulators/vice.yml
@@ -602,7 +602,7 @@ files:
     size: 16384
     sha1: a3d9e5be091b98de39a046ab167fb7632d053682
     md5: b3c450c36a88b491281d18fdf3e94ab0
-    crc32: 70295038
+    crc32: "70295038"
     note: "Plus/4 Kernal NTSC Rev 5 (default NTSC). Embedded in core."
     source_ref: "vice/src/plus4/plus4rom.h:37"
 
diff --git a/scripts/common.py b/scripts/common.py
index e4cfc93d..cb43d41d 100644
--- a/scripts/common.py
+++ b/scripts/common.py
@@ -768,11 +768,35 @@ def load_emulator_profiles(
             continue
         if skip_aliases and profile.get("type") == "alias":
             continue
+        _normalize_hash_fields(profile)
         profiles[f.stem] = profile
     _emulator_profiles_cache[cache_key] = profiles
     return profiles
 
 
+_HASH_FIELDS = ("crc32", "md5", "sha1", "sha256")
+
+
+def _normalize_hash_fields(node: object) -> None:
+    """Coerce hash values to str in place.
+
+    YAML types unquoted digit-only hashes (e.g. crc32: 70295038) as int,
+    which breaks string comparisons and formatting downstream.
+    """
+    if isinstance(node, dict):
+        for key, val in node.items():
+            if key in _HASH_FIELDS:
+                if isinstance(val, int):
+                    node[key] = str(val)
+                elif isinstance(val, list):
+                    node[key] = [str(v) if isinstance(v, int) else v for v in val]
+            else:
+                _normalize_hash_fields(val)
+    elif isinstance(node, list):
+        for item in node:
+            _normalize_hash_fields(item)
+
+
 def unique_emulator_profiles(profiles: dict[str, dict]) -> dict[str, dict]:
     """Profiles counting as distinct emulators (aliases and tests excluded)."""
     return {
diff --git a/scripts/exporter/systemdat_exporter.py b/scripts/exporter/systemdat_exporter.py
index 63c94228..dc9a07e2 100644
--- a/scripts/exporter/systemdat_exporter.py
+++ b/scripts/exporter/systemdat_exporter.py
@@ -96,7 +96,7 @@ class Exporter(BaseExporter):
                     rom_parts.append(f"size {size}")
                 crc = fe.get("crc32", "")
                 if crc:
-                    rom_parts.append(f"crc {crc.upper()}")
+                    rom_parts.append(f"crc {str(crc).upper()}")
                 md5 = fe.get("md5", "")
                 if isinstance(md5, list):
                     md5 = md5[0] if md5 else ""
diff --git a/tests/test_e2e.py b/tests/test_e2e.py
index a8f2588d..9c91bea8 100644
--- a/tests/test_e2e.py
+++ b/tests/test_e2e.py
@@ -935,6 +935,22 @@ class TestE2E(unittest.TestCase):
         profiles = load_emulator_profiles(self.emulators_dir)
         self.assertNotIn("test_alias", profiles)
 
+    def test_42b_profile_hash_fields_load_as_str(self):
+        """Unquoted digit-only hashes must load as strings, not YAML ints."""
+        with tempfile.TemporaryDirectory() as tmp:
+            Path(tmp, "numcrc.yml").write_text(
+                "emulator: NumCRC\n"
+                "type: libretro\n"
+                "files:\n"
+                "  - name: kernal.bin\n"
+                "    crc32: 70295038\n"
+                "    md5: [12345678901234567890123456789012]\n"
+            )
+            profiles = load_emulator_profiles(tmp)
+            entry = profiles["numcrc"]["files"][0]
+        self.assertEqual(entry["crc32"], "70295038")
+        self.assertEqual(entry["md5"], ["12345678901234567890123456789012"])
+
     def test_43_cross_ref_data_dir_does_not_suppress_files(self):
         config = load_platform_config("test_md5", self.platforms_dir)
         profiles = load_emulator_profiles(self.emulators_dir)
@@ -3842,6 +3858,34 @@ class TestE2E(unittest.TestCase):
         self.assertIn("retroarch", exporters)
         self.assertIs(exporters["retroarch"], SystemDatExporter)
 
+    def test_systemdat_exporter_numeric_crc(self):
+        """Digit-only crc32 typed as int by YAML must not break the export."""
+        from exporter.systemdat_exporter import Exporter as SystemDatExporter
+
+        truth = {
+            "platform": "retroarch",
+            "systems": {
+                "commodore-plus4": {
+                    "files": [
+                        {
+                            "name": "kernal-318005-05.bin",
+                            "path": "kernal-318005-05.bin",
+                            "size": 16384,
+                            "crc32": 70295038,
+                            "required": False,
+                        },
+                    ],
+                },
+            },
+        }
+
+        out_path = os.path.join(self.root, "System_numeric.dat")
+        SystemDatExporter().export(truth, out_path, scraped_data={"systems": {}})
+
+        with open(out_path) as fh:
+            content = fh.read()
+        self.assertIn("crc 70295038", content)
+
     # ---------------------------------------------------------------
     # Full truth + diff integration test
     # ---------------------------------------------------------------