fix: fail a pack member the collection cannot name

scripts/packverify.py

@@ -31,6 +31,25 @@ from packresolve import resolve_file
 from common import resolve_platform_cores
 from common import sanitize_pack_path
 import zipfile
+def _members_are_held(data: bytes, by_md5: dict, db: dict) -> bool:
+    """Whether every member of an archive is a dump the collection holds."""
+    import io
+
+    held_inside = build_zip_contents_index(db)
+    try:
+        with zipfile.ZipFile(io.BytesIO(data)) as archive:
+            members = [i for i in archive.infolist() if not i.is_dir()]
+            if not members:
+                return False
+            for info in members:
+                digest = hashlib.md5(archive.read(info)).hexdigest()
+                if digest not in by_md5 and digest not in held_inside:
+                    return False
+    except (zipfile.BadZipFile, OSError):
+        return False
+    return True
+
+
 def verify_pack(
     zip_path: str, db: dict, data_registry: dict | None = None
 ) -> tuple[bool, dict]:
@@ -183,6 +202,19 @@ def verify_pack(
                         except (zipfile.BadZipFile, OSError):
                             continue
 
+            # An archive the builder assembled (a MAME clone set): every
+            # member must be a dump the collection holds, loose or inside a
+            # romset.
+            if status == "untracked" and name.endswith(".zip"):
+                if _members_are_held(zf.read(name), by_md5, db):
+                    status = "verified_members"
+                    file_name = os.path.basename(name)
+
+            if status == "untracked":
+                # Bytes nothing recognises: written wrong, or a source that
+                # no longer matches. Counted and passed, it went unseen.
+                errors.append(f"{name}: content matches no collected file")
+
             manifest["files"].append(
                 {
                     "path": name,

tests/test_pack_member_identity.py

@@ -0,0 +1,46 @@
+"""Every pack member is bytes the collection recognises.
+
+A member whose sha1 and md5 matched nothing was counted `untracked` with
+zero errors and the pack passed, so bytes altered while the ZIP was written
+went unseen. Archives the builder assembles (MAME clone sets) are
+recognised by their members.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import io
+import sys
+import tempfile
+import unittest
+import zipfile
+from pathlib import Path
+
+REPO_ROOT = Path(__file__).resolve().parents[1]
+sys.path.insert(0, str(REPO_ROOT / "scripts"))
+
+
+class MembersMustBeKnown(unittest.TestCase):
+    def test_unknown_bytes_are_an_error_and_assembled_sets_pass(self):
+        from packverify import verify_pack
+
+        known = b"known rom" * 10
+        md5 = hashlib.md5(known).hexdigest()
+        inner = io.BytesIO()
+        with zipfile.ZipFile(inner, "w") as zf:
+            zf.writestr("rom.bin", known)
+        db = {"files": {}, "indexes": {"by_md5": {md5: "s"}, "by_name": {}}}
+        with tempfile.TemporaryDirectory(dir=REPO_ROOT / "tmp") as tmp:
+            pack = Path(tmp) / "P_BIOS_Pack.zip"
+            with zipfile.ZipFile(pack, "w") as zf:
+                zf.writestr("clone.zip", inner.getvalue())
+                zf.writestr("garbled.bin", b"no such dump")
+            ok, manifest = verify_pack(str(pack), db)
+        statuses = {f["path"]: f["status"] for f in manifest["files"]}
+        self.assertEqual(statuses["clone.zip"], "verified_members")
+        self.assertFalse(ok)
+        self.assertTrue(any("garbled.bin" in e for e in manifest["errors"]))
+
+
+if __name__ == "__main__":
+    unittest.main()