fix: agree on missing files in system packs

The emulator pack recorded only loose files as a path's occupant, so a
path an archive or data tree had filled was reported missing even with
the same bytes in the ZIP, and an archive several profiles ask for was
counted once per profile. verify registered neither, and credited one
profile's file at a path holding another's. Both now record archives
and tree members in pack order; --system snk-neogeo-mvs went from 113
missing in the pack and 51 in verify to 59 on both sides.

scripts/verify.py

@@ -1466,9 +1466,28 @@ def verify_emulator(
     sha1_by_path = {
         entry.get("path"): sha1 for sha1, entry in db.get("files", {}).items()
     }
+    # Destination -> (file packed there, its profile, the path as written).
+    # Archives and data trees take paths too, before and after the loose
+    # files of their profile, as the pack writes them.
     placed: dict[str, tuple[str | None, str, str]] = {}
     placed_files: set[str] = set()
     placed_parents: set[str] = set()
+    content_ids: dict[str, str | None] = {}
+
+    def content_id(path: str | None) -> str | None:
+        if not path:
+            return None
+        if path not in content_ids:
+            content_ids[path] = sha1_by_path.get(path) or (
+                compute_hashes(path)["sha1"] if os.path.isfile(path) else None
+            )
+        return content_ids[path]
+
+    def occupy(dest: str, path: str | None, owner: str) -> None:
+        if placement_block(dest, placed_files, set(placed), placed_parents, True):
+            return
+        placed[dest.lower()] = (path, owner, dest)
+        _register_path(dest, placed_files, placed_parents)
     for emu_name, profile in sorted(selected):
         files = expand_directory_entries(
             filter_files_by_mode(profile.get("files", []), standalone),
@@ -1559,6 +1578,8 @@ def verify_emulator(
                     archive, profile.get("archive_prefix", ""),
                     profile.get("pack_structure"), standalone,
                 )
+                if local_path:
+                    occupy(dest, local_path, emu_name)
                 dest_to_name[dest] = archive
                 cur = result["status"]
                 prev = file_status.get(dest)
@@ -1690,20 +1711,22 @@ def verify_emulator(
             )
             block_reason = ""
             if block and held is not None:
-                other_sha1, owner, other_dest = placed.get(dest.lower(), (None, "", ""))
+                other_path, owner, other_dest = placed.get(dest.lower(), (None, "", ""))
+                same = content_id(held) is not None and content_id(held) == content_id(
+                    other_path
+                )
                 if block == "path_conflict":
                     block_reason = "a packed file occupies its directory"
                 elif block == "case_twin":
                     block_reason = (
                         f"packed only as {other_dest}, from {owner}"
-                        if sha1_by_path.get(held) == other_sha1
+                        if same
                         else f"the pack holds {owner}'s {other_dest} at this path"
                     )
-                elif sha1_by_path.get(held) != other_sha1:
+                elif not same:
                     block_reason = f"the pack holds {owner}'s {other_dest} at this path"
             elif not block and packed is not None:
-                placed[dest.lower()] = (sha1_by_path.get(packed), emu_name, dest)
-                _register_path(dest, placed_files, placed_parents)
+                occupy(dest, packed, emu_name)
             for result, required, hle in alternatives:
                 if block_reason and result["status"] != Status.MISSING:
                     result["status"] = Status.MISSING
@@ -1726,6 +1749,25 @@ def verify_emulator(
                 ) > _SEVERITY_ORDER.get(prev_sev, 0):
                     file_severity[dest] = sev
 
+        # The profile's data trees follow its files in the pack.
+        for dd in filter_files_by_mode(profile.get("data_directories", []), standalone):
+            cache = (data_registry or {}).get(dd.get("ref", ""), {}).get("local_cache", "")
+            if not cache or not os.path.isdir(cache):
+                continue
+            dd_dest = dd.get("destination", "")
+            structure = profile.get("pack_structure")
+            prefix = (
+                structure.get("standalone" if standalone else "libretro", "")
+                if structure else ""
+            )
+            if prefix:
+                dd_dest = f"{prefix}/{dd_dest}" if dd_dest else prefix
+            for root, _dirs, names in os.walk(cache):
+                for fname in names:
+                    src = os.path.join(root, fname)
+                    rel = os.path.relpath(src, cache).replace(os.sep, "/")
+                    occupy(f"{dd_dest}/{rel}" if dd_dest else rel, src, emu_name)
+
     counts = {
         Severity.OK: 0,
         Severity.INFO: 0,