fix: choose a variant the same way on every run

scripts/validation.py

@@ -512,11 +512,13 @@ def find_validated_variant(
     seen: set[str] = set()
 
     def candidates():
+        # Sorted: the index holds sets, whose order follows the per-process
+        # string hash, and the first passing candidate is the one shipped.
         expected = validation_index[fname]
         for hash_type, index_key in (
             ("sha1", None), ("md5", "by_md5"), ("crc32", "by_crc32"), ("sha256", "by_sha256"),
         ):
-            for value in expected.get(hash_type) or []:
+            for value in sorted(expected.get(hash_type) or []):
                 if index_key is None:
                     yield value
                     continue

tests/test_variant_choice.py

@@ -10,7 +10,9 @@ never swapped in. Both sides now read find_validated_variant.
 from __future__ import annotations
 
 import hashlib
+import os
 import re
+import subprocess
 import sys
 import tempfile
 import unittest
@@ -87,5 +89,45 @@ class OneImplementation(unittest.TestCase):
                 self.assertGreaterEqual(len(calls), 2 if name == "generate_pack.py" else 3)
 
 
+_SEEDED = """
+import hashlib, sys, tempfile
+from pathlib import Path
+sys.path.insert(0, sys.argv[1])
+from validation import _build_validation_index, find_validated_variant
+root = Path(sys.argv[2])
+paths = [root / f"ctrl_{i}.rom" for i in range(8)]
+files = {}
+for i, path in enumerate(paths):
+    data = bytes([i]) * 16
+    path.write_bytes(data)
+    files[hashlib.sha1(data).hexdigest()] = {"path": str(path), "name": path.name}
+bad = root / "MT32_CONTROL.ROM"
+bad.write_bytes(b"x" * 16)
+index = _build_validation_index({"emu": {"files": [
+    {"name": "MT32_CONTROL.ROM", "validation": ["sha1"], "sha1": sha1} for sha1 in files]}})
+db = {"files": files, "indexes": {"by_name": {}}}
+print(find_validated_variant({"name": "MT32_CONTROL.ROM"}, db, str(bad), index))
+"""
+
+
+class TheChoiceDoesNotDependOnTheRun(unittest.TestCase):
+    """Eight MT-32 control ROMs pass dosbox_pure's sha1 check. The candidates
+    were walked in set order, which follows the per-process string hash: the
+    manifest named ctrl_mt32_2_07.rom in one run and ctrl_mt32_2_06.rom in
+    the next."""
+
+    def test_the_same_variant_under_every_hash_seed(self):
+        chosen = set()
+        for seed in ("1", "2", "3", "4", "5", "6"):
+            with tempfile.TemporaryDirectory() as tmp:
+                out = subprocess.run(
+                    [sys.executable, "-c", _SEEDED, str(REPO_ROOT / "scripts"), tmp],
+                    env={**os.environ, "PYTHONHASHSEED": seed},
+                    capture_output=True, text=True, check=True,
+                ).stdout.strip()
+                chosen.add(Path(out).name)
+        self.assertEqual(len(chosen), 1, chosen)
+
+
 if __name__ == "__main__":
     unittest.main()