feat: report external citations as unverifiable

scripts/profile_sync.py

@@ -32,7 +32,9 @@ DEFAULT_CACHE = ".cache/upstream"
 ANON_QUOTA = 60
 TRIAGE_PATH_SAMPLE = 5
 
-STATUS_ORDER = ("ANCHORED", "SHIFTED", "RENAMED", "AMBIGUOUS", "CHANGED", "GONE")
+STATUS_ORDER = (
+    "ANCHORED", "EXTERNAL", "SHIFTED", "RENAMED", "AMBIGUOUS", "CHANGED", "GONE",
+)
 REVIEW_STATUSES = ("CHANGED", "GONE", "AMBIGUOUS")
 REBASE_STATUSES = ("SHIFTED", "RENAMED")
 
@@ -171,6 +173,20 @@ def collect_tokens(entry: dict) -> list[str]:
     return tokens
 
 
+def is_external_citation(path: str) -> bool:
+    """True when a ref names a project rather than a path in a known repo.
+
+    `munt ROMInfo.cpp` and `Nuked-SC55-CLAP rom_io.cpp` cite where a ROM is
+    identified inside a dependency the profile does not declare, so no
+    revision of the declared repositories can confirm or deny them. Reporting
+    them as broken would be wrong: they are simply out of reach.
+    """
+    if " " not in path:
+        return False
+    head = path.split(" ", 1)[0]
+    return "/" not in head and "." not in head
+
+
 def _anchor_tokens(entry: dict) -> list[str]:
     """Values worth searching for when a cited line misses its subject.
 
@@ -350,6 +366,11 @@ def anchor_part(
     `describe(path)` returns (repo slug, raw URL at HEAD) for the repository
     that owns the path, or (None, None) when the caller does not track it.
     """
+    if is_external_citation(part.path):
+        return PartResult(
+            part, "EXTERNAL", None, None, None, [],
+            "names a project the profile does not declare",
+        )
     slug, url, actual = (
         describe(part.path) if describe else (None, None, part.path)
     )

tests/test_profile_sync.py

@@ -285,6 +285,38 @@ def _ambiguous(path, old, candidates):
     )
 
 
+class TestExternalCitation(unittest.TestCase):
+    def test_project_name_then_file(self):
+        self.assertTrue(profile_sync.is_external_citation("munt ROMInfo.cpp"))
+        self.assertTrue(
+            profile_sync.is_external_citation("Nuked-SC55-CLAP rom_io.cpp")
+        )
+        self.assertTrue(
+            profile_sync.is_external_citation("EmuDeck emuDeckares.sh")
+        )
+
+    def test_plain_paths_are_not_citations(self):
+        for path in ("src/midi/mt32.cpp", "libretro.c", "a/b/c.h"):
+            self.assertFalse(profile_sync.is_external_citation(path), path)
+
+    def test_a_path_with_a_directory_prefix_is_not_a_citation(self):
+        self.assertFalse(
+            profile_sync.is_external_citation("src/dir file.cpp")
+        )
+
+    def test_a_filename_prefix_is_not_a_citation(self):
+        self.assertFalse(profile_sync.is_external_citation("main.cpp note"))
+
+    def test_external_part_is_reported_not_reviewed(self):
+        result = anchor_part(
+            RefPart("munt ROMInfo.cpp", 59, 59),
+            make_fetch({}),
+            renamer(CompareResult([], False)),
+        )
+        self.assertEqual(result.status, "EXTERNAL")
+        self.assertNotIn("EXTERNAL", profile_sync.REVIEW_STATUSES)
+
+
 class TestAnchorTokens(unittest.TestCase):
     def test_archive_stem_is_searched_too(self):
         tokens = profile_sync._anchor_tokens({"name": "stvbios.zip"})