fix(import-audit): parse export aliases as C++ tokens (#4196)

Co-authored-by: lorkydey <lorky.contact@gmail.com>

tools/import_audit.py

@@ -14,7 +14,6 @@ import progress
 ROOT = Path(__file__).resolve().parent.parent
 PRX = ROOT / "core" / "libs" / "prx"
 NID = re.compile(r"[A-Za-z0-9+\-]{11}")
-EXPORT_ALIAS = re.compile(r'APS5_EXPORT\("([A-Za-z0-9+\-]{11})",\s*(\w+)\)')
 NID_POSTFIX = "_nid_postfix"
 CLASSES = ("implemented", "stub", "absent", "module")
 PE_DIRECTORY_OFFSET = {0x20b: 112, 0x10b: 96}
@@ -208,7 +207,7 @@ def source_stubs(source):
     for library in sorted(item for item in Path(source).iterdir() if item.is_dir()):
         aliases = {}
         for file in library.rglob("*.cpp"):
-            for nid, name in EXPORT_ALIAS.findall(file.read_text(errors="ignore")):
+            for nid, name in nid_names.source_aliases(file.read_text(errors="ignore")):
                 aliases[name] = nid
         group = progress.scan_library(library)
         done = {nid_of(name, aliases) for name in group["done_names"]}

tools/nid_names.py

@@ -81,7 +81,7 @@ def source_aliases(text):
             continue
         if re.fullmatch(r'"[A-Za-z0-9+\-]{11}"', literal) is None:
             continue
-        if re.fullmatch(r"[A-Za-z_][A-Za-z0-9_]*", name) and "unknown" in name.lower():
+        if re.fullmatch(r"[A-Za-z_][A-Za-z0-9_]*", name):
             yield literal[1:-1], name
 
 
@@ -97,7 +97,8 @@ def collect_unknowns():
     for path in source_files(PRX):
         text = path.read_text(encoding="utf-8")
         for nid, name in source_aliases(text):
-            found.setdefault(nid, "%s:%s" % (path.relative_to(ROOT).as_posix(), name))
+            if "unknown" in name.lower():
+                found.setdefault(nid, "%s:%s" % (path.relative_to(ROOT).as_posix(), name))
     return found
 
 

tools/tests/test_import_audit.py

@@ -191,6 +191,48 @@ class ImportAuditTests(unittest.TestCase):
         records, _ = import_audit.audit([(nid, "libB.prx")], exports, stubs, {"libA.prx", "libB.prx"}, set())
         self.assertEqual(records[0]["class"], "implemented")
 
+    def test_formatted_export_aliases_remain_stubs(self):
+        work = self.work
+        nid = "AAAAAAAAAAA"
+        (work.libs / "libSceAlias.prx").write_bytes(exporting(nid))
+        registry = work.root / "game.json"
+        result = work.root / "audit.json"
+        write_registry(registry, [(nid, "libSceAlias.prx")])
+        aliases = [
+            'APS5_EXPORT (\n "AAAAAAAAAAA",\n NAME\n);\n',
+            'APS5_EXPORT(/* nid */ "AAAAAAAAAAA", /* name */ NAME);\n',
+            'APS5_EX\\\nPORT("AAAAAAAAAAA", NAME);\n',
+        ]
+        for name in ("sceAliasUnknown", "sceAliasedCall"):
+            for alias in aliases:
+                with self.subTest(name=name, alias=alias):
+                    write_source(work.source, "libSceAlias", alias.replace("NAME", name) + STUB_SOURCE.format(name=name))
+                    code, _, err = work.run(registry, "--json", str(result))
+                    self.assertEqual((code, err), (0, ""))
+                    record, = json.loads(result.read_text())["imports"]
+                    self.assertEqual((record["class"], record["name"]), ("stub", name))
+
+    def test_comments_and_literals_do_not_override_export_aliases(self):
+        work = self.work
+        nid, name = "AAAAAAAAAAA", "sceAliasUnknown"
+        (work.libs / "libSceAlias.prx").write_bytes(exporting(nid))
+        registry = work.root / "game.json"
+        result = work.root / "audit.json"
+        write_registry(registry, [(nid, "libSceAlias.prx")])
+        source = 'APS5_EXPORT("AAAAAAAAAAA", sceAliasUnknown);\n' + STUB_SOURCE.format(name=name)
+        decoys = [
+            '// APS5_EXPORT("BBBBBBBBBBB", sceAliasUnknown);\n',
+            '/* APS5_EXPORT("BBBBBBBBBBB", sceAliasUnknown); */\n',
+            'auto example = R"text(APS5_EXPORT("BBBBBBBBBBB", sceAliasUnknown);)text";\n',
+        ]
+        for decoy in decoys:
+            with self.subTest(decoy=decoy):
+                write_source(work.source, "libSceAlias", source + decoy)
+                code, _, err = work.run(registry, "--json", str(result))
+                self.assertEqual((code, err), (0, ""))
+                record, = json.loads(result.read_text())["imports"]
+                self.assertEqual((record["class"], record["name"]), ("stub", name))
+
     @unittest.skipUnless((import_audit.ROOT / ".git").exists(), "needs a git checkout")
     def test_json_is_stamped_with_the_source_commit(self):
         work = self.work

tools/tests/test_nid_names.py

@@ -161,6 +161,11 @@ int APS5_VABI sceReal() { return 0; }
         self.write_source('APS5_EXPORT (\n "AAAAAAAAAAA" ,\n exampleUnknown00\n );\n')
         self.assertEqual(nid_names.collect_unknowns(), {"AAAAAAAAAAA": "core/libs/prx/libSceExample/Export.cpp:exampleUnknown00"})
 
+    def test_known_aliases_are_not_unknown_exports(self):
+        self.write_source('APS5_EXPORT("AAAAAAAAAAA", sceKnownCall);\n'
+                          'APS5_EXPORT("BBBBBBBBBBB", exampleUnknown00);\n')
+        self.assertEqual(nid_names.collect_unknowns(), {"BBBBBBBBBBB": "core/libs/prx/libSceExample/Export.cpp:exampleUnknown00"})
+
     def test_windows_source_paths_use_portable_separators(self):
         self.write_source('APS5_EXPORT("AAAAAAAAAAA", exampleUnknown00);\n')
         relative = PureWindowsPath("core/libs/prx/libSceExample/Export.cpp")