fix(findrefs,listclass): bound exact-member scan, decode names as MUTF-8

droidasc/asc_client/apk_handler.py

@@ -415,6 +415,10 @@ def get_class_dex(self, dalvik_class : str):
             fp.close()
 
     def list_classes(self, prefix : str = None):
+        # deferred like the ProcessPoolExecutor import in for_each_findrefs, so the
+        # getclass path does not pay for it
+        from droidasc.asc_core.utils.tinydex import _decode_mutf8_fallback
+
         if self.max_workers <= 0:
             raise ValueError("Worker count must be greater than zero")
         if prefix is not None:
@@ -425,6 +429,15 @@ def list_classes(self, prefix : str = None):
                 prefix = "L" + prefix.replace(".", "/")
         prefix_bytes = prefix.encode("ascii") if prefix and prefix.isascii() else None
 
+        def decode_name(raw_name : bytes) -> str:
+            # string_data_item is MUTF-8, so strict UTF-8 rejects a surrogate pair
+            # above the BMP; decode exactly as DEX.get_string does, or a listing
+            # names a class differently from every other command
+            try:
+                return raw_name.decode("utf-8")
+            except UnicodeDecodeError:
+                return _decode_mutf8_fallback(raw_name)
+
         def list_dex_classes(buf):
             if len(buf) < 0x70 or buf[:3] != b"dex":
                 raise ValueError("invalid DEX header")
@@ -464,9 +477,9 @@ def list_dex_classes(buf):
 
                 if prefix_bytes is not None:
                     if raw_name.startswith(prefix_bytes):
-                        append(raw_name.decode("utf-8", errors="replace"))
+                        append(decode_name(raw_name))
                 else:
-                    name = raw_name.decode("utf-8", errors="replace")
+                    name = decode_name(raw_name)
                     if prefix is None or name.startswith(prefix):
                         append(name)
             return names

droidasc/asc_core/findrefs/locator/field_locator.py

@@ -108,7 +108,12 @@ def locate(self, find : dict) -> set:
 
         if clz is None:
             if field_precise:
-                return {fid for fid, item in enumerate(self.dex.fields) if item.name == field}
+                # see MethodLocator.locate: dex.fields is a proxy without __iter__,
+                # so enumerate() walks off the field table instead of stopping at it
+                return {
+                    fid for fid in range(len(self.dex.fields))
+                    if self.dex.fields[fid].name == field
+                }
             name_idxs = self.str_locator.locate(field)
             field_maps = self.field_maps
             ret = set()

droidasc/asc_core/findrefs/locator/method_locator.py

@@ -107,7 +107,13 @@ def locate(self, find : dict) -> set:
 
         if clz is None:
             if method_precise:
-                return {mid for mid, item in enumerate(self.dex.methods) if item.name == method}
+                # dex.methods is a proxy without __iter__, so enumerate() falls back
+                # to the sequence protocol and never sees an IndexError - it walks off
+                # the method table into garbage. Bound the walk by the table size.
+                return {
+                    mid for mid in range(len(self.dex.methods))
+                    if self.dex.methods[mid].name == method
+                }
             name_idxs = self.str_locator.locate(method)
             method_maps = self.method_maps
             ret = set()

pyproject.toml

@@ -4,15 +4,15 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "droidasc"
-version = "0.1.1.post4"
+version = "0.1.1.post5"
 description = "ASC is a super FAST python Android Decompiler for Agents/Mobile Researchers"
 readme = "README.md"
 license = "Apache-2.0"
 license-files = ["LICENSE"]
 requires-python = ">=3.10"
 dependencies = [
     "axml>=0.0.2",
-    "mutf8>=1.1.0",
+    "mutf8>=1.0.6",
 ]
 
 [project.scripts]