add specialize_as(), use it to speed up .smali comparison in APKs

diffoscope/comparators/apk.py

@@ -35,9 +35,11 @@ from diffoscope.tools import (
 )
 from diffoscope.tempfiles import get_temporary_directory
 
-from .utils.archive import Archive
+from .text import TextFile
+from .utils.archive import Archive, ArchiveMember
 from .utils.command import Command
 from .utils.compare import compare_files
+from .utils.specialize import specialize_as
 from .zip import ZipContainer, zipinfo_differences, ZipFileBase
 from .missing_file import MissingFile
 
@@ -157,6 +159,14 @@ class ApkContainer(Archive):
     def get_member_names(self):
         return self._members
 
+    def get_member(self, member_name):
+        member = ArchiveMember(self, member_name)
+        if member_name.endswith(".smali") and member_name.startswith("smali"):
+            # smali{,_classesN}/**/*.smali files from apktool are always text,
+            # and using libmagic on thousands of these files takes minutes
+            return specialize_as(TextFile, member)
+        return member
+
     def extract(self, member_name, dest_dir):
         return os.path.join(self._tmpdir.name, member_name)
 

diffoscope/comparators/utils/specialize.py

@@ -28,9 +28,6 @@ logger = logging.getLogger(__name__)
 
 
 def try_recognize(file, cls, recognizes):
-    if isinstance(file, cls):
-        return True
-
     # Does this file class match?
     with profile("recognizes", file):
         # logger.debug("trying %s on %s", cls, file)
@@ -43,13 +40,22 @@ def try_recognize(file, cls, recognizes):
         format_class(cls, strip="diffoscope.comparators."),
         file.name,
     )
-    new_cls = type(cls.__name__, (cls, type(file)), {})
-    file.__class__ = new_cls
+    specialize_as(cls, file)
 
     return True
 
 
+def specialize_as(cls, file):
+    new_cls = type(cls.__name__, (cls, type(file)), {})
+    file.__class__ = new_cls
+    return file
+
+
 def specialize(file):
+    for cls in ComparatorManager().classes:
+        if isinstance(file, cls):
+            return file
+
     for cls in ComparatorManager().classes:
         if try_recognize(file, cls, cls.recognizes):
             return file