Fix #7488 (Cover extraction from cbr/cbz does not always extract correct file)

src/calibre/customize/builtins.py

@@ -2,9 +2,7 @@ import os.path
 __license__   = 'GPL v3'
 __copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
 
-import textwrap
-import os
-import glob
+import textwrap, os, glob, functools
 from calibre.customize import FileTypePlugin, MetadataReaderPlugin, \
     MetadataWriterPlugin, PreferencesPlugin, InterfaceActionBase
 from calibre.constants import numeric_version
@@ -95,10 +93,12 @@ class ComicMetadataReader(MetadataReaderPlugin):
 
     def get_metadata(self, stream, ftype):
         if ftype == 'cbr':
-            from calibre.libunrar import extract_member as extract_first
+            from calibre.libunrar import extract_first_alphabetically as extract_first
             extract_first
         else:
-            from calibre.libunzip import extract_member as extract_first
+            from calibre.libunzip import extract_member
+            extract_first = functools.partial(extract_member,
+                    sort_alphabetically=True)
         from calibre.ebooks.metadata import MetaInformation
         ret = extract_first(stream)
         mi = MetaInformation(None, None)

src/calibre/libunrar.py

@@ -269,3 +269,16 @@ def extract_member(path, match=re.compile(r'\.(jpg|jpeg|gif|png)\s*$', re.I),
                 path = _extract_member(path, match, name)
                 return path, open(path, 'rb').read()
 
+def extract_first_alphabetically(path):
+    if hasattr(path, 'read'):
+        data = path.read()
+        f = NamedTemporaryFile(suffix='.rar')
+        f.write(data)
+        f.flush()
+        path = f.name
+
+    names_ = [x for x in names(path) if os.path.splitext(x)[1][1:].lower() in
+            ('png', 'jpg', 'jpeg', 'gif')]
+    names_.sort()
+    return extract_member(path, name=names_[0], match=None)
+

src/calibre/libunzip.py

@@ -43,9 +43,12 @@ def extract(filename, dir):
     zf = zipfile.ZipFile( filename )
     zf.extractall(dir)
 
-def extract_member(filename, match=re.compile(r'\.(jpg|jpeg|gif|png)\s*$', re.I)):
+def extract_member(filename, match=re.compile(r'\.(jpg|jpeg|gif|png)\s*$',
+    re.I), sort_alphabetically=False):
     zf = zipfile.ZipFile(filename)
-    names = zf.namelist()
+    names = list(zf.namelist())
+    if sort_alphabetically:
+        names.sort()
     for name in names:
         if match.search(name):
             return name, zf.read(name)