py3: misc fixes for conversion pipeline

src/calibre/ebooks/conversion/plugins/html_output.py

@@ -151,7 +151,7 @@ class HTMLOutput(OutputFormatPlugin):
                         pass
                 else:
                     with open(path, 'wb') as f:
-                        f.write(str(item))
+                        f.write(item.bytes_representation)
                     item.unload_data_from_memory(memory=path)
 
             for item in oeb_book.spine:

src/calibre/ebooks/conversion/plugins/oeb_output.py

@@ -51,7 +51,7 @@ class OEBOutput(OutputFormatPlugin):
                     if key == OPF_MIME:
                         # Needed as I can't get lxml to output opf:role and
                         # not output <opf:metadata> as well
-                        raw = re.sub(r'(<[/]{0,1})opf:', r'\1', raw)
+                        raw = re.sub(br'(<[/]{0,1})opf:', br'\1', raw)
                     with open(href, 'wb') as f:
                         f.write(raw)
 
@@ -65,7 +65,7 @@ class OEBOutput(OutputFormatPlugin):
                 if not os.path.exists(dir):
                     os.makedirs(dir)
                 with open(path, 'wb') as f:
-                    f.write(str(item))
+                    f.write(item.bytes_representation)
                 item.unload_data_from_memory(memory=path)
 
     def workaround_nook_cover_bug(self, root):  # {{{

src/calibre/ebooks/oeb/base.py

@@ -1095,6 +1095,10 @@ class Manifest(object):
                 return unicode_type(data.cssText, 'utf-8', 'replace')
             return unicode_type(data)
 
+        @property
+        def bytes_representation(self):
+            return serialize(self.data, self.media_type, pretty_print=self.oeb.pretty_print)
+
         if ispy3:
             def __str__(self):
                 return self.unicode_representation
@@ -1103,7 +1107,7 @@ class Manifest(object):
                 return self.unicode_representation
 
             def __str__(self):
-                return serialize(self.data, self.media_type, pretty_print=self.oeb.pretty_print)
+                return self.bytes_representation
 
         def __eq__(self, other):
             return self is other
@@ -1111,6 +1115,9 @@ class Manifest(object):
         def __ne__(self, other):
             return self is not other
 
+        def __hash__(self):
+            return id(self)
+
         @property
         def sort_key(self):
             href = self.href

src/calibre/ebooks/oeb/stylizer.py

@@ -9,6 +9,7 @@ __license__   = 'GPL v3'
 __copyright__ = '2008, Marshall T. Vandegrift <llasram@gmail.com>'
 
 import os, re, logging, copy, unicodedata, numbers
+from operator import itemgetter
 from weakref import WeakKeyDictionary
 from xml.dom import SyntaxErr as CSSSyntaxError
 from css_parser.css import (CSSStyleRule, CSSPageRule, CSSFontFaceRule,
@@ -215,7 +216,7 @@ class Stylizer(object):
                 else:
                     rules.extend(self.flatten_rule(rule, href, index, is_user_agent_sheet=sheet_index==0))
                     index = index + 1
-        rules.sort()
+        rules.sort(key=itemgetter(0))  # sort by specificity
         self.rules = rules
         self._styles = {}
         pseudo_pat = re.compile(u':{1,2}(%s)' % ('|'.join(INAPPROPRIATE_PSEUDO_CLASSES)), re.I)

src/calibre/ebooks/oeb/writer.py

@@ -62,7 +62,7 @@ class OEBWriter(object):
             os.mkdir(path)
         output = DirContainer(path, oeb.log)
         for item in oeb.manifest.values():
-            output.write(item.href, str(item))
+            output.write(item.href, item.bytes_representation)
 
         if version == 1:
             metadata = oeb.to_opf1()