IGN:...

src/calibre/ebooks/mobi/writer.py

@@ -452,6 +452,13 @@ class MobiWriter(object):
             code = EXTH_CODES[term]
             for item in oeb.metadata[term]:
                 data = self.COLLAPSE_RE.sub(' ', unicode(item))
+                if term == 'identifier':
+                    if data.lower().startswith('urn:isbn:'):
+                        data = data[9:]
+                    elif item.get('scheme', '').lower() == 'isbn':
+                        pass
+                    else:
+                        continue
                 data = data.encode('utf-8')
                 exth.write(pack('>II', code, len(data) + 8))
                 exth.write(data)
@@ -468,7 +475,7 @@ class MobiWriter(object):
             nrecs += 3
         exth = exth.getvalue()
         trail = len(exth) % 4
-        pad = '' if not trail else '\0' * (4 - trail)
+        pad = '\0' * (4 - trail) # Always pad w/ at least 1 byte
         exth = ['EXTH', pack('>II', len(exth) + 12, nrecs), exth, pad]
         return ''.join(exth)
 

src/calibre/ebooks/oeb/base.py

@@ -21,6 +21,7 @@ from lxml import etree
 from lxml import html
 from calibre import LoggingInterface
 from calibre.translations.dynamic import translate
+from calibre.startup import get_lang
 
 XML_PARSER = etree.XMLParser(recover=True)
 XML_NS = 'http://www.w3.org/XML/1998/namespace'
@@ -30,6 +31,7 @@ OPF2_NS = 'http://www.idpf.org/2007/opf'
 DC09_NS = 'http://purl.org/metadata/dublin_core'
 DC10_NS = 'http://purl.org/dc/elements/1.0/'
 DC11_NS = 'http://purl.org/dc/elements/1.1/'
+DC_NSES = set([DC09_NS, DC10_NS, DC11_NS])
 XSI_NS = 'http://www.w3.org/2001/XMLSchema-instance'
 DCTERMS_NS = 'http://purl.org/dc/terms/'
 NCX_NS = 'http://www.daisy.org/z3986/2005/ncx/'
@@ -194,8 +196,12 @@ class Metadata(object):
             if term == OPF('meta') and not value:
                 term = self.fq_attrib.pop('name')
                 value = self.fq_attrib.pop('content')
-            elif term in Metadata.TERMS and not namespace(term):
-                term = DC(term)
+            elif barename(term).lower() in Metadata.TERMS and \
+                 (not namespace(term) or namespace(term) in DC_NSES):
+                # Anything looking like Dublin Core is coerced
+                term = DC(barename(term).lower())
+            elif namespace(term) == OPF2_NS:
+                term = barename(term)
             self.term = term
             self.value = value
             self.attrib = attrib = {}
@@ -217,6 +223,15 @@ class Metadata(object):
                     '%r object has no attribute %r' \
                         % (self.__class__.__name__, name))
         
+        def __getitem__(self, key):
+            return self.attrib[key]
+        
+        def __contains__(self, key):
+            return key in self.attrib
+        
+        def get(self, key, default=None):
+            return self.attrib.get(key, default)
+        
         def __repr__(self):
             return 'Item(term=%r, value=%r, attrib=%r)' \
                 % (barename(self.term), self.value, self.attrib)
@@ -814,7 +829,7 @@ class OEBBook(object):
                     break
         if not metadata.language:
             self.logger.warn(u'Language not specified.')
-            metadata.add('language', 'en')
+            metadata.add('language', get_lang())
         if not metadata.creator:
             self.logger.warn(u'Creator not specified.')
             metadata.add('creator', _('Unknown'))
@@ -857,6 +872,8 @@ class OEBBook(object):
         extras.sort()
         for item in extras:
             spine.add(item, False)
+        if len(spine) == 0:
+            raise OEBError("Spine is empty")
 
     def _guide_from_opf(self, opf):
         self.guide = guide = Guide(self)
@@ -886,8 +903,11 @@ class OEBBook(object):
             if len(result) != 1:
                 return False
         id = result[0]
-        ncx = self.manifest[id].data
-        self.manifest.remove(id)
+        if id not in self.manifest.ids:
+            return False
+        item = self.manifest.ids[id]
+        ncx = item.data
+        self.manifest.remove(item)
         title = xpath(ncx, 'ncx:docTitle/ncx:text/text()')[0]
         self.toc = toc = TOC(title)
         navmaps = xpath(ncx, 'ncx:navMap')