Convert self closed <div> and <a> tags to empty closed tags as many browser based renderers refuse to recognize them as closed without a proper DOCTYPE. Fixes #3095 (Some epub files can only be advanced chapter to chapter)

src/calibre/ebooks/oeb/base.py

@@ -1010,7 +1010,10 @@ class Manifest(object):
         def __str__(self):
             data = self.data
             if isinstance(data, etree._Element):
-                return xml2str(data, pretty_print=self.oeb.pretty_print)
+                ans = xml2str(data, pretty_print=self.oeb.pretty_print)
+                if self.media_type in OEB_DOCS:
+                    ans = re.sub(r'<(div|a)([^>]*)/>', r'<\1\2></\1>', ans)
+                return ans
             if isinstance(data, unicode):
                 return data.encode('utf-8')
             if hasattr(data, 'cssText'):

src/calibre/ebooks/oeb/transforms/structure.py

@@ -78,6 +78,7 @@ class DetectStructure(object):
             page_break_after = 'display: block; page-break-after: always'
             for item, elem in self.detected_chapters:
                 text = u' '.join([t.strip() for t in elem.xpath('descendant::text()')])
+                text = text.strip()
                 self.log('\tDetected chapter:', text[:50])
                 if chapter_mark == 'none':
                     continue

src/calibre/manual/conversion.rst

@@ -4,3 +4,5 @@
 
 E-book Conversion
 ===================
+
+This section is under construction. In the meantime, you can see some documentation of the command line interface to conversion at :ref:`ebook-convert`.