diff --git a/src/calibre/ebooks/html/input.py b/src/calibre/ebooks/html/input.py
index f566714878..4385291f69 100644
--- a/src/calibre/ebooks/html/input.py
+++ b/src/calibre/ebooks/html/input.py
@@ -233,7 +233,7 @@ class HTMLInput(InputFormatPlugin):
name = 'HTML Input'
author = 'Kovid Goyal'
description = 'Convert HTML and OPF files to an OEB'
- file_types = set(['opf', 'html', 'htm', 'xhtml', 'xhtm'])
+ file_types = set(['opf', 'html', 'htm', 'xhtml', 'xhtm', 'shtm', 'shtml'])
options = set([
OptionRecommendation(name='breadth_first',
diff --git a/src/calibre/ebooks/mobi/reader.py b/src/calibre/ebooks/mobi/reader.py
index c1ea4d8f2f..686a6efdf9 100644
--- a/src/calibre/ebooks/mobi/reader.py
+++ b/src/calibre/ebooks/mobi/reader.py
@@ -315,18 +315,24 @@ class MobiReader(object):
htmls = list(root.xpath('//html'))
if len(htmls) > 1:
- self.log.warn('Markup contains multiple tags')
- # Keep only the largest head and body
+ self.log.warn('Markup contains multiple tags, merging.')
+ # Merge all and sections
+ for h in htmls:
+ p = h.getparent()
+ if hasattr(p, 'remove'):
+ p.remove(h)
bodies, heads = root.xpath('//body'), root.xpath('//head')
- def sz(x): return len(list(x.iter()))
- def scmp(x, y): return cmp(sz(x), sz(y))
- body = list(sorted(bodies, cmp=scmp))
- head = list(sorted(heads, cmp=scmp))
for x in root: root.remove(x)
- if head:
- root.append(head[-1])
- if body:
- root.append(body[-1])
+ head, body = map(root.makeelement, ('head', 'body'))
+ for h in heads:
+ for x in h:
+ h.remove(x)
+ head.append(x)
+ for b in bodies:
+ for x in b:
+ b.remove(x)
+ body.append(x)
+ root.append(head), root.append(body)
for x in root.xpath('//script'):
x.getparent().remove(x)
diff --git a/src/calibre/ebooks/oeb/base.py b/src/calibre/ebooks/oeb/base.py
index 711dce0b8d..67a8f04cc8 100644
--- a/src/calibre/ebooks/oeb/base.py
+++ b/src/calibre/ebooks/oeb/base.py
@@ -823,6 +823,7 @@ class Manifest(object):
for key in list(body.attrib.keys()):
if key == 'lang' or key.endswith('}lang'):
body.attrib.pop(key)
+
return data
def _parse_css(self, data):