diff --git a/src/calibre/customize/builtins.py b/src/calibre/customize/builtins.py
index 8f6c597ee5..d5957eb70a 100644
--- a/src/calibre/customize/builtins.py
+++ b/src/calibre/customize/builtins.py
@@ -625,8 +625,9 @@ if test_eight_code:
     from calibre.ebooks.metadata.sources.google import GoogleBooks
     from calibre.ebooks.metadata.sources.amazon import Amazon
     from calibre.ebooks.metadata.sources.openlibrary import OpenLibrary
+    from calibre.ebooks.metadata.sources.isbndb import ISBNDB
 
-    plugins += [GoogleBooks, Amazon, OpenLibrary]
+    plugins += [GoogleBooks, Amazon, OpenLibrary, ISBNDB]
 
 # }}}
 else:
diff --git a/src/calibre/ebooks/__init__.py b/src/calibre/ebooks/__init__.py
index 7776be5e28..a56abb907e 100644
--- a/src/calibre/ebooks/__init__.py
+++ b/src/calibre/ebooks/__init__.py
@@ -26,7 +26,7 @@ class ParserError(ValueError):
     pass
 
 BOOK_EXTENSIONS = ['lrf', 'rar', 'zip', 'rtf', 'lit', 'txt', 'txtz', 'text', 'htm', 'xhtm',
-                   'html', 'xhtml', 'pdf', 'pdb', 'pdr', 'prc', 'mobi', 'azw', 'doc',
+                   'html', 'htmlz', 'xhtml', 'pdf', 'pdb', 'pdr', 'prc', 'mobi', 'azw', 'doc',
                    'epub', 'fb2', 'djvu', 'lrx', 'cbr', 'cbz', 'cbc', 'oebzip',
                    'rb', 'imp', 'odt', 'chm', 'tpz', 'azw1', 'pml', 'pmlz', 'mbp', 'tan', 'snb']
 
diff --git a/src/calibre/ebooks/htmlz/output.py b/src/calibre/ebooks/htmlz/output.py
index 7cdf04bcdb..03fe12c89e 100644
--- a/src/calibre/ebooks/htmlz/output.py
+++ b/src/calibre/ebooks/htmlz/output.py
@@ -12,7 +12,7 @@ from lxml import etree
 
 from calibre.customize.conversion import OutputFormatPlugin, \
     OptionRecommendation
-from calibre.ebooks.oeb.base import OEB_IMAGES
+from calibre.ebooks.oeb.base import OEB_IMAGES, SVG_MIME
 from calibre.ptempfile import TemporaryDirectory
 from calibre.utils.zipfile import ZipFile
 
@@ -71,9 +71,13 @@ class HTMLZOutput(OutputFormatPlugin):
                     os.makedirs(os.path.join(tdir, 'images'))
                 for item in oeb_book.manifest:
                     if item.media_type in OEB_IMAGES and item.href in images:
+                        if item.media_type == SVG_MIME:
+                            data = unicode(etree.tostring(item.data, encoding=unicode)) 
+                        else:
+                            data = item.data
                         fname = os.path.join(tdir, 'images', images[item.href])
                         with open(fname, 'wb') as img:
-                            img.write(item.data)
+                            img.write(data)
 
             # Metadata
             with open(os.path.join(tdir, 'metadata.opf'), 'wb') as mdataf:
diff --git a/src/calibre/ebooks/metadata/extz.py b/src/calibre/ebooks/metadata/extz.py
index 0ecdbe9ea6..18069b2a6a 100644
--- a/src/calibre/ebooks/metadata/extz.py
+++ b/src/calibre/ebooks/metadata/extz.py
@@ -8,12 +8,13 @@ Read meta information from extZ (TXTZ, HTMLZ...) files.
 '''
 
 import os
+import posixpath
 
 from cStringIO import StringIO
 
 from calibre.ebooks.metadata import MetaInformation
-from calibre.ebooks.metadata.opf2 import OPF, metadata_to_opf
-from calibre.ptempfile import TemporaryDirectory
+from calibre.ebooks.metadata.opf2 import OPF
+from calibre.ptempfile import PersistentTemporaryFile
 from calibre.utils.zipfile import ZipFile, safe_replace
 
 def get_metadata(stream, extract_cover=True):
@@ -23,16 +24,75 @@ def get_metadata(stream, extract_cover=True):
     mi = MetaInformation(_('Unknown'), [_('Unknown')])
     stream.seek(0)
 
-    with TemporaryDirectory('_untxtz_mdata') as tdir:
-        try:
-            zf = ZipFile(stream)
-            zf.extract('metadata.opf', tdir)
-            with open(os.path.join(tdir, 'metadata.opf'), 'rb') as opff:
-                mi = OPF(opff).to_book_metadata()
-        except:
-            return mi
+    try:
+        with ZipFile(stream) as zf:
+            opf_name = get_first_opf_name(zf)
+            opf_stream = StringIO(zf.read(opf_name))
+            opf = OPF(opf_stream)
+            mi = opf.to_book_metadata()
+            if extract_cover:
+                cover_name = opf.raster_cover
+                if cover_name:
+                    mi.cover_data = ('jpg', zf.read(cover_name))
+    except:
+        return mi
     return mi
 
 def set_metadata(stream, mi):
-    opf = StringIO(metadata_to_opf(mi))
-    safe_replace(stream, 'metadata.opf', opf)
+    replacements = {}
+
+    # Get the OPF in the archive.
+    with ZipFile(stream) as zf:
+        opf_path = get_first_opf_name(zf)
+        opf_stream = StringIO(zf.read(opf_path))
+    opf = OPF(opf_stream)
+
+    # Cover.
+    new_cdata = None
+    try:
+        new_cdata = mi.cover_data[1]
+        if not new_cdata:
+            raise Exception('no cover')
+    except:
+        try:
+            new_cdata = open(mi.cover, 'rb').read()
+        except:
+            pass
+    if new_cdata:
+        raster_cover = opf.raster_cover
+        if not raster_cover:
+            raster_cover = 'cover.jpg'
+        cpath = posixpath.join(posixpath.dirname(opf_path), raster_cover)
+        new_cover = _write_new_cover(new_cdata, cpath)
+        replacements[cpath] = open(new_cover.name, 'rb')
+
+    # Update the metadata.
+    opf.smart_update(mi, replace_metadata=True)
+    newopf = StringIO(opf.render())
+    safe_replace(stream, opf_path, newopf, extra_replacements=replacements)
+
+    # Cleanup temporary files.
+    try:
+        if cpath is not None:
+            replacements[cpath].close()
+            os.remove(replacements[cpath].name)
+    except:
+        pass
+
+def get_first_opf_name(zf):
+    names = zf.namelist()
+    opfs = []
+    for n in names:
+        if n.endswith('.opf') and '/' not in n:
+            opfs.append(n)
+    if not opfs:
+        raise Exception('No OPF found')
+    opfs.sort()
+    return opfs[0]
+
+def _write_new_cover(new_cdata, cpath):
+    from calibre.utils.magick.draw import save_cover_data_to
+    new_cover = PersistentTemporaryFile(suffix=os.path.splitext(cpath)[1])
+    new_cover.close()
+    save_cover_data_to(new_cdata, new_cover.name)
+    return new_cover
diff --git a/src/calibre/ebooks/metadata/sources/base.py b/src/calibre/ebooks/metadata/sources/base.py
index 5089d8951b..d9144fdf34 100644
--- a/src/calibre/ebooks/metadata/sources/base.py
+++ b/src/calibre/ebooks/metadata/sources/base.py
@@ -181,6 +181,10 @@ class Source(Plugin):
     #: construct the configuration widget for this plugin
     options = ()
 
+    #: A string that is displayed at the top of the config widget for this
+    #: plugin
+    config_help_message = None
+
 
     def __init__(self, *args, **kwargs):
         Plugin.__init__(self, *args, **kwargs)
diff --git a/src/calibre/ebooks/metadata/sources/covers.py b/src/calibre/ebooks/metadata/sources/covers.py
index cf6ec90c54..d28ce146c6 100644
--- a/src/calibre/ebooks/metadata/sources/covers.py
+++ b/src/calibre/ebooks/metadata/sources/covers.py
@@ -76,6 +76,11 @@ def run_download(log, results, abort,
         (plugin, width, height, fmt, bytes)
 
     '''
+    if title == _('Unknown'):
+        title = None
+    if authors == [_('Unknown')]:
+        authors = None
+
     plugins = [p for p in metadata_plugins(['cover']) if p.is_configured()]
 
     rq = Queue()
@@ -145,7 +150,7 @@ def download_cover(log,
     Synchronous cover download. Returns the "best" cover as per user
     prefs/cover resolution.
 
-    Return cover is a tuple: (plugin, width, height, fmt, data)
+    Returned cover is a tuple: (plugin, width, height, fmt, data)
 
     Returns None if no cover is found.
     '''
diff --git a/src/calibre/ebooks/metadata/sources/identify.py b/src/calibre/ebooks/metadata/sources/identify.py
index fad810c26e..b494e05e1a 100644
--- a/src/calibre/ebooks/metadata/sources/identify.py
+++ b/src/calibre/ebooks/metadata/sources/identify.py
@@ -253,6 +253,10 @@ def merge_identify_results(result_map, log):
 
 def identify(log, abort, # {{{
         title=None, authors=None, identifiers={}, timeout=30):
+    if title == _('Unknown'):
+        title = None
+    if authors == [_('Unknown')]:
+        authors = None
     start_time = time.time()
     plugins = [p for p in metadata_plugins(['identify']) if p.is_configured()]
 
diff --git a/src/calibre/ebooks/metadata/sources/isbndb.py b/src/calibre/ebooks/metadata/sources/isbndb.py
index ab9342c6cb..b8deea56df 100644
--- a/src/calibre/ebooks/metadata/sources/isbndb.py
+++ b/src/calibre/ebooks/metadata/sources/isbndb.py
@@ -7,7 +7,19 @@ __license__   = 'GPL v3'
 __copyright__ = '2011, Kovid Goyal <kovid@kovidgoyal.net>'
 __docformat__ = 'restructuredtext en'
 
-from calibre.ebooks.metadata.sources.base import Source
+from urllib import quote
+
+from lxml import etree
+
+from calibre.ebooks.metadata import check_isbn
+from calibre.ebooks.metadata.sources.base import Source, Option
+from calibre.ebooks.chardet import xml_to_unicode
+from calibre.utils.cleantext import clean_ascii_chars
+from calibre.utils.icu import lower
+from calibre.ebooks.metadata.book.base import Metadata
+
+BASE_URL = 'http://isbndb.com/api/books.xml?access_key=%s&page_number=1&results=subjects,authors,texts&'
+
 
 class ISBNDB(Source):
 
@@ -18,6 +30,20 @@ class ISBNDB(Source):
     touched_fields = frozenset(['title', 'authors',
         'identifier:isbn', 'comments', 'publisher'])
     supports_gzip_transfer_encoding = True
+    # Shortcut, since we have no cached cover URLS
+    cached_cover_url_is_reliable = False
+
+    options = (
+            Option('isbndb_key', 'string', None, _('IsbnDB key:'),
+                _('To use isbndb.com you have to sign up for a free account'
+                    'at isbndb.com and get an access key.')),
+            )
+
+    config_help_message = '<p>'+_('To use metadata from isbndb.com you must sign'
+            ' up for a free account and get an isbndb key and enter it below.'
+            ' Instructions to get the key are '
+            '<a href="http://isbndb.com/docs/api/30-keys.html">here</a>.')
+
 
     def __init__(self, *args, **kwargs):
         Source.__init__(self, *args, **kwargs)
@@ -35,9 +61,186 @@ class ISBNDB(Source):
             except:
                 pass
 
-        self.isbndb_key = prefs['isbndb_key']
+    @property
+    def isbndb_key(self):
+        return self.prefs['isbndb_key']
 
     def is_configured(self):
         return self.isbndb_key is not None
 
+    def create_query(self, title=None, authors=None, identifiers={}): # {{{
+        base_url = BASE_URL%self.isbndb_key
+        isbn = check_isbn(identifiers.get('isbn', None))
+        q = ''
+        if isbn is not None:
+            q = 'index1=isbn&value1='+isbn
+        elif title or authors:
+            tokens = []
+            title_tokens = list(self.get_title_tokens(title))
+            tokens += title_tokens
+            author_tokens = self.get_author_tokens(authors,
+                    only_first_author=True)
+            tokens += author_tokens
+            tokens = [quote(t) for t in tokens]
+            q = '+'.join(tokens)
+            q = 'index1=combined&value1='+q
+
+        if not q:
+            return None
+        if isinstance(q, unicode):
+            q = q.encode('utf-8')
+        return base_url + q
+    # }}}
+
+    def identify(self, log, result_queue, abort, title=None, authors=None, # {{{
+            identifiers={}, timeout=30):
+        if not self.is_configured():
+            return
+        query = self.create_query(title=title, authors=authors,
+                identifiers=identifiers)
+        if not query:
+            err = 'Insufficient metadata to construct query'
+            log.error(err)
+            return err
+
+        results = []
+        try:
+            results = self.make_query(query, abort, title=title, authors=authors,
+                    identifiers=identifiers, timeout=timeout)
+        except:
+            err = 'Failed to make query to ISBNDb, aborting.'
+            log.exception(err)
+            return err
+
+        if not results and identifiers.get('isbn', False) and title and authors and \
+                not abort.is_set():
+            return self.identify(log, result_queue, abort, title=title,
+                    authors=authors, timeout=timeout)
+
+        for result in results:
+            self.clean_downloaded_metadata(result)
+            result_queue.put(result)
+
+    def parse_feed(self, feed, seen, orig_title, orig_authors, identifiers):
+
+        def tostring(x):
+            if x is None:
+                return ''
+            return etree.tostring(x, method='text', encoding=unicode).strip()
+
+        orig_isbn = identifiers.get('isbn', None)
+        title_tokens = list(self.get_title_tokens(orig_title))
+        author_tokens = list(self.get_author_tokens(orig_authors))
+        results = []
+
+        def ismatch(title, authors):
+            authors = lower(' '.join(authors))
+            title = lower(title)
+            match = not title_tokens
+            for t in title_tokens:
+                if lower(t) in title:
+                    match = True
+                    break
+            amatch = not author_tokens
+            for a in author_tokens:
+                if lower(a) in authors:
+                    amatch = True
+                    break
+            if not author_tokens: amatch = True
+            return match and amatch
+
+        bl = feed.find('BookList')
+        if bl is None:
+            err = tostring(etree.find('errormessage'))
+            raise ValueError('ISBNDb query failed:' + err)
+        total_results = int(bl.get('total_results'))
+        shown_results = int(bl.get('shown_results'))
+        for bd in bl.xpath('.//BookData'):
+            isbn = check_isbn(bd.get('isbn13', bd.get('isbn', None)))
+            if not isbn:
+                continue
+            if orig_isbn and isbn != orig_isbn:
+                continue
+            title = tostring(bd.find('Title'))
+            if not title:
+                continue
+            authors = []
+            for au in bd.xpath('.//Authors/Person'):
+                au = tostring(au)
+                if au:
+                    if ',' in au:
+                        ln, _, fn = au.partition(',')
+                        au = fn.strip() + ' ' + ln.strip()
+                authors.append(au)
+            if not authors:
+                continue
+            comments = tostring(bd.find('Summary'))
+            if not comments:
+                # Require comments, since without them the result is useless
+                # anyway
+                continue
+            id_ = (title, tuple(authors))
+            if id_ in seen:
+                continue
+            seen.add(id_)
+            if not ismatch(title, authors):
+                continue
+            publisher = tostring(bd.find('PublisherText'))
+            if not publisher: publisher = None
+            if publisher and 'audio' in publisher.lower():
+                continue
+            mi = Metadata(title, authors)
+            mi.isbn = isbn
+            mi.publisher = publisher
+            mi.comments = comments
+            results.append(mi)
+        return total_results, shown_results, results
+
+    def make_query(self, q, abort, title=None, authors=None, identifiers={},
+            max_pages=10, timeout=30):
+        page_num = 1
+        parser = etree.XMLParser(recover=True, no_network=True)
+        br = self.browser
+
+        seen = set()
+
+        candidates = []
+        total_found = 0
+        while page_num <= max_pages and not abort.is_set():
+            url = q.replace('&page_number=1&', '&page_number=%d&'%page_num)
+            page_num += 1
+            raw = br.open_novisit(url, timeout=timeout).read()
+            feed = etree.fromstring(xml_to_unicode(clean_ascii_chars(raw),
+                strip_encoding_pats=True)[0], parser=parser)
+            total, found, results = self.parse_feed(
+                    feed, seen, title, authors, identifiers)
+            total_found += found
+            candidates += results
+            if total_found >= total or len(candidates) > 9:
+                break
+
+        return candidates
+    # }}}
+
+if __name__ == '__main__':
+    # To run these test use:
+    # calibre-debug -e src/calibre/ebooks/metadata/sources/isbndb.py
+    from calibre.ebooks.metadata.sources.test import (test_identify_plugin,
+            title_test, authors_test)
+    test_identify_plugin(ISBNDB.name,
+        [
+
+
+            (
+                {'title':'Great Gatsby',
+                    'authors':['Fitzgerald']},
+                [title_test('The great gatsby', exact=True),
+                    authors_test(['F. Scott Fitzgerald'])]
+            ),
+
+            (
+                {'title': 'Flatland', 'authors':['Abbott']},
+                [title_test('Flatland', exact=False)]
+            ),
+    ])
 
diff --git a/src/calibre/ebooks/mobi/writer.py b/src/calibre/ebooks/mobi/writer.py
index 5f4c47cdf3..fc47b26c02 100644
--- a/src/calibre/ebooks/mobi/writer.py
+++ b/src/calibre/ebooks/mobi/writer.py
@@ -310,10 +310,11 @@ class Serializer(object):
             if href not in id_offsets:
                 self.logger.warn('Hyperlink target %r not found' % href)
                 href, _ = urldefrag(href)
-            ioff = self.id_offsets[href]
-            for hoff in hoffs:
-                buffer.seek(hoff)
-                buffer.write('%010d' % ioff)
+            if href in self.id_offsets:
+                ioff = self.id_offsets[href]
+                for hoff in hoffs:
+                    buffer.seek(hoff)
+                    buffer.write('%010d' % ioff)
 
 class MobiWriter(object):
     COLLAPSE_RE = re.compile(r'[ \t\r\n\v]+')
diff --git a/src/calibre/gui2/__init__.py b/src/calibre/gui2/__init__.py
index 22aaabf592..e39427021e 100644
--- a/src/calibre/gui2/__init__.py
+++ b/src/calibre/gui2/__init__.py
@@ -357,6 +357,7 @@ class FileIconProvider(QFileIconProvider):
              'bmp'     : 'bmp',
              'svg'     : 'svg',
              'html'    : 'html',
+             'htmlz'   : 'html',
              'htm'     : 'html',
              'xhtml'   : 'html',
              'xhtm'    : 'html',
diff --git a/src/calibre/gui2/actions/edit_metadata.py b/src/calibre/gui2/actions/edit_metadata.py
index 9f2cacb177..18a73fb282 100644
--- a/src/calibre/gui2/actions/edit_metadata.py
+++ b/src/calibre/gui2/actions/edit_metadata.py
@@ -94,7 +94,7 @@ class EditMetadataAction(InterfaceAction):
 
     def bulk_metadata_downloaded(self, job):
         if job.failed:
-            self.job_exception(job, dialog_title=_('Failed to download metadata'))
+            self.gui.job_exception(job, dialog_title=_('Failed to download metadata'))
             return
         from calibre.gui2.metadata.bulk_download2 import proceed
         proceed(self.gui, job)
diff --git a/src/calibre/gui2/metadata/bulk_download2.py b/src/calibre/gui2/metadata/bulk_download2.py
index 19cd3df9d4..5f0af1b316 100644
--- a/src/calibre/gui2/metadata/bulk_download2.py
+++ b/src/calibre/gui2/metadata/bulk_download2.py
@@ -54,6 +54,8 @@ def start_download(gui, ids, callback, identify, covers):
             _('Download metadata for %d books')%len(ids),
             download, (ids, gui.current_db, identify, covers), {}, callback)
     gui.job_manager.run_threaded_job(job)
+    gui.status_bar.show_message(_('Metadata download started'), 3000)
+
 
 class ViewLog(QDialog): # {{{
 
@@ -75,7 +77,7 @@ class ViewLog(QDialog): # {{{
         self.copy_button.clicked.connect(self.copy_to_clipboard)
         l.addWidget(self.bb)
         self.setModal(False)
-        self.resize(QSize(500, 400))
+        self.resize(QSize(700, 500))
         self.setWindowTitle(_('Download log'))
         self.setWindowIcon(QIcon(I('debug.png')))
         self.show()
@@ -110,25 +112,27 @@ class ApplyDialog(QDialog):
         self.bb.accepted.connect(self.accept)
         l.addWidget(self.bb)
 
-        self.db = gui.current_db
+        self.gui = gui
         self.id_map = list(id_map.iteritems())
         self.current_idx = 0
 
         self.failures = []
+        self.ids = []
         self.canceled = False
 
         QTimer.singleShot(20, self.do_one)
-        self.exec_()
 
     def do_one(self):
         if self.canceled:
             return
         i, mi = self.id_map[self.current_idx]
+        db = self.gui.current_db
         try:
             set_title = not mi.is_null('title')
             set_authors = not mi.is_null('authors')
-            self.db.set_metadata(i, mi, commit=False, set_title=set_title,
+            db.set_metadata(i, mi, commit=False, set_title=set_title,
                     set_authors=set_authors)
+            self.ids.append(i)
         except:
             import traceback
             self.failures.append((i, traceback.format_exc()))
@@ -156,9 +160,10 @@ class ApplyDialog(QDialog):
             return
         if self.failures:
             msg = []
+            db = self.gui.current_db
             for i, tb in self.failures:
-                title = self.db.title(i, index_is_id=True)
-                authors = self.db.authors(i, index_is_id=True)
+                title = db.title(i, index_is_id=True)
+                authors = db.authors(i, index_is_id=True)
                 if authors:
                     authors = [x.replace('|', ',') for x in authors.split(',')]
                     title += ' - ' + authors_to_string(authors)
@@ -169,6 +174,12 @@ class ApplyDialog(QDialog):
                     ' in your library. Click "Show Details" to see '
                     'details.'), det_msg='\n\n'.join(msg), show=True)
         self.accept()
+        if self.ids:
+            cr = self.gui.library_view.currentIndex().row()
+            self.gui.library_view.model().refresh_ids(
+                self.ids, cr)
+            if self.gui.cover_flow:
+                self.gui.cover_flow.dataChanged()
 
 _amd = None
 def apply_metadata(job, gui, q, result):
@@ -177,7 +188,7 @@ def apply_metadata(job, gui, q, result):
     q.finished.disconnect()
     if result != q.Accepted:
         return
-    id_map, failed_ids = job.result
+    id_map, failed_ids, failed_covers, title_map = job.result
     id_map = dict([(k, v) for k, v in id_map.iteritems() if k not in
         failed_ids])
     if not id_map:
@@ -207,23 +218,32 @@ def apply_metadata(job, gui, q, result):
             return
 
     _amd = ApplyDialog(id_map, gui)
+    _amd.exec_()
 
 def proceed(gui, job):
-    id_map, failed_ids = job.result
+    gui.status_bar.show_message(_('Metadata download completed'), 3000)
+    id_map, failed_ids, failed_covers, title_map = job.result
     fmsg = det_msg = ''
-    if failed_ids:
-        fmsg = _('Could not download metadata for %d of the books. Click'
+    if failed_ids or failed_covers:
+        fmsg = '<p>'+_('Could not download metadata and/or covers for %d of the books. Click'
                 ' "Show details" to see which books.')%len(failed_ids)
-        det_msg = '\n'.join([id_map[i].title for i in failed_ids])
+        det_msg = []
+        for i in failed_ids | failed_covers:
+            title = title_map[i]
+            if i in failed_ids:
+                title += (' ' + _('(Failed metadata)'))
+            if i in failed_covers:
+                title += (' ' + _('(Failed cover)'))
+            det_msg.append(title)
     msg = '<p>' + _('Finished downloading metadata for <b>%d book(s)</b>. '
         'Proceed with updating the metadata in your library?')%len(id_map)
     q = MessageBox(MessageBox.QUESTION, _('Download complete'),
-            msg + fmsg, det_msg=det_msg, show_copy_button=bool(failed_ids),
+            msg + fmsg, det_msg='\n'.join(det_msg), show_copy_button=bool(failed_ids),
             parent=gui)
     q.vlb = q.bb.addButton(_('View log'), q.bb.ActionRole)
     q.vlb.setIcon(QIcon(I('debug.png')))
     q.vlb.clicked.connect(partial(view_log, job, q))
-    q.det_msg_toggle.setVisible(bool(failed_ids))
+    q.det_msg_toggle.setVisible(bool(failed_ids | failed_covers))
     q.setModal(False)
     q.show()
     q.finished.connect(partial(apply_metadata, job, gui, q))
@@ -242,12 +262,18 @@ def merge_result(oldmi, newmi):
             if (not newmi.is_null(f) and getattr(newmi, f) == getattr(oldmi, f)):
                 setattr(newmi, f, getattr(dummy, f))
 
+    newmi.last_modified = oldmi.last_modified
+
+    return newmi
+
 def download(ids, db, do_identify, covers,
         log=None, abort=None, notifications=None):
     ids = list(ids)
     metadata = [db.get_metadata(i, index_is_id=True, get_user_categories=False)
         for i in ids]
     failed_ids = set()
+    failed_covers = set()
+    title_map = {}
     ans = {}
     count = 0
     for i, mi in izip(ids, metadata):
@@ -255,6 +281,7 @@ def download(ids, db, do_identify, covers,
             log.error('Aborting...')
             break
         title, authors, identifiers = mi.title, mi.authors, mi.identifiers
+        title_map[i] = title
         if do_identify:
             results = []
             try:
@@ -265,22 +292,29 @@ def download(ids, db, do_identify, covers,
             if results:
                 mi = merge_result(mi, results[0])
                 identifiers = mi.identifiers
+                if not mi.is_null('rating'):
+                    # set_metadata expects a rating out of 10
+                    mi.rating *= 2
             else:
                 log.error('Failed to download metadata for', title)
-                failed_ids.add(mi)
+                failed_ids.add(i)
+                # We don't want set_metadata operating on anything but covers
+                mi = merge_result(mi, mi)
         if covers:
             cdata = download_cover(log, title=title, authors=authors,
                     identifiers=identifiers)
-            if cdata:
+            if cdata is not None:
                 with PersistentTemporaryFile('.jpg', 'downloaded-cover-') as f:
-                    f.write(cdata)
+                    f.write(cdata[-1])
                     mi.cover = f.name
+            else:
+                failed_covers.add(i)
         ans[i] = mi
         count += 1
         notifications.put((count/len(ids),
             _('Downloaded %d of %d')%(count, len(ids))))
     log('Download complete, with %d failures'%len(failed_ids))
-    return (ans, failed_ids)
+    return (ans, failed_ids, failed_covers, title_map)
 
 
 
diff --git a/src/calibre/gui2/metadata/config.py b/src/calibre/gui2/metadata/config.py
index 68c935061d..abb45faa46 100644
--- a/src/calibre/gui2/metadata/config.py
+++ b/src/calibre/gui2/metadata/config.py
@@ -56,7 +56,12 @@ class ConfigWidget(QWidget):
         self.setLayout(l)
 
         self.gb = QGroupBox(_('Downloaded metadata fields'), self)
-        l.addWidget(self.gb, 0, 0, 1, 2)
+        if plugin.config_help_message:
+            self.pchm = QLabel(plugin.config_help_message)
+            self.pchm.setWordWrap(True)
+            self.pchm.setOpenExternalLinks(True)
+            l.addWidget(self.pchm, 0, 0, 1, 2)
+        l.addWidget(self.gb, l.rowCount(), 0, 1, 2)
         self.gb.l = QGridLayout()
         self.gb.setLayout(self.gb.l)
         self.fields_view = v = QListView(self)
@@ -81,7 +86,7 @@ class ConfigWidget(QWidget):
             widget.setValue(val)
         elif opt.type == 'string':
             widget = QLineEdit(self)
-            widget.setText(val)
+            widget.setText(val if val else '')
         elif opt.type == 'bool':
             widget = QCheckBox(opt.label, self)
             widget.setChecked(bool(val))
diff --git a/src/calibre/gui2/preferences/metadata_sources.py b/src/calibre/gui2/preferences/metadata_sources.py
index 4500a03b30..17a70bcc33 100644
--- a/src/calibre/gui2/preferences/metadata_sources.py
+++ b/src/calibre/gui2/preferences/metadata_sources.py
@@ -10,7 +10,7 @@ __docformat__ = 'restructuredtext en'
 from operator import attrgetter
 
 from PyQt4.Qt import (QAbstractTableModel, Qt, QAbstractListModel, QWidget,
-        pyqtSignal, QVBoxLayout, QDialogButtonBox, QFrame, QLabel)
+        pyqtSignal, QVBoxLayout, QDialogButtonBox, QFrame, QLabel, QIcon)
 
 from calibre.gui2.preferences import ConfigWidgetBase, test_widget
 from calibre.gui2.preferences.metadata_sources_ui import Ui_Form
@@ -67,6 +67,13 @@ class SourcesModel(QAbstractTableModel): # {{{
             return self.enabled_overrides.get(plugin, orig)
         elif role == Qt.UserRole:
             return plugin
+        elif (role == Qt.DecorationRole and col == 0 and not
+                    plugin.is_configured()):
+            return QIcon(I('list_remove.png'))
+        elif role == Qt.ToolTipRole:
+            if plugin.is_configured():
+                return _('This source is configured and ready to go')
+            return _('This source needs configuration')
         return NONE
 
     def setData(self, index, val, role):
diff --git a/src/calibre/gui2/preferences/metadata_sources.ui b/src/calibre/gui2/preferences/metadata_sources.ui
index 546120f628..b515f13ba1 100644
--- a/src/calibre/gui2/preferences/metadata_sources.ui
+++ b/src/calibre/gui2/preferences/metadata_sources.ui
@@ -48,6 +48,16 @@
             </property>
            </widget>
           </item>
+          <item>
+           <widget class="QLabel" name="label_5">
+            <property name="text">
+             <string>Sources with a red X next to their names must be configured before they will be used. </string>
+            </property>
+            <property name="wordWrap">
+             <bool>true</bool>
+            </property>
+           </widget>
+          </item>
           <item>
            <widget class="QPushButton" name="configure_plugin_button">
             <property name="text">
diff --git a/src/calibre/gui2/threaded_jobs.py b/src/calibre/gui2/threaded_jobs.py
index f98488da79..9c791c5b0d 100644
--- a/src/calibre/gui2/threaded_jobs.py
+++ b/src/calibre/gui2/threaded_jobs.py
@@ -189,7 +189,11 @@ class ThreadedJobServer(Thread):
 
     def run(self):
         while self.keep_going:
-            self.run_once()
+            try:
+                self.run_once()
+            except:
+                import traceback
+                traceback.print_exc()
             time.sleep(0.1)
 
     def run_once(self):
diff --git a/src/calibre/manual/server.rst b/src/calibre/manual/server.rst
index 82ec5c2927..aa98ba57df 100644
--- a/src/calibre/manual/server.rst
+++ b/src/calibre/manual/server.rst
@@ -22,6 +22,8 @@ First start the |app| content server as shown below::
 
     calibre-server --url-prefix /calibre --port 8080 
 
+The key parameter here is ``--url-prefix /calibre``. This causes the content server to serve all URLs prefixed by calibre. To see this in action, visit ``http://localhost:8080/calibre`` in your browser. You should see the normal content server website, but now it will run under /calibre.
+
 Now suppose you are using Apache as your main server. First enable the proxy modules in apache, by adding the following to :file:`httpd.conf`::
 
     LoadModule proxy_module modules/mod_proxy.so