Fix various regressions

src/calibre/ebooks/mobi/palmdoc.c

@@ -118,9 +118,9 @@ cpalmdoc_do_compress(buffer *b, char *output) {
             found = false;
             for (chunk_len = 10; chunk_len > 2; chunk_len--) {
                 j = cpalmdoc_rfind(b->data, i, chunk_len);
-                if (j < i) {
+                dist = i - j;
+                if (j < i && dist <= 2047) {
                     found = true;
-                    dist = i - j;
                     compound = (dist << 3) + chunk_len-3;
                     *(output++) = CHAR(0x80 + (compound >> 8 ));
                     *(output++) = CHAR(compound & 0xFF);

src/calibre/ebooks/mobi/reader.py

@@ -226,7 +226,7 @@ class MobiReader(object):
                     page-break-after: always; margin: 0; display: block
                 }
                 ''')
-        self.tag_css_rules = []
+        self.tag_css_rules = {}
 
         if hasattr(filename_or_stream, 'read'):
             stream = filename_or_stream
@@ -328,10 +328,10 @@ class MobiReader(object):
 
         with open('styles.css', 'wb') as s:
             s.write(self.base_css_rules+'\n\n')
-            for rule in self.tag_css_rules:
+            for cls, rule in self.tag_css_rules.items():
                 if isinstance(rule, unicode):
                     rule = rule.encode('utf-8')
-                s.write(rule+'\n\n')
+                s.write('.%s { %s }\n\n'%(cls, rule))
 
 
         if self.book_header.exth is not None or self.embedded_mi is not None:
@@ -389,6 +389,7 @@ class MobiReader(object):
                     'xx-large' : '6',
                     }
         mobi_version = self.book_header.mobi_version
+        style_map = {}
         for i, tag in enumerate(root.iter(etree.Element)):
             if tag.tag in ('country-region', 'place', 'placetype', 'placename',
                            'state', 'city', 'street', 'address', 'content'):
@@ -455,9 +456,18 @@ class MobiReader(object):
                 except ValueError:
                     pass
             if styles:
-                attrib['id'] = attrib.get('id', 'calibre_mr_gid%d'%i)
+                cls = None
-                self.tag_css_rules.append('#%s {%s}'%(attrib['id'],
+                rule = '; '.join(styles)
-                                                      '; '.join(styles)))
+                for sel, srule in self.tag_css_rules.items():
+                    if srule == rule:
+                        cls = sel
+                        break
+                if cls is None:
+                    ncls = 'calibre_%d'%i
+                    self.tag_css_rules[ncls] = rule
+                cls = attrib.get('class', '')
+                cls = cls + (' ' if cls else '') + ncls
+                attrib['class'] = cls
 
     def create_opf(self, htmlfile, guide=None, root=None):
         mi = getattr(self.book_header.exth, 'mi', self.embedded_mi)

src/calibre/ebooks/oeb/base.py

@@ -726,6 +726,7 @@ class Manifest(object):
                 % (self.id, self.href, self.media_type)
 
         def _parse_xhtml(self, data):
+            self.oeb.log.debug('Parsing', self.href, '...')
             # Convert to Unicode and normalize line endings
             data = self.oeb.decode(data)
             data = self.oeb.html_preprocessor(data)
@@ -804,6 +805,7 @@ class Manifest(object):
             return data
 
         def _parse_css(self, data):
+            self.oeb.log.debug('Parsing', self.href, '...')
             data = self.oeb.decode(data)
             data = self.oeb.css_preprocessor(data)
             data = XHTML_CSS_NAMESPACE + data

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

@@ -41,10 +41,12 @@ class Clean(object):
 
         for x in list(self.oeb.guide):
             href = urldefrag(self.oeb.guide[x].href)[0]
-            if x.lower() != ('cover', 'titlepage'):
+            if x.lower() not in ('cover', 'titlepage'):
                 try:
                     if href not in protected_hrefs:
-                        self.oeb.manifest.remove(self.oeb.manifest.hrefs[href])
+                        item = self.oeb.manifest.hrefs[href]
+                        if item not in self.oeb.spine:
+                            self.oeb.manifest.remove(self.oeb.manifest.hrefs[href])
                 except KeyError:
                     pass
                 self.oeb.guide.remove(x)

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

@@ -44,14 +44,14 @@ class Split(object):
         self.split_on_page_breaks = split_on_page_breaks
         self.page_breaks_xpath = page_breaks_xpath
         self.max_flow_size = max_flow_size
+        self.page_break_selectors = None
         if self.page_breaks_xpath is not None:
-            self.page_breaks_xpath = XPath(self.page_breaks_xpath)
+            self.page_break_selectors = [(XPath(self.page_breaks_xpath), False)]
 
     def __call__(self, oeb, context):
         self.oeb = oeb
         self.log = oeb.log
         self.map = {}
-        self.page_break_selectors = None
         for item in list(self.oeb.manifest.items):
             if item.spine_position is not None and etree.iselement(item.data):
                 self.split_item(item)
@@ -60,10 +60,7 @@ class Split(object):
 
     def split_item(self, item):
         if self.split_on_page_breaks:
-            if self.page_breaks_xpath is None:
+            page_breaks, page_break_ids = self.find_page_breaks(item)
-                page_breaks, page_break_ids = self.find_page_breaks(item)
-            else:
-                page_breaks, page_break_ids = self.page_breaks_xpath(item.data)
 
         splitter = FlowSplitter(item, page_breaks, page_break_ids,
                 self.max_flow_size, self.oeb)