Fix #355

src/libprs500/ebooks/lrf/__init__.py

@@ -313,8 +313,10 @@ def Book(options, logger, font_delta=0, header=None,
             raise ConversionError, 'Could not find the normal version of the ' + family + ' font'
     return book, fonts
 
-def entity_to_unicode(match):
+def entity_to_unicode(match, exceptions=[]):
     ent = match.group(1)
+    if ent in exceptions:
+        return '&'+ent+';'
     if ent.startswith(u'#x'):
         return unichr(int(ent[2:], 16))
     if ent.startswith(u'#'):
@@ -322,4 +324,4 @@ def entity_to_unicode(match):
     try:
         return unichr(name2codepoint[ent])
     except KeyError:
-        return ent
+        return '&'+ent+';'

src/libprs500/ebooks/lrf/html/convert_from.py

@@ -25,6 +25,8 @@ from collections import deque
 from urllib import unquote
 from urlparse import urlparse
 from math import ceil, floor
+from functools import partial
+
 try:
     from PIL import Image as PILImage
 except ImportError:
@@ -63,7 +65,6 @@ def munge_paths(basepath, url):
         path = os.path.join(os.path.dirname(basepath), path)
     return os.path.normpath(path), fragment
 
-                              
 class HTMLConverter(object):
     SELECTOR_PAT   = re.compile(r"([A-Za-z0-9\-\_\:\.]+[A-Za-z0-9\-\_\:\.\s\,]*)\s*\{([^\}]*)\}")
     PAGE_BREAK_PAT = re.compile(r'page-break-(?:after|before)\s*:\s*(\w+)', re.IGNORECASE)
@@ -84,7 +85,7 @@ class HTMLConverter(object):
                         # Workaround bug in BeautifulSoup   handling
                         (re.compile(u' | | |\xa0', re.IGNORECASE), lambda match : u'\uffff'),
                         # Replace entities
-                        (re.compile(ur'&(\S+?);'), entity_to_unicode),
+                        (re.compile(ur'&(\S+?);'), partial(entity_to_unicode, exceptions=['lt', 'gt'])),
                         ]
     # Fix Baen markup
     BAEN = [