Fix #5666 (coverter UnicodeDecodeError: 'utf8' codec can't decode byte 0xff in position 57: unexpected code byte)

src/calibre/ebooks/html/input.py

@@ -416,9 +416,9 @@ class HTMLInput(InputFormatPlugin):
         link = unquote(link).replace('/', os.sep)
         if not link.strip():
             return link_
-        if base and not os.path.isabs(link):
-            link = os.path.join(base, link)
         try:
+            if base and not os.path.isabs(link):
+                link = os.path.join(base, link)
             link = os.path.abspath(link)
         except:
             return link_

src/calibre/ebooks/metadata/html.py

@@ -11,7 +11,7 @@ import re
 
 from calibre.ebooks.metadata import MetaInformation
 from calibre.ebooks.chardet import xml_to_unicode
-
+from calibre import entity_to_unicode
 
 def get_metadata(stream):
     src = stream.read()
@@ -43,6 +43,10 @@ def get_metadata_(src, encoding=None):
     if match:
         author = match.group(2).replace(',', ';')
 
+    ent_pat = re.compile(r'&(\S+)?;')
+    title = ent_pat.sub(entity_to_unicode, title)
+    if author:
+        author = ent_pat.sub(entity_to_unicode, author)
     mi = MetaInformation(title, [author] if author else None)
 
     # Publisher

src/calibre/utils/magick_draw.py

@@ -51,6 +51,8 @@ class FontMetrics(object):
 
 
 def get_font_metrics(image, d_wand, text):
+    if isinstance(text, unicode):
+        text = text.encode('utf-8')
     ret = p.MagickQueryFontMetrics(image, d_wand, text)
     return FontMetrics(ret)