Parse <meta charset="...">

[bookmarks_db.git] / parse_html / bkmk_ph_lxml.py

diff --git a/parse_html/bkmk_ph_lxml.py b/parse_html/bkmk_ph_lxml.py
index b14be408571ba53e0cfa5f60997394658ad2061b..222f11697340f32381f846ab55cf753ffa15444a 100644 (file)
--- a/parse_html/bkmk_ph_lxml.py
+++ b/parse_html/bkmk_ph_lxml.py
@@ -4,7 +4,7 @@ This file is a part of Bookmarks database and Internet robot.
 """
 
 __author__ = "Oleg Broytman <phd@phdru.name>"
-__copyright__ = "Copyright (C) 2010-2012 PhiloSoft Design"
+__copyright__ = "Copyright (C) 2010-2013 PhiloSoft Design"
 __license__ = "GNU GPL"
 
 __all__ = ['parse_html']
@@ -35,6 +35,9 @@ def parse_html(filename, charset=None, log=None):
                     break
                 except IndexError:
                     meta_charset = False
+        elif m.get('charset', ''):
+           meta_charset = m.get('charset').lower()
+           break
     else:
         meta_charset = False