Fix a bug: html_tree may have no childNodes

[bookmarks_db.git] / parse_html / bkmk_ph_html5.py

diff --git a/parse_html/bkmk_ph_html5.py b/parse_html/bkmk_ph_html5.py
index a4906288af13fe5da0cf3f24f39e30a010f7cbe7..6d0d38094bea6bd8935b4ddb29455ca458cd7cbf 100644 (file)
--- a/parse_html/bkmk_ph_html5.py
+++ b/parse_html/bkmk_ph_html5.py
@@ -1,10 +1,11 @@
 """HTML Parser using html5
 
 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) 2010-2014 PhiloSoft Design"
 __license__ = "GNU GPL"
 
 __all__ = ['parse_html']
@@ -14,19 +15,17 @@ from html5lib import HTMLParser as HTML5Parser
 from .bkmk_ph_util import HTMLParser
 
 
-def parse_html(filename, charset=None, log=None):
+def parse_html(html_text, charset=None, log=None):
     parser = HTML5Parser()
-    fp = open(filename)
-    parser._parse(fp, encoding=charset, parseMeta=bool(charset))
-    fp.close()
-    html_tree = parser.tree.getDocument()
-
-    for node in html_tree.childNodes:
-        if (node.name == 'html') and (node.type != 3): # Skip DocType element
-            html = node
-            break
-    else:
-        html = None
+    html_tree = parser.parse(html_text, encoding=charset, parseMeta=bool(charset))
+
+    html = None
+    if hasattr(html_tree, 'childNodes'):
+        for node in html_tree.childNodes:
+            # Skip DocType element
+            if (node.name == 'html') and (node.type != 3):
+                html = node
+                break
 
     if not html:
         return None
@@ -53,7 +52,7 @@ def parse_html(filename, charset=None, log=None):
                     title = ''
 
         for node in head.childNodes:
-            if node.name == 'meta' and \
+            if (node.name == 'meta') and \
                     ('http-equiv' in node.attributes) and \
                     (node.attributes['http-equiv'] == 'content-type'):
                 meta_content = node.attributes['content']
@@ -65,6 +64,9 @@ def parse_html(filename, charset=None, log=None):
                         meta_charset = False
                     else:
                         break
+            elif (node.name == 'meta') and ('charset' in node.attributes):
+                meta_charset = node.attributes['charset'].lower()
+                break
 
         if not charset:
             charset = parser.tokenizer.stream.charEncoding[0]