# HG changeset patch
# User paulb
# Date 1173730818 0
# Node ID e631f2826b8ca191b973bc32be83d53c495ea2ab
# Parent  f7db5375fa9b3d69cc7bca3aa328b7d8e853f618
[project @ 2007-03-12 20:20:18 by paulb]
Fixed getElementsByTagName(NS).
Made parseURI work for HTML documents.

diff -r f7db5375fa9b -r e631f2826b8c libxml2dom/__init__.py
--- a/libxml2dom/__init__.py	Mon Mar 12 20:20:01 2007 +0000
+++ b/libxml2dom/__init__.py	Mon Mar 12 20:20:18 2007 +0000
@@ -29,6 +29,7 @@
     parseFile as Node_parseFile, \
     toString as Node_toString, toStream as Node_toStream, \
     toFile as Node_toFile
+import urllib # for parseURI in HTML mode
 
 class Implementation(object):
 
@@ -345,10 +346,10 @@
             Node_removeChild(self._node, tmp)
 
     def getElementsByTagName(self, tagName):
-        return self.xpath("//" + tagName)
+        return self.xpath(".//" + tagName)
 
     def getElementsByTagNameNS(self, namespaceURI, localName):
-        return self.xpath("//ns:" + localName, namespaces={"ns" : namespaceURI})
+        return self.xpath(".//ns:" + localName, namespaces={"ns" : namespaceURI})
 
     def normalize(self):
         text_nodes = []
@@ -540,16 +541,29 @@
     is set to a true value, the content to be parsed will be treated as being
     HTML rather than XML.
 
-    The parseURI does not currently work with HTML. Use parse with a stream
-    object instead. For example:
+    XML documents are retrieved using libxml2's own network capabilities; HTML
+    documents are retrieved using the urllib module provided by Python. To
+    retrieve either kind of document using Python's own modules for this purpose
+    (such as urllib), open a stream and pass it to the parse function:
 
-    d = parse(urllib.urlopen("http://www.python.org"), html=1)
+    f = urllib.urlopen(uri)
+    try:
+        doc = libxml2dom.parse(f, html)
+    finally:
+        f.close()
 
     A document object is returned by this function.
     """
 
-    impl = impl or default_impl
-    return Document(Node_parseURI(uri, html), impl)
+    if html:
+        f = urllib.urlopen(uri)
+        try:
+            return parse(f, html, impl)
+        finally:
+            f.close()
+    else:
+        impl = impl or default_impl
+        return Document(Node_parseURI(uri, html), impl)
 
 def toString(node, encoding=None, prettyprint=0):