def parse(self) :
if (isinstance(self.source,unicode)) :
# Create a string source
file = io.StringIO(self.source)
input = InputSource(file)
input.setEncoding("utf-8")
input.setCharacterStream(file)
# There is a bug in xml.sax.saxutils.prepare_input_source
input.setByteStream(file)
input.setSystemId(None)
elif (isinstance(self.source,InputSource)):
input = self.source
else:
raise Exception("Parse source must be either string or InputSource")
# Create the parser/xmlreader
parser = xml.sax.make_parser()
# Tell the parser to use our handler(s)
parser.setContentHandler(self)
#parser.setErrorHandler(self)
#parser.setFeature(xml.sax.handler.feature_namespaces,True)
# Shut off dtd validation
parser.setFeature(xml.sax.handler.feature_validation,False)
parser.setFeature(xml.sax.handler.feature_external_ges, False)
# Parse the document
parser.parse(input)
def test_expat_inpsource_location():
parser = create_parser()
parser.setContentHandler(ContentHandler()) # do nothing
source = InputSource()
source.setByteStream(StringIO("<foo bar foobar>")) #ill-formed
name = "a file name"
source.setSystemId(name)
try:
parser.parse(source)
except SAXException, e:
return e.getSystemId() == name
def test_expat_inpsource_location():
parser = make_parser()
parser.setContentHandler(ContentHandler()) # do nothing
source = InputSource()
source.setByteStream(StringIO("<foo bar foobar>")) #ill-formed
name = "a file name"
source.setSystemId(name)
try:
parser.parse(source)
except SAXException, e:
return e.getSystemId() == name
def test_expat_inpsource_location(self):
parser = create_parser()
parser.setContentHandler(ContentHandler()) # do nothing
source = InputSource()
source.setByteStream(BytesIO(b"<foo bar foobar>")) #ill-formed
name = "a file name"
source.setSystemId(name)
try:
parser.parse(source)
self.fail()
except SAXException as e:
self.assertEqual(e.getSystemId(), name)
def test_expat_inpsource_location(self):
parser = create_parser()
parser.setContentHandler(ContentHandler()) # do nothing
source = InputSource()
source.setByteStream(StringIO("<foo bar foobar>")) #ill-formed
name = "a file name"
source.setSystemId(name)
try:
parser.parse(source)
self.fail()
except SAXException as e:
self.assertEqual(e.getSystemId(), name)
def resolveEntity(self, publicId, systemId):
source = InputSource()
source.setSystemId(systemId)
try:
dtdPath = self.knownDTDs[systemId]
except KeyError:
raise process.ProcessingFailure(
"Invalid DTD system identifier (%r) in %s. Only "
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd "
"is allowed." % (systemId, self.filename))
source.setByteStream(dtdPath.open())
return source
def resolveEntity(self, publicId, systemId):
source = InputSource()
source.setSystemId(systemId)
try:
dtdPath = self.knownDTDs[systemId]
except KeyError:
raise process.ProcessingFailure(
"Invalid DTD system identifier (%r) in %s. Only "
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd "
"is allowed." % (systemId, self.filename))
source.setByteStream(dtdPath.open())
return source
def test_expat_inpsource_location(self):
parser = create_parser()
parser.setContentHandler(ContentHandler())
source = InputSource()
source.setByteStream(BytesIO(b'<foo bar foobar>'))
name = 'a file name'
source.setSystemId(name)
try:
parser.parse(source)
self.fail()
except SAXException as e:
self.assertEqual(e.getSystemId(), name)
def caching_prepare_input_source(source, base=None):
if isinstance(source, InputSource):
return source
full_uri = urljoin(base or "", source)
# Convert StringIO to string
if hasattr(full_uri, "getvalue"):
full_uri = full_uri.getvalue()
if not full_uri.startswith("http:"):
args = (source,) if base is None else (source, base)
return old_prepare_input_source(*args)
document = urlopen(full_uri, proxies=proxies)
input_source = InputSource()
input_source.setSystemId(source)
input_source.setByteStream(document)
return input_source
def resolveEntity(self, publicId, systemId):
self.entities.append((publicId, systemId))
source = InputSource()
source.setPublicId(publicId)
source.setSystemId(systemId)
return source
def resolveEntity(self, publicId, systemId):
self.entities.append((publicId, systemId))
source = InputSource()
source.setPublicId(publicId)
source.setSystemId(systemId)
return source
# regression test for SAX 2.0
