def test_etree_parse_io_error(self):
# this is a directory name that contains characters beyond latin-1
dirnameEN = _str('Directory')
dirnameRU = _str('Каталог')
filename = _str('nosuchfile.xml')
dn = tempfile.mkdtemp(prefix=dirnameEN)
try:
self.assertRaises(IOError, self.etree.parse,
os.path.join(dn, filename))
finally:
os.rmdir(dn)
dn = tempfile.mkdtemp(prefix=dirnameRU)
try:
self.assertRaises(IOError, self.etree.parse,
os.path.join(dn, filename))
finally:
os.rmdir(dn)
def test_parse_utf8_bom(self):
utext = _str('Søk på nettet')
uxml = '<?xml version="1.0" encoding="UTF-8"?><p>%s</p>' % utext
bom = _bytes('\\xEF\\xBB\\xBF').decode("unicode_escape").encode(
"latin1")
self.assertEqual(3, len(bom))
f = tempfile.NamedTemporaryFile(delete=False)
try:
try:
f.write(bom)
f.write(uxml.encode("utf-8"))
finally:
f.close()
tree = self.etree.parse(f.name)
finally:
os.unlink(f.name)
self.assertEqual(utext, tree.getroot().text)
def test_tree_io_latin1(self):
Element = self.etree.Element
ElementTree = self.etree.ElementTree
element = Element('top')
element.text = _str("qwrtioüöäßá")
tree = ElementTree(element)
self.buildNodes(element, 10, 3)
with open(self.getTestFilePath('testdump.xml'), 'wb') as f:
tree.write(f, encoding='iso-8859-1')
with open(self.getTestFilePath('testdump.xml'), 'rb') as f:
tree = ElementTree(file=f)
with open(self.getTestFilePath('testdump2.xml'), 'wb') as f:
tree.write(f, encoding='iso-8859-1')
with open(self.getTestFilePath('testdump.xml'), 'rb') as f:
data1 = f.read()
with open(self.getTestFilePath('testdump2.xml'), 'rb') as f:
data2 = f.read()
self.assertEqual(data1, data2)
def test_iterparse_utf16_bom(self):
utext = _str('Søk på nettet')
uxml = '<?xml version="1.0" encoding="UTF-16"?><p>%s</p>' % utext
boms = _bytes('\\xFE\\xFF \\xFF\\xFE').decode("unicode_escape").encode(
"latin1")
self.assertEqual(5, len(boms))
xml = uxml.encode("utf-16")
self.assertTrue(xml[:2] in boms, repr(xml[:2]))
f = tempfile.NamedTemporaryFile(delete=False)
try:
try:
f.write(xml)
finally:
f.close()
elements = [el for _, el in self.etree.iterparse(f.name)]
self.assertEqual(1, len(elements))
root = elements[0]
finally:
os.unlink(f.name)
self.assertEqual(utext, root.text)
def test_write_compressed_text(self):
Element = self.etree.Element
SubElement = self.etree.SubElement
ElementTree = self.etree.ElementTree
text = _str("qwrtioüöä")
root = Element('root')
root.text = text
child = SubElement(root, 'sub')
child.text = 'TEXT'
child.tail = 'TAIL'
SubElement(root, 'sub').text = text
tree = ElementTree(root)
out = BytesIO()
tree.write(out, method='text', encoding='utf8', compression=9)
out.seek(0)
f = gzip.GzipFile(fileobj=out)
try:
result = f.read().decode('utf8')
finally:
f.close()
self.assertEqual(text + 'TEXTTAIL' + text, result)