Skip to content

Commit

Permalink
pythongh-91810: Expand ElementTree.write() tests to use non-ASCII data (
Browse files Browse the repository at this point in the history 
pythonGH-91989)
  • Loading branch information
@serhiy-storchaka
serhiy-storchaka authored Apr 27, 2022
1 parent ad9f817 commit f60b4c3
Showing 1 changed file with 80 additions and 17 deletions.
97 changes: 80 additions & 17 deletions Lib/test/test_xml_etree.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -130,6 +130,9 @@ def newtest(*args, **kwargs):
return newtest
return decorator

def convlinesep(data):
return data.replace(b'\n', os.linesep.encode())


class ModuleTest(unittest.TestCase):
def test_sanity(self):
Expand Down Expand Up @@ -3713,48 +3716,108 @@ def test_encoding(self):

def test_write_to_filename(self):
self.addCleanup(os_helper.unlink, TESTFN)
tree = ET.ElementTree(ET.XML('''<site />'''))
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
tree.write(TESTFN)
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(), b'''<site />''')
self.assertEqual(f.read(), b'''<site>&#248;</site>''')

def test_write_to_filename_with_encoding(self):
self.addCleanup(os_helper.unlink, TESTFN)
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
tree.write(TESTFN, encoding='utf-8')
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(), b'''<site>\xc3\xb8</site>''')

tree.write(TESTFN, encoding='ISO-8859-1')
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(), convlinesep(
b'''<?xml version='1.0' encoding='ISO-8859-1'?>\n'''
b'''<site>\xf8</site>'''))

def test_write_to_filename_as_unicode(self):
self.addCleanup(os_helper.unlink, TESTFN)
with open(TESTFN, 'w') as f:
encoding = f.encoding
os_helper.unlink(TESTFN)

try:
'\xf8'.encode(encoding)
except UnicodeEncodeError:
self.skipTest(f'default file encoding {encoding} not supported')

tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
tree.write(TESTFN, encoding='unicode')
with open(TESTFN, 'rb') as f:
data = f.read()
expected = "<site>\xf8</site>".encode(encoding, 'xmlcharrefreplace')
self.assertEqual(data, expected)

def test_write_to_text_file(self):
self.addCleanup(os_helper.unlink, TESTFN)
tree = ET.ElementTree(ET.XML('''<site />'''))
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
with open(TESTFN, 'w', encoding='utf-8') as f:
tree.write(f, encoding='unicode')
self.assertFalse(f.closed)
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(), b'''<site />''')
self.assertEqual(f.read(), b'''<site>\xc3\xb8</site>''')

with open(TESTFN, 'w', encoding='ascii', errors='xmlcharrefreplace') as f:
tree.write(f, encoding='unicode')
self.assertFalse(f.closed)
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(), b'''<site>&#248;</site>''')

with open(TESTFN, 'w', encoding='ISO-8859-1') as f:
tree.write(f, encoding='unicode')
self.assertFalse(f.closed)
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(), b'''<site>\xf8</site>''')

def test_write_to_binary_file(self):
self.addCleanup(os_helper.unlink, TESTFN)
tree = ET.ElementTree(ET.XML('''<site />'''))
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
with open(TESTFN, 'wb') as f:
tree.write(f)
self.assertFalse(f.closed)
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(), b'''<site />''')
self.assertEqual(f.read(), b'''<site>&#248;</site>''')

def test_write_to_binary_file_with_encoding(self):
self.addCleanup(os_helper.unlink, TESTFN)
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
with open(TESTFN, 'wb') as f:
tree.write(f, encoding='utf-8')
self.assertFalse(f.closed)
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(), b'''<site>\xc3\xb8</site>''')

with open(TESTFN, 'wb') as f:
tree.write(f, encoding='ISO-8859-1')
self.assertFalse(f.closed)
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(),
b'''<?xml version='1.0' encoding='ISO-8859-1'?>\n'''
b'''<site>\xf8</site>''')

def test_write_to_binary_file_with_bom(self):
self.addCleanup(os_helper.unlink, TESTFN)
tree = ET.ElementTree(ET.XML('''<site />'''))
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
# test BOM writing to buffered file
with open(TESTFN, 'wb') as f:
tree.write(f, encoding='utf-16')
self.assertFalse(f.closed)
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(),
'''<?xml version='1.0' encoding='utf-16'?>\n'''
'''<site />'''.encode("utf-16"))
'''<site>\xf8</site>'''.encode("utf-16"))
# test BOM writing to non-buffered file
with open(TESTFN, 'wb', buffering=0) as f:
tree.write(f, encoding='utf-16')
self.assertFalse(f.closed)
with open(TESTFN, 'rb') as f:
self.assertEqual(f.read(),
'''<?xml version='1.0' encoding='utf-16'?>\n'''
'''<site />'''.encode("utf-16"))
'''<site>\xf8</site>'''.encode("utf-16"))

def test_read_from_stringio(self):
tree = ET.ElementTree()
Expand All @@ -3763,10 +3826,10 @@ def test_read_from_stringio(self):
self.assertEqual(tree.getroot().tag, 'site')

def test_write_to_stringio(self):
tree = ET.ElementTree(ET.XML('''<site />'''))
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
stream = io.StringIO()
tree.write(stream, encoding='unicode')
self.assertEqual(stream.getvalue(), '''<site />''')
self.assertEqual(stream.getvalue(), '''<site>\xf8</site>''')

def test_read_from_bytesio(self):
tree = ET.ElementTree()
Expand All @@ -3775,10 +3838,10 @@ def test_read_from_bytesio(self):
self.assertEqual(tree.getroot().tag, 'site')

def test_write_to_bytesio(self):
tree = ET.ElementTree(ET.XML('''<site />'''))
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
raw = io.BytesIO()
tree.write(raw)
self.assertEqual(raw.getvalue(), b'''<site />''')
self.assertEqual(raw.getvalue(), b'''<site>&#248;</site>''')

class dummy:
pass
Expand All @@ -3792,12 +3855,12 @@ def test_read_from_user_text_reader(self):
self.assertEqual(tree.getroot().tag, 'site')

def test_write_to_user_text_writer(self):
tree = ET.ElementTree(ET.XML('''<site />'''))
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
stream = io.StringIO()
writer = self.dummy()
writer.write = stream.write
tree.write(writer, encoding='unicode')
self.assertEqual(stream.getvalue(), '''<site />''')
self.assertEqual(stream.getvalue(), '''<site>\xf8</site>''')

def test_read_from_user_binary_reader(self):
raw = io.BytesIO(b'''<?xml version="1.0"?><site></site>''')
Expand All @@ -3809,12 +3872,12 @@ def test_read_from_user_binary_reader(self):
tree = ET.ElementTree()

def test_write_to_user_binary_writer(self):
tree = ET.ElementTree(ET.XML('''<site />'''))
tree = ET.ElementTree(ET.XML('''<site>\xf8</site>'''))
raw = io.BytesIO()
writer = self.dummy()
writer.write = raw.write
tree.write(writer)
self.assertEqual(raw.getvalue(), b'''<site />''')
self.assertEqual(raw.getvalue(), b'''<site>&#248;</site>''')

def test_write_to_user_binary_writer_with_bom(self):
tree = ET.ElementTree(ET.XML('''<site />'''))
Expand Down

0 comments on commit f60b4c3

Please sign in to comment.