def test_filename_star():
# NOTE(vytas): Generated by requests_toolbelt 0.9.1 on Py2
# (interestingly, one gets a "normal" filename on Py3.7).
data = (
b'--a0d738bcdb30449eb0d13f4b72c2897e\r\n'
b'Content-Disposition: form-data; name="file"; '
b"filename*=utf-8''%E2%AC%85%20Arrow.txt\r\n\r\n"
b'A unicode arrow in the filename.\r\n'
b'--a0d738bcdb30449eb0d13f4b72c2897e--\r\n'
)
handler = media.MultipartFormHandler()
content_type = ('multipart/form-data; boundary=' +
'a0d738bcdb30449eb0d13f4b72c2897e')
stream = BufferedReader(io.BytesIO(data).read, len(data))
form = handler.deserialize(stream, content_type, len(data))
for part in form:
assert part.filename == '⬅ Arrow.txt'
assert part.secure_filename == '__Arrow.txt'
data = data.replace(b'*=utf-8', b'*=esoteric')
stream = BufferedReader(io.BytesIO(data).read, len(data))
form = handler.deserialize(stream, content_type, len(data))
for part in form:
with pytest.raises(falcon.HTTPBadRequest):
part.filename
def test_readline_with_size():
source = (b'Hello, world! This is a short village name in Wales.\n'
b'Llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogoch')
stream = BufferedReader(io.BytesIO(source).read, len(source))
assert stream.readline(37) == b'Hello, world! This is a short village'
assert stream.readline(37) == b' name in Wales.\n'
assert stream.readline(8) == b'Llanfair'
assert stream.readline(16) == b'pwllgwyngyllgoge'
assert stream.readline(64) == b'rychwyrndrobwllllantysiliogogogoch'
def test_peek_eof():
source = b'Hello, world!\n'
source_stream = io.BytesIO(source)
stream = BufferedReader(source_stream.read, len(source) - 1)
assert stream.peek(0) == b''
assert stream.peek(1) == b'H'
assert stream.peek(2) == b'He'
assert stream.peek(16) == b'Hello, world!'
assert stream.peek(32) == b'Hello, world!'
assert source_stream.read() == b'\n'
def __init__(self, stream, boundary, content_length, parse_options):
# NOTE(vytas): More lenient check whether the provided stream is not
# already an instance of BufferedReader.
# This approach makes testing both the Cythonized and pure-Python
# streams easier within the same test/benchmark suite.
if not hasattr(stream, 'read_until'):
if isinstance(stream, BoundedStream):
stream = BufferedReader(stream.stream.read, content_length)
else:
stream = BufferedReader(stream.read, content_length)
self._stream = stream
self._boundary = boundary
# NOTE(vytas): Here self._dash_boundary is not prepended with CRLF
# (yet) for parsing the prologue. The CRLF will be prepended later to
# construct the inter-part delimiter as per RFC 7578, section 4.1
# (see the note below).
self._dash_boundary = b'--' + boundary
self._parse_options = parse_options
def test_empty_filename():
data = (b'--a0d738bcdb30449eb0d13f4b72c2897e\r\n'
b'Content-Disposition: form-data; name="file"; filename=\r\n\r\n'
b'An empty filename.\r\n'
b'--a0d738bcdb30449eb0d13f4b72c2897e--\r\n')
handler = media.MultipartFormHandler()
content_type = 'multipart/form-data; boundary=' + 'a0d738bcdb30449eb0d13f4b72c2897e'
stream = BufferedReader(io.BytesIO(data).read, len(data))
form = handler.deserialize(stream, content_type, len(data))
for part in form:
assert part.filename == ''
with pytest.raises(falcon.MediaMalformedError):
part.secure_filename
def test_from_buffered_stream():
data = (b'--BOUNDARY\r\n'
b'Content-Disposition: form-data; name="empty"\r\n'
b'Content-Coolness: fair\r\n'
b'Content-Type: text/plain\r\n\r\n'
b'\r\n'
b'--BOUNDARY--\r\n')
handler = media.MultipartFormHandler()
stream = BufferedReader(io.BytesIO(data).read, len(data))
form = handler.deserialize(stream,
'multipart/form-data; boundary=BOUNDARY',
len(data))
for part in form:
assert part.data == b''
def test_parsing_correctness(buffer_size, chunk_size):
example = EXAMPLES['boundary']
handler = media.MultipartFormHandler()
stream = BufferedReader(io.BytesIO(example).read, len(example),
buffer_size)
form = handler.deserialize(
stream, 'multipart/form-data; boundary=boundary', len(example))
for part in form:
if part.name in ('lorem1', 'lorem2'):
part_stream = part.stream
result = []
while True:
chunk = part_stream.read(chunk_size)
if not chunk:
break
result.append(chunk)
assert b''.join(result) == LOREM_IPSUM
def test_readline():
source = (b'Hello, world!\n'
b'A line.\n'
b'\n'
b'A longer line... \n' + b'SPAM ' * 7 + b'\n' + b'\n')
stream = BufferedReader(io.BytesIO(source).read, len(source))
assert stream.readline() == b'Hello, world!\n'
assert stream.readline() == b'A line.\n'
assert stream.readline() == b'\n'
assert stream.readline() == b'A longer line... \n'
assert stream.readline() == b'SPAM SPAM SPAM SPAM SPAM SPAM SPAM \n'
assert stream.readline() == b'\n'
assert stream.readline() == b''
assert stream.readline() == b''
def test_headers_edge_cases(max_headers_size):
data = (
b'--a0d738bcdb30449eb0d13f4b72c2897e\r\n'
b'X-Falcon: Peregrine\r\n'
b'Content-Type: application/vnd.oasis.opendocument.text\r\n'
b'Junk\r\n'
b'Content-Disposition: form-data; name="file"; filename=hd.txt\r\n\r\n'
b'No, it is not an ODT document...\r\n'
b'--a0d738bcdb30449eb0d13f4b72c2897e--\r\n'
)
handler = media.MultipartFormHandler()
handler.parse_options.max_body_part_headers_size = max_headers_size
content_type = ('multipart/form-data; boundary=' +
'a0d738bcdb30449eb0d13f4b72c2897e')
stream = BufferedReader(io.BytesIO(data).read, len(data))
form = handler.deserialize(stream, content_type, len(data))
if max_headers_size < 142:
with pytest.raises(falcon.HTTPBadRequest):
list(form)
else:
assert len(list(form)) == 1
def test_bounded_read():
stream = io.BytesIO(b'Hello, world!')
buffered = BufferedReader(stream.read, len('Hello, world'))
buffered.read()
assert stream.read() == b'!'
def fragmented_stream():
stream = FragmentedStream(TEST_DATA)
return BufferedReader(stream.read, len(TEST_DATA))
def shorter_stream():
TEST_BYTES_IO.seek(0)
return BufferedReader(TEST_BYTES_IO.read, 1024, 128)
def _reader_fixture(chunk_size=None):
TEST_BYTES_IO.seek(0)
return BufferedReader(TEST_BYTES_IO.read, len(TEST_DATA), chunk_size)
def test_read_until_shared_boundary(chunk_size):
source = b'-boundary-like-' * 4 + b'--some junk--\n' + b'\n' * 1024
source_stream = io.BytesIO(source)
stream = BufferedReader(source_stream.read, len(source), chunk_size)
assert stream.read_until(b'-boundary-like---') == b'-boundary-like-' * 3
assert stream.peek(17) == b'-boundary-like---'