Silent fail of long url parsing

[euri10]

greetings, while I see the following issue: #43 has been closed I wondered since the parser was changed if this could be reevaluated or if we should handle that downstream in uvicorn,

this added test will fail if the url us 2^16-1 bytes long and the returned url is b''

    def test_parser_url_11(self):
        LARGE_URL = b'/' + b'a' * (2**16-1)
        self.assertEqual(self.parse(LARGE_URL), (None, None, None, LARGE_URL, None, None, None))

Failure
Traceback (most recent call last):
  File "/home/lotso/PycharmProjects/httptools/tests/test_parser.py", line 630, in test_parser_url_11
    self.assertEqual(self.parse(LARGE_URL), (None, None, None, LARGE_URL, None, None, None))
AssertionError: Tuples differ: (None, None, None, b'', None, None, None) != (None, None, None, b'/aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa[65510 chars]None)

First differing element 3:
b''
b'/aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa[65491 chars]aaaa'

Diff is 71676 characters long. Set self.maxDiff to None to see it.

note the issue is slightly different here we get b'' while in old issue the url was trimmed

[euri10]

ha now I read the readme and I'm confused, http-parse is still used for url parsing ?

[euri10]

ok it is, and llhttp does not do it nodejs/llhttp#7
sorry for the noise