Nov 3, 2022 · Nov 4, 2022 · Nov 8, 2022 · Nov 8, 2022 · Nov 8, 2022 · Nov 9, 2022
diff --git a/tests/test_url.py b/tests/test_url.py
    _C0_CONTROL_PERCENT_ENCODE_SET,
    _parse_url,
    _percent_encode_after_encoding,
    _serialize_host,
    # _serialize_url,
    _SPECIAL_SCHEMES,
 )
    assert url.scheme == (protocol[:-1] if protocol else None)
    assert url.username == username
    assert url.password == password
 #assert url.hostname == hostname
    assert_serialize_host(url.hostname) == hostname
    # TODO: Cover additional fields
    # assert _serialize_url(url) == href

        and chr(value) not in ":/?#\\"
    )
 )
 USERNAME_ENCODED = "".join(f"%{ord(char):X}" for char in USERNAME_TO_ENCODE)
 USERNAME_ENCODED = "".join(f"%{ord(char):02X}" for char in USERNAME_TO_ENCODE)
 PASSWORD_TO_ENCODE = USERNAME_TO_ENCODE + ":"
 PASSWORD_ENCODED = "".join(f"%{ord(char):X}" for char in PASSWORD_TO_ENCODE)
 PASSWORD_ENCODED = "".join(f"%{ord(char):02X}" for char in PASSWORD_TO_ENCODE)

 # Path characters that do not need escaping.
 # Removed for RFC 2396 and RFC 3986: %[\]^|
        and chr(value) not in "?#\\"
    )
 )
 PATH_ENCODED = "".join(f"%{ord(char):X}" for char in PATH_TO_ENCODE)
 PATH_ENCODED = "".join(f"%{ord(char):02X}" for char in PATH_TO_ENCODE)

 # Query characters that do not need escaping.
 # Removed for RFC 2396 and RFC 3986: %[\]^`{|}
        and chr(value) not in "#"
    )
 )
 QUERY_ENCODED = "".join(f"%{ord(char):X}" for char in QUERY_TO_ENCODE)
 QUERY_ENCODED = "".join(f"%{ord(char):02X}" for char in QUERY_TO_ENCODE)
 SPECIAL_QUERY_SAFE = QUERY_SAFE.replace("'", "")
 SPECIAL_QUERY_TO_ENCODE = "".join(
    chr(value)
        and chr(value) not in "#"
    )
 )
 SPECIAL_QUERY_ENCODED = "".join(f"%{ord(char):X}" for char in SPECIAL_QUERY_TO_ENCODE)
 SPECIAL_QUERY_ENCODED = "".join(f"%{ord(char):02X}" for char in SPECIAL_QUERY_TO_ENCODE)

 # Fragment characters that do not need escaping.
 # Removed for RFC 2396 and RFC 3986: #%[\\]^{|}
    for value in range(0x80)
    if (chr(value) not in _C0_CONTROL_OR_SPACE and chr(value) not in FRAGMENT_SAFE)
 )
 FRAGMENT_ENCODED = "".join(f"%{ord(char):X}" for char in FRAGMENT_TO_ENCODE)
 FRAGMENT_ENCODED = "".join(f"%{ord(char):02X}" for char in FRAGMENT_TO_ENCODE)


 # Test cases for URL-to-safe-URL conversions with only a URL as input parameter
diff --git a/w3lib/_url.py b/w3lib/_url.py
    scheme: str = ""
    username: str = ""
    password: str = ""
    hostname: Union[None,int, List[int], str] =None
    hostname: Union[int, List[int], str] =""
    port: Optional[int] = None
    path: Union[str, List[str]]
    query: Optional[str] = None
                else:
                    output += "%"
            else:
                output += f"%{byte[0]:X}"
                output += f"%{byte[0]:02X}"
        if potential_error is not None:
            assert isinstance(potential_error, _PotentialError)
            assert isinstance(potential_error.code_point, (bytes, str))
        elif state == _State.RELATIVE:
            assert isinstance(base, _URL)
            url.scheme = base.scheme
            if c == "/":
                state = _State.RELATIVE
            elif url.is_special() and c == "\\":
            if (
                c == "/"
                or url.is_special() and c == "\\"
            ):
                state = _State.RELATIVE_SLASH
            else:
                url.username = base.username
Original file line number	Diff line number	Diff line change
Expand Up		@@ -20,6 +20,7 @@
		_C0_CONTROL_PERCENT_ENCODE_SET,
		_parse_url,
		_percent_encode_after_encoding,
		_serialize_host,
		# _serialize_url,
		_SPECIAL_SCHEMES,
		)
Expand DownExpand Up		@@ -79,7 +80,7 @@ def test_parse_url(input, base, failure, href, protocol, username, password, hos
		assert url.scheme == (protocol[:-1] if protocol else None)
		assert url.username == username
		assert url.password == password
		#assert url.hostname == hostname
		assert_serialize_host(url.hostname) == hostname
		# TODO: Cover additional fields
		# assert _serialize_url(url) == href

Expand DownExpand Up		@@ -182,9 +183,9 @@ def test_percent_encode_after_encoding(input, output):
		and chr(value) not in ":/?#\\"
		)
		)
		USERNAME_ENCODED = "".join(f"%{ord(char):X}" for char in USERNAME_TO_ENCODE)
		USERNAME_ENCODED = "".join(f"%{ord(char):02X}" for char in USERNAME_TO_ENCODE)
		PASSWORD_TO_ENCODE = USERNAME_TO_ENCODE + ":"
		PASSWORD_ENCODED = "".join(f"%{ord(char):X}" for char in PASSWORD_TO_ENCODE)
		PASSWORD_ENCODED = "".join(f"%{ord(char):02X}" for char in PASSWORD_TO_ENCODE)

		# Path characters that do not need escaping.
		# Removed for RFC 2396 and RFC 3986: %[\]^\|
Expand All		@@ -198,7 +199,7 @@ def test_percent_encode_after_encoding(input, output):
		and chr(value) not in "?#\\"
		)
		)
		PATH_ENCODED = "".join(f"%{ord(char):X}" for char in PATH_TO_ENCODE)
		PATH_ENCODED = "".join(f"%{ord(char):02X}" for char in PATH_TO_ENCODE)

		# Query characters that do not need escaping.
		# Removed for RFC 2396 and RFC 3986: %[\]^`{\|}
Expand All		@@ -213,7 +214,7 @@ def test_percent_encode_after_encoding(input, output):
		and chr(value) not in "#"
		)
		)
		QUERY_ENCODED = "".join(f"%{ord(char):X}" for char in QUERY_TO_ENCODE)
		QUERY_ENCODED = "".join(f"%{ord(char):02X}" for char in QUERY_TO_ENCODE)
		SPECIAL_QUERY_SAFE = QUERY_SAFE.replace("'", "")
		SPECIAL_QUERY_TO_ENCODE = "".join(
		chr(value)
Expand All		@@ -224,7 +225,7 @@ def test_percent_encode_after_encoding(input, output):
		and chr(value) not in "#"
		)
		)
		SPECIAL_QUERY_ENCODED = "".join(f"%{ord(char):X}" for char in SPECIAL_QUERY_TO_ENCODE)
		SPECIAL_QUERY_ENCODED = "".join(f"%{ord(char):02X}" for char in SPECIAL_QUERY_TO_ENCODE)

		# Fragment characters that do not need escaping.
		# Removed for RFC 2396 and RFC 3986: #%[\\]^{\|}
Expand All		@@ -234,7 +235,7 @@ def test_percent_encode_after_encoding(input, output):
		for value in range(0x80)
		if (chr(value) not in _C0_CONTROL_OR_SPACE and chr(value) not in FRAGMENT_SAFE)
		)
		FRAGMENT_ENCODED = "".join(f"%{ord(char):X}" for char in FRAGMENT_TO_ENCODE)
		FRAGMENT_ENCODED = "".join(f"%{ord(char):02X}" for char in FRAGMENT_TO_ENCODE)


		# Test cases for URL-to-safe-URL conversions with only a URL as input parameter
Expand Down
Original file line number	Diff line number	Diff line change
Expand Up		@@ -97,7 +97,7 @@ class _URL:
		scheme: str = ""
		username: str = ""
		password: str = ""
		hostname: Union[None,int, List[int], str] =None
		hostname: Union[int, List[int], str] =""
		port: Optional[int] = None
		path: Union[str, List[str]]
		query: Optional[str] = None
Expand DownExpand Up		@@ -192,7 +192,7 @@ def _percent_encode_after_encoding(
		else:
		output += "%"
		else:
		output += f"%{byte[0]:X}"
		output += f"%{byte[0]:02X}"
		if potential_error is not None:
		assert isinstance(potential_error, _PotentialError)
		assert isinstance(potential_error.code_point, (bytes, str))
Expand DownExpand Up		@@ -665,9 +665,10 @@ def _parse_url(
		elif state == _State.RELATIVE:
		assert isinstance(base, _URL)
		url.scheme = base.scheme
		if c == "/":
		state = _State.RELATIVE
		elif url.is_special() and c == "\\":
		if (
		c == "/"
		or url.is_special() and c == "\\"
		):
		state = _State.RELATIVE_SLASH
		else:
		url.username = base.username
Expand Down