Complete removal of SplitResult as an internal (#1396)

CHANGES/1396.misc.rst

@@ -0,0 +1 @@
+Improved performance many :class:`~yarl.URL` methods -- by :user:`bdraco`.

tests/test_pickle.py

@@ -18,6 +18,7 @@ def test_pickle():
 def test_default_style_state():
     u = URL("test")
     hash(u)
-    u.__setstate__((None, {"_val": "test", "_strict": False, "_cache": {"hash": 1}}))
+    val = ("test", "test", "test", "test", "test")
+    u.__setstate__((None, {"_val": val, "_strict": False, "_cache": {"hash": 1}}))
     assert not u._cache
-    assert u._val == "test"
+    assert u._val == val

tests/test_url.py

@@ -132,13 +132,13 @@ def test_scheme():
 def test_raw_user():
     url = URL("http://[email protected]")
     assert "user" == url.raw_user
-    assert url.raw_user == url._val.username
+    assert url.raw_user == SplitResult(*url._val).username
 
 
 def test_raw_user_non_ascii():
     url = URL("http://бажан@example.com")
     assert "%D0%B1%D0%B0%D0%B6%D0%B0%D0%BD" == url.raw_user
-    assert url.raw_user == url._val.username
+    assert url.raw_user == SplitResult(*url._val).username
 
 
 def test_no_user():
@@ -154,13 +154,13 @@ def test_user_non_ascii():
 def test_raw_password():
     url = URL("http://user:[email protected]")
     assert "password" == url.raw_password
-    assert url.raw_password == url._val.password
+    assert url.raw_password == SplitResult(*url._val).password
 
 
 def test_raw_password_non_ascii():
     url = URL("http://user:пароль@example.com")
     assert "%D0%BF%D0%B0%D1%80%D0%BE%D0%BB%D1%8C" == url.raw_password
-    assert url.raw_password == url._val.password
+    assert url.raw_password == SplitResult(*url._val).password
 
 
 def test_password_non_ascii():
@@ -179,7 +179,7 @@ def test_empty_password_without_user():
     assert url.user is None
     assert url.password == ""
     assert url.raw_password == ""
-    assert url.raw_password == url._val.password
+    assert url.raw_password == SplitResult(*url._val).password
 
 
 def test_user_empty_password():
@@ -191,7 +191,7 @@ def test_user_empty_password():
 def test_raw_host():
     url = URL("http://example.com")
     assert "example.com" == url.raw_host
-    assert url.raw_host == url._val.hostname
+    assert url.raw_host == SplitResult(*url._val).hostname
 
 
 @pytest.mark.parametrize(
@@ -244,7 +244,7 @@ def test_invalid_idna_a_label_encoding():
 def test_raw_host_non_ascii():
     url = URL("http://оун-упа.укр")
     assert "xn----8sb1bdhvc.xn--j1amh" == url.raw_host
-    assert url.raw_host == url._val.hostname
+    assert url.raw_host == SplitResult(*url._val).hostname
 
 
 def test_host_non_ascii():
@@ -265,19 +265,19 @@ def test_host_with_underscore():
 def test_raw_host_when_port_is_specified():
     url = URL("http://example.com:8888")
     assert "example.com" == url.raw_host
-    assert url.raw_host == url._val.hostname
+    assert url.raw_host == SplitResult(*url._val).hostname
 
 
 def test_raw_host_from_str_with_ipv4():
     url = URL("http://127.0.0.1:80")
     assert url.raw_host == "127.0.0.1"
-    assert url.raw_host == url._val.hostname
+    assert url.raw_host == SplitResult(*url._val).hostname
 
 
 def test_raw_host_from_str_with_ipv6():
     url = URL("http://[::1]:80")
     assert url.raw_host == "::1"
-    assert url.raw_host == url._val.hostname
+    assert url.raw_host == SplitResult(*url._val).hostname
 
 
 def test_authority_full() -> None:
@@ -311,21 +311,21 @@ def test_lowercase():
     url = URL("http://gitHUB.com")
     assert url.raw_host == "github.com"
     assert url.host == url.raw_host
-    assert url.raw_host == url._val.hostname
+    assert url.raw_host == SplitResult(*url._val).hostname
 
 
 def test_lowercase_nonascii():
     url = URL("http://Слава.Укр")
     assert url.raw_host == "xn--80aaf8a3a.xn--j1amh"
-    assert url.raw_host == url._val.hostname
+    assert url.raw_host == SplitResult(*url._val).hostname
     assert url.host == "слава.укр"
 
 
 def test_compressed_ipv6():
     url = URL("http://[1DEC:0:0:0::1]")
     assert url.raw_host == "1dec::1"
     assert url.host == url.raw_host
-    assert url.raw_host == url._val.hostname
+    assert url.raw_host == SplitResult(*url._val).hostname
 
 
 def test_ipv6_missing_left_bracket():
@@ -353,19 +353,19 @@ def test_ipv4_zone():
     url = URL("http://1.2.3.4%тест%42:123")
     assert url.raw_host == "1.2.3.4%тест%42"
     assert url.host == url.raw_host
-    assert url.raw_host == url._val.hostname
+    assert url.raw_host == SplitResult(*url._val).hostname
 
 
 def test_port_for_explicit_port():
     url = URL("http://example.com:8888")
     assert 8888 == url.port
-    assert url.explicit_port == url._val.port
+    assert url.explicit_port == SplitResult(*url._val).port
 
 
 def test_port_for_implicit_port():
     url = URL("http://example.com")
     assert 80 == url.port
-    assert url.explicit_port == url._val.port
+    assert url.explicit_port == SplitResult(*url._val).port
 
 
 def test_port_for_relative_url():
@@ -383,25 +383,25 @@ def test_port_for_unknown_scheme():
 def test_explicit_port_for_explicit_port():
     url = URL("http://example.com:8888")
     assert 8888 == url.explicit_port
-    assert url.explicit_port == url._val.port
+    assert url.explicit_port == SplitResult(*url._val).port
 
 
 def test_explicit_port_for_implicit_port():
     url = URL("http://example.com")
     assert url.explicit_port is None
-    assert url.explicit_port == url._val.port
+    assert url.explicit_port == SplitResult(*url._val).port
 
 
 def test_explicit_port_for_relative_url():
     url = URL("/path/to")
     assert url.explicit_port is None
-    assert url.explicit_port == url._val.port
+    assert url.explicit_port == SplitResult(*url._val).port
 
 
 def test_explicit_port_for_unknown_scheme():
     url = URL("unknown://example.com")
     assert url.explicit_port is None
-    assert url.explicit_port == url._val.port
+    assert url.explicit_port == SplitResult(*url._val).port
 
 
 def test_raw_path_string_empty():
@@ -1563,7 +1563,7 @@ def test_is_default_port_for_unknown_scheme():
 def test_handling_port_zero():
     url = URL("http://example.com:0")
     assert url.explicit_port == 0
-    assert url.explicit_port == url._val.port
+    assert url.explicit_port == SplitResult(*url._val).port
     assert str(url) == "http://example.com:0"
     assert not url.is_default_port()
 
@@ -1767,13 +1767,13 @@ def test_parent_for_empty_url():
 def test_parent_for_relative_url_with_child():
     url = URL("path/to")
     assert url.parent == URL("path")
-    assert url.parent._val.path == "path"
+    assert SplitResult(*url.parent._val).path == "path"
 
 
 def test_parent_for_relative_url():
     url = URL("path")
     assert url.parent == URL("")
-    assert url.parent._val.path == ""
+    assert SplitResult(*url.parent._val).path == ""
 
 
 def test_parent_for_no_netloc_url():
@@ -1784,7 +1784,7 @@ def test_parent_for_no_netloc_url():
 def test_parent_for_top_level_no_netloc_url():
     url = URL("/")
     assert url.parent == URL("/")
-    assert url.parent._val.path == "/"
+    assert SplitResult(*url.parent._val).path == "/"
 
 
 def test_parent_for_absolute_url():
@@ -1795,7 +1795,7 @@ def test_parent_for_absolute_url():
 def test_parent_for_top_level_absolute_url():
     url = URL("http://go.to/")
     assert url.parent == URL("http://go.to/")
-    assert url.parent._val.path == "/"
+    assert SplitResult(*url.parent._val).path == "/"
 
 
 def test_empty_value_for_query():

tests/test_url_parsing.py

@@ -1,3 +1,5 @@
+from urllib.parse import SplitResult
+
 import pytest
 
 from yarl import URL
@@ -616,7 +618,7 @@ def test_url_round_trips(
 ) -> None:
     """Verify that URLs round-trip correctly."""
     parsed = URL(url)
-    assert parsed._val.hostname == hostname_without_brackets
+    assert SplitResult(*parsed._val).hostname == hostname_without_brackets
     assert parsed.raw_host == hostname_without_brackets
     assert parsed.host_subcomponent == hostname
     assert str(parsed) == url

yarl/_parse.py

@@ -3,7 +3,7 @@
 import re
 import unicodedata
 from functools import lru_cache
-from typing import Union
+from typing import Final, Union
 from urllib.parse import scheme_chars, uses_netloc
 
 from ._quoters import QUOTER
@@ -19,8 +19,15 @@
 UNSAFE_URL_BYTES_TO_REMOVE = ["\t", "\r", "\n"]
 USES_AUTHORITY = frozenset(uses_netloc)
 
+SplitURL = tuple[str, str, str, str, str]
+SCHEME: Final[int] = 0
+NETLOC: Final[int] = 1
+PATH: Final[int] = 2
+QUERY: Final[int] = 3
+FRAGMENT: Final[int] = 4
 
-def split_url(url: str) -> tuple[str, str, str, str, str]:
+
+def split_url(url: str) -> SplitURL:
     """Split URL into parts."""
     # Adapted from urllib.parse.urlsplit
     # Only lstrip url as some applications rely on preserving trailing space.