Remove URL validation from requirement parsing

The scheme and path validation logic limits how users of the library can
provide URL support. This limitation is lifted, and dependants now need
to implement their own URL validation logic they see fit.

CHANGELOG.rst

@@ -4,7 +4,7 @@ Changelog
 *unreleased*
 ~~~~~~~~~~~~
 
-No unreleased changes.
+* Requirement parsing no longer automatically validates the URL (:issue:`120`)
 
 23.0 - 2023-01-08
 ~~~~~~~~~~~~~~~~~

docs/requirements.rst

@@ -56,6 +56,13 @@ Usage
     >>> requirements1 == requirements2
     True
 
+.. versionchanged:: 23.1
+
+    When a requirement is specified with a URL, the :class:`Requirement` class
+    used to check the URL and reject values containing invalid scheme and
+    netloc combinations. This is no longer performed since PEP 508 does not
+    specify such rules, and the check incorrectly disallows valid requirement
+    strings from being parsed.
 
 Reference
 ---------

src/packaging/requirements.py

@@ -2,7 +2,6 @@
 # 2.0, and the BSD License. See the LICENSE file in the root of this repository
 # for complete details.
 
-import urllib.parse
 from typing import Any, List, Optional, Set
 
 from ._parser import parse_requirement as _parse_requirement
@@ -37,18 +36,7 @@ def __init__(self, requirement_string: str) -> None:
             raise InvalidRequirement(str(e)) from e
 
         self.name: str = parsed.name
-        if parsed.url:
-            parsed_url = urllib.parse.urlparse(parsed.url)
-            if parsed_url.scheme == "file":
-                if urllib.parse.urlunparse(parsed_url) != parsed.url:
-                    raise InvalidRequirement("Invalid URL given")
-            elif not (parsed_url.scheme and parsed_url.netloc) or (
-                not parsed_url.scheme and not parsed_url.netloc
-            ):
-                raise InvalidRequirement(f"Invalid URL: {parsed.url}")
-            self.url: Optional[str] = parsed.url
-        else:
-            self.url = None
+        self.url: Optional[str] = parsed.url or None
         self.extras: Set[str] = set(parsed.extras if parsed.extras else [])
         self.specifier: SpecifierSet = SpecifierSet(parsed.specifier)
         self.marker: Optional[Marker] = None

tests/test_requirements.py

@@ -78,6 +78,7 @@
         ("git+https://git.example.com/MyProject.git@master", ""),
         ("git+https://git.example.com/MyProject.git@v1.0", ""),
         ("git+https://git.example.com/MyProject.git@refs/pull/123/head", ""),
+        ("gopher:/foo/com", ""),
         (None, "==={ws}arbitrarystring"),
         (None, "({ws}==={ws}arbitrarystring{ws})"),
         (None, "=={ws}1.0"),
@@ -164,6 +165,8 @@ def test_valid_marker(self, marker: str) -> None:
         [
             "file:///absolute/path",
             "file://.",
+            "file:.",
+            "file:/.",
         ],
     )
     def test_file_url(self, url: str) -> None:
@@ -503,25 +506,6 @@ def test_error_invalid_marker_with_invalid_op(self) -> None:
             "                ^"
         )
 
-    @pytest.mark.parametrize(
-        "url",
-        [
-            "gopher:/foo/com",
-            "file:.",
-            "file:/.",
-        ],
-    )
-    def test_error_on_invalid_url(self, url: str) -> None:
-        # GIVEN
-        to_parse = f"name @ {url}"
-
-        # WHEN
-        with pytest.raises(InvalidRequirement) as ctx:
-            Requirement(to_parse)
-
-        # THEN
-        assert "Invalid URL" in ctx.exconly()
-
     def test_error_on_legacy_version_outside_triple_equals(self) -> None:
         # GIVEN
         to_parse = "name==1.0.org1"