Fixed colon in filename bug

tests/test_url.py

@@ -302,3 +302,25 @@ def test_percent_encoding_spaces() -> None:
     # Test that actual spaces get encoded properly
     url_with_spaces = URL("https://somepath.com/test") / "Test path" / "my test file.txt"
     assert str(url_with_spaces) == "https://somepath.com/test/Test%20path/my%20test%20file.txt"
+
+
+def test_colon_in_filename() -> None:
+    """Test that colons in filenames are not treated as scheme separators."""
+    # Reported bug: URL('http://www.example.com/abc:def.html') was truncated to 'http://www.example.com/def.html'
+    url = URL("http://www.example.com/abc:def.html")
+    assert str(url) == "http://www.example.com/abc:def.html"
+    assert url.name == "abc:def.html"
+    assert url.path == "/abc:def.html"
+
+    # Test various positions and uses of colons
+    assert str(URL("http://www.example.com/file:name.txt")) == "http://www.example.com/file:name.txt"
+    assert str(URL("http://www.example.com/path/to/file:v2.html")) == "http://www.example.com/path/to/file:v2.html"
+    assert str(URL("http://www.example.com/:colon.txt")) == "http://www.example.com/:colon.txt"
+    assert str(URL("http://www.example.com/colon:.txt")) == "http://www.example.com/colon:.txt"
+
+    # Test with query and fragment
+    url = URL("http://www.example.com/abc:def.html?key=value#frag")
+    assert url.name == "abc:def.html"
+    assert url.query == "key=value"
+    assert url.fragment == "frag"
+    assert str(url) == "http://www.example.com/abc:def.html?key=value#frag"

urlpath/__init__.py

@@ -479,6 +479,41 @@ def path(self) -> str:
             + self.trailing_sep
         )
 
+    @property
+    @cached_property
+    def _name_parts(self) -> tuple[str, str, str]:
+        """Parse super().name into (path, query, fragment) without using urlsplit.
+
+        We can't use urlsplit here because it treats colons as scheme separators,
+        which breaks filenames like 'abc:def.html'.
+
+        Parsing order: fragment first (after #), then query (after ?), then path.
+
+        Returns:
+            Tuple of (path, query, fragment) strings.
+        """
+        full_name = super().name
+
+        # Fragment takes priority - everything after # is fragment
+        fragment_idx = full_name.find("#")
+        if fragment_idx != -1:
+            fragment = full_name[fragment_idx + 1 :]
+            before_fragment = full_name[:fragment_idx]
+        else:
+            fragment = ""
+            before_fragment = full_name
+
+        # Query is everything after ? (but before #)
+        query_idx = before_fragment.find("?")
+        if query_idx != -1:
+            query = before_fragment[query_idx + 1 :]
+            path = before_fragment[:query_idx]
+        else:
+            query = ""
+            path = before_fragment
+
+        return path, query, fragment
+
     @property
     @cached_property
     def name(self) -> str:
@@ -487,7 +522,7 @@ def name(self) -> str:
         Returns:
             The decoded filename or last path segment.
         """
-        return urllib.parse.unquote(urllib.parse.urlsplit(super().name).path.rstrip(self._flavour.sep))
+        return urllib.parse.unquote(self._name_parts[0].rstrip(self._flavour.sep))
 
     @property
     @cached_property
@@ -497,7 +532,7 @@ def query(self) -> str:
         Returns:
             The raw query string (without the leading '?').
         """
-        return urllib.parse.urlsplit(super().name).query
+        return self._name_parts[1]
 
     @property
     @cached_property
@@ -507,7 +542,7 @@ def fragment(self) -> str:
         Returns:
             The fragment string (without the leading '#').
         """
-        return urllib.parse.urlsplit(super().name).fragment
+        return self._name_parts[2]
 
     @property
     @cached_property
@@ -517,7 +552,7 @@ def trailing_sep(self) -> str:
         Returns:
             The trailing '/' characters, or empty string if none.
         """
-        match = re.search("(" + re.escape(self._flavour.sep) + "*)$", urllib.parse.urlsplit(super().name).path)
+        match = re.search("(" + re.escape(self._flavour.sep) + "*)$", self._name_parts[0])
         assert match is not None
         return match.group(0)