Strip leading UTF-8 BOM when reading INI files

src/iniconfig/__init__.py

@@ -112,6 +112,9 @@ def __init__(
             if data is None:
                 with open(self.path, encoding=encoding) as fp:
                     data = fp.read()
+            # Strip a leading UTF-8 BOM so editors that write one still parse.
+            if data.startswith(""):
+                data = data.removeprefix("")
 
             # Use old behavior (no stripping) for backward compatibility
             sections_data, sources = _parse.parse_ini_data(
@@ -166,6 +169,9 @@ def parse(
         if data is None:
             with open(fspath, encoding=encoding) as fp:
                 data = fp.read()
+        # Strip a leading UTF-8 BOM so editors that write one still parse.
+        if data.startswith(""):
+            data = data.removeprefix("")
 
         sections_data, sources = _parse.parse_ini_data(
             fspath,

testing/test_iniconfig.py

@@ -412,3 +412,23 @@ def test_unicode_whitespace_in_key_names() -> None:
     )
     assert "key" in config["section"]
     assert config["section"]["key"] == "value"
+
+
+def test_utf8_bom_file(tmp_path):
+    """Files saved with a UTF-8 BOM (common on Windows editors) must parse."""
+    path = tmp_path / "bom.ini"
+    path.write_bytes(b"\xef\xbb\xbf[section]\nkey = value\n")
+    config = IniConfig(path)
+    assert config["section"]["key"] == "value"
+
+
+def test_utf8_bom_in_data_string():
+    config = IniConfig("x.ini", data="\ufeff[section]\nkey = value\n")
+    assert config["section"]["key"] == "value"
+
+
+def test_parse_utf8_bom_file(tmp_path):
+    path = tmp_path / "bom.ini"
+    path.write_bytes(b"\xef\xbb\xbf[section]\nkey = value\n")
+    config = IniConfig.parse(path)
+    assert config["section"]["key"] == "value"