From ad631099729021b9711fb160d5c7e2cbb6a6463b Mon Sep 17 00:00:00 2001
From: Pitchfork-and-Torch <pitchfork-and-torch@users.noreply.github.com>
Date: Fri, 18 Sep 2026 02:15:38 +0000
Subject: [PATCH 1/2] Strip leading UTF-8 BOM when reading INI files

Windows editors often write a UTF-8 BOM; with encoding=utf-8 that left
U+FEFF on the first section line and raised ParseError. Strip the BOM
after read for both IniConfig() and IniConfig.parse().
---
 src/iniconfig/__init__.py |  6 ++++++
 testing/test_iniconfig.py | 20 ++++++++++++++++++++
 2 files changed, 26 insertions(+)

diff --git a/src/iniconfig/__init__.py b/src/iniconfig/__init__.py
index b84809f..e9fd907 100644
--- a/src/iniconfig/__init__.py
+++ b/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,
diff --git a/testing/test_iniconfig.py b/testing/test_iniconfig.py
index 85193c5..488178f 100644
--- a/testing/test_iniconfig.py
+++ b/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"

From 270e11952c01dbec408f5e636f1177d56cb9b912 Mon Sep 17 00:00:00 2001
From: Pitchfork-and-Torch
 <297513015+Pitchfork-and-Torch@users.noreply.github.com>
Date: Mon, 5 Oct 2026 20:23:34 -0400
Subject: [PATCH 2/2] Move UTF-8 BOM stripping into parse_ini_data

Strip once in parse_ini_data, name the mark with \N{BYTE ORDER MARK}, and annotate the new tests.
---
 src/iniconfig/__init__.py | 6 ------
 src/iniconfig/_parse.py   | 4 ++++
 testing/test_iniconfig.py | 6 +++---
 3 files changed, 7 insertions(+), 9 deletions(-)

diff --git a/src/iniconfig/__init__.py b/src/iniconfig/__init__.py
index e9fd907..b84809f 100644
--- a/src/iniconfig/__init__.py
+++ b/src/iniconfig/__init__.py
@@ -112,9 +112,6 @@ 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(
@@ -169,9 +166,6 @@ 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,
diff --git a/src/iniconfig/_parse.py b/src/iniconfig/_parse.py
index 57b9b44..e7a1a8c 100644
--- a/src/iniconfig/_parse.py
+++ b/src/iniconfig/_parse.py
@@ -1,9 +1,11 @@
 from collections.abc import Mapping
+from typing import Final
 from typing import NamedTuple
 
 from .exceptions import ParseError
 
 COMMENTCHARS = "#;"
+UTF8_BOM: Final = "\N{BYTE ORDER MARK}"
 
 
 class ParsedLine(NamedTuple):
@@ -34,6 +36,8 @@ def parse_ini_data(
         - sections_data: mapping of section -> {name -> value}
         - sources: mapping of (section, name) -> line number
     """
+    data = data.removeprefix(UTF8_BOM)
+
     tokens = parse_lines(
         path,
         data.splitlines(True),
diff --git a/testing/test_iniconfig.py b/testing/test_iniconfig.py
index 488178f..6162256 100644
--- a/testing/test_iniconfig.py
+++ b/testing/test_iniconfig.py
@@ -414,7 +414,7 @@ def test_unicode_whitespace_in_key_names() -> None:
     assert config["section"]["key"] == "value"
 
 
-def test_utf8_bom_file(tmp_path):
+def test_utf8_bom_file(tmp_path: Path) -> None:
     """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")
@@ -422,12 +422,12 @@ def test_utf8_bom_file(tmp_path):
     assert config["section"]["key"] == "value"
 
 
-def test_utf8_bom_in_data_string():
+def test_utf8_bom_in_data_string() -> None:
     config = IniConfig("x.ini", data="\ufeff[section]\nkey = value\n")
     assert config["section"]["key"] == "value"
 
 
-def test_parse_utf8_bom_file(tmp_path):
+def test_parse_utf8_bom_file(tmp_path: Path) -> None:
     path = tmp_path / "bom.ini"
     path.write_bytes(b"\xef\xbb\xbf[section]\nkey = value\n")
     config = IniConfig.parse(path)