From 78cebcbbdc463730f2adaeb290345dfa4c4d8de7 Mon Sep 17 00:00:00 2001
From: Leo <leocasti2@gmail.com>
Date: Mon, 1 Dec 2025 02:41:54 -0500
Subject: [PATCH] fix: prioritize utf-8 over system locale to prevent crashes
 on Windows

---
 src/gitingest/utils/file_utils.py |  2 +-
 tests/test_windows_encoding.py    | 48 +++++++++++++++++++++++++++++++
 2 files changed, 49 insertions(+), 1 deletion(-)
 create mode 100644 tests/test_windows_encoding.py

diff --git a/src/gitingest/utils/file_utils.py b/src/gitingest/utils/file_utils.py
index 2c6ef74d..066195bc 100644
--- a/src/gitingest/utils/file_utils.py
+++ b/src/gitingest/utils/file_utils.py
@@ -27,7 +27,7 @@ def _get_preferred_encodings() -> list[str]:
         platform's default encoding followed by common fallback encodings.
 
     """
-    encodings = [locale.getpreferredencoding(), "utf-8", "utf-16", "utf-16le", "utf-8-sig", "latin"]
+    encodings = ["utf-8", locale.getpreferredencoding(), "utf-16", "utf-16le", "utf-8-sig", "latin"]
     if platform.system() == "Windows":
         encodings += ["cp1252", "iso-8859-1"]
     return list(dict.fromkeys(encodings))
diff --git a/tests/test_windows_encoding.py b/tests/test_windows_encoding.py
new file mode 100644
index 00000000..65706bf3
--- /dev/null
+++ b/tests/test_windows_encoding.py
@@ -0,0 +1,48 @@
+"""Tests for Windows encoding handling."""
+
+from pathlib import Path
+from unittest.mock import patch
+
+from gitingest.schemas.filesystem import FileSystemNode, FileSystemNodeType
+from gitingest.utils.file_utils import _CHUNK_SIZE
+
+
+def test_utf8_priority_on_windows(tmp_path: Path) -> None:
+    """Ensure UTF-8 files are read correctly even on Windows systems defaulting to cp1252.
+
+    This test reproduces a specific crash scenario where:
+    1. A file is valid UTF-8.
+    2. The first 1024 bytes are safe ASCII (passing the initial CP1252 check).
+    3. Subsequent bytes contain characters undefined in CP1252 (e.g., smart quotes),
+       causing a UnicodeDecodeError if CP1252 is preferred over UTF-8.
+    """
+    file_path = tmp_path / "test_encoding_crash.md"
+
+    # The right double quotation mark (”) is:
+    # - UTF-8: 0xE2 0x80 0x9D
+    # - CP1252: Byte 0x9D is UNDEFINED and causes a crash during full read.
+    poison_char = "”"
+
+    # Fill buffer to bypass the initial chunk read (which checks the first 1024 bytes)
+    content = ("a" * (_CHUNK_SIZE + 50)) + poison_char + "\nEnd."
+
+    file_path.write_text(content, encoding="utf-8")
+
+    node = FileSystemNode(
+        name=file_path.name,
+        type=FileSystemNodeType.FILE,
+        path_str=str(file_path),
+        path=file_path,
+        size=file_path.stat().st_size,
+    )
+
+    # Mock the environment to simulate Windows with CP1252 locale
+    with patch("locale.getpreferredencoding", return_value="cp1252"), patch(
+        "platform.system",
+        return_value="Windows",
+    ):
+        read_content = node.content
+
+        assert "Error reading file" not in read_content, "Failed to read valid UTF-8 file on Windows/CP1252 simulation"
+        assert poison_char in read_content, "Failed to correctly decode the special character"
+        assert read_content.endswith("End."), "Content appears truncated or malformed"