Fix destructive file header detection (#835)

Deleted: broad unanchored header matching that removed source content.

actions/__init__.py

@@ -28,4 +28,4 @@
 #     ├── test_summarize_pr.py
 #     └── ...
 
-__version__ = "0.2.36"
+__version__ = "0.2.37"

actions/update_file_headers.py

@@ -120,18 +120,32 @@ def update_file(file_path, prefix, block_start, block_end, base_header):
 
     # Check for special first line
     special_line_index = -1
-    if lines and (lines[0].startswith("#!") or lines[0].startswith("<?xml") or lines[0].startswith("<!DOCTYPE")):
+    first_line = lines[0].lstrip("\ufeff") if lines else ""
+    encoding_cookie = r"^[ \t\f]*#.*?coding[:=][ \t]*[-_.a-zA-Z0-9]+"
+    if lines and (
+        first_line.startswith(("#!", "<?xml"))
+        or first_line.lower().startswith("<!doctype")
+        or re.match(encoding_cookie, first_line)
+    ):
         special_line_index = 0
         prefix_lines.append(lines[0])
+    if first_line.startswith("#!") and len(lines) > 1 and re.match(encoding_cookie, lines[1]):
+        special_line_index = 1
+        prefix_lines.append(lines[1])
 
     start_idx = special_line_index + 1 if special_line_index >= 0 else 0
     end_idx = min(start_idx + 5, len(lines))  # Look in first few lines
 
-    # Find existing header
-    header_index = next(
-        (i for i in range(start_idx, end_idx) if re.search(r"AGPL-3.0|CONFIDENTIAL|Ultralytics|©\s*\d{4}", lines[i])),
-        -1,
+    # An existing header must be the first non-empty line and use the file's comment syntax.
+    candidate_index = next((i for i in range(start_idx, end_idx) if lines[i].strip()), -1)
+    candidate = lines[candidate_index].strip() if candidate_index >= 0 else ""
+    comment_start = (prefix or block_start or "# ").strip()
+    known_header = candidate.startswith(comment_start) and re.search(
+        r"AGPL-3\.0 License|CONFIDENTIAL: Unauthorized use|©\s*2014[–-]\d{4}\s+Ultralytics Inc\.|"
+        r"Ultralytics Inc\..*Copyright ©\s*2014[–-]\d{4}\s*-\s*CONFIDENTIAL\s*-",
+        candidate,
     )
+    header_index = candidate_index if candidate == formatted_header.strip() or known_header else -1
 
     # Add the formatted header to prefix lines
     prefix_lines.append(formatted_header)

tests/test_file_headers.py

@@ -76,6 +76,73 @@ def test_update_file_with_shebang():
         assert "print('Hello World')" in content
 
 
+def test_update_file_preserves_early_ultralytics_source():
+    """Test that Ultralytics references in source code are not mistaken for headers."""
+    with TemporaryDirectory() as tmp_dir:
+        tmp_path = Path(tmp_dir)
+        docstring_file = tmp_path / "api.py"
+        docstring_file.write_text('"""Ultralytics Platform API client."""\n\nVALUE = 1\n')
+        import_file = tmp_path / "__init__.py"
+        import_file.write_text("from ultralytics import YOLO\n")
+
+        for test_file, spacing in ((docstring_file, ""), (import_file, "\n")):
+            original = test_file.read_text()
+            assert update_file(test_file, "# ", None, None, "Ultralytics 🚀 Test Header") is True
+            assert test_file.read_text() == f"# Ultralytics 🚀 Test Header\n{spacing}{original}"
+
+
+def test_update_file_with_lowercase_doctype():
+    """Test that Prettier's lowercase HTML doctype remains first."""
+    with TemporaryDirectory() as tmp_dir:
+        test_file = Path(tmp_dir) / "index.html"
+        test_file.write_text("<!doctype html>\n<html></html>\n")
+
+        assert update_file(test_file, None, "<!-- ", " -->", "Ultralytics 🚀 Test Header") is True
+        assert test_file.read_text().startswith("<!doctype html>\n<!-- Ultralytics 🚀 Test Header -->\n\n")
+
+
+def test_update_file_with_bom_xml_declaration():
+    """Test that a BOM-prefixed XML declaration remains first."""
+    with TemporaryDirectory() as tmp_dir:
+        test_file = Path(tmp_dir) / "layout.xml"
+        test_file.write_text('\ufeff<?xml version="1.0" encoding="utf-8"?>\n<root />\n')
+
+        assert update_file(test_file, None, "<!-- ", " -->", "Ultralytics 🚀 Test Header") is True
+        assert test_file.read_text().startswith(
+            '\ufeff<?xml version="1.0" encoding="utf-8"?>\n<!-- Ultralytics 🚀 Test Header -->\n\n'
+        )
+
+
+def test_update_file_with_shebang_and_encoding_cookie():
+    """Test that a Python encoding cookie remains on line two and the header stays idempotent."""
+    with TemporaryDirectory() as tmp_dir:
+        test_file = Path(tmp_dir) / "script.py"
+        test_file.write_text(
+            "#!/usr/bin/env python3\n# -*- coding: utf-8 -*-\n# Ultralytics 🚀 AGPL-3.0 License\n\nprint('hello')\n"
+        )
+
+        assert update_file(test_file, "# ", None, None, "Ultralytics 🚀 Test Header") is True
+        assert test_file.read_text().startswith(
+            "#!/usr/bin/env python3\n# -*- coding: utf-8 -*-\n# Ultralytics 🚀 Test Header\n\n"
+        )
+        assert update_file(test_file, "# ", None, None, "Ultralytics 🚀 Test Header") is False
+
+
+def test_update_file_replaces_legacy_private_header():
+    """Test that the previously emitted private header is replaced rather than duplicated."""
+    with TemporaryDirectory() as tmp_dir:
+        test_file = Path(tmp_dir) / "app.py"
+        test_file.write_text(
+            "# Ultralytics Inc. 🚀 Copyright © 2014-2025 - CONFIDENTIAL - "
+            "https://ultralytics.com - All Rights Reserved\n\n"
+            "print('hello')\n"
+        )
+
+        assert update_file(test_file, "# ", None, None, "© 2014-2026 Ultralytics Inc. CONFIDENTIAL") is True
+        assert test_file.read_text().startswith("# © 2014-2026 Ultralytics Inc. CONFIDENTIAL\n\n")
+        assert "Copyright" not in test_file.read_text()
+
+
 def test_update_file_no_changes():
     """Test updating file with no changes needed."""
     with TemporaryDirectory() as tmp_dir: