From d51ccbf6ce2758dc24a16a5dc4ffcd113f9eb5cc Mon Sep 17 00:00:00 2001
From: xuu33030 <268775543+xuu33030@users.noreply.github.com>
Date: Sat, 19 Sep 2026 01:04:05 +0800
Subject: [PATCH] fix(cli): handle non-UTF-8 source files

---
 CHANGELOG.md           |  4 ++++
 maithili_dsl/cli.py    | 12 ++++++++++--
 tests/test_cli.py      | 29 +++++++++++++++++++++++++++++
 tests/test_security.py | 14 ++++++++++++++
 4 files changed, 57 insertions(+), 2 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index f310d88..430cc2d 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Fixed
+- Non-UTF-8 `.dmai` files now return a dedicated CLI error instead of a
+  raw Python traceback. Closes #27.
+
 ## [0.4.0] — 2026-07-23
 
 First release published to PyPI since 0.2.0 (0.3.0 was an internal
diff --git a/maithili_dsl/cli.py b/maithili_dsl/cli.py
index 6516e5a..0e31cde 100644
--- a/maithili_dsl/cli.py
+++ b/maithili_dsl/cli.py
@@ -103,6 +103,7 @@ def _validate_imports(python_code, translated_modules):
 EXIT_LINT_ERROR = 4
 EXIT_IMPORT_BLOCKED = 5
 EXIT_RUNTIME_ERROR = 6
+EXIT_BAD_ENCODING = 7
 
 
 def run_dmai_file(file_path):
@@ -116,8 +117,15 @@ def run_dmai_file(file_path):
         print(f"Error: File '{file_path}' not found.")
         return EXIT_NOT_FOUND
 
-    with open(path, 'r', encoding='utf-8') as file:
-        maithili_code = file.read()
+    try:
+        with open(path, 'r', encoding='utf-8') as file:
+            maithili_code = file.read()
+    except UnicodeDecodeError:
+        print("⚠️ त्रुटि: फाइल UTF-8 में नहि अछि।")
+        return EXIT_BAD_ENCODING
+    except OSError:
+        print("⚠️ त्रुटि: फाइल पढ़ल नहि जा सकल।")
+        return EXIT_NOT_FOUND
 
     errors = lint_maithili_code(maithili_code)
     if errors:
diff --git a/tests/test_cli.py b/tests/test_cli.py
index ce1d080..02bd71d 100644
--- a/tests/test_cli.py
+++ b/tests/test_cli.py
@@ -8,6 +8,7 @@
 import pytest
 
 from maithili_dsl.cli import (
+    EXIT_BAD_ENCODING,
     EXIT_IMPORT_BLOCKED,
     EXIT_LINT_ERROR,
     EXIT_NOT_FOUND,
@@ -30,6 +31,34 @@ def test_missing_file_returns_not_found(capsys):
     assert "not found" in captured.out.lower()
 
 
+def test_non_utf8_file_returns_bad_encoding(tmp_path, capsys):
+    path = tmp_path / "invalid.dmai"
+    path.write_bytes(b"\xff\xfe")
+
+    code = run_dmai_file(str(path))
+    captured = capsys.readouterr()
+
+    assert code == EXIT_BAD_ENCODING
+    assert "UTF-8" in captured.out
+    assert "Traceback" not in captured.out
+
+
+def test_unreadable_file_returns_not_found(tmp_path, monkeypatch, capsys):
+    path = tmp_path / "unreadable.dmai"
+    path.write_text("छपाउ(\"hello\")\n", encoding="utf-8")
+
+    def deny_read(*args, **kwargs):
+        raise PermissionError("private path details")
+
+    monkeypatch.setattr("builtins.open", deny_read)
+    code = run_dmai_file(str(path))
+    captured = capsys.readouterr()
+
+    assert code == EXIT_NOT_FOUND
+    assert "फाइल पढ़ल नहि जा सकल" in captured.out
+    assert "private path details" not in captured.out
+
+
 def test_lint_error_returns_lint_exit_code(tmp_dmai_file, capsys):
     # The leading = triggers the "'=' चिह्नक पहिले" lint error.
     path = tmp_dmai_file("= ५\n")
diff --git a/tests/test_security.py b/tests/test_security.py
index 8620de5..0a2c105 100644
--- a/tests/test_security.py
+++ b/tests/test_security.py
@@ -22,6 +22,7 @@
     _validate_imports,
     run_dmai_file,
     translate_imports,
+    EXIT_BAD_ENCODING,
     EXIT_OK,
     EXIT_IMPORT_BLOCKED,
     EXIT_RUNTIME_ERROR,
@@ -291,3 +292,16 @@ def test_actual_dangerous_code_outside_strings_still_blocked():
     transpiled = transpile_maithili_code(code)
     errors = _validate_imports(transpiled, translated_modules=set())
     assert errors, f"raw 'import os' must be blocked by import validator: {errors}"
+
+
+@pytest.mark.security
+def test_invalid_encoding_is_rejected_before_execution(tmp_path, capsys):
+    """Malformed source must fail closed without reaching the sandbox."""
+    path = tmp_path / "invalid.dmai"
+    path.write_bytes(b"\xff\xfe")
+
+    code = run_dmai_file(str(path))
+    captured = capsys.readouterr()
+
+    assert code == EXIT_BAD_ENCODING
+    assert "UTF-8" in captured.out