Fix support for unicode paths in target-shell (fox-it#440)

This fixes ls and cat in target-shell when paths contain unicode characters.

Author: pyrco

dissect/target/tools/shell.py

@@ -171,6 +171,7 @@ def _exec(self, func: Callable[[list[str], TextIO], bool], command_args_str: str
         if command_args_str is not None:
             lexer = shlex.shlex(command_args_str, posix=True, punctuation_chars=True)
             lexer.wordchars += "$"
+            lexer.whitespace_split = True
             argparts = list(lexer)
 
         try:

tests/_data/tools/shell/unicode.tar

@@ -1,7 +1,7 @@
 import argparse
 import platform
 import sys
-from io import BytesIO
+from io import BytesIO, StringIO
 from pathlib import Path
 from unittest.mock import MagicMock
 
@@ -17,6 +17,8 @@
     build_pipe,
     build_pipe_stdout,
 )
+from dissect.target.tools.shell import main as target_shell
+from tests._utils import absolute_path
 
 GREP_MATCH = "test1 and test2"
 GREP_MISSING = "test2 alone"
@@ -247,3 +249,32 @@ def _map_function(path: Path) -> str:
     tree = "|".join(sorted(path_map))
 
     assert tree == expected
+
+
+@pytest.mark.parametrize(
+    "provided_input, expected_output",
+    [
+        ("hello", "world.txt"),  # Latin
+        ("ħēļľŏ", "ŵőřŀđ.txt"),  # Latin Extended-A
+        ("مرحبًا", "عالم.txt"),  # Arabic
+        ("你好", "世界.txt"),  # Chineese Simplified
+        ("привет", "мир.txt"),  # Cyrillic
+        ("🕵🕵🕵", "👀👀👀.txt"),  # Emoji
+    ],
+)
+def test_target_cli_unicode_argparse(
+    capsys: pytest.CaptureFixture,
+    monkeypatch: pytest.MonkeyPatch,
+    provided_input: str,
+    expected_output: str,
+) -> None:
+    with monkeypatch.context() as m:
+        target_file = absolute_path("_data/tools/shell/unicode.tar")
+        m.setattr("sys.argv", ["target-shell", target_file])
+        m.setattr("sys.stdin", StringIO(f"ls unicode/charsets/{provided_input}"))
+        target_shell()
+        out, err = capsys.readouterr()
+        out = out.replace("unicode.tar />", "").strip()
+
+        assert out == expected_output
+        assert "unrecognized arguments" not in err