From 3788139d26556b39896de329e6a2f2b2b1a2d406 Mon Sep 17 00:00:00 2001
From: deeplow <deeplower@protonmail.com>
Date: Fri, 28 Jul 2023 18:55:58 +0300
Subject: [PATCH] Add utility for sanitizing strings

Add `replace_control_chars()` function in `util.py`, which can be used
to sanitize strings from ANSI escape sequences or weird Unicode symbols.
---
 dangerzone/util.py | 10 ++++++++++
 tests/__init__.py  | 21 +++++++++++++++++++++
 tests/test_util.py | 11 ++++++++++-
 3 files changed, 41 insertions(+), 1 deletion(-)

diff --git a/dangerzone/util.py b/dangerzone/util.py
index c9b0be0..dc188f2 100644
--- a/dangerzone/util.py
+++ b/dangerzone/util.py
@@ -1,5 +1,6 @@
 import pathlib
 import platform
+import string
 import subprocess
 import sys
 from typing import Optional
@@ -62,3 +63,12 @@ def get_subprocess_startupinfo():  # type: ignore [no-untyped-def]
         return startupinfo
     else:
         return None
+
+
+def replace_control_chars(untrusted_str: str) -> str:
+    """Remove control characters from string. Protects a terminal emulator
+    from obcure control characters"""
+    sanitized_str = ""
+    for char in untrusted_str:
+        sanitized_str += char if char in string.printable else "_"
+    return sanitized_str
diff --git a/tests/__init__.py b/tests/__init__.py
index 32168ff..03bbb06 100644
--- a/tests/__init__.py
+++ b/tests/__init__.py
@@ -37,3 +37,24 @@ def unreadable_pdf(tmp_path: Path) -> str:
     file_path = tmp_path / "document.pdf"
     file_path.touch(mode=0o000)
     return str(file_path)
+
+
+@pytest.fixture
+def uncommon_text() -> str:
+    """Craft a string with Unicode characters that are considered not common.
+
+    Create a string that contains the following uncommon characters:
+
+    * ANSI escape sequences: \033[31;1;4m and \033[0m
+    * A Unicode character that resembles an English character: greek "X" (U+03A7)
+    * A Unicode control character that is not part of ASCII: zero-width joiner
+      (U+200D)
+    * An emoji: Cross Mark (U+274C)
+    """
+    return "\033[31;1;4m BaD TeΧt \u200d ❌ \033[0m"
+
+
+@pytest.fixture
+def sanitized_text() -> str:
+    """Return a sanitized version of the uncommon_text."""
+    return "_[31;1;4m BaD Te_t _ _ _[0m"
diff --git a/tests/test_util.py b/tests/test_util.py
index e0cf2b6..04eecb2 100644
--- a/tests/test_util.py
+++ b/tests/test_util.py
@@ -4,7 +4,9 @@ from pathlib import Path
 
 import pytest
 
-import dangerzone.util as util
+from dangerzone import util
+
+from . import sanitized_text, uncommon_text
 
 VERSION_FILE_NAME = "version.txt"
 
@@ -21,3 +23,10 @@ def test_get_resource_path() -> None:
 def test_get_subprocess_startupinfo() -> None:
     startupinfo = util.get_subprocess_startupinfo()
     assert isinstance(startupinfo, subprocess.STARTUPINFO)  # type: ignore[attr-defined]
+
+
+def test_replace_control_chars(uncommon_text: str, sanitized_text: str) -> None:
+    """Test that the replace_control_chars() function works properly."""
+    assert util.replace_control_chars(uncommon_text) == sanitized_text
+    assert util.replace_control_chars("normal text") == "normal text"
+    assert util.replace_control_chars("") == ""