Merge pull request #2620 from sopel-irc/irc.utils.safe-remove-nulls

irc.utils, test: also remove null (\x00) in safe()

sopel/irc/utils.py

@@ -17,34 +17,40 @@
 from sopel.lifecycle import deprecated
 
 
-def safe(string):
-    """Remove newlines from a string.
+def safe(string: str) -> str:
+    """Remove disallowed bytes from a string, and ensure Unicode.
 
-    :param str string: input text to process
-    :return: the string without newlines
-    :rtype: str
+    :param string: input text to process
+    :return: the string as Unicode without characters prohibited in IRC messages
     :raises TypeError: when ``string`` is ``None``
 
-    This function removes newlines from a string and always returns a unicode
-    string (``str``), but doesn't strip or alter it in any other way::
+    This function removes newlines and null bytes from a string. It will always
+    return a Unicode ``str``, even if given non-Unicode input, but doesn't strip
+    or alter the string in any other way::
 
-        >>> safe('some text\\r\\n')
+        >>> safe('some \\x00text\\r\\n')
         'some text'
 
-    This is useful to ensure a string can be used in a IRC message.
+    This is useful to ensure a string can be used in a IRC message. Parameters
+    can **never** contain NUL, CR, or LF octets, per :rfc:`2812#section-2.3.1`.
 
     .. versionchanged:: 7.1
 
         This function now raises a :exc:`TypeError` instead of an unpredictable
         behaviour when given ``None``.
 
+    .. versionchanged:: 8.0.1
+
+        Also remove NUL (``\\x00``) in addition to CR/LF.
+
     """
     if string is None:
         raise TypeError('safe function requires a string, not NoneType')
     if isinstance(string, bytes):
         string = string.decode("utf8")
     string = string.replace('\n', '')
     string = string.replace('\r', '')
+    string = string.replace('\x00', '')
     return string
 
 

test/irc/test_irc_utils.py

@@ -1,43 +1,52 @@
 """Tests for core ``sopel.irc.utils``"""
 from __future__ import annotations
 
+from itertools import permutations
+
 import pytest
 
 from sopel.irc import utils
 
 
-def test_safe():
+@pytest.mark.parametrize('s1, s2, s3', permutations(('\n', '\r', '\x00')))
+def test_safe(s1, s2, s3):
     text = 'some text'
-    assert utils.safe(text + '\r\n') == text
-    assert utils.safe(text + '\n') == text
-    assert utils.safe(text + '\r') == text
-    assert utils.safe('\r\n' + text) == text
-    assert utils.safe('\n' + text) == text
-    assert utils.safe('\r' + text) == text
-    assert utils.safe('some \r\ntext') == text
-    assert utils.safe('some \ntext') == text
-    assert utils.safe('some \rtext') == text
+    seq = ''.join((s1, s2, s3))
+
+    assert utils.safe(text + seq) == text
+    assert utils.safe(seq + text) == text
+    assert utils.safe('some ' + seq + 'text') == text
+    assert utils.safe(
+        s1
+        + 'some '
+        + s2
+        + 'text'
+        + s3
+    ) == text
 
 
 def test_safe_empty():
     text = ''
     assert utils.safe(text) == text
 
 
-def test_safe_null():
+def test_safe_none():
     with pytest.raises(TypeError):
         utils.safe(None)
 
 
-def test_safe_bytes():
+@pytest.mark.parametrize('b1, b2, b3', permutations((b'\n', b'\r', b'\x00')))
+def test_safe_bytes(b1, b2, b3):
     text = b'some text'
-    assert utils.safe(text) == text.decode('utf-8')
-    assert utils.safe(text + b'\r\n') == text.decode('utf-8')
-    assert utils.safe(text + b'\n') == text.decode('utf-8')
-    assert utils.safe(text + b'\r') == text.decode('utf-8')
-    assert utils.safe(b'\r\n' + text) == text.decode('utf-8')
-    assert utils.safe(b'\n' + text) == text.decode('utf-8')
-    assert utils.safe(b'\r' + text) == text.decode('utf-8')
-    assert utils.safe(b'some \r\ntext') == text.decode('utf-8')
-    assert utils.safe(b'some \ntext') == text.decode('utf-8')
-    assert utils.safe(b'some \rtext') == text.decode('utf-8')
+    seq = b''.join((b1, b2, b3))
+
+    assert utils.safe(text + seq) == text.decode('utf-8')
+    assert utils.safe(seq + text) == text.decode('utf-8')
+    assert utils.safe(b'some ' + seq + b'text') == text.decode('utf-8')
+    assert utils.safe(
+        b1
+        + b'some '
+        + b2
+        + b'text'
+        + b3
+    ) == text.decode('utf-8')