fix(ciphers): operate on raw bytes in base32 to fix corruption and empty encoding

ciphers/base32.py

@@ -2,42 +2,83 @@
 Base32 encoding and decoding
 
 https://en.wikipedia.org/wiki/Base32
+https://www.rfc-editor.org/rfc/rfc4648.html#section-6
 """
 
 B32_CHARSET = "ABCDEFGHIJKLMNOPQRSTUVWXYZ234567"
 
 
 def base32_encode(data: bytes) -> bytes:
     """
+    Encode bytes into Base32 bytes per RFC 4648.
+
+    >>> base32_encode(b"")
+    b''
     >>> base32_encode(b"Hello World!")
     b'JBSWY3DPEBLW64TMMQQQ===='
     >>> base32_encode(b"123456")
     b'GEZDGNBVGY======'
     >>> base32_encode(b"some long complex string")
     b'ONXW2ZJANRXW4ZZAMNXW24DMMV4CA43UOJUW4ZY='
+    >>> base32_encode(b"\\xff\\xfe\\x01")
+    b'777AC==='
+    >>> import base64
+    >>> all(
+    ...     base32_encode(bytes([i])) == base64.b32encode(bytes([i]))
+    ...     for i in range(256)
+    ... )
+    True
     """
-    binary_data = "".join(bin(ord(d))[2:].zfill(8) for d in data.decode("utf-8"))
-    binary_data = binary_data.ljust(5 * ((len(binary_data) // 5) + 1), "0")
-    b32_chunks = map("".join, zip(*[iter(binary_data)] * 5))
+    if not isinstance(data, (bytes, bytearray)):
+        msg = f"data must be bytes or bytearray, not {type(data).__name__}"
+        raise TypeError(msg)
+
+    if not data:
+        return b""
+
+    binary_data = "".join(bin(byte)[2:].zfill(8) for byte in data)
+    pad_len = (5 - len(binary_data) % 5) % 5
+    binary_data += "0" * pad_len
+    b32_chunks = [binary_data[i : i + 5] for i in range(0, len(binary_data), 5)]
     b32_result = "".join(B32_CHARSET[int(chunk, 2)] for chunk in b32_chunks)
-    return bytes(b32_result.ljust(8 * ((len(b32_result) // 8) + 1), "="), "utf-8")
+    pad_chars = (8 - len(b32_result) % 8) % 8
+    return (b32_result + "=" * pad_chars).encode("ascii")
 
 
 def base32_decode(data: bytes) -> bytes:
     """
+    Decode Base32 bytes back into raw bytes.
+
+    >>> base32_decode(b"")
+    b''
     >>> base32_decode(b'JBSWY3DPEBLW64TMMQQQ====')
     b'Hello World!'
     >>> base32_decode(b'GEZDGNBVGY======')
     b'123456'
     >>> base32_decode(b'ONXW2ZJANRXW4ZZAMNXW24DMMV4CA43UOJUW4ZY=')
     b'some long complex string'
+    >>> base32_decode(b'777AC===')
+    b'\\xff\\xfe\\x01'
+    >>> import base64
+    >>> all(
+    ...     base32_decode(base64.b32encode(bytes([i]))) == bytes([i])
+    ...     for i in range(256)
+    ... )
+    True
     """
+    if not isinstance(data, (bytes, bytearray)):
+        msg = f"data must be bytes or bytearray, not {type(data).__name__}"
+        raise TypeError(msg)
+
+    cleaned = data.decode("ascii").rstrip("=")
+    if not cleaned:
+        return b""
+
     binary_chunks = "".join(
-        bin(B32_CHARSET.index(_d))[2:].zfill(5)
-        for _d in data.decode("utf-8").strip("=")
+        bin(B32_CHARSET.index(char))[2:].zfill(5) for char in cleaned
     )
-    binary_data = list(map("".join, zip(*[iter(binary_chunks)] * 8)))
-    return bytes("".join([chr(int(_d, 2)) for _d in binary_data]), "utf-8")
+    num_bytes = len(binary_chunks) // 8
+    return bytes(int(binary_chunks[i * 8 : (i + 1) * 8], 2) for i in range(num_bytes))
 
 
 if __name__ == "__main__":