From b287bd01836357fa6358c1eac102bd1db8b05a11 Mon Sep 17 00:00:00 2001
From: Rossen <rossengeorgiev@users.noreply.github.com>
Date: Wed, 24 Jul 2019 17:55:40 +0100
Subject: [PATCH 1/2] fix #20; replace invalid unicode chars instead of raising
 UnicodeDecodeError

---
 vdf/__init__.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/vdf/__init__.py b/vdf/__init__.py
index 6c793c6..6620c1a 100644
--- a/vdf/__init__.py
+++ b/vdf/__init__.py
@@ -300,14 +300,14 @@ def read_string(s, idx, wide=False):
             raise SyntaxError("Unterminated cstring, index: %d" % idx)
         result = s[idx:end]
         if wide:
-            result = result.decode('utf-16')
+            result = result.decode('utf-16', 'replace')
         elif bytes is not str:
-            result = result.decode('utf-8')
+            result = result.decode('utf-8', 'replace')
         else:
             try:
                 result.decode('ascii')
             except:
-                result = result.decode('utf-8')
+                result = result.decode('utf-8', 'replace')
         return result, end + (2 if wide else 1)
 
     stack = [mapper()]

From 01d63dd38b3825d565fbf6f6e4b539aa2eac51f9 Mon Sep 17 00:00:00 2001
From: Rossen Georgiev <rossen@rgp.io>
Date: Fri, 26 Jul 2019 23:48:39 +0100
Subject: [PATCH 2/2] binary: utf16 decode + tests

---
 tests/test_binary_vdf.py |  5 +++++
 vdf/__init__.py          | 10 ++++++++--
 2 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/tests/test_binary_vdf.py b/tests/test_binary_vdf.py
index f64d10e..6ef93ea 100644
--- a/tests/test_binary_vdf.py
+++ b/tests/test_binary_vdf.py
@@ -123,3 +123,8 @@ def test_vbkv_loads_invalid_checksum(self):
         with self.assertRaises(ValueError):
             vdf.vbkv_loads(b'VBKV\x01\x02\x03\x04\x00a\x00\x0b\x0b')
 
+    def test_loads_utf8_invalmid(self):
+        self.assertEqual({'aaa': b'bb\xef\xbf\xbdbb'.decode('utf-8')}, vdf.binary_loads(b'\x01aaa\x00bb\xffbb\x00\x08'))
+
+    def test_loads_utf16(self):
+        self.assertEqual({'aaa': b'b\x00b\x00\xff\xffb\x00b\x00'.decode('utf-16le')}, vdf.binary_loads(b'\x05aaa\x00b\x00b\x00\xff\xffb\x00b\x00\x00\x00\x08'))
diff --git a/vdf/__init__.py b/vdf/__init__.py
index 6620c1a..74fdd83 100644
--- a/vdf/__init__.py
+++ b/vdf/__init__.py
@@ -295,12 +295,18 @@ def binary_loads(s, mapper=dict, merge_duplicate_keys=True, alt_format=False):
     float32 = struct.Struct('<f')
 
     def read_string(s, idx, wide=False):
-        end = s.find(b'\x00\x00' if wide else b'\x00', idx)
+        if wide:
+            end = s.find(b'\x00\x00', idx)
+            if (end - idx) % 2 != 0:
+                end += 1
+        else:
+            end = s.find(b'\x00', idx)
+
         if end == -1:
             raise SyntaxError("Unterminated cstring, index: %d" % idx)
         result = s[idx:end]
         if wide:
-            result = result.decode('utf-16', 'replace')
+            result = result.decode('utf-16')
         elif bytes is not str:
             result = result.decode('utf-8', 'replace')
         else: