Merge pull request #2415 from anatawa12/validate-bson-string

chore: throw exception when encounter unpaired surrogate instead of replace with U+FFFD
litedb-org · Feb 13, 2024 · 31e0ff7 · 31e0ff7
2 parents 4e856d8 + 289e9b1
commit 31e0ff7
Show file tree

Hide file tree

Showing 5 changed files with 26 additions and 14 deletions.
diff --git a/LiteDB/Document/BsonValue.cs b/LiteDB/Document/BsonValue.cs
@@ -648,7 +648,7 @@ internal virtual int GetBytesCount(bool recalc)
                 case BsonType.Double: return 8;
                 case BsonType.Decimal: return 16;
 
-                case BsonType.String: return Encoding.UTF8.GetByteCount(this.AsString);
+                case BsonType.String: return StringEncoding.UTF8.GetByteCount(this.AsString);
 
                 case BsonType.Binary: return this.AsBinary.Length;
                 case BsonType.ObjectId: return 12;
@@ -674,7 +674,7 @@ protected int GetBytesCountElement(string key, BsonValue value)
 
             return
                 1 + // element type
-                Encoding.UTF8.GetByteCount(key) + // CString
+                StringEncoding.UTF8.GetByteCount(key) + // CString
                 1 + // CString \0
                 value.GetBytesCount(true) +
                 (variant ? 5 : 0); // bytes.Length + 0x??

diff --git a/LiteDB/Engine/Disk/Serializer/BufferReader.cs b/LiteDB/Engine/Disk/Serializer/BufferReader.cs
@@ -154,7 +154,7 @@ public string ReadString(int count)
             // if fits in current segment, use inner array - otherwise copy from multiples segments
             if (_currentPosition + count <= _current.Count)
             {
-                value = Encoding.UTF8.GetString(_current.Array, _current.Offset + _currentPosition, count);
+                value = StringEncoding.UTF8.GetString(_current.Array, _current.Offset + _currentPosition, count);
 
                 this.MoveForward(count);
             }
@@ -165,7 +165,7 @@ public string ReadString(int count)
 
                 this.Read(buffer, 0, count);
 
-                value = Encoding.UTF8.GetString(buffer, 0, count);
+                value = StringEncoding.UTF8.GetString(buffer, 0, count);
 
                 BufferPool.Return(buffer);
             }
@@ -204,7 +204,7 @@ public string ReadCString()
 
                     this.MoveForward(1); // +1 to '\0'
 
-                    return Encoding.UTF8.GetString(mem.ToArray());
+                    return StringEncoding.UTF8.GetString(mem.ToArray());
                 }
             }
         }
@@ -220,7 +220,7 @@ private bool TryReadCStringCurrentSegment(out string value)
             {
                 if (_current[pos] == 0x00)
                 {
-                    value = Encoding.UTF8.GetString(_current.Array, _current.Offset + _currentPosition, count);
+                    value = StringEncoding.UTF8.GetString(_current.Array, _current.Offset + _currentPosition, count);
                     this.MoveForward(count + 1); // +1 means '\0'	
                     return true;
                 }

diff --git a/LiteDB/Engine/Disk/Serializer/BufferWriter.cs b/LiteDB/Engine/Disk/Serializer/BufferWriter.cs
@@ -152,13 +152,13 @@ public void WriteCString(string value)
         {
             if (value.IndexOf('\0') > -1) throw LiteException.InvalidNullCharInString();
 
-            var bytesCount = Encoding.UTF8.GetByteCount(value);
+            var bytesCount = StringEncoding.UTF8.GetByteCount(value);
             var available = _current.Count - _currentPosition; // avaiable in current segment
 
             // can write direct in current segment (use < because need +1 \0)
             if (bytesCount < available)
             {
-                Encoding.UTF8.GetBytes(value, 0, value.Length, _current.Array, _current.Offset + _currentPosition);
+                StringEncoding.UTF8.GetBytes(value, 0, value.Length, _current.Array, _current.Offset + _currentPosition);
 
                 _current[_currentPosition + bytesCount] = 0x00;
 
@@ -168,7 +168,7 @@ public void WriteCString(string value)
             {
                 var buffer = BufferPool.Rent(bytesCount);
 
-                Encoding.UTF8.GetBytes(value, 0, value.Length, buffer, 0);
+                StringEncoding.UTF8.GetBytes(value, 0, value.Length, buffer, 0);
 
                 this.Write(buffer, 0, bytesCount);
 
@@ -186,7 +186,7 @@ public void WriteCString(string value)
         /// </summary>
         public void WriteString(string value, bool specs)
         {
-            var count = Encoding.UTF8.GetByteCount(value);
+            var count = StringEncoding.UTF8.GetByteCount(value);
 
             if (specs)
             {
@@ -195,7 +195,7 @@ public void WriteString(string value, bool specs)
 
             if (count <= _current.Count - _currentPosition)
             {
-                Encoding.UTF8.GetBytes(value, 0, value.Length, _current.Array, _current.Offset + _currentPosition);
+                StringEncoding.UTF8.GetBytes(value, 0, value.Length, _current.Array, _current.Offset + _currentPosition);
 
                 this.MoveForward(count);
             }
@@ -204,7 +204,7 @@ public void WriteString(string value, bool specs)
                 // rent a buffer to be re-usable
                 var buffer = BufferPool.Rent(count);
 
-                Encoding.UTF8.GetBytes(value, 0, value.Length, buffer, 0);
+                StringEncoding.UTF8.GetBytes(value, 0, value.Length, buffer, 0);
 
                 this.Write(buffer, 0, count);
 

diff --git a/LiteDB/Engine/Structures/CollectionIndex.cs b/LiteDB/Engine/Structures/CollectionIndex.cs
@@ -121,8 +121,8 @@ public static int GetLength(string name, string expr)
             return
                 1 + // Slot
                 1 + // IndexType
-                Encoding.UTF8.GetByteCount(name) + 1 + // Name + \0
-                Encoding.UTF8.GetByteCount(expr) + 1 + // Expression + \0
+                StringEncoding.UTF8.GetByteCount(name) + 1 + // Name + \0
+                StringEncoding.UTF8.GetByteCount(expr) + 1 + // Expression + \0
                 1 + // Unique
                 PageAddress.SIZE + // Head
                 PageAddress.SIZE + // Tail

diff --git a/LiteDB/Utils/Encoding.cs b/LiteDB/Utils/Encoding.cs
@@ -0,0 +1,12 @@
+using System.Text;
+
+namespace LiteDB
+{
+    internal class StringEncoding
+    {
+        // Original Encoding.UTF8 will replace unpaired surrogate with U+FFFD, which is not suitable for database
+        // so, we need to use new UTF8Encoding(false, true) to make throw exception when unpaired surrogate is found
+        //public static System.Text.Encoding UTF8 = new UTF8Encoding(false, true);
+        public static Encoding UTF8 = new UTF8Encoding(false, true);
+    }
+}