Add Thrift string size limit reader property (#564)

Author: haianhng31

cpp/ReaderProperties.cpp

@@ -84,4 +84,14 @@ extern "C"
 			*memory_pool = reader_properties->memory_pool();
 		)
 	}
+
+	PARQUETSHARP_EXPORT ExceptionInfo *ReaderProperties_Thrift_String_Size_Limit(const ReaderProperties *reader_properties, int *size)
+	{
+		TRYCATCH(*size = reader_properties->thrift_string_size_limit();)
+	}
+
+	PARQUETSHARP_EXPORT ExceptionInfo *ReaderProperties_Set_Thrift_String_Size_Limit(ReaderProperties *reader_properties, int size)
+	{
+		TRYCATCH(reader_properties->set_thrift_string_size_limit(size);)
+	}
 }

csharp.test/TestReaderProperties.cs

@@ -1,4 +1,5 @@
 using NUnit.Framework;
+using ParquetSharp.IO;
 
 namespace ParquetSharp.Test
 {
@@ -35,6 +36,45 @@ public static void TestModifyProperties()
             Assert.That(p.IsBufferedStreamEnabled, Is.True);
             p.DisableBufferedStream();
             Assert.That(p.IsBufferedStreamEnabled, Is.False);
+
+            p.SetThriftStringSizeLimit(2048576);
+            Assert.That(p.ThriftStringSizeLimit, Is.EqualTo(2048576));
+        }
+
+        [Test]
+        public static void TestSetThriftStringSizeLimit_ReturnException()
+        {
+            using var buffer = new ResizableBuffer();
+
+            using (var output = new BufferOutputStream(buffer))
+            {
+                var longColumnName = new string('X', 100); // 100 chars
+                var schema = new Column[] { new Column<string>(longColumnName) };
+
+                using var writer = new ParquetFileWriter(output, schema);
+                using (var rowGroup = writer.AppendRowGroup())
+                {
+                    using var colWriter = rowGroup.NextColumn().LogicalWriter<string>();
+                    colWriter.WriteBatch(new[] { "hello" });
+                }
+                writer.Close();
+            }
+
+            // Configure reader with a small thrift string size limit
+            using var props = ReaderProperties.GetDefaultReaderProperties();
+            props.SetThriftStringSizeLimit(10);
+
+            var ex = Assert.Throws<ParquetException>(() =>
+            {
+                using var input = new BufferReader(buffer);
+                using var reader = new ParquetFileReader(input, props);
+                var rg = reader.RowGroup(0); // Force metadata read
+            });
+
+            // Validate the exception is related to the thrift string size limit
+            Assert.That(ex?.Message,
+                Does.Contain("Couldn't deserialize thrift: TProtocolException: Exceeded size limit")
+                    .IgnoreCase);
         }
 
         [TestCaseSource(typeof(MemoryPools), nameof(MemoryPools.NonNullTestCases))]

csharp/PublicAPI.Unshipped.txt

@@ -1 +1,3 @@
 #nullable enable
+ParquetSharp.ReaderProperties.ThriftStringSizeLimit.get -> int
+ParquetSharp.ReaderProperties.SetThriftStringSizeLimit(int size) -> void

csharp/ReaderProperties.cs

@@ -137,6 +137,26 @@ public MemoryPool MemoryPool
             }
         }
 
+        /// <summary>
+        /// Return the size limit on thrift strings.
+        ///
+        /// This limit helps prevent space and time bombs in files, 
+        /// but may need to be increased in order to read files with especially large headers.
+        /// </summary>
+        public int ThriftStringSizeLimit
+        {
+            get => ExceptionInfo.Return<int>(Handle, ReaderProperties_Thrift_String_Size_Limit);
+        }
+
+        /// <summary>
+        /// Set the size limit on thrift strings.
+        /// </summary>
+        public void SetThriftStringSizeLimit(int size)
+        {
+            ExceptionInfo.Check(ReaderProperties_Set_Thrift_String_Size_Limit(Handle.IntPtr, size));
+            GC.KeepAlive(Handle);
+        }
+
         [DllImport(ParquetDll.Name)]
         private static extern IntPtr ReaderProperties_Get_Default_Reader_Properties(out IntPtr readerProperties);
 
@@ -179,6 +199,12 @@ public MemoryPool MemoryPool
         [DllImport(ParquetDll.Name)]
         private static extern IntPtr ReaderProperties_Get_Memory_Pool(IntPtr readerProperties, out IntPtr memoryPool);
 
+        [DllImport(ParquetDll.Name)]
+        private static extern IntPtr ReaderProperties_Thrift_String_Size_Limit(IntPtr readerProperties, out int size);
+
+        [DllImport(ParquetDll.Name)]
+        private static extern IntPtr ReaderProperties_Set_Thrift_String_Size_Limit(IntPtr readerProperties, int size);
+
         internal readonly ParquetHandle Handle;
     }
 }