[MLIR][Bytecode] Add missing field initializer in constructor initializer list

Leaving this field unitialized could led to crashes when it'll diverge from the
IRNumbering phase.

Differential Revision: https://reviews.llvm.org/D156965

Author: joker-eph

mlir/lib/Bytecode/Writer/BytecodeWriter.cpp

@@ -344,8 +344,8 @@ class DialectWriter : public DialectBytecodeWriter {
   DialectWriter(int64_t bytecodeVersion, EncodingEmitter &emitter,
                 IRNumberingState &numberingState,
                 StringSectionBuilder &stringSection)
-      : emitter(emitter), numberingState(numberingState),
-        stringSection(stringSection) {}
+      : bytecodeVersion(bytecodeVersion), emitter(emitter),
+        numberingState(numberingState), stringSection(stringSection) {}
 
   //===--------------------------------------------------------------------===//
   // IR

mlir/test/Bytecode/operand_segment_sizes.mlir

@@ -0,0 +1,8 @@
+// RUN: mlir-opt -emit-bytecode %s | mlir-opt | FileCheck %s
+
+
+func.func @roundtripOperandSizeAttr(%arg0: i32) {
+  // CHECK: odsOperandSegmentSizes = array<i32: 0, 2, 1, 1>}>
+  "test.attr_sized_operands"(%arg0, %arg0, %arg0, %arg0) <{odsOperandSegmentSizes = array<i32: 0, 2, 1, 1>}> : (i32, i32, i32, i32) -> ()
+  return
+}