GH-704: Fix initialization of offset buffer when exporting VarChar vectors through C Data Interface (#705)

## What's Changed

This patch fixes the initialization of offset buffers when exporting
variable width arrays through Arrow C Data Interface. The original code
incorrectly mess up with the member `this.offsetBuffer` while we should
actually initialize the newly allocated offsetBuffer. I think the diff
itself will be quite self-explanatory.

Closes #704 and probably #88 .

Author: Kontinuation

c/src/test/java/org/apache/arrow/c/RoundtripTest.java

@@ -38,6 +38,8 @@
 import org.apache.arrow.memory.BufferAllocator;
 import org.apache.arrow.memory.RootAllocator;
 import org.apache.arrow.memory.util.hash.ArrowBufHasher;
+import org.apache.arrow.vector.BaseLargeVariableWidthVector;
+import org.apache.arrow.vector.BaseVariableWidthVector;
 import org.apache.arrow.vector.BigIntVector;
 import org.apache.arrow.vector.BitVector;
 import org.apache.arrow.vector.DateDayVector;
@@ -181,6 +183,13 @@ boolean roundtrip(FieldVector vector, Class<?> clazz) {
           clazz.isInstance(imported),
           String.format("expected %s but was %s", clazz, imported.getClass()));
       result = VectorEqualsVisitor.vectorEquals(vector, imported);
+
+      if (imported instanceof BaseVariableWidthVector
+          || imported instanceof BaseLargeVariableWidthVector) {
+        ArrowBuf offsetBuffer = imported.getOffsetBuffer();
+        assertTrue(offsetBuffer.capacity() > 0);
+        assertEquals(0, offsetBuffer.getInt(0));
+      }
     }
 
     // Check that the ref counts of the buffers are the same after the roundtrip
@@ -602,6 +611,13 @@ public void testVarCharVector() {
     }
   }
 
+  @Test
+  public void testEmptyVarCharVector() {
+    try (final VarCharVector vector = new VarCharVector("v", allocator)) {
+      assertTrue(roundtrip(vector, VarCharVector.class));
+    }
+  }
+
   @Test
   public void testLargeVarBinaryVector() {
     try (final LargeVarBinaryVector vector = new LargeVarBinaryVector("", allocator)) {
@@ -635,6 +651,13 @@ public void testLargeVarCharVector() {
     }
   }
 
+  @Test
+  public void testEmptyLargeVarCharVector() {
+    try (final LargeVarCharVector vector = new LargeVarCharVector("v", allocator)) {
+      assertTrue(roundtrip(vector, LargeVarCharVector.class));
+    }
+  }
+
   @Test
   public void testListVector() {
     try (final ListVector vector = ListVector.empty("v", allocator)) {

vector/src/main/java/org/apache/arrow/vector/BaseLargeVariableWidthVector.java

@@ -496,7 +496,7 @@ private void allocateBytes(final long valueBufferSize, final int valueCount) {
   private ArrowBuf allocateOffsetBuffer(final long size) {
     ArrowBuf offsetBuffer = allocator.buffer(size);
     offsetBuffer.readerIndex(0);
-    initOffsetBuffer();
+    offsetBuffer.setZero(0, offsetBuffer.capacity());
     return offsetBuffer;
   }
 

vector/src/main/java/org/apache/arrow/vector/BaseVariableWidthVector.java

@@ -514,7 +514,7 @@ private ArrowBuf allocateOffsetBuffer(final long size) {
     final int curSize = (int) size;
     ArrowBuf offsetBuffer = allocator.buffer(curSize);
     offsetBuffer.readerIndex(0);
-    initOffsetBuffer();
+    offsetBuffer.setZero(0, offsetBuffer.capacity());
     return offsetBuffer;
   }