[CASSANDRA-16218] Simplify the almost duplicated code for calculating serialization size and serialization

• Type: Improvement
• Status: Resolved
• Priority: Normal
• Resolution: Abandoned
• Fix Version/s: 4.0, 4.0-beta3
• Component/s: Messaging/Client, Messaging/Internode
• Change Category: Code Clarity
• Complexity: Normal
• Platform: All
• Assignee: Yifan Cai Assign to me
• Reporter: Yifan Cai

Authors

Yifan Cai

• Created: 20/Oct/20 20:14
• Updated: 21/Dec/20 09:29
• Resolved: 29/Oct/20 16:38

Description

The current pattern of counting the serialization size and the actual serialization in the codebase is error-prone and hard to maintain. Those 2 code paths have almost the same code repeated. 
 
The pattern can be found in org.apache.cassandra.net.Message#Serializer and numerous locations that use org.apache.cassandra.transport.CBCodec. 
 
I am proposing a new approach that lets both methods share the same code path. The code basically looks like the below (in the case of org.apache.cassandra.net.Message#Serializer).

// A fake DataOutputPlus that simply increment the size when invoking write* methods
public class SizeCountingDataOutput implements DataOutputPlus
{
 private int size = 0;

 public int getSize()
 {
   return size;
 }

 public void write(int b)
 {
   size += 1;
 }

 public void write(byte[] b)
 {
   size += b.length;
 }
 ...
}

Therefore, in the size calculation, we can supply the fake data output and call serialize to get the size.

private <T> int serializedSize(Message<T> message, int version)
{
 SizeCountingDataOutput out = new SizeCountingDataOutput();
 try
 {
   serialize(message, out, version);
 }
 catch (IOException exception)
 {
   throw new AssertionError("It should not happen!", exception);
 }
 return out.getSize();
// The original implementation
// return version >= VERSION_40 ? serializedSizePost40(message, version) : serializedSizePre40(message, version);
}

Issue Links

• is related to: CASSANDRA-16103 Invalid serialized size for responses