LinShunKang commented on code in PR #12545:
URL: https://github.com/apache/kafka/pull/12545#discussion_r1174332358
##########
clients/src/main/java/org/apache/kafka/common/serialization/DoubleDeserializer.java:
##########
@@ -35,4 +41,22 @@ public Double deserialize(String topic, byte[] data) {
}
return Double.longBitsToDouble(value);
}
+
+ @Override
+ public Double deserialize(String topic, Headers headers, ByteBuffer data) {
+ if (data == null) {
+ return null;
+ }
+
+ if (data.remaining() != 8) {
+ throw new SerializationException("Size of data received by
DoubleDeserializer is not 8");
+ }
+
+ final ByteOrder srcOrder = data.order();
+ data.order(BIG_ENDIAN);
+
+ final double value = data.getDouble(data.position());
Review Comment:
> @LinShunKang
>
> > Because DoubleDeserializer and other Number Deserializers use BIG_ENDIAN
byte order to read from byte[], and the current byte order of ByteBuffer may
not be BIG_ENDIAN, we set the byte order of ByteBuffer to BIG_ENDIAN to be
consistent with the byte order used when reading from byte[].
>
> Again, from the
[javadoc](https://docs.oracle.com/javase/8/docs/api/java/nio/ByteBuffer.html#getDouble--):
>
> > Reads the next eight bytes at this buffer's current position, composing
them into a double value according to the current byte order, and then
increments the position by eight.
>
> The byte order is already considered while reading from the byte buffer.
That means, this is not true:
>
> > Because DoubleDeserializer and other Number Deserializers use BIG_ENDIAN
byte order to read from byte[]
>
> It will read from byte[] using the current byte order (BIG_ENDIAN or
LITTLE_ENDIAN). Did I miss anything?
In fact, the essence of this issue is which of the following methods we
should use for data reading:
1. Force users to write data into ByteBuffer in BIG_ENDIAN mode, and then we
read data from ByteBuffer in BIG_ENDIAN mode;
2. We read data according to the byte order of the ByteBuffer passed in by
the user by default.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: jira-unsubscr...@kafka.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
