chesnokoff commented on code in PR #12634:
URL: https://github.com/apache/ignite/pull/12634#discussion_r2730556935
##########
modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/TcpDiscoveryIoSession.java:
##########
@@ -163,63 +176,123 @@ void writeMessage(TcpDiscoveryAbstractMessage msg)
throws IgniteCheckedException
* @throws IgniteCheckedException If deserialization fails.
*/
<T> T readMessage() throws IgniteCheckedException, IOException {
- byte serMode = (byte)in.read();
+ return reader.readMessage();
+ }
- if (JAVA_SERIALIZATION == serMode)
- return U.unmarshal(spi.marshaller(), in, clsLdr);
+ /** Switches reader to fast mode after initial validated read. */
+ public void switchToFastReader() {
+ reader = fastReader;
+ }
- try {
- if (MESSAGE_SERIALIZATION != serMode) {
Review Comment:
Initially `detectJavaObjectStreamHeader` was placed after `detectSslAlert`.
The suggestion was to not make check for `serMode` after connection
establishments.
To do that we need to notify session that we do not need that checks. So now
higher levels of code call `switchToFastReader` for notification and we change
reader. As a result, we optimized one `if` check
But on the other hand for me it looks like really small optimization because
work with IO takes much more resources than one simple if. Moreover, it seems
to break DIP because now higher levels of code have to know details about
implementation of `TcpDiscoveryIoSession`
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
