Github user uce commented on the pull request:
https://github.com/apache/incubator-flink/pull/254#issuecomment-66360684
Thanks for the feedback so far. :-) I replied inline where possible.
**NettyMessage**: that's a very good idea. I will also use the String
serialization utilities you suggested.
**IntermediateResultManager**: I agree, but left it as it is as it doesn't
hurt yet. I will change it.
**BufferFuture**: I initially thought that we might use them in a case like
that. But not at the moment and therefore it indeed complicates things. I will
revert that as well. ;)
**Buffer references**: we don't use it at the moment, but we will need that
in any case for the "adaptive" spilling or multiple consumers. It would be OK
to remove it for time being. It would be easy to add back.
**Exceptions**: I will remove the stringify calls and add proper messages.
Maybe we should completely remove them from the other code as well?
**PipelinedPartitionQueue**: your proposed name is better. I have a simple
synchronized version as well, which should work fine for the single consumer
case.
Can you have a special look at the partition queue interface? We need some
kind of asynchrony (subscribe-notify). Maybe you have another idea about how to
expose this.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
