Sanha Lee

Results 19 issues of Sanha Lee

It seems that `FixtureMonkey` always returns `null` for an interface arbitrary builder. Is there any chance to support interface arbitrary builder? If it is hard to scan implementation candidates automatically,...

enhancement

At now, join operation is implemented very simply to reuse the implementation of windowing operation. However, we assumed that we would receive a lots of queries containing join operation because...

enhancement

For union operator, two streams to be unified must have same type. The problem is, they contains their output data type as a Java generic. At now, we provide a...

client

At now, the local kafka server and local zookeeper in `KafkaLocalBroker`, which is used in `KafkaSourceTest`, throws every logs to stdout. We need to change this logging way because of...

minor

At now, in `KafkaSourceTest`, there are a sleep code between the close of `KafkaSource` and the collpase of `LocalKafkaBroker`. This is because that `KafkaDataGenerator` waits until it's `pollTimeout` before it...

minor

At now, if there are many Kafka sources accessing single Kafka broker (with identical configuration), MIST creates separate Kafka data consumers to each of them. We might need to improve...

enhancement

We have to distribute the MIST master.

scale-out

If we union two conditional branches right after the branching, during the optimization, the conditional branch vertex should have two edges toward the union operator. However, this cannot be represented...

minor

At now, we do not delete an empty group. We need to implement the group deletion during query deletion.

We need to introduce a new query placement algorithm

optimization
query placement