Yuto Kawamura

Yeah.. I know few users who've tried to do that and turns out it can't. While the needs of this feature is totally understandable, as currently decaton uses one consumer...

Ugh, that's true. Then solution here would be either one of: * Decaton supports retry topic in a different cluster or * #118 + option to disable retry "consumer" in...

Thanks for starting this 👍 Read your proposal and can agree with most of the parts. Some comments inline: > Rename DecatonTaskRequest protobuf message to LegacyDecatonTaskRequest and marked it as...

> By the way maybe bit off-topic, will adding a method to TaskExtractor be breaking change ? Not sure if we should really avoid that, but we might possibly workaround...

Can you rebase onto latest master branch then? I'll start looking into detail then.

`wasmtime-java` uses Slf4j for all loggings so depending on which logger backend (log4j, logback, etc) is used in your application, you should be able to configure your backend to suppress...

Yeah, because of `JObject`s are expected to be used in JNI methods locally in their callstack, it doesn't implement `Send` hence it cannot be passed to any other contexts which...

Can you try spotting which `JNIEnv` call exactly causing NPE? btw I saw your code uses `byte_array_from_slice` for Read implementation as well https://github.com/Superice666/wasmtime-java/blob/4765caef3ff51a5dc31ca14e11ef32d54b385b48/wasmtime-jni/src/wstream.rs#L47 , but I think it doesn't work...

It's easy to support if github actions supports that arch. Perhaps we can use this? https://github.com/marketplace/actions/run-on-architecture

I think that's because your wasm module is recognized as a "command" rather than a "reactor". Please read https://github.com/WebAssembly/WASI/blob/2ccb606926bee332a7b0a486e150d1053a493977/design/application-abi.md#current-unstable-abi this document for the detail. In short, when your wasm module...