Document why this isn't merged back into capnproto-java (yet?)

[Flowdalic]

I am maybe missing something, but I wonder why this hasn't been merged back into capnproto-java. Is there a particular reason for that? Or is it planned? It sure would be great to have the answers to this questions documented in the project' README (or somewhere else).

[keturn]

https://github.com/vaci/capnproto-java-rpc/blob/master/website/index.md#future-work provides some information on this, but I do not know if all of those things are necessary for the minimum viable product.

[vaci]

Hi,

Thank you for your interest!

I would be very pleased for this RPC implementation to be merged into the mainline repository, but unfortunately I don't have have enough free time to contribute to the project at the moment, and this implementation is sufficient for my limited requirements.

The impact on the existing codebase is fairly minimal, and mostly involves adding a capability table member to various data structures, and extending the (C++) code generator.

There are a few issues that need to be addressed before the implementation could be considered production quality.

• The code generator currently generates incorrect Java code for generic interfaces. This is fixable, but the generator is not particularly fun to extend. It might be sufficient to use AnyPointers in place of generic parameters as an intermediate solution.
• The event loop implementation (i.e. RpcSystem::runOnce) is a hack, and a cleaner integration with CompletableFuture would be nice.
• There are a couple of places where the implementation relied on CompletableFuture features that were only introduced in Java 17, and the changes to make the code build under Java 11 need work: 950ba82