Sean Gilligan

> Does that mean we could implement proxy support for NIO, them remove the blocking I/O implementation? I don't know. We'd have to look into it. > > > create...

I'm hoping that after we migrate Peer/PeerGroup I/O to CompletableFuture we can work to improve these tests.

> I think we should look at this again once the Message hierarchy refactorings are finished. OK. I'll mark it as DRAFT.

I added an additional commit that detects the invalid index a few lines earlier.

For a Java 8/CompletionStage refactor of bitcoinj `StreamConnection` interface this [sample code from IBM](https://github.com/IBM/java-async-util/blob/master/asyncutil/src/test/java/com/ibm/asyncutil/examples/nio/nio.md) might be helpful if we want to use `AsynchronousChannel`. See also: https://codereview.stackexchange.com/questions/47354/echo-server-with-completablefuture

Note: I created Issue #2254 which proposes migrating the entire bitcoinj API to `CompletableFuture`

Issue #2254 is closed/complete. I'm going to leave this issue open because it also encompasses other improvements to `StreamConnection` and perhaps other classes.

Related Issue #2794 was recently opened.

Related: * https://github.com/bitcoinj/bitcoinj/pull/1840 * https://github.com/bitcoinj/bitcoinj/issues/2697 * https://github.com/bitcoinj/bitcoinj/pull/2743

> It's documented in `Address.getHash()`: > > > > > Get either the public key hash or script hash that is encoded in the address. > I wasn't sure whether...