Fix wrong stdin queue order

[Exidnus]

Because of unsafeRunAndForget was used, order of bytes from stdin and order of bytes in stdinQ: Queue[F, ByteVector] could be different. It doesn't look right, so I switched to unsafeRunSync.
And up sbt version to the latest.

[fmonniot]

That's a good point, I haven't seen this issue because most of my usages have low throughput. One question I have though: do you know what thread unsafeRunSync block? And is it ok to block it?

[Exidnus]

Sorry for late response, I missed your comment.
Well, good question. Looks like, unsafeRunSync would block on executionContext from Async. So it's a runtime loop from IO and block on it is not a good idea, I think.

[Exidnus]

Well, I thought a little more. :)

1. ExecutionContext from Async would be used for unsafeRunSync and unsafeRunAndForget, it's no difference here.
2. Calling unsafeRunSync would block some thread from NuProcess, which we shouldn't block according to javadoc from NuProcessHandler.
   "Note that the processing thread that executes these callbacks on behalf of a NuProcess is the same thread that handles processing for other instances of NuProcess, and as a result you should never perform a blocking or time-consuming operation on the callback thread. Doing so will block or severely hinder the processing of other NuProcess instances. Any operation that requires more than single-digit millisecond processing should be handed off to another thread so that the processing thread can return to immediately service other NuProcess instances." It's from here

So blocking is not an option.

[Exidnus]

But from the other side: there is a Queue.unbounded for stdout and stderr, it's just operation in memory, it's not blocking. So unsafeRunSync looks fine here, no?

[fmonniot]

I believe you are correct, and the offer function should not have blocking underlying it. Let's merge it then.

[fmonniot]

0.4.0+3-f3574c73-SNAPSHOT have been released to https://oss.sonatype.org. Let me know if that fixed your issue, and if yes, I'll tag 0.4.1.

[Exidnus]

0.4.0+3-f3574c73-SNAPSHOT have been released to https://oss.sonatype.org. Let me know if that fixed your issue, and if yes, I'll tag 0.4.1.

Thank you. Yes, it solves the problem.

[fmonniot]

v0.4.4 have been released to maven central.