Adding a simple Stream as JSON array circe encoder

[BalmungSan]

This adds some functions to create a simple encoder for Streams, which encodes it as a single JSON array. The specs show how they work.

I also added the option to specify which charset to use when encoding a JSON to all the previous functions, since the method has a default argument, I guess it is binary compatible?
I am actually pretty bad at that. If the change is dangerous I may remove it.

I am open to any feedback you may have about these changes.
(btw, I tried to follow the style conventions that were already there).

[BalmungSan]

As a side note, I have read somewhere that mixing Context Bounds with normal implicit parameters is not recommended. However, my changes do that, but just because the rest of the file also does it, and I tried to be consistent.
But, IMHO, it would be good to rewrite all that to just use implicit parameters.

[BalmungSan]

It seems a test failed, but as far as I can tell, the test that failed has nothing to do with my change.
Also, I am pretty sure all test passed in my first commit (I amened some scalafmt issues), so it seems the problem is non-deterministic.
Did you already know about this problem?

[bfdes]

I think I have a similar problem in #3054.

[rossabaker]

Thanks!

We have had a problem with a few flaky tests, particularly in blaze, on Travis CI for a long time. I think it's related to excessive use of unsafeRunSync in the tests and manifests itself on lesser powered hardware, but I've had a hard time reproducing outside Travis. I restarted the test, so not your problem.

I agree that mixing context bounds and implicits is bad (I wish there were a scala flag for this), and would support fixing those as well.

[BalmungSan]

I agree that mixing context bounds and implicits is bad (I wish there were a scala flag for this), and would support fixing those as well.

Would you prefer that on this PR, or in a following PR with only aesthetic changes?
Also, I am not sure if that mix happens elsewhere? Maybe it would be worth creating a new issue to track all files with such mixture and all PRs fixing those.

[rossabaker]

I'd bet it does happen elsewhere, but I don't know how to find them. I'm sure there's some scalameta wizardry that would help.

A separate PR sounds good for fixing this file.

[BalmungSan]

@rossabaker I just pushed the change removing the use of Charset and adding the constant chunks.

Let me know if you have any other suggestions.

[rossabaker]

Looks good. Thanks!

[BalmungSan]

@rossabaker I moved the constant byte chunks into the companion object, so they aren't instantiated for each instance. And I also cached the creation of the open and close brace streams.

[rossabaker]

Looks great. The test failure is unrelated, which I'm trying to mitigate in #3058, because I don't think it describes a real issue.