Json Protocol support

[reisub]

I started from the official Thrift implementation, made it work and tried to clean up the style in the process. Let me know if there's anything you'd like me to change, happy to do it.

I already signed the contributor agreement for an earlier PR.

Closes #154.

[codecov-io]

Codecov Report

Merging #156 into master will increase coverage by 0.61%.
The diff coverage is 73.17%.

@@             Coverage Diff              @@
##             master     #156      +/-   ##
============================================
+ Coverage     65.51%   66.13%   +0.61%     
- Complexity      938     1012      +74     
============================================
  Files            89       90       +1     
  Lines          4843     5268     +425     
  Branches        562      611      +49     
============================================
+ Hits           3173     3484     +311     
- Misses         1431     1524      +93     
- Partials        239      260      +21

Impacted Files	Coverage Δ	Complexity Δ
...java/com/microsoft/thrifty/testing/TestServer.java	76.62% <100%> (+0.3%)	18 <0> (+1)	⬆️
.../com/microsoft/thrifty/testing/ServerProtocol.java	100% <100%> (ø)	1 <0> (ø)	⬇️
...a/com/microsoft/thrifty/protocol/JsonProtocol.java	73.04% <73.04%> (ø)	73 <73> (?)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6817551...2833b4d. Read the comment docs.

[reisub]

@benjamin-bader did you have time to look at this code?

[benjamin-bader]

I think we should exercise the base-64 coder and decoder more thoroughly here. Let's generate an array of bytes from 0..255 and encode that, so that we hit all of the encoding/decoding table entries.

I'm assuming that those are copied verbatim from Apache, and so are almost certainly correct, but we shouldn't assume such things in tests :)

[reisub]

Yes, your assumption is correct. And yeah, I'll add tests for the decoding.

[benjamin-bader]

Yes, sorry for the delay - I've been an overloaded maintainer on quite a few projects for the past while. This change is nice and self-contained, which is great, but is also large. Am I correct in thinking that this is mainly a copy-and-adapt implementation based on Apache's?

[reisub]

@benjamin-bader I understand - don't worry about it, it's not urgent.

You're right, Apache's implementation was the basis for this PR.

[reisub]

I wrote a test for the base64 decoding (see e33a66d).

But then I realized we can use the base64 decoding from the Okio ByteString class. So I changed the implementation to use that instead, making it simpler.

@benjamin-bader let me know if you want anything else changed.

[reisub]

@benjamin-bader is there anything I can do to help you review and merge this work?

[benjamin-bader]

Thanks for contributing!