poc: wraparound check

[junyechen1996]

• added a simple test that test validity circuit, because it takes too long to finish one test with the default num_wr_reps and alpha.
• had to customize code to run PINE FLP because wraparound joint rand is part of FLP encoding.
• with l2_norm_bound of 1.0, and num_frac_bits of 4, encoding time: 0.01 seconds, single verifier eval time: 0.01 seconds, flp.prove: 22.1 seconds, flp.query total time: 48.3 seconds
• stacked PR on #27

[junyechen1996]

@cjpatton Thanks for the review. I responded to all your comments (hopefully), let me know if I miss any. Things that I would like you to pay particular attention to:

• https://github.com/junyechen1996/draft-chen-cfrg-vdaf-pine/pull/28#discussion_r1320357881
• https://github.com/junyechen1996/draft-chen-cfrg-vdaf-pine/pull/28#discussion_r1320358418 ( I don't have a great solution other than re-writing run_flp to take in the joint randomness).

• 30, I will work on it for the VDAF draft, because Prio3Sum and Prio3SumVec are also using it.

Responding to the high-level comments:

Please remove every variable from PineValid that is not used by any of the methods.

Done. I will add them back once I implement L2-norm check.

Suggestion: I prefer the name "wraparound test" instead of "repetition". "Repetition" begs the question "what is being repeated"? the only answer to which is "the wraparound test".

I start calling all of them "checks" now, and removed repetitions. I'm not sure if it's worth introducing another term "test", but I'd welcome your feedback.

The wraparound subcircuit should include more comments that help connect the reader to the properties we intend to enforce. (See inline comments.)

I addressed the inline comments and added more code comments. Please take a look again.

[junyechen1996]

Looking really good! I know it looks like a lot of comments, but they're mostly editorial. However I spotted one potential bug I wanted you to look at, in the wraparound rand derivation.

@cjpatton Thanks for the thoughtful reviews! That is a bug, I've fixed it now. Here are the threads that I'd like you to take a look again. As always, feel free to look over the revised PR again.

• unused l2_norm_bound: https://github.com/junyechen1996/draft-chen-cfrg-vdaf-pine/pull/28#discussion_r1326519399.
• XOF output length: https://github.com/junyechen1996/draft-chen-cfrg-vdaf-pine/pull/28#discussion_r1326548617, I added comments to clarify the computation for XOF output length.
• sample_wraparound_joint_rand bug: https://github.com/junyechen1996/draft-chen-cfrg-vdaf-pine/pull/28#discussion_r1326552908, should be fixed now.
• make encode_f64_into_field into instance method in PineValid: https://github.com/junyechen1996/draft-chen-cfrg-vdaf-pine/pull/28#discussion_r1326530095

[junyechen1996]

@cjpatton Great suggestion about testing bit chunking. Please take a look at the updated code, which contains some of your code, but with some simplification.

The remaining thread is the one on whether l2_norm_bound should be stored: https://github.com/junyechen1996/draft-chen-cfrg-vdaf-pine/pull/28#discussion_r1326519399