[ZKS-04] Track committee size.

[d0cd]

This PR:

• tracks the committee size in a metadata mapping.
• enforces that the committee size does not exceed 200.
• enforces that the number of delegators does not exceed 100_000.

The benches/rewards branch measures the performance of atomic_post_ratify for various numbers of delegators distributed uniformly over 200 validators.
On a 2021 M1 MAX:

DB: `atomic_post_ratify` with 0 delegators and 200 validators
                        time:   [18.821 ms 19.708 ms 20.456 ms]

DB: `atomic_post_ratify` with 1000 delegators and 200 validators
                        time:   [88.696 ms 104.41 ms 131.43 ms]

DB: `atomic_post_ratify` with 10000 delegators and 200 validators
                        time:   [750.47 ms 793.26 ms 838.26 ms]

DB: `atomic_post_ratify` with 100000 delegators and 200 validators
                        time:   [4.0215 s 4.3443 s 4.7269 s]

DB: `atomic_post_ratify` with 1000000 delegators and 200 validators
                        time:   [37.494 s 37.753 s 38.058 s]

Using memory based storage, we found:

Memory: `atomic_post_ratify` with 0 delegators and 200 validators
                        time:   [53.368 ms 53.403 ms 53.433 ms]

Memory: `atomic_post_ratify` with 1000 delegators and 200 validators
                        time:   [181.87 ms 182.05 ms 182.29 ms]

Memory: `atomic_post_ratify` with 10000 delegators and 200 validators
                        time:   [1.3540 s 1.3547 s 1.3553 s]

Memory: `atomic_post_ratify` with 100000 delegators and 200 validators
                        time:   [13.088 s 13.224 s 13.480 s]

Memory: `atomic_post_ratify` with 1000000 delegators and 200 validators
                        time:   [131.51 s 131.59 s 131.68 s]

Helps with addressing Audit Finding: [zksecurity 04] Garbage Collection Can Block Commits From Happening

[randomsleep]

Hi, I recently submitted a bug report (id: 2289066) at HackerOne that is directly related to the committee update. The report shows that updating the committee at every block will cause some consensus issues. I suggest to review the report and take it into consideration.

[vicsn]

Hi @randomsleep thank you for the extra heads up. Just FYI we've seen the issue, great finding, we will circle back on HackerOne.