serde_v8: restore ser/de benches

[AaronO]

This restores the serde_v8 benches from the original repo, using bencher like core/benches/op_baseline.

Notes

The original benches were compared to serde_json working on existing strings. Whilst useful, these _json benches can be misleading and lead people to believe that serde_v8 is slower than using JSON for the op-layer, forgetting that using JSON requires the extra steps of JSON.parse()/.stringify() and co, not measured here.

Since they can't be meaningfully compared in absolute, it might make sense to remove them

Output

running 15 tests
test de_array_json        ... bench:         488 ns/iter (+/- 9)
test de_array_v8          ... bench:         601 ns/iter (+/- 19)
test de_bool_json         ... bench:          21 ns/iter (+/- 1)
test de_bool_v8           ... bench:           7 ns/iter (+/- 0)
test de_int_json          ... bench:          43 ns/iter (+/- 2)
test de_int_v8            ... bench:           7 ns/iter (+/- 1)
test de_str_json          ... bench:          49 ns/iter (+/- 5)
test de_str_v8            ... bench:          79 ns/iter (+/- 3)
test de_struct_json       ... bench:         142 ns/iter (+/- 3)
test de_struct_json_deopt ... bench:         466 ns/iter (+/- 15)
test de_struct_v8         ... bench:         231 ns/iter (+/- 11)
test de_struct_v8_opt     ... bench:          97 ns/iter (+/- 3)
test de_tuple_json        ... bench:          57 ns/iter (+/- 3)
test de_tuple_v8          ... bench:          69 ns/iter (+/- 0)
test de_tuple_v8_opt      ... bench:          68 ns/iter (+/- 9)

test result: ok. 0 passed; 0 failed; 0 ignored; 15 measured

     Running target/release/deps/ser-7e48eeb04c97e76a

running 13 tests
test ser_array_json       ... bench:         171 ns/iter (+/- 5)
test ser_array_v8         ... bench:         459 ns/iter (+/- 37)
test ser_bool_json        ... bench:          32 ns/iter (+/- 0)
test ser_bool_v8          ... bench:          60 ns/iter (+/- 1)
test ser_int_json         ... bench:          35 ns/iter (+/- 1)
test ser_int_v8           ... bench:          61 ns/iter (+/- 1)
test ser_str_json         ... bench:          48 ns/iter (+/- 1)
test ser_str_v8           ... bench:         119 ns/iter (+/- 2)
test ser_struct_json      ... bench:         155 ns/iter (+/- 1)
test ser_struct_v8        ... bench:         459 ns/iter (+/- 61)
test ser_struct_v8_manual ... bench:         538 ns/iter (+/- 90)
test ser_tuple_json       ... bench:          57 ns/iter (+/- 3)
test ser_tuple_v8         ... bench:         217 ns/iter (+/- 12)

[ry]

LGTM