add hvm benchmark && recode inline_hvm_book && hvm_book_show_pretty

[Liberxue]

hvm_book_show_pretty

• Uses write!() and writeln!() macros instead of push_str() and format!()
• pre-allocates string capacity: String::with_capacity(book.defs.len())

bench_optimized                 ... bench:  64,930,066.60 ns/iter (+/- 5,466,772.94) 

bench_original                  ... bench:  74,655,083.30 ns/iter (+/- 5,307,304.10) 

add hvm benchmark

├── benches
│   ├── hvm_common.rs
│   ├── hvm_eta_reduce_benchmarks.rs
│   ├── hvm_inline_benchmarks.rs
│   ├── hvm_mutual_recursion_bench.rs
│   ├── hvm_net_size_benchmarks.rs
│   ├── hvm_prune_book_benchmark.rs
│   └── hvm_utils_benchmarks.rs   

[developedby]

We're in the process of moving the repo to Agda, so I won't accept optimizations or benchmarks unless they simplify the existing code.

Still, thank you very much for your interest