fix: use memset in bitset_set and bitset_clear

[theVerySharpFlat]

Hello again!
The current implementations of bitset_set_range and bitset_clear_range use while loops and set bytes individually. Often, the system's standard c library will include system-specific optimizations with memset.

Speedups are a lot more humble than my previous suggestion, but the implementation is cleaner!

(buddy_bench numbers from Release configuration, zen4)
before: 11.580 with around 0.050s variance over 10 runs
after:  11.280 with around 0.050s variance over 10 runs

[spaskalev]

Thanks. Can you make it != 0 so that CodeQL doesn't complain ?

[theVerySharpFlat]

yeah, just force pushed. Oh, do you prefer != 0 over just empty?

[spaskalev]

This works too :))