Optimize {write,take}Leb128

[NicoElbers]

Old (Base impl) vs this (New impl) performance based on a [micro-benchmark](https://zigbin.io/b8b396) compiled with zig `0.16.0-dev.1220+95c76b1b4` on ReleaseFast. The 'small' category is items of at most 7 bits, the 'medium' category is items of at most 32 bits and the 'full' category is arbitrary items of `@bitSizeOf(T)` bits.

Rewrite writeLeb128 to no longer use writeMultipleOf7Leb128,
instead:

• Make use of byte aligned ints
• Use writeableSliceGreedy instead of an array
• Special case small numbers (<= 7 bits)

Rewrite Reader.takeLeb128 to not use takeMultipleOf7Leb128 and
instead:

• Use byte aligned integers
• Turn the decoding into a finite state machine
• Turn the main reading loop into an inlined loop of static length
• Special case small integers (<= 7 bits)

For writing across larger integers (> 7 bits) this roughly doubles performance in a micro-benchmark.

For reading across larger integers (> 7 bits) this roughly doubles performance in a micro-benchmark, however with some notable exceptions:

• i50 sees barely any speedup and u50 even sees a minor slowdown. I
  have genuinely no clue what could be causing this, it's the exception
  to the rule.
• 8, 16 and 32 bit integers see a significantly higher speedup