Primitives for common bytes operations

[JukkaL]

These operations related to bytes values are either common or "fundamental" and worth dedicated primitives if they help with performance:

• + (concatenation)
• len(b)
• b[n] (indexing)
• ==, != ([mypyc] Implement bytes equality optimizations python/mypy#10928)
• <, <=, etc.
• b[:n] etc. (slicing) ([mypyc] Add primitive for bytes slicing python/mypy#10966)
• s.encode(<literal>) (for utf8, ascii and latin1)
• s.encode(x) (for an arbitrary encoding)
• b.decode(<literal>) (for utf8, ascii and latin1)
• b.decode(x) (for an arbitrary encoding)
• % formatting
• b.join(...)
• b.split(...)

These operations in six are also quite common in many codebases:

• ensure_str
• ensure_binary

These are somewhat common and fairly easy to implement:

• b.startswith(...)
• b.endswith(...)
• b.lower()
• b.upper()
• b.strip()

Maybe also support these (fundamental but not very common operations):

• bytes([n]) (corresponds to chr(x))
• ord(b)
• b * n

This is part of the wider issue #644.

[ilevkivskyi]

Something that we discussed in the meeting that may be worth supporting (if there is time) is special-casing six.ensure_xxx() functions.

[SanjanaSogimatt]

Hey, I would like to work on this issue. I am new to open source so it would be great can contribute to it.

[TH3CHARLie]

@SanjanaSogimatt feel free to pick one that has been implemented yet. You can use @97littleleaf11 PRs as examples.