Optimizes lualib functions.
Also fixes #1130.
Depends on #1213, however it can be made independent.

Array.push with only 1 argument gets even more optimization, as it is very common:

• Pushing a single element results in array[#array + 1] = el
• Pushing a spread array of elements uses a new function ArrayPushArray to avoid unpacking/repacking the array.

Benchmarks on my machine show 10-70% time reduction for most functions on lua5.3 and luajit. As can be expected, lua5.3 has better relative improvement for many functions compared to luajit.

Some optimization methods

• caching global functions in local variables (when used more than once)
• using a lua for i in range loop to iterate arrays