Properly store strings in bytecode and on 64-bit
Currently, strings in the bytecode take twice more space than needed: they are stored as 64-bit numbers, but half is zero because 32-bit systems discard that half. We need to have a proper way of dealing with strings, e.g. in a separate 8-bit data segment.
This is not urgent; for 64-bit there are currently hackfixes in string printing functions to skip over half of the numbers.