Add 'utf8Size' opcode

[lock9]

Hi,

The compiler / VM is not processing strings properly. It is always considering one byte per character, but this is not true for UTF-8 strings. We should probably add an 'utf8Size' opcode in addition to the current 'size' one. The compiler must use the correct opcode when dealing with strings.

There may be other places and opcodes to be added, like the 'substring' method (I don know how it is implemented, but I can confirm that it considers 1 byte per character).

Related to neo-project/neo-devpack-dotnet#681

[shargon]

If we use another encoding we will have the same error

[lock9]

What do you mean by another encoding? We only support UTF-8, right? We can call it unicodeSize if that is the matter, but it does have a problem that has to be fixed.

[erikzhang]

I'm not sure if it's necessary. Because it is also possible to operate by byte length.

[devhawk]

This seems more appropriate to add to neo than neo-vm. NeoVM doesn't have a string type, so it seems odd to have a string length opcode in the VM.

Since Neo leverages NeoVM ByteStrings and Buffers to store UTF8 strings, it makes sense to me that there should be mechanism in Neo.dll to calculate the length of a UTF8 string that is multibyte character aware