Perform extra skipping if possible when decoding a string

[John-John Tedro]

When decoding a string from memory, for UTF-8 we currently scan for a null byte to act as a termination. After that, we decode the found string from the entire buffer.

Instead, we should decode the UTF-8 string incrementally as we scan through memory and stop when we fail to decode or we encounter \0.

[John-John Tedro]

Fixed in master