Missing string extraction if first characters decode to multlingual sequence

[glmcdona]

There is an issue where occasional strings will be missed.

Eg. assume utf8 "abcdef". Sometimes when "abcd.." decodes into a valid wchar sequence at the start it will decode a wchar junk string.

Fix is when extracting a wchar string, do not increase the offset skipping the rest of the string from being processed as a utf8 option.

See issue here: glmcdona/strings2#10