UTF string validation

[Barry Schwartz]

UTF-7/8/16/32 strings need to checked for validity when they are first inputted into the program. Right now there probably are cases where ff would abort with a ‘memory exhausted’ message, but actually it was an invalid string. This is as a consequence of the switch to using external libraries for UTF-handling. Actually it shouldn’t be possible for an invalid string to get that far. (Thus, these should be assertion errors, if anything.)