[Rpm-metadata] unicode conversion

[seth vidal]

> try:
>     # this will validate UTF-8
>     string = unicode(string, 'utf-8', 'strict')
> except UnicodeError, e:
>     # replaces invalid chars with '?'
>     string = unicode(string, 'iso-8859-1', 'replace')
> 
> return string

Thanks,
 I wasn't aware that unicode.encode() was able to do the replacement
internally.

-sv