Non-ASCII Characters Breaks OCUnitReportParser

[blakewatters]

My test suite generates a number of NSError objects in tests without an NSLocalizedDescriptionKey configured for the userInfo of the NSError instance, which results in a default error message of "The operation couldn’t be completed. (Test error 0.)". The apostrophe within the text causes an exception to be raised during parsing:

#<ArgumentError: invalid byte sequence in US-ASCII>