Roger Leigh <rleigh@codelibre.net> writes:
> The language in SUSv2 in fact explicitly states that this is
> allowed. In fact, I've seen documentation that some UNIX systems such
> as HPUX already do have a UTF-8 C locale as an option.
I don't argue with the concept of a C.UTF8 locale --- in fact I think
it sounds pretty useful. What I think is 100% broken is trying to make
C locale work that way. C locale is supposed to be the traditional
locale-ignorant, characters-are-bytes behavior.
regards, tom lane