On Wednesday, May 19, 2004 6:14 PM, Markus Kuhn wrote: > "UTF-16" UTF-16 (same byte order as C's short) > "UTF-32" UTF-32 (same byte order as C's long)
Why are you (do you seem to be, in fact) requiring that short should be 16 bits or long being 32 (the latter is a particular nuisance these days, of course). I understand the below idea is the peculiar (and atypical) byte order of Dennis Ritchie's implementation of long on the PDP-11. However, even htonl or ntohs do no imply relationship between short/long and 16/32 bitness, except in the acronym (this is similar to gmtime). Certainly a better point would be to have direct mapping with uint16_t[] and uint32_t[]. > This minimum requirement would therefore not put > any unreasonable burden on the implementors of even low-memory > footprint implementations. By the way: <iconv.h> is not "really" POSIX in the strict sense, it is XSI (which does not target low-memory footprint, as I understand things.) Antoine -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/
