I've been using the iconv routine with some success to convert from utf-8 encoded Japanese to cp932 encoded Japanese. However, some characters do not convert. For instance 0xe7 0xab 0x8e is a valid utf-8 character that iconv cannot handle. I've searched the web in vain for insight into what is going wrong. I find that regardless of whether iconv returns a correctly translated string or not, I always get an error number of ESPIPE (errno = 29) which is the error code for illegal seek. That suggests to me that there is a truncated file somewhere that iconv relies upon. Does anyone have any ideas?