Date: Mon, 12 Jan 1998 23:45:46 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: jfieber@indiana.edu Cc: daniel_sobral@voga.com.br, hackers@FreeBSD.ORG Subject: Re: Wide characters on tcp connections Message-ID: <199801122345.QAA23572@usr08.primenet.com> In-Reply-To: <Pine.BSF.3.96.980112101833.1419S-100000@fallout.campusview.indiana.edu> from "John Fieber" at Jan 12, 98 10:32:16 am
next in thread | previous in thread | raw e-mail | index | archive | help
> > I'd like to know if there is any provision on TCP protocol for > > wide-character tcp streams. > > I don't know the specific answer, but the definition of a wide > character itself is pretty vague. For example ISO10646 > (~Unicode) has specification for both 16 and 32 bit characters, > both of which would be considered "wide". Unix wide character > implementations generally define "wide" as large enough to handle > any locale the vendor plans to support, and it may vary from > implementation to implementation. > > For network communications using ISO10646/Unicode, byte encodings > such as UTF-7 or UTF-8 make things such as wide-character tcp > streams irrelevant, and they head off endian debates in the > process. The big issue would probably be resolved by htons() on the sender, and ntohs() on the receiver. Good reasons for DCE RPC... We can argue about whether ISO 10646 specified wchar_t should be 16 or 32 bits... however, it's 16 bits on Windows 95/NT, and so if we ever want to interoperate, we should be 16 bits as well. It should be noted that only code page 0x00000000/0x???????? is defined, so all code pages other than the first 16 bits with high 16 bits all zero are undefined (and unlikely to be defined in the near future, if ever). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199801122345.QAA23572>