From owner-freebsd-hackers Mon Oct 16 13:40:40 1995 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id NAA20158 for hackers-outgoing; Mon, 16 Oct 1995 13:40:40 -0700 Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id NAA20150 for ; Mon, 16 Oct 1995 13:40:35 -0700 Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id NAA25305; Mon, 16 Oct 1995 13:34:10 -0700 From: Terry Lambert Message-Id: <199510162034.NAA25305@phaeton.artisoft.com> Subject: Re: A couple problems in FreeBSD 2.1.0-950922-SNAP To: joerg_wunsch@uriah.heep.sax.de Date: Mon, 16 Oct 1995 13:34:10 -0700 (MST) Cc: kaleb@x.org, hackers@freefall.freebsd.org In-Reply-To: <199510152303.AAA22305@uriah.heep.sax.de> from "J Wunsch" at Oct 16, 95 00:03:59 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 2786 Sender: owner-hackers@FreeBSD.org Precedence: bulk > As Kaleb S. KEITHLEY wrote: > > > > As near as I can tell the SVR4 ls doesn't change its locale, yet still > > manages to do the right thing, probably because for most SVR4-en the C > > locale is full ISO8859-1. This leads me to believe that FreeBSD's ls > > probably doesn't need to change its locale either if the default chartype > > table is fully populated. > > So SVR4 would still break on koi8-r, for example. Either make it > right, or let it be. But not on 8859-x. For Coptic alphabets, that's 8859-9. The problem with KOI-8 is that KOI-8 is a defacto standard, and is not accepted by international standards bodies. Mostly because the most popular BBS software in the area picked it up instead of 8859-9. The problem is not in the blank areas of the locale. In point of fact, the ANSI standards for terminal control sequences after ANSI 3.64 leave the codes in columns 0x80 and 0x90 to be used to represent 8 bit command sequence introducers, which are the same as an escape character followed by a character in columns 0x20 or 0x30. Because of this, KOI-8 as a character set is not compatible with post 3.64 ANSI terminal control sequence standardization. It is not unreasonable, then, for the code to function correctly in the 8-bit case for ISO 8859-x but not function correctly without an environment or (more preferrably) code change in the application for KOI-8. Really, they should be using the 8859 character set instead of KOI-8, but there is understood to be a large historical investment in the non-standard KOI-8 representation (unfortunately). > isctype() is not necessarily related to message catalogs. The > extensive (i.e. blatant) use of message catalogs (AIX, IRIX) leads to > very undesirable results, e.g. SMTP daemons throwing their error > messages in German. :-( This is an issue of when the message is translated: a function of the interface violation of an internationalized system. Actually, the error presentation between the dameons should be abstract and token based: ie: a number, which your smtp agen translates into the correct error. The problem is that the translation to a human representation should occur in the locale of the recipient of the human-form message rather than in the locale of the human who ran the program. That is, the error reporting protocol is a half-circuit and the abstraction is thereby compromised. The only correct procedure for dealing with this would be to change the error reporting in the SMTP protocol to encapsulate it as well as the transactions for which the error is being reported, and then translate it locally before presentation to the user. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.