Subject: Re: fmt and high bit characters
To: David Brownlee <abs@netbsd.org>
From: None <itojun@iijlab.net>
List: tech-userlevel
Date: 09/28/2000 23:55:00
>	Will this help for the case when you are sharing files with
>	someone from a different locale? - eg fix it for the current
>	locale so isprint() does the right thing, but allow a '-8' overide
>	(or similar) to allow all highbit characters through...

	please also note that there are text encoding with multibyte 8bit chars
	(like euc-jp).

	for the moment, i think allowing 8bit chars to go through (either -8 or
	without option) is sufficient workaround for western 8bit encodings.
	stripping off all 8bit chars is, i think, not useful.

	the right way is to have a full locale support and use that.
	even in this case, we have some issues:
	- if locale setting ($LANG) and the file encoding does not match,
	  strange output may result (anyway i can say that it is a pilot error).
	- for non-western languages how should fmt behave?
	  for example, japanese/chinese/korean text does not have space between
	  chars.

itojun