Re: EUC-JP and the Yen sign

[Paul Eggert <eggert at twinsun dot com>]

Ulrich Drepper <drepper@redhat.com> writes:

> Paul Eggert <eggert@twinsun.com> writes:
> > The following implementations support both methods:
> > 
> >    DEC
> >    Fujitsu
> >    NEC
> >    nkf (Network Kanji Filter)*
> >    Jcode <http://openlab.ring.gr.jp/Jcode/>*
> 
> How are they doing this?

Jcode has a PEDANTIC flag.  Normally Jcode treats 0x5C as backslash,
but if you set the flag, it treats it as a Yen sign.

nkf has a command-line option.

Fujitsu uses a different locale name.  ja_JP.S90 uses backslash, and
ja_JP.U90 and ja_JP.eucJP use Yen sign.

DEC and NEC use some unspecified mechanism to let the user specify the
character shape.  My guess is that DEC always translates 0x5C to
backslash but can display it either way.  Historically DEC's old
(pre-EUC) systems used backslash.


I think Fujitsu's method is a reasonable one, though I would use the
IANA-registered name ("EUC-JP" for backslash, and I dunno what for yen).