Gerd Knorr and me have added multibyte support to the
talk client using the wide character support in libncursesw.
'talk' now works also in UTF-8 locales in UTF-8 capable terminals
where it didn't work at all before. In other multibyte locales where
the old 'talk' already worked partly by accident, it works better now.
In other multibyte locales, for example ja_JP.eucJP, 'talk' happened
to work partly even without proper multibyte support because one
Japanese character needs two bytes in EUC-JP encoding and uses 2
columns on the screen which happens to even out. But there were the
usual problems like having to type 'Backspace' twice to delete one
Japanese character etc.. This should work much better now.
New packages for SuSE Linux 8.1 are here:
ftp://ftp.suse.com/pub/people/mfabian/8.1-i586/talk-0.17-318.i586.rpm
ftp://ftp.suse.com/pub/people/mfabian/8.1-i586/talk-server-0.17-318.i586.rpm
ftp://ftp.suse.com/pub/people/mfabian/8.1-src/talk-0.17-318.src.rpm
'talk' has a new option now to choose the encoding:
SYNOPSIS
talk [-p encoding] person [ttyname]
DESCRIPTION
Talk is a visual communication program which copies lines from your ter
minal to that of another user.
Options available:
encoding The charset encoding sent by your peer (i.e. UTF-8, ISO-8859-1,
EUC-JP, whatever). Default is some guesswork based on the
incoming data and your current locate.
Usually it is not necessary to use that option, the autodetection
apparently works well enough in most cases.
'talk' will always use the encoding of the current locale for the text
send to the other person. By default, the new talk on the receiving
side will try to detect whether it receives UTF-8 and if it isn't
UTF-8 try to use the encoding of the current locale instead.
This autodetection seems to work rather reliable, i.e. it appears to
work correctly out of the box even if the persons using talk use
different locales. For example it appears to work fine when one person
is using ja_JP.eucJP and the other ja_JP.UTF-8 or one person is using
de_DE@euro and the other any de_DE.UTF-8.
Please test, we would be happy to receive feedback.
--
Mike Fabian http://www.suse.de/~mfabian
睡眠不足はいい仕事の敵だ。
