Bug#38077: [Konsole-devel] Bug#38077: konsole is completely broken for multibyte languages

[Takumi ASAKI]

On 2002 February15 Friday 06:37, Waldo Bastian wrote:

> Can you explain me a bit more how Kanji is supposed to work in a terminal?
> Does a single Kanji character take up two logical screen positions?
> Can you point me to a terminal program which does this right so that I
> could use it as example?
> Which locale should I set to try this out?

I know only Japanese.
I don't know Chinese, Korean, and other.

Almost Kanji char was used as two 8bit ASCII char.
In Japan mose Kanji's width is these code's length.

ex. `HIRAGANA LETTER A'(U+3042)
`HIRAGANA LETTER A's EUC-JP(Japanese locale) code is 0xa4 0xa2.
So Japanise terminal application output 2 byte, 0xa4 and 0xa2.

Terminal recognize these 2 byte as `HIRAGANA LETTER A'.
And show it in 2 columns.

If cursor is on Kanji, cursor width is twice.

when cursor move from Kanji to right,
Japanese terminal application move cursor to 2 columns right, not 1 column.

http://www.kde.gr.jp/patch/kdebase-2.2.2-konsole-ja-20011126.diff
is very adhoc Japanese patch for kdebase-2.2.2.
It's not perfect. but work better than original konsole.

Maybe xterm or mlterm is useful example.
http://dic45key.his.com/xterm/
http://mlterm.sourceforge.net/

-- 
  Che Che - Bye Bye
        From: Takumi ASAKI <asataku at osk3.3web.ne.jp>
	URL: http://www3.osk.3web.ne.jp/~asataku/