Is GNU Emacs Unicode compliant, or is it just Multi-Lingual aware via Mule?

Question

I've read that GNU Emacs handles international scripts via a system called Mule (Multi-Lingual-Environment).

I also read (somewhere, a while back) that there was a move afoot (in recent years) to make it Unicode compliant, and that this was being implemented in version 23 (-ish?).

I typed some Unicode text (eg.कैसे) into gedit, and successfully copied it into Emacs. The copy-paste works in both directions.

However, in Emacs, using Ctrl+Shift+U to type a single Unicode-Codepoint does not work, nor is the IBus Input Method Editor recognized by Emacs.

What is the state of play with Emacs and Unicode?
... and is it possible to use the IBus IME with Emacs?

Answer 1

Short answer: IBus with ibus.el; Gnome Ctrl+Shift+U ⇒ Emacs C-x 8 RET

---

Emacs's support for more-than-8-bit character sets comes from MULE, which is about as old as Unicode (it derives from support for Japanese), and much older than widespread support for Unicode (in terms of programs supporting Unicode text, fonts, etc.). Good Unicode support came with GNU Emacs 23, which is the default version as of Ubuntu 10.04. Acceptable Unicode support came with GNU Emacs 22, which was the default version as of Ubuntu 8.04. Prior versions did have some support for Unicode, but it was more common to run into bugs or limitations.

Emacs doesn't use Gnome, so you can't use Gnome input methods in it. For example, Ctrl+Shift+U to insert a code point by its hexadecimal code is a Gnome feature, so it doesn't work in Emacs. The native Emacs equivalent is Ctrl+Q, which for historical reasons must be followed by the octal code; if you prefer hexadecimal, customize read-quoted-char-radix to the value 16. A similar Emacs command is C-x 8 RET (ucs-insert; if you use it often, you may want to bind it to a shorter key sequence), which can be followed by a hexadecimal code point or the name of a unicode character, with completion.

Emacs has its own set of input methods, and can also use system input methods such as UIM (install uim-el ). IBus support is relatively new; you need ibus.el, which isn't in Lucid and I think not in Maverick either; see also IBus on the Emacs wiki.

Answer 2

GNU Emacs 23.x has built-in Unicode support for UTF-8-encoded text. Its also contains support for UTF-16. I really recommend you to install emacs23 package to avoid several bugs on Emacs related with Mule problems.

And to try to avoid Intelligent Input Bus problem, you can install Anthy input systems. I saw it works very well for Japanese:

sudo apt-get install anthy anthy-el

After install it, edit your .emacsrc to include the text below:

; Set load path for anthy.el
(push "/usr/share/emacs/site-lisp/anthy/" load-path)

; Load anthy.el
(load-library "leim-list")
(load-library "anthy")

;; Workarounds and customizations

; Fix slow input response in emacs23
(if (>= emacs-major-version 23)
(setq anthy-accept-timeout 1))

; Set japanese-anthy as the default input-method
(setq default-input-method "japanese-anthy")

; Shift-space to toggle anthy-mode (default is Ctrl-\)
(global-set-key (kbd "S-SPC") 'anthy-mode)

; Map wide-space to hankaku-space
(setq anthy-wide-space " ")

Probably will works for you; or you found a bug on Emacs regarding its supporting iBus. Thanks!