doomemacs/modules/input/bidi/README.org

#+TITLE:   input/bidi
#+DATE:    April 22, 2022
#+SINCE:   v3.0.0
#+STARTUP: inlineimages nofold

* Table of Contents :TOC_3:noexport:
- [[#description][Description]]
  - [[#maintainers][Maintainers]]
  - [[#module-flags][Module Flags]]
  - [[#plugins][Plugins]]
- [[#prerequisites][Prerequisites]]
- [[#features][Features]]
- [[#configuration][Configuration]]
  - [[#using-bidi-mode][Using ~+bidi-mode~]]
  - [[#force-rtl-text-alignment][Force RTL text alignment]]
  - [[#input-methods][Input Methods]]
  - [[#fonts][Fonts]]
    - [[#smart-fontify][Smart Fontify]]
  - [[#change-dictionary-language-on-bidi-buffers][Change Dictionary Language On Bidi Buffers]]
  - [[#automatic-input-mode-switching][Automatic input mode switching]]
- [[#troubleshooting][Troubleshooting]]
  - [[#nastaliq-font-display-bug][Nastaliq font display bug]]

* Description

This module improves support for bidi (bidirectional text). It should be enabled
if you regularly write in languages that write right-to-left. It also provides
some added configuration instructions in the README, since a lot of it is user
specific.

** Maintainers
+ [[https://github.com/iyefrat][@iyefrat]]
+ [[https://github.com/ymarco][@ymarco]]

** Module Flags
This module provides no flags.

** Plugins
This module provides no plugins.

* Prerequisites
This module has no prerequisites.

* Features
This module provides ~+bidi-mode~, a minor mode that improves the display of RTL
text by right-aligning lines that start with an RTL language, on a per-line
basis. Since exact use cases vary, turning on this mode is left to the user.

It also provides easy font configuration for Hebrew and Arabic-derived scripts
(Arabic, Persian, Urdu, etc.) in ~+bidi-hebrew-font~ and ~+bidi-arabic-font~.
See [[Fonts]] for more information. If you use an RTL language that isn't covered by
these characters, open an issue requesting support for it.

* Configuration
** Using ~+bidi-mode~
~+bidi-mode~ is a local minor mode, meaning it has to be turned on a per-buffer
basis.

If you want to have it on for all buffers, use ~+bidi-global-mode~:

#+begin_src emacs-lisp
;; ~/.doom.d/config.el
(+bidi-global-mode 1)
#+end_src

If you only need it for specific purposes, e.g. editing LaTeX
documents, you probably want to enable it through a hook:

#+begin_src emacs-lisp
(add-hook 'TeX-mode-hook #'+bidi-mode)
#+end_src

This is also useful for adding specific functionality for when ~+bidi-mode~ is on.

** Force RTL text alignment
By default, ~+bidi-mode~ will align paragraphs by the first character with
strong directionality. If you want to force all paragraphs to be aligned
right-to-left when ~+bidi-mode~ is on, add the following to your config:

#+begin_src emacs-lisp
(setq +bidi-paragraph-direction 'right-to-left)
#+end_src

*Warning:* do not do this if you are using ~+bidi-global-mode~, it will mess up
all of the buffers in Emacs that use English, including things like the =M-x= buffer.

** Input Methods
If you need bidi support, it's likely that you want to easily switch between
English and your favorite RTL language. In order to be able to do this without
losing access to all of the keybindings require English letters, you should use
[[https://www.gnu.org/software/emacs/manual/html_node/emacs/Input-Methods.html][input methods]] to switch languages instead of changing the system keyboard
language. If you use a non-qwerty layout, you will need extra configuration to
keep the input method consistent, see [[https://github.com/ymarco/doom-emacs-config/blob/2d655adb6a35c5cd3afcba24e76327f5444cf774/dvorak-config.el#L3-L18][here]] for an example for dvorak.

Toggling the input method bound to =C-\=. It prompts you to choose an input
method the first time you do this in a session, but you bypass this by setting
the default input method:

#+begin_src emacs-lisp
(setq default-input-method "hebrew")
#+end_src

** Fonts
Many good English fonts do not have great coverage for RTL languages, especially
for Hebrew and monospace fonts. To this end, we provide ~+bidi-hebrew-font~ and
~+bidi-arabic-font~ as an easy way to override the default fonts but only for
Hebrew and Arabic characters. They are set by default to =DejaVu Sans=, since
it has pretty decent looking Hebrew and Arabic characters.

Note, that if you are writing in an Arabic-derived script, such as Persian,
Urdu, or Pashto, you may want to change ~+bidi-arabic-font~ to one specific to
your language, especially if you want your script to be written in the Nastaliq
style.

If you use an RTL language the script of which isn't covered by the =hebrew= or
=arabic= unicode blocks, you can set a font override manually. For example:

#+begin_src emacs-lisp
(add-hook 'after-setting-font-hook
          (lambda () (set-fontset-font t 'syriac (font-spec :family "DejaVu Sans"))))
#+end_src

Make sure to use the correct unicode block name, see the documentation of
~set-fontset-font~ for more details.

*** Smart Fontify
Since good bidi fonts are often not monospace (as is the default =DejaVu Sans=),
It usually looks better to have the surrounding spaces and punctuation in the
use the bidi font as well. This is the default behaviour, but you can turn this
off by setting:

#+begin_src emacs-lisp
(setq +bidi-want-smart-fontify nil)
#+end_src

** Change Dictionary Language On Bidi Buffers
If you are only using ~+bidi-mode~ in specific buffers, you might want to
automatically change the dictionary language there. For example:

#+begin_src emacs-lisp
(add-hook! '+bidi-mode-hook
  (if +bidi-mode
      (ispell-change-dictionary "hebrew")
    (ispell-change-dictionary "default")))
#+end_src

** Automatic input mode switching
You may want to Emacs to try and guess when you want it to switch input methods.
See [[https://github.com/ymarco/doom-emacs-config/blob/2d655adb6a35c5cd3afcba24e76327f5444cf774/hebrew-latex-config.el#L7-L21][here]] and [[https://github.com/ymarco/doom-emacs-config/blob/2d655adb6a35c5cd3afcba24e76327f5444cf774/hebrew-latex-config.el#L99-L102][here]] for an example of how to get Emacs to switch to hebrew when
entering insert mode after a hebrew character, in LaTeX buffers.

* Troubleshooting

** Nastaliq font display bug
If Emacs is having trouble properly displaying a Nastaliq font, try using one of
[[https://urdufonts.net/fonts/jameel-noori-nastaleeq-regular][these]] [[https://urdufonts.net/fonts/alvi-nastaleeq-regular][two]] fonts for ~+bidi-arabic-font~.