Bidirectional text

From ArchWiki

From Wikipedia:Bidirectional text, is text that contains two text directionalities, right-to-left (RTL) and left-to-right (LTR).

Concepts

Text direction
while editing, setting text direction changes the interaction of the editor with the text. It affects text navigation, selection, and insertion.
Text alignment
a stylized form of presenting text. E.g: English text can be aligned to the right despite being LTR language, but this does not change the direction of the text which is LTR.
Fake Bidi
a fake locale displaying completely reversed English sentences to mimic a RTL context. This helps who are not familiar with RTL language to visualize RTL issues. For example:
This English sentence contains SDROW CIBARA EMOS and ends with English words.
Poor man's bidi mode
automatic visual text order right-to-left input in normal terminal.
Logical navigation
words are traversed as they are read.
Visual navigation
words are traversed as they are displayed. This GIF demonstrates the difference between the two.

Logical navigation is the correct way of navigation that programs should implement to fully support Bidi.

Unicode support
In Unicode-supported text editors, text direction can be controlled by inserting special formatting characters in front of it: Wikipedia:RLM (U+200F), Wikipedia:LRM (U+200E), which is a workaround for text editors not supporting switching text direction.
Bidirectional Algorithm
is the set of rules applied to display text in the correct order.

Some implementations:

  • fribidi is an open source implementations of Bidi Algorithm.
  • python-bidiAUR is a python library for related functions.
  • SheenBidi: Improved Unicode Implementation in C.

Support

In order for a program to fully support Bidi text, it must have:

Text shaping (RTL shaping)
correctly apply letter rendering and ligatures. See W:harfbuzz and Complete Text Layout.
Unicode Bidirectional Algorithm (UBA)
determine the directionality of text segments and apply the appropriate rendering rules.
Editing and input (Nav)
the ability to force text direction and correct behavior of cursor movement, selection, deletion, and insertion (logical navigation).

Browsers

App RTL UBA Nav Notes
chromium Yes Yes No
firefox Yes Yes No
qtwebbrowser Yes Yes No
webkit Yes Yes No

Editors

App RTL UBA Nav Notes
AbiWord Yes No No
emacs Yes No No With extension. See also: emacs bidi editing.
gedit Yes Yes No For Gedit <= 2.1, workaround by extension.
ghostwriter Yes Yes No Qtwebbrowser-based
kate Yes Yes No
LibreOffice Yes Yes Yes See: LibreOffice#Bidirectional support
marktextAUR Yes No No Blink-based
neovim No No Yes See Vim#Bidirectional support
texworks Yes Yes No
vim No No Yes see Vim#Bidirectional support
featherpad Yes Yes No
Visual Studio Code Yes No No
mousepad Yes Yes No
GTK Yes Yes No
xed Yes Yes No

Terminal

App RTL UBA Nav Notes
alacritty No No No Tested with alacritty version 0.13.0-3, the 2024-01-03. Does not support RLO (U+202E).
contourAUR No No No Tested with contourAUR version 0.3.12, the 2024-01-03. Does not support RLO (U+202E).
dvtm No No No
extraterm No No No Tested with extraterm-binAUR version 0.74.0-1, the 2024-01-03. Does not support RLO (U+202E).
kitty Yes No Tested with kitty version 0.31.0-1, the 2024-01-02. Does not support RLO (U+202E).
konsole Yes No Tested with konsole version 23.08.4-2, the 2024-01-02.
libvte Yes No including libvte-based like Gnome, sakura, xfce4, terminator... Minor issues: 1
mltermAUR Yes No Tested with mltermAUR version 3.9.3-2, the 2024-01-03.
neatvi-gitAUR No
pymux No
qterminal No No Tested with qterminal version 1.4.0-1, the 2024-01-02. Does not support RLO (U+202E).
radare2 No
Terminator Yes No Tested with terminator version 2.1.3-3, the 2024-01-02.
tmux No won't fix
Urxvt No No Tested with rxvt-unicode version 9.31-4, the 2024-01-03. Does not support RLO (U+202E).
wezterm No
xstAUR No
less No

Other

App RTL UBA Notes
electronjs No
groff No No
irssi No
latex Yes Yes Using LuaTeX or XeLaTeX with polyglossia
Wikipedia:HTML Yes Yes
wine Yes No

Troubleshooting

Verify UTF-8 Encoding is properly configured first. Make sure you install the fonts corresponding to your language: Arabic, Persian, Hebrew. For some terminals, (bicon-gitAUR) is required to properly display RTL text.

See also

Internal links

External links