-
-
Notifications
You must be signed in to change notification settings - Fork 2.6k
Keyboard layouts
Keyboard layouts are .json files containing a list of keys to form the on-screen keyboard.
You can quickly switch layouts by navigating to the userData/keyboards
folder and clicking/tapping a layout file on the filesystem display, but you will need to edit your settings file to make any change permanent.
Keyboard layouts are identified by a code with the format language-REGION
*, eg: en-US
for American English.
A "valid" code is any provided layouts' filename (without the .json extension), and/or the name of any user-provided layout you might have copied in the userData/keyboards
folder.
*In practice, this rule is not enforced.
The layout is organized by rows (see below). Each row contains an array of keys that will be parsed in order by keyboard.class.js
, from left to right.
{
"row_numbers": [],
"row_1": [],
"row_2": [],
"row_3": [],
"row_space": []
}
Each key contains name
and cmd
properties, and sometimes prefixed name
and cmd
properties. Valid prefixes are:
fn_
altshift_
alt_
ctrl_
shift_
shift_
is the "less important" prefix, and fn_
the "most important".
When a special key (Shift,Control,Alt...) is in use, the matching prefixed cmd property will be used. If multiple special keys are active and the key has multiple matching prefixed properties, the "most important" one will be used.
There is also and additional prefix capslck_
that will replace the shift_
property if the shift command is locked (caps-lock). It is mostly used to allow typing uppercase diacritics like "É".
Prefixed or not, name
is the text to display on the key, and cmd
is the sequence to send to the shell.
Since sequences for special keys like delete, enter, or control+A are invalid JSON characters, special ~~~CTRLSEQx~~~
sequences (where x
is an integer from 1 to 21) are used instead. They are replaced by the correct UTF8 character by the parser. The list of all CTRLSEQ characters is on the fifth line of keyboard.class.js. You might need a special text editor to see them correctly.
Another type of parsed sequences are ESCAPED|--
commands. They are not forwarded to the shell and instead trigger special actions on the keyboard, eg: ESCAPED|-- SHIFT: LEFT
.
Here is a key example:
{
"name": "r",
"cmd": "r",
"shift_name": "R",
"shift_cmd": "R",
"ctrl_cmd": "~~~CTRLSEQ8~~~",
"alt_cmd": "~~~CTRLSEQ1~~~r"
}
To write special latin characters such as "ô" or "Ë", greek letters or other specific characters, the keyboard parser understand some escape sequences as "modifiers" that will trigger the next key to be mapped to a different character subset:
ESCAPED|-- CIRCUM
ESCAPED|-- TREMA
ESCAPED|-- ACUTE
ESCAPED|-- BREVE
ESCAPED|-- CARON
- etc... (about 13 modifiers available at time of writing)
Please refer to the source file to see available modifiers.
The layout parser supports placing icons instead of text on the keys, by using an ESCAPED|-- ICON: (icon name)
command in the key's name
property.
Currently icon support is limited to arrows (ARROW_UP
, ARROW_DOWN
, ARROW_LEFT
and ARROW_RIGHT
).
Icons are hard-coded into the parser.