gtk/docs/reference/gdk/keys.md

Title: Keyboard Codes, Groups, And Modifiers

Functions for manipulating keyboard codes

Key values are the codes which are sent whenever a key is pressed or released. They are included in the data contained in a key press or release GdkEvent. The complete list of key values can be found in the gdk/gdkkeysyms.h header file.

Key values are regularly updated from the upstream X.org X11 implementation, so new values are added regularly. They will be prefixed with GDK_KEY_ rather than XF86XK_ or XK_ (for older symbols).

Key values can be converted into a string representation using [func@Gdk.keyval_name]. The reverse function, converting a string to a key value, is provided by [func@Gdk.keyval_from_name].

The case of key values can be determined using [func@Gdk.keyval_is_upper] and [func@Gdk.keyval_is_lower]. Key values can be converted to upper or lower case using [func@Gdk.keyval_to_upper] and [func@Gdk.keyval_to_lower].

When it makes sense, key values can be converted to and from Unicode characters with [func@Gdk.keyval_to_unicode] and [func@Gdk.unicode_to_keyval].

Key groups

At the lowest level, physical keys on the keyboard are represented by numeric keycodes, and GDK knows how to translate these keycodes into key values according to the configured keyboard layout and the current state of the keyboard. In the GDK api, the mapping from keycodes to key values is available via [method@Gdk.Display.map_keycode], and the reverse mapping is available via [method@Gdk.Display.map_keyval]. The results of these functions are returned in [struct@Gdk.KeymapKey] structures.

You can think of a [struct@Gdk.KeymapKey] as a representation of a symbol printed on a physical keyboard key. That is, it contains three pieces of information:

1. first, it contains the hardware keycode; this is an identifying number for a physical key
2. second, it contains the “level” of the key. The level indicates which symbol on the key will be used, in a vertical direction. So on a standard US keyboard, the key with the number “1“ on it also has the exclamation point (”!”) character on it. The level indicates whether to use the “1” or the “!” symbol. The letter keys are considered to have a lowercase letter at level 0, and an uppercase letter at level 1, though normally only the uppercase letter is printed on the key
3. third, the [struct@Gdk.KeymapKey] contains a group; groups are not used on standard US keyboards, but are used in many other countries. On a keyboard with groups, there can be 3 or 4 symbols printed on a single key. The group indicates movement in a horizontal direction. Usually groups are used for two different languages. In group 0, a key might have two English characters, and in group 1 it might have two Hebrew characters. The Hebrew characters will be printed on the key next to the English characters.

When GDK creates a key event in order to deliver a key press or release, it first converts the current keyboard state into an effective group and level. This is done via a set of rules that varies widely according to type of keyboard and user configuration. The input to this translation consists of the hardware keycode pressed, the active modifiers, and the active group. It then applies the appropriate rules, and returns the group/level to be used to index the keymap, along with the modifiers which did not affect the group and level. i.e. it returns “unconsumed modifiers.” The keyboard group may differ from the effective group used for lookups because some keys don't have multiple groups - e.g. the Enter key is always in group 0 regardless of keyboard state.

The results of the translation, including the keyval, are all included in the key event and can be obtained via [class@Gdk.KeyEvent] getters.

Consumed modifiers

The consumed_modifiers in a key event are modifiers that should be masked out from @state when comparing this key press to a hot key. For instance, on a US keyboard, the plus symbol is shifted, so when comparing a key press to a <Control>plus accelerator <Shift> should be masked out.

// We want to ignore irrelevant modifiers like ScrollLock
#define ALL_ACCELS_MASK (GDK_CONTROL_MASK | GDK_SHIFT_MASK | GDK_ALT_MASK)
state = gdk_event_get_modifier_state (event);
gdk_keymap_translate_keyboard_state (keymap,
                                     gdk_key_event_get_keycode (event),
                                     state,
                                     gdk_key_event_get_group (event),
                                     &keyval, NULL, NULL, &consumed);
if (keyval == GDK_PLUS &&
    (state & ~consumed & ALL_ACCELS_MASK) == GDK_CONTROL_MASK)
  // Control was pressed

An older interpretation of consumed_modifiers was that it contained all modifiers that might affect the translation of the key; this allowed accelerators to be stored with irrelevant consumed modifiers, by doing:

// XXX Don’t do this XXX
if (keyval == accel_keyval &&
    (state & ~consumed & ALL_ACCELS_MASK) == (accel_mods & ~consumed))
  // Accelerator was pressed

However, this did not work if multi-modifier combinations were used in the keymap, since, for instance, <Control> would be masked out even if only <Control><Alt> was used in the keymap. To support this usage as well as well as possible, all single modifier combinations that could affect the key for any combination of modifiers will be returned in consumed_modifiers; multi-modifier combinations are returned only when actually found in state. When you store accelerators, you should always store them with consumed modifiers removed. Store <Control>plus, not <Control><Shift>plus.