mirror of
https://codeberg.org/dnkl/foot.git
synced 2026-02-04 04:06:06 -05:00
input: kitty: merge handling of plain-text and composed characters
All plain-text and composed characters are now printed as-is, in a single place. Also fix handling of “generic” keys when emitted as escapes; don’t use the raw XKB symbol as key in the escape, convert it to a unicode code point first. For many symbols, these are the same. But not all. For now, we fallback to using the symbol as is if XKB fails to convert it to a codepoint. Not sure if we should simply drop the key press instead. Composed characters also need special treatment; we can’t use the symbol as is, since it typically refers to the last key pressed (i.e. not the composed character). And, that key is also (usually) a special “dead” key, which cannot be converted to a unicode codepoint. So, what we do is convert the generated utf8 string, and (try to) convert it to a wchar. If it succeeds, use that. If not, fallback to using the XKB symbol (as above).
This commit is contained in:
Daniel Eklöf
parent
660626118a
commit
93a8f51b75
1 changed files with 92 additions and 93 deletions
185
input.c
185
input.c
|
|
@ -1151,6 +1151,7 @@ kitty_kbd_protocol(struct seat *seat, struct terminal *term,
|
|||
const bool repeating = seat->kbd.repeat.dont_re_repeat;
|
||||
const bool pressed = ctx->key_state == WL_KEYBOARD_KEY_STATE_PRESSED && !repeating;
|
||||
const bool released = ctx->key_state == WL_KEYBOARD_KEY_STATE_RELEASED;
|
||||
const bool composed = ctx->compose_status == XKB_COMPOSE_COMPOSED;
|
||||
|
||||
const enum kitty_kbd_flags flags = term->grid->kitty_kbd.flags[term->grid->kitty_kbd.idx];
|
||||
const bool disambiguate = flags & KITTY_KBD_DISAMBIGUATE;
|
||||
|
|
@ -1159,6 +1160,9 @@ kitty_kbd_protocol(struct seat *seat, struct terminal *term,
|
|||
if (!report_events && released)
|
||||
return false;
|
||||
|
||||
if (composed && released)
|
||||
return false;
|
||||
|
||||
/* TODO: should we even bother with this, or just say it’s not supported? */
|
||||
if (!disambiguate && pressed)
|
||||
return legacy_kbd_protocol(seat, term, ctx);
|
||||
|
|
@ -1183,21 +1187,10 @@ kitty_kbd_protocol(struct seat *seat, struct terminal *term,
|
|||
}
|
||||
}
|
||||
|
||||
if (ctx->compose_status == XKB_COMPOSE_COMPOSED && !released) {
|
||||
term_to_slave(term, utf8, count);
|
||||
return true;
|
||||
}
|
||||
|
||||
/*
|
||||
* Printables without any modifiers are printed as is.
|
||||
*
|
||||
* TODO: plain text keys (a-z, 0-9 etc) are still printed as text,
|
||||
* even when NumLock is active, despite NumLock being a
|
||||
* significant modifier, *and* despite NumLock affecting other
|
||||
* keys, like Return and Backspace; figure out if there’s some
|
||||
* better magic than filtering out Caps- and Num-Lock here..
|
||||
*/
|
||||
if (iswprint(utf32) && (effective & ~caps_num) == 0 && !released) {
|
||||
/* Plain-text without modifiers, or commposed text, is emitted as-is */
|
||||
if (((iswprint(utf32) && (effective & ~caps_num) == 0) || composed)
|
||||
&& !released)
|
||||
{
|
||||
term_to_slave(term, utf8, count);
|
||||
return true;
|
||||
}
|
||||
|
|
@ -1319,90 +1312,96 @@ kitty_kbd_protocol(struct seat *seat, struct terminal *term,
|
|||
case XKB_KEY_XF86AudioRaiseVolume: key = 57439; final = 'u'; break;
|
||||
case XKB_KEY_XF86AudioMute: key = 57440; final = 'u'; break;
|
||||
|
||||
#if 0 /* TODO: enable when “Report all keys as escape codes” is enabled */
|
||||
case XKB_KEY_Caps_Lock: key = 57358; final = 'u'; break;
|
||||
case XKB_KEY_Num_Lock: key = 57360; final = 'u'; break;
|
||||
case XKB_KEY_Caps_Lock: if (false) {key = 57358; final = 'u';} break;
|
||||
case XKB_KEY_Num_Lock: if (false) {key = 57360; final = 'u';} break;
|
||||
|
||||
case XKB_KEY_Shift_L: key = 57441; final = 'u'; break;
|
||||
case XKB_KEY_Control_L: key = 57442; final = 'u'; break;
|
||||
case XKB_KEY_Alt_L: key = 57443; final = 'u'; break;
|
||||
case XKB_KEY_Super_L: key = 57444; final = 'u'; break;
|
||||
case XKB_KEY_Hyper_L: key = 57445; final = 'u'; break;
|
||||
case XKB_KEY_Meta_L: key = 57446; final = 'u'; break;
|
||||
case XKB_KEY_Shift_R: key = 57447; final = 'u'; break;
|
||||
case XKB_KEY_Control_R: key = 57448; final = 'u'; break;
|
||||
case XKB_KEY_Alt_R: key = 57449; final = 'u'; break;
|
||||
case XKB_KEY_Super_R: key = 57450; final = 'u'; break;
|
||||
case XKB_KEY_Hyper_R: key = 57451; final = 'u'; break;
|
||||
case XKB_KEY_Meta_R: key = 57452; final = 'u'; break;
|
||||
#endif
|
||||
case XKB_KEY_Shift_L: if (false) {key = 57441; final = 'u';} break;
|
||||
case XKB_KEY_Control_L: if (false) {key = 57442; final = 'u';} break;
|
||||
case XKB_KEY_Alt_L: if (false) {key = 57443; final = 'u';} break;
|
||||
case XKB_KEY_Super_L: if (false) {key = 57444; final = 'u';} break;
|
||||
case XKB_KEY_Hyper_L: if (false) {key = 57445; final = 'u';} break;
|
||||
case XKB_KEY_Meta_L: if (false) {key = 57446; final = 'u';} break;
|
||||
case XKB_KEY_Shift_R: if (false) {key = 57447; final = 'u';} break;
|
||||
case XKB_KEY_Control_R: if (false) {key = 57448; final = 'u';} break;
|
||||
case XKB_KEY_Alt_R: if (false) {key = 57449; final = 'u';} break;
|
||||
case XKB_KEY_Super_R: if (false) {key = 57450; final = 'u';} break;
|
||||
case XKB_KEY_Hyper_R: if (false) {key = 57451; final = 'u';} break;
|
||||
case XKB_KEY_Meta_R: if (false) {key = 57452; final = 'u';} break;
|
||||
|
||||
default:
|
||||
if (count > 0) {
|
||||
if (effective == 0 && !released) {
|
||||
term_to_slave(term, utf8, count);
|
||||
return true;
|
||||
default: {
|
||||
/*
|
||||
* Use keysym (typically its Unicode codepoint value).
|
||||
*
|
||||
* If the keysym is shifted, use its unshifted codepoint
|
||||
* instead. In other words, ctrl+a and ctrl+shift+a should
|
||||
* both use the same value for ‘key’ (97 - i.a. ‘a’).
|
||||
*
|
||||
* However, if a non-significant modifier was used to
|
||||
* generate the symbol. This is needed since we cannot
|
||||
* encode non-significant modifiers, and thus the “extra”
|
||||
* modifier(s) would get lost.
|
||||
*
|
||||
* Example:
|
||||
*
|
||||
* the Swedish layout has ‘2’, QUOTATION MARK (“double
|
||||
* quote”), ‘@’, and ‘²’ on the same key. ‘2’ is the base
|
||||
* symbol.
|
||||
*
|
||||
* Shift+2 results in QUOTATION MARK
|
||||
* AltGr+2 results in ‘@’
|
||||
* AltGr+Shift+2 results in ‘²’
|
||||
*
|
||||
* The kitty kbd protocol can’t encode AltGr. So, if we
|
||||
* always used the base symbol (‘2’), Alt+Shift+2 would
|
||||
* result in the same escape sequence as
|
||||
* AltGr+Alt+Shift+2.
|
||||
*
|
||||
* (yes, this matches what kitty does, as of 0.23.1)
|
||||
*/
|
||||
|
||||
/* Get the key’s shift level */
|
||||
xkb_level_index_t lvl = xkb_state_key_get_level(
|
||||
seat->kbd.xkb_state, ctx->key, ctx->layout);
|
||||
|
||||
/* And get all modifier combinations that, combined with
|
||||
* the pressed key, results in the current shift level */
|
||||
xkb_mod_mask_t masks[32];
|
||||
size_t mask_count = xkb_keymap_key_get_mods_for_level(
|
||||
seat->kbd.xkb_keymap, ctx->key, ctx->layout, lvl,
|
||||
masks, ALEN(masks));
|
||||
|
||||
/* Check modifier combinations - if a combination has
|
||||
* modifiers not in our set of ‘significant’ modifiers,
|
||||
* use key sym as-is */
|
||||
bool use_level0_sym = true;
|
||||
for (size_t i = 0; i < mask_count; i++) {
|
||||
if ((masks[i] & ~seat->kbd.kitty_significant) > 0) {
|
||||
use_level0_sym = false;
|
||||
break;
|
||||
}
|
||||
|
||||
/*
|
||||
* Use keysym (typically its Unicode codepoint value).
|
||||
*
|
||||
* If the keysym is shifted, use its unshifted codepoint
|
||||
* instead. In other words, ctrl+a and ctrl+shift+a should
|
||||
* both use the same value for ‘key’ (97 - i.a. ‘a’).
|
||||
*
|
||||
* However, if a non-significant modifier was used to
|
||||
* generate the symbol. This is needed since we cannot
|
||||
* encode non-significant modifiers, and thus the “extra”
|
||||
* modifier(s) would get lost.
|
||||
*
|
||||
* Example:
|
||||
*
|
||||
* the Swedish layout has ‘2’, QUOTATION MARK (“double
|
||||
* quote”), ‘@’, and ‘²’ on the same key. ‘2’ is the base
|
||||
* symbol.
|
||||
*
|
||||
* Shift+2 results in QUOTATION MARK
|
||||
* AltGr+2 results in ‘@’
|
||||
* AltGr+Shift+2 results in ‘²’
|
||||
*
|
||||
* The kitty kbd protocol can’t encode AltGr. So, if we
|
||||
* always used the base symbol (‘2’), Alt+Shift+2 would
|
||||
* result in the same escape sequence as
|
||||
* AltGr+Alt+Shift+2.
|
||||
*
|
||||
* (yes, this matches what kitty does, as of 0.23.1)
|
||||
*/
|
||||
|
||||
/* Get the key’s shift level */
|
||||
xkb_level_index_t lvl = xkb_state_key_get_level(
|
||||
seat->kbd.xkb_state, ctx->key, ctx->layout);
|
||||
|
||||
/* And get all modifier combinations that, combined with
|
||||
* the pressed key, results in the current shift level */
|
||||
xkb_mod_mask_t masks[32];
|
||||
size_t mask_count = xkb_keymap_key_get_mods_for_level(
|
||||
seat->kbd.xkb_keymap, ctx->key, ctx->layout, lvl,
|
||||
masks, ALEN(masks));
|
||||
|
||||
/* Check modifier combinations - if a combination has
|
||||
* modifiers not in our set of ‘significant’ modifiers,
|
||||
* use key sym as-is */
|
||||
bool use_level0_sym = true;
|
||||
for (size_t i = 0; i < mask_count; i++) {
|
||||
if ((masks[i] & ~seat->kbd.kitty_significant) > 0) {
|
||||
use_level0_sym = false;
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
key = use_level0_sym && ctx->level0_syms.count > 0
|
||||
? ctx->level0_syms.syms[0]
|
||||
: sym;
|
||||
final = 'u';
|
||||
}
|
||||
|
||||
xkb_keysym_t sym_to_use = use_level0_sym && ctx->level0_syms.count > 0
|
||||
? ctx->level0_syms.syms[0]
|
||||
: sym;
|
||||
|
||||
if (composed) {
|
||||
wchar_t wc;
|
||||
if (mbtowc(&wc, (const char *)utf8, count) == count) {
|
||||
xassert(false);
|
||||
key = wc;
|
||||
}
|
||||
}
|
||||
|
||||
if (key < 0) {
|
||||
key = xkb_keysym_to_utf32(sym_to_use);
|
||||
if (key == 0)
|
||||
key = sym_to_use;
|
||||
}
|
||||
final = 'u';
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
xassert(encoded_mods >= 1);
|
||||
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue