mirror of
https://codeberg.org/dnkl/foot.git
synced 2026-06-01 21:38:16 -04:00
unicode-combining: store seen combining chains "globally" in the term struct
Instead of storing combining data per cell, realize that most combinations are re-occurring and that there's lots of available space left in the unicode range, and store seen base+combining combinations chains in a per-terminal array. When we encounter a combining character, we first try to pre-compose, like before. If that fails, we then search for the current base+combining combo in the list of previously seen combinations. If not found there either, we allocate a new combo and add it to the list. Regardless, the result is an index into this array. We store this index, offsetted by COMB_CHARS_LO=0x40000000ul in the cell. When rendering, we need to check if the cell character is a plain character, or if it's a composed character (identified by checking if the cell character is >= COMB_CHARS_LO). Then we render the grapheme pretty much like before.
This commit is contained in:
parent
ae7383189a
commit
62e0774319
8 changed files with 97 additions and 92 deletions
23
grid.c
23
grid.c
|
|
@ -34,17 +34,10 @@ grid_row_alloc(int cols, bool initialize)
|
||||||
|
|
||||||
if (initialize) {
|
if (initialize) {
|
||||||
row->cells = calloc(cols, sizeof(row->cells[0]));
|
row->cells = calloc(cols, sizeof(row->cells[0]));
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
|
||||||
row->comb_chars = calloc(cols, sizeof(row->comb_chars[0]));
|
|
||||||
#endif
|
|
||||||
for (size_t c = 0; c < cols; c++)
|
for (size_t c = 0; c < cols; c++)
|
||||||
row->cells[c].attrs.clean = 1;
|
row->cells[c].attrs.clean = 1;
|
||||||
} else {
|
} else
|
||||||
row->cells = malloc(cols * sizeof(row->cells[0]));
|
row->cells = malloc(cols * sizeof(row->cells[0]));
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
|
||||||
row->comb_chars = malloc(cols * sizeof(row->comb_chars[0]));
|
|
||||||
#endif
|
|
||||||
}
|
|
||||||
|
|
||||||
return row;
|
return row;
|
||||||
}
|
}
|
||||||
|
|
@ -55,9 +48,6 @@ grid_row_free(struct row *row)
|
||||||
if (row == NULL)
|
if (row == NULL)
|
||||||
return;
|
return;
|
||||||
|
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
|
||||||
free(row->comb_chars);
|
|
||||||
#endif
|
|
||||||
free(row->cells);
|
free(row->cells);
|
||||||
free(row);
|
free(row);
|
||||||
}
|
}
|
||||||
|
|
@ -214,17 +204,6 @@ grid_reflow(struct grid *grid, int new_rows, int new_cols,
|
||||||
new_row->cells[new_col_idx] = *old_cell;
|
new_row->cells[new_col_idx] = *old_cell;
|
||||||
new_row->cells[new_col_idx].attrs.clean = 1;
|
new_row->cells[new_col_idx].attrs.clean = 1;
|
||||||
|
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
|
||||||
struct combining_chars *old_comb_chars
|
|
||||||
= &old_row->comb_chars[c - empty_count + i];
|
|
||||||
struct combining_chars *new_comb_chars
|
|
||||||
= &new_row->comb_chars[new_col_idx];
|
|
||||||
|
|
||||||
new_comb_chars->count = old_comb_chars->count;
|
|
||||||
for (size_t j = 0; j < ALEN(new_comb_chars->chars); j++)
|
|
||||||
new_comb_chars->chars[j] = old_comb_chars->chars[j];
|
|
||||||
#endif
|
|
||||||
|
|
||||||
/* Translate tracking point(s) */
|
/* Translate tracking point(s) */
|
||||||
if (is_tracking_point && i >= empty_count) {
|
if (is_tracking_point && i >= empty_count) {
|
||||||
tll_foreach(tracking_points, it) {
|
tll_foreach(tracking_points, it) {
|
||||||
|
|
|
||||||
10
meson.build
10
meson.build
|
|
@ -57,11 +57,8 @@ wayland_client = dependency('wayland-client')
|
||||||
wayland_cursor = dependency('wayland-cursor')
|
wayland_cursor = dependency('wayland-cursor')
|
||||||
xkb = dependency('xkbcommon')
|
xkb = dependency('xkbcommon')
|
||||||
|
|
||||||
add_project_arguments('-DFOOT_UNICODE_MAX_COMBINING_CHARS=@0@'.format(
|
|
||||||
get_option('unicode-max-combining-chars')), language: 'c')
|
|
||||||
add_project_arguments('-DFOOT_UNICODE_PRECOMPOSE=@0@'.format(
|
add_project_arguments('-DFOOT_UNICODE_PRECOMPOSE=@0@'.format(
|
||||||
get_option('unicode-max-combining-chars') > 0 and get_option('unicode-precompose')),
|
get_option('unicode-precompose')), language: 'c')
|
||||||
language: 'c')
|
|
||||||
|
|
||||||
tllist = dependency('tllist', version: '>=1.0.1', fallback: 'tllist')
|
tllist = dependency('tllist', version: '>=1.0.1', fallback: 'tllist')
|
||||||
fcft = dependency('fcft', version: ['>=2.0.0', '<2.1.0'], fallback: 'fcft')
|
fcft = dependency('fcft', version: ['>=2.0.0', '<2.1.0'], fallback: 'fcft')
|
||||||
|
|
@ -95,7 +92,7 @@ foreach prot : [
|
||||||
command: [wscanner_prog, 'private-code', '@INPUT@', '@OUTPUT@'])
|
command: [wscanner_prog, 'private-code', '@INPUT@', '@OUTPUT@'])
|
||||||
endforeach
|
endforeach
|
||||||
|
|
||||||
if get_option('unicode-max-combining-chars') > 0 and get_option('unicode-precompose')
|
if get_option('unicode-precompose')
|
||||||
generate_unicode_precompose_sh = files('scripts/generate-unicode-precompose.sh')
|
generate_unicode_precompose_sh = files('scripts/generate-unicode-precompose.sh')
|
||||||
unicode_data = custom_target(
|
unicode_data = custom_target(
|
||||||
'unicode-data',
|
'unicode-data',
|
||||||
|
|
@ -167,8 +164,7 @@ subdir('doc')
|
||||||
|
|
||||||
summary(
|
summary(
|
||||||
{
|
{
|
||||||
'Unicode max combining chars': get_option('unicode-max-combining-chars'),
|
'Unicode precompose': get_option('unicode-precompose'),
|
||||||
'Unicode precompose': get_option('unicode-max-combining-chars') > 0 and get_option('unicode-precompose'),
|
|
||||||
},
|
},
|
||||||
bool_yn: true
|
bool_yn: true
|
||||||
)
|
)
|
||||||
|
|
|
||||||
|
|
@ -1,5 +1,2 @@
|
||||||
option('unicode-max-combining-chars', type: 'integer', value: 2,
|
|
||||||
description: 'Maximum number of combining characters to track per cell. A value of 0 completely disables unicode combining (this reduces the runtime memory footprint)')
|
|
||||||
|
|
||||||
option('unicode-precompose', type: 'boolean', value: true,
|
option('unicode-precompose', type: 'boolean', value: true,
|
||||||
description: 'Convert decomposed characters to precomposed. Ignored if "unicode-combining" has been disabled')
|
description: 'Convert decomposed characters to precomposed. Ignored if "unicode-combining" has been disabled')
|
||||||
|
|
|
||||||
45
render.c
45
render.c
|
|
@ -402,9 +402,20 @@ render_cell(struct terminal *term, pixman_image_t *pix,
|
||||||
|
|
||||||
struct fcft_font *font = attrs_to_font(term, &cell->attrs);
|
struct fcft_font *font = attrs_to_font(term, &cell->attrs);
|
||||||
const struct fcft_glyph *glyph = NULL;
|
const struct fcft_glyph *glyph = NULL;
|
||||||
|
const struct composed *composed = NULL;
|
||||||
|
|
||||||
if (cell->wc != 0)
|
if (cell->wc != 0) {
|
||||||
glyph = fcft_glyph_rasterize(font, cell->wc, term->font_subpixel);
|
wchar_t base = cell->wc;
|
||||||
|
|
||||||
|
if (base >= COMB_CHARS_LO &&
|
||||||
|
base < (COMB_CHARS_LO + term->composed_count))
|
||||||
|
{
|
||||||
|
composed = &term->composed[base - COMB_CHARS_LO];
|
||||||
|
base = composed->base;
|
||||||
|
}
|
||||||
|
|
||||||
|
glyph = fcft_glyph_rasterize(font, base, term->font_subpixel);
|
||||||
|
}
|
||||||
|
|
||||||
int cell_cols = glyph != NULL ? max(1, glyph->cols) : 1;
|
int cell_cols = glyph != NULL ? max(1, glyph->cols) : 1;
|
||||||
|
|
||||||
|
|
@ -442,25 +453,25 @@ render_cell(struct terminal *term, pixman_image_t *pix,
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
|
||||||
/* Combining characters */
|
/* Combining characters */
|
||||||
const struct combining_chars *comb_chars = &row->comb_chars[col];
|
if (composed != NULL) {
|
||||||
for (size_t i = 0; i < comb_chars->count; i++) {
|
for (size_t i = 0; i < composed->count; i++) {
|
||||||
const struct fcft_glyph *g = fcft_glyph_rasterize(
|
const struct fcft_glyph *g = fcft_glyph_rasterize(
|
||||||
font, comb_chars->chars[i], term->font_subpixel);
|
font, composed->combining[i], term->font_subpixel);
|
||||||
|
|
||||||
if (g == NULL)
|
if (g == NULL)
|
||||||
continue;
|
continue;
|
||||||
|
|
||||||
pixman_image_composite32(
|
pixman_image_composite32(
|
||||||
PIXMAN_OP_OVER, clr_pix, g->pix, pix, 0, 0, 0, 0,
|
PIXMAN_OP_OVER, clr_pix, g->pix, pix, 0, 0, 0, 0,
|
||||||
/* Some fonts use a negative offset, while others use a
|
/* Some fonts use a negative offset, while others use a
|
||||||
* "normal" offset */
|
* "normal" offset */
|
||||||
x + (g->x < 0 ? term->cell_width : 0) + g->x,
|
x + (g->x < 0 ? term->cell_width : 0) + g->x,
|
||||||
y + font_baseline(term) - g->y,
|
y + font_baseline(term) - g->y,
|
||||||
g->width, g->height);
|
g->width, g->height);
|
||||||
|
}
|
||||||
}
|
}
|
||||||
#endif
|
|
||||||
pixman_image_unref(clr_pix);
|
pixman_image_unref(clr_pix);
|
||||||
|
|
||||||
/* Underline */
|
/* Underline */
|
||||||
|
|
|
||||||
24
selection.c
24
selection.c
|
|
@ -142,12 +142,7 @@ min_bufsize_for_extraction(const struct terminal *term)
|
||||||
{
|
{
|
||||||
const struct coord *start = &term->selection.start;
|
const struct coord *start = &term->selection.start;
|
||||||
const struct coord *end = &term->selection.end;
|
const struct coord *end = &term->selection.end;
|
||||||
const size_t chars_per_cell =
|
const size_t chars_per_cell = 1 + ALEN(term->composed[0].combining);
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
|
||||||
1 + ALEN(term->grid->cur_row->comb_chars[0].chars);
|
|
||||||
#else
|
|
||||||
1;
|
|
||||||
#endif
|
|
||||||
|
|
||||||
switch (term->selection.kind) {
|
switch (term->selection.kind) {
|
||||||
case SELECTION_NONE:
|
case SELECTION_NONE:
|
||||||
|
|
@ -239,16 +234,17 @@ extract_one(struct terminal *term, struct row *row, struct cell *cell,
|
||||||
ctx->empty_count = 0;
|
ctx->empty_count = 0;
|
||||||
|
|
||||||
assert(ctx->idx + 1 <= ctx->size);
|
assert(ctx->idx + 1 <= ctx->size);
|
||||||
ctx->buf[ctx->idx++] = cell->wc;
|
|
||||||
|
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
if (cell->wc >= COMB_CHARS_LO && cell->wc < (COMB_CHARS_LO + term->composed_count)) {
|
||||||
const struct combining_chars *comb_chars = &row->comb_chars[col];
|
const struct composed *composed = &term->composed[cell->wc - COMB_CHARS_LO];
|
||||||
|
|
||||||
assert(cell->wc != 0);
|
ctx->buf[ctx->idx++] = composed->base;
|
||||||
assert(ctx->idx + comb_chars->count <= ctx->size);
|
|
||||||
for (size_t i = 0; i < comb_chars->count; i++)
|
assert(ctx->idx + composed->count <= ctx->size);
|
||||||
ctx->buf[ctx->idx++] = comb_chars->chars[i];
|
for (size_t i = 0; i < composed->count; i++)
|
||||||
#endif
|
ctx->buf[ctx->idx++] = composed->combining[i];
|
||||||
|
} else
|
||||||
|
ctx->buf[ctx->idx++] = cell->wc;
|
||||||
|
|
||||||
ctx->last_row = row;
|
ctx->last_row = row;
|
||||||
ctx->last_cell = cell;
|
ctx->last_cell = cell;
|
||||||
|
|
|
||||||
|
|
@ -822,6 +822,8 @@ term_init(const struct config *conf, struct fdm *fdm, struct wayland *wayl,
|
||||||
.normal = {.damage = tll_init(), .scroll_damage = tll_init(), .sixel_images = tll_init()},
|
.normal = {.damage = tll_init(), .scroll_damage = tll_init(), .sixel_images = tll_init()},
|
||||||
.alt = {.damage = tll_init(), .scroll_damage = tll_init(), .sixel_images = tll_init()},
|
.alt = {.damage = tll_init(), .scroll_damage = tll_init(), .sixel_images = tll_init()},
|
||||||
.grid = &term->normal,
|
.grid = &term->normal,
|
||||||
|
.composed_count = 0,
|
||||||
|
.composed = NULL,
|
||||||
.meta = {
|
.meta = {
|
||||||
.esc_prefix = true,
|
.esc_prefix = true,
|
||||||
.eight_bit = true,
|
.eight_bit = true,
|
||||||
|
|
@ -1086,6 +1088,8 @@ term_destroy(struct terminal *term)
|
||||||
tll_free(term->normal.scroll_damage);
|
tll_free(term->normal.scroll_damage);
|
||||||
tll_free(term->alt.scroll_damage);
|
tll_free(term->alt.scroll_damage);
|
||||||
|
|
||||||
|
free(term->composed);
|
||||||
|
|
||||||
free(term->window_title);
|
free(term->window_title);
|
||||||
tll_free_and_free(term->window_title_stack, free);
|
tll_free_and_free(term->window_title_stack, free);
|
||||||
|
|
||||||
|
|
@ -2295,10 +2299,6 @@ term_print(struct terminal *term, wchar_t wc, int width)
|
||||||
cell->wc = term->vt.last_printed = wc;
|
cell->wc = term->vt.last_printed = wc;
|
||||||
cell->attrs = term->vt.attrs;
|
cell->attrs = term->vt.attrs;
|
||||||
|
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
|
||||||
row->comb_chars[term->grid->cursor.point.col].count = 0;
|
|
||||||
#endif
|
|
||||||
|
|
||||||
row->dirty = true;
|
row->dirty = true;
|
||||||
cell->attrs.clean = 0;
|
cell->attrs.clean = 0;
|
||||||
|
|
||||||
|
|
|
||||||
17
terminal.h
17
terminal.h
|
|
@ -77,21 +77,16 @@ struct damage {
|
||||||
int lines;
|
int lines;
|
||||||
};
|
};
|
||||||
|
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
struct composed {
|
||||||
struct combining_chars {
|
wchar_t base;
|
||||||
|
wchar_t combining[5];
|
||||||
uint8_t count;
|
uint8_t count;
|
||||||
wchar_t chars[FOOT_UNICODE_MAX_COMBINING_CHARS];
|
};
|
||||||
} __attribute__((packed));
|
|
||||||
#endif
|
|
||||||
|
|
||||||
struct row {
|
struct row {
|
||||||
struct cell *cells;
|
struct cell *cells;
|
||||||
bool dirty;
|
bool dirty;
|
||||||
bool linebreak;
|
bool linebreak;
|
||||||
|
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
|
||||||
struct combining_chars *comb_chars;
|
|
||||||
#endif
|
|
||||||
};
|
};
|
||||||
|
|
||||||
struct sixel {
|
struct sixel {
|
||||||
|
|
@ -221,6 +216,10 @@ struct terminal {
|
||||||
struct grid alt;
|
struct grid alt;
|
||||||
struct grid *grid;
|
struct grid *grid;
|
||||||
|
|
||||||
|
#define COMB_CHARS_LO 0x40000000ul
|
||||||
|
size_t composed_count;
|
||||||
|
struct composed *composed;
|
||||||
|
|
||||||
struct fcft_font *fonts[4];
|
struct fcft_font *fonts[4];
|
||||||
int font_dpi;
|
int font_dpi;
|
||||||
int font_adjustments;
|
int font_adjustments;
|
||||||
|
|
|
||||||
59
vt.c
59
vt.c
|
|
@ -571,8 +571,6 @@ action_utf8_print(struct terminal *term, uint8_t c)
|
||||||
|
|
||||||
int width = wcwidth(wc);
|
int width = wcwidth(wc);
|
||||||
|
|
||||||
#if FOOT_UNICODE_MAX_COMBINING_CHARS > 0
|
|
||||||
|
|
||||||
/*
|
/*
|
||||||
* Is this is combining character? The basic assumption is that if
|
* Is this is combining character? The basic assumption is that if
|
||||||
* wcwdith() returns 0, then it *is* a combining character.
|
* wcwdith() returns 0, then it *is* a combining character.
|
||||||
|
|
@ -606,6 +604,15 @@ action_utf8_print(struct terminal *term, uint8_t c)
|
||||||
base = row->cells[base_col].wc;
|
base = row->cells[base_col].wc;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
const struct composed *composed =
|
||||||
|
(base >= COMB_CHARS_LO &&
|
||||||
|
base < (COMB_CHARS_LO + term->composed_count))
|
||||||
|
? &term->composed[base - COMB_CHARS_LO]
|
||||||
|
: NULL;
|
||||||
|
|
||||||
|
if (composed != NULL)
|
||||||
|
base = composed->base;
|
||||||
|
|
||||||
int base_width = wcwidth(base);
|
int base_width = wcwidth(base);
|
||||||
|
|
||||||
if (base != 0 && base_width > 0) {
|
if (base != 0 && base_width > 0) {
|
||||||
|
|
@ -619,35 +626,55 @@ action_utf8_print(struct terminal *term, uint8_t c)
|
||||||
* pre-composed character, as that is likely to produce a
|
* pre-composed character, as that is likely to produce a
|
||||||
* better looking result.
|
* better looking result.
|
||||||
*/
|
*/
|
||||||
|
term->grid->cursor.point.col = base_col;
|
||||||
struct combining_chars *comb_chars = &row->comb_chars[base_col];
|
term->grid->cursor.lcf = false;
|
||||||
|
|
||||||
#if FOOT_UNICODE_PRECOMPOSE
|
#if FOOT_UNICODE_PRECOMPOSE
|
||||||
if (comb_chars->count == 0) {
|
if (composed == NULL) {
|
||||||
wchar_t precomposed = precompose(base, wc);
|
wchar_t precomposed = precompose(base, wc);
|
||||||
int precomposed_width = wcwidth(precomposed);
|
int precomposed_width = wcwidth(precomposed);
|
||||||
if (precomposed != (wchar_t)-1 && precomposed_width == base_width) {
|
if (precomposed != (wchar_t)-1 && precomposed_width == base_width) {
|
||||||
term->grid->cursor.point.col = base_col;
|
|
||||||
term->grid->cursor.lcf = false;
|
|
||||||
term_print(term, precomposed, precomposed_width);
|
term_print(term, precomposed, precomposed_width);
|
||||||
return;
|
return;
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
#endif
|
#endif
|
||||||
|
|
||||||
if (comb_chars->count < ALEN(comb_chars->chars))
|
size_t wanted_count = composed != NULL ? composed->count + 1 : 1;
|
||||||
comb_chars->chars[comb_chars->count++] = wc;
|
|
||||||
else {
|
/* Look for existing combining chain */
|
||||||
LOG_WARN("combining character overflow:");
|
for (size_t i = 0; i < term->composed_count; i++) {
|
||||||
LOG_WARN(" 0x%04x", base);
|
const struct composed *cc = &term->composed[i];
|
||||||
for (size_t i = 0; i < comb_chars->count; i++)
|
if (cc->base != base)
|
||||||
LOG_WARN(" 0x%04x", comb_chars->chars[i]);
|
continue;
|
||||||
LOG_ERR(" 0x%04x", wc);
|
|
||||||
|
if (cc->count != wanted_count)
|
||||||
|
continue;
|
||||||
|
|
||||||
|
if (cc->combining[wanted_count - 1] != wc)
|
||||||
|
continue;
|
||||||
|
|
||||||
|
term_print(term, COMB_CHARS_LO + i, base_width);
|
||||||
|
return;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
/* Allocate new chain */
|
||||||
|
|
||||||
|
struct composed new_cc;
|
||||||
|
new_cc.base = base;
|
||||||
|
new_cc.count = wanted_count;
|
||||||
|
for (size_t i = 0; i < wanted_count - 1; i++)
|
||||||
|
new_cc.combining[i] = composed->combining[i];
|
||||||
|
new_cc.combining[wanted_count - 1] = wc;
|
||||||
|
|
||||||
|
term->composed_count++;
|
||||||
|
term->composed = realloc(term->composed, term->composed_count * sizeof(term->composed[0]));
|
||||||
|
term->composed[term->composed_count - 1] = new_cc;
|
||||||
|
|
||||||
|
term_print(term, COMB_CHARS_LO + term->composed_count - 1, base_width);
|
||||||
return;
|
return;
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
#endif /* FOOT_UNICODE_MAX_COMBINING_CHARS > 0 */
|
|
||||||
|
|
||||||
term_print(term, wc, width);
|
term_print(term, wc, width);
|
||||||
}
|
}
|
||||||
|
|
|
||||||
Loading…
Add table
Add a link
Reference in a new issue