mirrors/foot
1
0
Fork 0
mirror of https://codeberg.org/dnkl/foot.git synced 2026-02-04 04:06:06 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

vt: improve key calculation for compose sequences

Browse source 
* Don’t assume 32 bits when rotating the old key. Use the number of
  actual bits available, as determined by CELL_COMB_CHARS_{HI,LO}
* Multiply with magic hash constant

This greatly reduces the number of collisions seen. For example, the
Emoji test file (from the Unicode specification), now has zero
collisions.
This commit is contained in:
Daniel Eklöf 2022-06-13 13:11:56 +02:00
parent edd68732ad
commit fbcb30bf98
No known key found for this signature in database
GPG key ID: 5BBD4992C116573F
1 changed files with 19 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

23
vt.c
View file

@ -622,8 +622,21 @@ action_put(struct terminal *term, uint8_t c)
static inline uint32_t
chain_key(uint32_t old_key, uint32_t new_wc)
{
/* Rotate left 8 bits, xor with new char */
return ((old_key << 8) | (old_key >> (32 - 8))) ^ new_wc;
unsigned bits = 32 - __builtin_clz(CELL_COMB_CHARS_HI - CELL_COMB_CHARS_LO);
/* Rotate old key 8 bits */
uint32_t new_key = (old_key << 8) | (old_key >> (bits - 8));
/* xor with new char */
new_key ^= new_wc;
/* Multiply with magic hash constant */
new_key *= 2654435761;
/* And mask, to ensure the new value is within range */
new_key &= CELL_COMB_CHARS_HI - CELL_COMB_CHARS_LO;
return new_key;
}
static void
@ -668,8 +681,6 @@ action_utf8_print(struct terminal *term, char32_t wc)
} else
key = chain_key(base, wc);
key &= CELL_COMB_CHARS_HI - CELL_COMB_CHARS_LO;
#if defined(FOOT_GRAPHEME_CLUSTERING)
if (grapheme_clustering) {
/* Check if we're on a grapheme cluster break */
@ -767,6 +778,10 @@ action_utf8_print(struct terminal *term, char32_t wc)
cc->count != wanted_count ||
cc->chars[wanted_count - 1] != wc)
{
#if 0
LOG_WARN("COLLISION: base: %04x/%04x, count: %d/%zu, last: %04x/%04x",
cc->chars[0], base, cc->count, wanted_count, cc->chars[wanted_count - 1], wc);
#endif
key++;
collision_count++;
continue;