Commit graph

8 commits

Author SHA1 Message Date
Artem Lytkin
ef1c0fd416 subseq: Score matches with dynamic programming
score_item tried every combination of positions of the needle chars in the
item, which is exponential in the length of the query. Typing the full name
of a theme like "Solarized Dark Higher Contrast" in the themes kitten froze it
for about half a minute.

Instead find the best score for each position of each needle char from the
best scores for the previous char. The per char scores are added in the same
order as before, so the scores are bit for bit identical. When several sets
of positions have exactly the same score, the highlighted positions can be
different from before.
2026-09-25 16:46:19 +03:00
Kovid Goyal
b627d2e4ab
Move error handling code into its own library 2025-10-12 13:51:16 +05:30
Kovid Goyal
3059c87bd0
Use a lock and channel free parallelization for subseq scoring 2025-06-06 06:39:09 +05:30
Kovid Goyal
cc500893f0
... 2025-05-21 12:16:25 +05:30
Kovid Goyal
c861259e3b
Rename go module from kitty -> github.com/kovidgoyal/kitty
Makes the code more easily re-useable in other projects
2025-05-16 08:43:39 +05:30
Kovid Goyal
405f5ce148
Use stdlib maps/slices 2024-06-24 07:54:13 +05:30
Kovid Goyal
341d845b9a
Port calls to slices.Sort functions since they now need a cmp() function rather than a less() function
Also rename os.SEEK_* to io.Seek* as the former has been deprecated
2023-08-04 22:50:13 +05:30
Kovid Goyal
29dd2438c9
Port the subseq matcher to Go 2023-03-14 12:29:44 +05:30