You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For now sequences are looked up in the order provided by the user. According to my testing, this beats the the original PR's performance, mainly due to how we filter down and track the state of candidate sequences between key presses.
Nevertheless, it may be worthwhile creating a sorted pointer array in data during initialization.
Sort by sequence length for small performance boost
Reverse-sort identical keycodes by modifiers so if both LS(A) and A are specified we first test for the former
Sort by usage page and keycodes and abort lookup if we exceed the current press? This speeds up termination on failed lookups but is otherwise somewhat arbitrary
The text was updated successfully, but these errors were encountered:
For now sequences are looked up in the order provided by the user. According to my testing, this beats the the original PR's performance, mainly due to how we filter down and track the state of candidate sequences between key presses.
Nevertheless, it may be worthwhile creating a sorted pointer array in
data
during initialization.LS(A)
andA
are specified we first test for the formerThe text was updated successfully, but these errors were encountered: