Use hard-coded limit to keep it simple. An alternative is to dynamically
calculate the width of the visible area and use it as the limit, but it
can cause unwanted truncation of the query on screen resize/split.
- Make structs smaller
- Introduce Result struct and use it to represent matched items instead of
reusing Item struct for that purpose
- Avoid unnecessary memory allocation
- Avoid growing slice from the initial capacity
- Code cleanup
Sort performance increases as the size of each sublist decreases (n in
nlog(n) decreases). Merger is then responsible for merging the sorted
lists in order, and since in most cases we are only interesed in the
matches in the first page on the screen so the overhead in the process
is negligible.
In the best case (all ascii), this reduces the memory footprint by 60%
and the response time by 15% to 20%. In the worst case (every line has
non-ascii characters), 3 to 4% overhead is observed.
When we prepend a single quote to our query in --exact mode, we are not
supposed to limit the scope of the new search to the previous
exact-match result.
Based on the patch by Matt Westcott (@mjwestcott).
But with a more conservative approach:
- Does not use linearly increasing penalties; It is agreed upon that we
should prefer matching characters at the beginnings of the words, but
it's not always clear that the relevance is inversely proportional to
the distance from the beginning.
- The approach here is more conservative in that the bonus is never
large enough to override the matchlen, so it can be thought of as the
first implicit tiebreak criterion.
- One may argue the change breaks the contract of --tiebreak, but the
judgement depends on the definition of "tie".
- Slightly more efficient processing of Options
- Do not return reference type arguments that are mutated inside the
function
- Use util.Constrain function when appropriate
Note that $SHELL only points to the default shell instead of the current
shell. If you're on a non-default shell, you might want to override the
value like follows.
SHELL=zsh fzf --bind 'enter:execute:echo $ZSH_VERSION; sleep 1'
fzf defers the initial rendering of the screen up to 100ms if the input
stream is ongoing to prevent unnecessary redraw during the initial
phase. However, 100ms delay is quite noticeable and might give the
impression that fzf is not snappy enough. This commit reduces the
maximum delay down to 20ms when --tac is not specified, in which case
the input list quickly fills the entire screen.
Related: #452
When `--multi` is set, tab key will bring your cursor down, and
shift-tab up. But since fzf by default draws the screen in bottom-up
fashion, one may feel that the opposite of the behavior is more
desirable and choose to customize the key bindings as follows.
export FZF_DEFAULT_OPTS="--bind tab:toggle-up,shift-tab:toggle-down"
This configuration, however, becomes no longer straightforward when
`--reverse` is set and fzf switches to top-down layout. To address the
requirement, this commit adds `toggle-in` and `toggle-out` option which
switch direction depending on `--reverse`-ness.
export FZF_DEFAULT_OPTS="--bind tab:toggle-out,shift-tab:toggle-in"
Having submodules causes vim-plug or other vim plugin managers to clone
them with no real benefit to the end-users. There's currently no
compelling reason for me to use submodules.
