Problem: pre-inserted text not exposed in complete_info()
Solution: Add the pre-inserted text to the complete_info() Vim script
function (Girish Palya)
closes: #18571
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Feat: expose preinserted text in complete_info()
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: whitespace not cleared with 'ai'
Solution: Remove spaces added by 'autoindent' when autocomplete is set
and restore did_ai in ins_compl_start() (Maxim Kim)
fixes#18560closes: #18582
Signed-off-by: Maxim Kim <habamax@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: A few typos in the code
Solution: Fix typos (zeertzjq)
closes: #18550
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Generating prototype files does not work on all platforms
Solution: Rework prototypes generation using python instead of cproto,
enable it in CI to test it for each PR (Hirohito Higashi).
closes: #18045
Signed-off-by: Hirohito Higashi <h.east.727@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: flicker when LSP server is slow
Solution: reinsert leader text before invoking user function
(Girish Palya)
Reference:
https://github.com/girishji/vimcomplete/issues/101#issuecomment-3343063245
In insert-mode completion, the leader text is temporarily removed while
searching for candidates. When the LSP server responds slowly, the
client may call `:sleep` to wait, which triggers `out_flush()`. This
causes the deleted text to disappear briefly before being redrawn,
resulting in visible flicker.
This commit reinserts the leader text before invoking the user function,
and removes it again afterward to eliminate flicker.
closes: #18468
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: some issues with 'acl' when "preinsert" and
"longest" is in 'completeopt' (musonius, after v9.1.1638)
Solution: Fix various issues (see details below) (Girish Palya)
This commit addresses multiple issues in the 'autocompletedelay' behavior with
"preinsert" and "longest":
- Prevents spurious characters from being inserted.
- Ensures the completion menu is not shown until `autocompletedelay` has
expired.
- Shows the "preinsert" effect immediately.
- Keeps the "preinsert" effect visible even when a character is deleted.
fixes: #18443closes: #18460
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: flicker with slow LSPs
Solution: Disable flushing the changed text (Girish Palya).
In insert-mode completion, the leader text is temporarily deleted while
searching for completion candidates. If the LSP server responds slowly,
the client may call `:sleep` to wait, which triggers `out_flush()`. This
causes the deleted text to briefly disappear before being redrawn when
results arrive, producing a visible flicker.
There are two possible fixes:
1. Suppress flushing while a user function (e.g. LSP client) is waiting.
2. Reinsert the deleted text before invoking the user function.
This Commit implements (1), which is the simpler solution, though somewhat
heavy-handed. If you think this may introduce unwanted side effects, I
can rework it to use (2).
closes: #18439
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: "longest" doesn't work for manual completion when
'autocomplete' is on (after 9.1.1800).
Solution: Only reset compl_get_longest when enabling autocompletion
(zeertzjq).
closes: #18430
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: autocompletion can be improved
Solution: Add support for "longest" and "preinsert" in 'autocomplete';
add preinserted() (Girish Palya)
* Add support for "longest" in 'completeopt' when 'autocomplete'
is enabled. (Note: the cursor position does not change automatically
when 'autocomplete' is enabled.)
* Add support for "preinsert" when 'autocomplete' is enabled. Ensure
"preinsert" works the same with and without 'autocomplete'
* introduce the preinserted() Vim script function, useful for defining
custom key mappings.
fixes: #18314closes: #18387
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: Enter does not insert match with "noinsert".
(Sergey Vlasov)
Solution: Check for compl_shown_match instead of compl_selected_item
(zeertzjq).
fixes: #18386
related: #1653closes: #18395
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: ins_compl_leader() defined too far away from
ins_compl_leader_len() (after 9.1.1781).
Solution: Move ins_compl_leader() just before ins_compl_leader_len().
Add missing "static" to eval0_simple_funccal() definition.
(zeertzjq).
closes: #18364
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: some functions could be defined static
Solution: Change function definitions and make them static
(Yegappan Lakshmanan)
closes: #18354
Signed-off-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: 'autocomplete' cannot be enabled per buffer
(Tomasz N)
Solution: Make 'autocomplete' global or local to buffer (Girish Palya)
fixes: #18320closes: #18333
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: "adding" expansion doesn't work with cfc=keyword.
Solution: Remove incorrect assignment to compl_cont_status. Don't use
fuzzy collection in "adding" status (zeertzjq).
related: neovim/neovim#35842
closes: #18351
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: complete: some redraw issues with 'autocomplete'
Solution: Fix the issues (Girish Palya)
This commit contains the following changes:
* Fix that wildtrigger() might leave opened popupmenu around #18298
* Remove blinking message on the command line when a menu item from a loaded
buffer is selected during 'autocomplete'
* Add a test for PR #18265 to demonstrate why the PR is required for correct
'autocomplete' behavior
fixes: #18298closes: #18328
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: "preinsert" insert wrong word with 'smartcase' and
'autocomplete'
Solution: Add compare completed item with the leader (Girish Palya)
closes: #18313
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: selected item not cleared on backspace when
'autocomplete' is set
Solution: Clear the selected item (Girish Palya)
closes: #18260
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: preinserted text highlighed using ComplMatchIns
Solution: Use highlighting group PreInsert and update the documentation
(Girish Palya).
When "preinsert" is included in 'completeopt', only the PreInsert
highlight group should be applied, whether autocompletion is active or not.
Previously, ComplMatchIns was used when autocompletion was not enabled.
Related to https://github.com/vim/vim/pull/18213.
closes: #18254
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: redo (.) broken with preinsert and autocompletion
Solution: Make redo (.) work with preinsert and autocompletion
(Girish Palya)
closes: #18253
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: complete: preinsert does not work well with preinsert
Solution: Make "preinsert" completeopt value work with autocompletion
(Girish Palya)
This change extends Insert mode autocompletion so that 'preinsert' also
works when 'autocomplete' is enabled.
Try: `:set ac cot=preinsert`
See `:help 'cot'` for more details.
closes: #18213
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: not possible to delay the autcompletion
Solution: add the 'autocompletedelay' option value (Girish Palya).
This patch introduces a new global option 'autocompletedelay'/'acl' that
specifies the delay, in milliseconds, before the autocomplete menu
appears after typing.
When set to a non-zero value, Vim waits for the specified time before
showing the completion popup, allowing users to reduce distraction from
rapid suggestion pop-ups or to fine-tune the responsiveness of
completion.
The default value is 0, which preserves the current immediate-popup
behavior.
closes: #17960
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: fuzzy-matching can be improved
Solution: Implement a better fuzzy matching algorithm
(Girish Palya)
Replace fuzzy matching algorithm with improved fzy-based implementation
The
[current](https://www.forrestthewoods.com/blog/reverse_engineering_sublime_texts_fuzzy_match/)
fuzzy matching algorithm has several accuracy issues:
* It struggles with CamelCase
* It fails to prioritize matches at the beginning of strings, often
ranking middle matches higher.
After evaluating alternatives (see my comments
[here](https://github.com/vim/vim/issues/17531#issuecomment-3112046897)
and
[here](https://github.com/vim/vim/issues/17531#issuecomment-3121593900)),
I chose to adopt the [fzy](https://github.com/jhawthorn/fzy) algorithm,
which:
* Resolves the aforementioned issues.
* Performs better.
Implementation details
This version is based on the original fzy
[algorithm](https://github.com/jhawthorn/fzy/blob/master/src/match.c),
with one key enhancement: **multibyte character support**.
* The original implementation supports only ASCII.
* This patch replaces ascii lookup tables with function calls, making it
compatible with multibyte character sets.
* Core logic (`match_row()` and `match_positions()`) remains faithful to
the original, but now operates on codepoints rather than single-byte
characters.
Performance
Tested against a dataset of **90,000 Linux kernel filenames**. Results
(in milliseconds) show a **\~2x performance improvement** over the
current fuzzy matching algorithm.
```
Search String Current Algo FZY Algo
-------------------------------------------------
init 131.759 66.916
main 83.688 40.861
sig 98.348 39.699
index 109.222 30.738
ab 72.222 44.357
cd 83.036 54.739
a 58.94 62.242
b 43.612 43.442
c 64.39 67.442
k 40.585 36.371
z 34.708 22.781
w 38.033 30.109
cpa 82.596 38.116
arz 84.251 23.964
zzzz 35.823 22.75
dimag 110.686 29.646
xa 43.188 29.199
nha 73.953 31.001
nedax 94.775 29.568
dbue 79.846 25.902
fp 46.826 31.641
tr 90.951 55.883
kw 38.875 23.194
rp 101.575 55.775
kkkkkkkkkkkkkkkkkkkkkkkkkkkkk 48.519 30.921
```
```vim
vim9script
var haystack = readfile('/Users/gp/linux.files')
var needles = ['init', 'main', 'sig', 'index', 'ab', 'cd', 'a', 'b',
'c', 'k',
'z', 'w', 'cpa', 'arz', 'zzzz', 'dimag', 'xa', 'nha', 'nedax',
'dbue',
'fp', 'tr', 'kw', 'rp', 'kkkkkkkkkkkkkkkkkkkkkkkkkkkkk']
for needle in needles
var start = reltime()
var tmp = matchfuzzy(haystack, needle)
echom $'{needle}' (start->reltime()->reltimefloat() * 1000)
endfor
```
Additional changes
* Removed the "camelcase" option from both matchfuzzy() and
matchfuzzypos(), as it's now obsolete with the improved algorithm.
related: neovim/neovim#34101
fixes#17531closes: #17900
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Autocompletion slow with include- and tag-completion
Solution: Refactor ins_compl_interrupted() to also check for timeout,
further refactor code to skip outputting message when
performing autocompletion (Girish Palya).
Running `vim *` in `vim/src` was slower than expected when
'autocomplete' was enabled. Include-file and tag-file completion
sources were not subject to the timeout check, causing unnecessary
delays.
So apply the timeout check to these sources as well, improving
autocompletion responsiveness, refactor find_pattern_in_path() to take
an additional "silent" argument, to suppress any messages.
closes: #17966
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: incorrect selected index returned from
complete_info()
Solution: Return the index into "items" and restore the previous
behaviour (Robert Muir).
complete_info() returned an incorrect selected index after
0ac1eb3555 (Patch v9.1.1311). Effectively
it became an index into "matches" instead of "items". Return the index
into "items" by default to restore the previous behavior, unless
"matches" was requested.
closes: #17952
Co-authored-by: Hirohito Higashi <h.east.727@gmail.com>
Signed-off-by: Robert Muir <rmuir@apache.org>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: hang (after 9.1.1471) or E684 (after 9.1.1410)
when 'tagfunc' calls complete().
Solution: Check if complete() has been called immediately after getting
matches instead of in the next loop iteration (zeertzjq).
related: #1668
related: neovim/neovim#34416
related: neovim/neovim#35163closes: #17929
Signed-off-by: zeertzjq <zeertzjq@outlook.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: complete: Heap-buffer overflow with complete function
(zeertzjq)
Solution: Do not let startcol become negative (Girish Palya).
fixes: #17907closes: #17934
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Co-authored-by: Hirohito Higashi <h.east.727@gmail.com>
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: cannot use autoloaded funcs in 'complete' F{func}
(Maxim Kim)
Solution: Make it work (Girish Palya)
fixes: #17869closes: #17885
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: cannot perform autocompletion
Solution: Add the 'autocomplete' option value
(Girish Palya)
This change introduces the 'autocomplete' ('ac') boolean option to
enable automatic popup menu completion during insert mode. When enabled,
Vim shows a completion menu as you type, similar to pressing |i\_CTRL-N|
manually. The items are collected from sources defined in the
'complete' option.
To ensure responsiveness, this feature uses a time-sliced strategy:
- Sources earlier in the 'complete' list are given more time.
- If a source exceeds its allocated timeout, it is interrupted.
- The next source is then started with a reduced timeout (exponentially
decayed).
- A small minimum ensures every source still gets a brief chance to
contribute.
The feature is fully compatible with other |i_CTRL-X| completion modes,
which can temporarily suspend automatic completion when triggered.
See :help 'autocomplete' and :help ins-autocompletion for more details.
To try it out, use :set ac
You should see a popup menu appear automatically with suggestions. This
works seamlessly across:
- Large files (multi-gigabyte size)
- Massive codebases (:argadd thousands of .c or .h files)
- Large dictionaries via the `k` option
- Slow or blocking LSP servers or user-defined 'completefunc'
Despite potential slowness in sources, the menu remains fast,
responsive, and useful.
Compatibility: This mode is fully compatible with existing completion
methods. You can still invoke any CTRL-X based completion (e.g.,
CTRL-X CTRL-F for filenames) at any time (CTRL-X temporarily
suspends 'autocomplete'). To specifically use i_CTRL-N, dismiss the
current popup by pressing CTRL-E first.
---
How it works
To keep completion snappy under all conditions, autocompletion uses a
decaying time-sliced algorithm:
- Starts with an initial timeout (80ms).
- If a source does not complete within the timeout, it's interrupted and
the timeout is halved for the next source.
- This continues recursively until a minimum timeout (5ms) is reached.
- All sources are given a chance, but slower ones are de-prioritized
quickly.
Most of the time, matches are computed well within the initial window.
---
Implementation details
- Completion logic is mostly triggered in `edit.c` and handled in
insexpand.c.
- Uses existing inc_compl_check_keys() mechanism, so no new polling
hooks are needed.
- The completion system already checks for user input periodically; it
now also checks for timer expiry.
---
Design notes
- The menu doesn't continuously update after it's shown to prevent
visual distraction (due to resizing) and ensure the internal list
stays synchronized with the displayed menu.
- The 'complete' option determines priority—sources listed earlier get
more time.
- The exponential time-decay mechanism prevents indefinite collection,
contributing to low CPU usage and a minimal memory footprint.
- Timeout values are intentionally not configurable—this system is
optimized to "just work" out of the box. If autocompletion feels slow,
it typically indicates a deeper performance bottleneck (e.g., a slow
custom function not using `complete_check()`) rather than a
configuration issue.
---
Performance
Based on testing, the total roundtrip time for completion is generally
under 200ms. For common usage, it often responds in under 50ms on an
average laptop, which falls within the "feels instantaneous" category
(sub-100ms) for perceived user experience.
| Upper Bound (ms) | Perceived UX
|----------------- |-------------
| <100 ms | Excellent; instantaneous
| <200 ms | Good; snappy
| >300 ms | Noticeable lag
| >500 ms | Sluggish/Broken
---
Why this belongs in core:
- Minimal and focused implementation, tightly integrated with existing
Insert-mode completion logic.
- Zero reliance on autocommands and external scripting.
- Makes full use of Vim’s highly composable 'complete' infrastructure
while avoiding the complexity of plugin-based solutions.
- Gives users C native autocompletion with excellent responsiveness and
no configuration overhead.
- Adds a key UX functionality in a simple, performant, and Vim-like way.
closes: #17812
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: repeated insertion and deletion of complete
functions
Solution: Remove unnecessary insertion and deletion of leader text
('compl_orig_text') during expansion of function present in
'complete' option (Girish Palya).
closes: #17738
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: menu state wrong on interruption
(Maxim Kim)
Solution: Call show_pum() if completion was interrupted
(Girish Palya).
Popup menu was not built after fuzzy sorting, so internal state wasn't
reflected on screen.
No test. Couldn't get the terminal test to trigger both interruption and
refresh together.
fixes: #17725closes: #17736
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: messages don't respect 'shm' setting
Solution: Turn off completion messages when 'shortmess' includes "c"
(Girish Palya).
`:set shortmess+=c` is intended to reduce noise during completion by
suppressing messages.
Previously, some completion messages still appeared regardless of this setting.
This change ensures that **all** completion-related messages are suppressed
when `'c'` is present in `'shortmess'`.
Not entirely sure if the original behavior was intentional. If there's a
reason certain messages were always shown, feel free to close this without
merging.
closes: #17737
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: can only complete from keyword characters
Solution: remove this restriction, allow completion functions when
called from i_CTRL-N/i_CTRL-P to be triggered from non-keyword
characters (Girish Palya)
Previously, functions specified in the `'complete'` option were
restricted to starting completion only from keyword characters (as
introduced in PR 17065). This change removes that restriction.
With this change, user-defined functions (e.g., `omnifunc`, `userfunc`)
used in `'complete'` can now initiate completion even when triggered
from non-keyword characters. This makes it easier to reuse existing
functions alongside other sources without having to consider whether the
cursor is on a keyword or non-keyword character, or worry about where
the replacement should begin (i.e., the `findstart=1` return value).
The logic for both the “collection” and “filtering” phases now fully
respects each source’s specified start column. This also extends to
fuzzy matching, making completions more predictable.
Internally, this builds on previously merged infrastructure that tracks
per-source metadata. This PR focuses on applying that metadata to
compute the leader string and insertion text appropriately for each
match.
Also, a memory corruption has been fixed in prepare_cpt_compl_funcs().
closes: #17651
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: completion: Functions specified in the 'complete' option did
not have the leader string removed when called with findstart = 0,
unlike 'omnifunc' behavior
Solution: update behaviour and make behaviour consistent (Girish Palya)
closes: #17636
Signed-off-by: Girish Palya <girishji@gmail.com>
Signed-off-by: Christian Brabandt <cb@256bit.org>
