Problem: Nvim has many Lua APIs that start callback-driven work: timers,
jobs, libuv handles, and other event-loop tasks. Callers that need to
sequence or cancel that work have to build their own coroutine wrappers,
task bookkeeping, and cleanup rules. This makes async control flow hard
to share, test, and document.
Solution: Add `vim.async`, a structured-concurrency module vendored from
async.nvim. It provides task handles, await/pawait helpers,
sleep/timeout helpers, completion-order iteration, and semaphores on top
of Nvim's event loop.
The API follows the same broad model as Trio: async work has an owner,
tasks are awaited explicitly, and cancellation is cooperative. Include
generated vimdoc with an introductory overview and examples, a news
entry, and functional tests for the new module.
AI-assisted
Problem:
'guicursor' is not respected when extra cursors exist.
Per-cursor selection toggles Visual off, and the sandbox restores it
only after restore_current_state(), whose ui_cursor_shape() caches the
Normal shape.
Solution:
Restore Visual before restore_current_state().
Problem: there should be exactly one cell of padding between sections,
and exactly one cell of minimum padding between the left and the right.
- Spaces between sections waste space when a section is empty.
According to a comment in #33036, this was the reason to avoid `%k`
and implement the keymap section with a vim expression, but other
sections still have this problem.
- The diagnostics section wastes space because it is not entirely empty
when there are diagnostics in another buffer.
- The terminal exit code section can touch the right side, e.g. the
ruler, even though it belongs to the left side.
Solution:
- Use auto-hiding item groups (`%(` without width fields) to get rid of
unneeded spaces when a section shows no information.
This simplifies the 'showcmd' and 'keymap' sections in particular.
- As a slight simplification, `term_exitcode` is moved into the flags
section since it is formatted with square brackets like a flag.
- Count the diagnostics for the current buffer specifically.
- Ensure at least one cell of padding between the left and the right
side by adding a space next to the separator `%=`.
Problem: some sections are implemented with `%{%`, even though
reevaluation of the expression result is not needed.
This leads to otherwise needless %-escaping in `progress_status`.
Solution: use `%{` instead.
Problem: `%{` and `%{%` (without items) replace spaces with fillchars.
This looks out-of-place inside the terminal exit code section, and in
contrast to all other sections, the 'busy' section is surrounded by
fillchars, which looks inconsistent, and with some terminal-font
combinations, ◐ overlaps the fillchar, e.g. Alacritty & JetBrains Mono.
Solution: use non-breaking spaces U+202F to avoid fillchar substitution.
Problem: sections that appear/disappear frequently can make otherwise
more stable sections jump around a lot.
Solution: sort the sections on the right roughly by volatility:
'showcmd' in first place, 'keymap' next to the ruler.
Problem:
`shada_read()` rebuilds `v:oldfiles` for any "forced" read, which first
*clears* the list. For an in-memory caller such as multicursor, the list
is never repopulated. So `v:oldfiles` is empty after `Q` + edit.
Solution:
Rebuild `v:oldfiles` only when `kShaDaGetOldfiles` is requested.
`shada_read_everything()` (`:rshada[!]`, startup) always requests it.
old bug from 411a06c8b6
Problem:
During 'autocomplete', a completion-ending `<BS>` restarts completion
immediately. Multicursor insert-cascade skips replay (`edit()` refuses
to nest if compl/pum is active), so the mcursors do not update until
ESC.
Solution:
Flush pending keys in `TRIGGER_AUTOCOMPLETE`, just before autocomplete
restarts again.
Problem:
Insert entered by a Visual-mode operator mapping (`xnoremap c c`) does
not live-mirror at cursors.
Solution:
Decide `typed` from the Visual session's own provenance (`vatom`, which
spans the selection and knows whether it was user input) instead of only
negating it.
Problem:
`parse_ssh_config()` compares tables against a freshly allocated empty
table.
- In `parse_multiple_values()`, the guard which avoids flushing an empty
accumulator never applies. Runs of separators and trailing whitespace
push empty strings into the results, and `is_valid()` does not filter
them. `Host alpha beta ` parses as `{ 'alpha', '', 'beta' }`.
- In `parse_value()`, the condition reduces to `chr == '"' and quoted`.
`quoted` starts false and only that branch sets it, so it can never
become true: quotes are never recognised and are inserted literally,
and the unterminated-quote check is unreachable.
Solution:
Compare `#val` instead. Add a test for repeated and trailing separators.
AI-assisted
Problem: Drawing 'statuscolumn' leads to a heap-buffer-overflow if a
sign/number column comes after many items.
Solution: Avoid curitem >= stl_items_len when writing to stl_items.
Problem:
`IterArray:take()` iterates up to `self._tail`, but `_tail` is
exclusive. When no element fails the predicate, the loop reads one index
past the last element and calls the predicate with `nil`.
Solution:
Stop at `self._tail - inc`, which is the last in-range index for both
iteration directions. Add tests using a predicate which dereferences its
argument and matches every element, forward and reversed.
AI-assisted
Problem:
With follow-mode ("q="), an 'operatorfunc' that only moves the cursor
does not cascade.
Solution:
Update the `effect` condition to include cursor-moves if follow-mode is
active.
Problem:
With mcursors and 'autocomplete', a non-literal key (`<BS>`) during an
active completion attempts to insert-cascade, which attempts a nested
`edit()`, which refuses with E565.
Solution:
Defer the flush while completion is active; the pending keys will be
handled later.
Problem:
An atom queued in buffer A cascades on B's cursors if the mapping ends
in B ("nnoremap X x:bnext<CR>").
Solution:
Check the atom's origin buffer (`CmdAtom.origin.buf`).
Note: This does not preclude mappings etc from doing work in temporary
throwaway buffers, as long as they return to the origin buffer.
Problem:
`vim.version.intersect()` returns tables with `VersionRange` (method
table) as their metatable. Although the calculated bounds are correct,
the results do not expose `VersionRange` methods.
Solution:
Construct intersection tables with `range_mt`, matching
`vim.version.range()`. Add assertions covering method access and
repeated intersection.
Problem:
When disabling, the loader list is traversed forwards while entries are
removed. As a result, some entries are skipped and remain active. Thus,
disabling `vim.loader` restores `_G.loadfile`, but does not fully
restore the original package loader chain.
Solution:
Traverse `package.loaders` in reverse when removing cached loaders.
Capture the original `package.loaders` list in the test, and assert that
the list is restored properly.
Problem:
If an edit cascades to cursor(s) that happen to be within a closed fold,
the edit acts on the fold itself (`:h fold-behavior`). E.g. "dd" deletes
the entire fold, not the cursor's line within the fold.
Solution:
Check `mc_replaying()` in `hasAnyFolding()`, so folds are ignored during
a multicursor cascade. This means multicursor always replays relative to
text *within* the fold.
Problem:
Coordinates-to-position conversion for the finish column in
`vim.hl.range()` checks the start column. This results in an incorrect
finish column when only one of the column coordinates is `vim.v.maxcol`.
Solution:
Check the finish column when converting the finish position. Update the
existing screen test so that a `vim.v.maxcol` finish highlights the
end-of-line marker, matching the existing `-1` behavior.
Problem: With 'wildmode' set to list:full the matches are listed but the
wildmenu is not shown, although it is "full" that starts
wildmenu mode (zeertzjq).
Solution: List the matches and show the menu, as the two behaviors in
the same phase ask for. The menu is left to the phases that
ask for it, so that "list" on its own still only lists
(Hirohito Higashi).
fixes: vim/vim#21196closes: vim/vim#21205fa1ddffcce
Co-authored-by: Hirohito Higashi <h.east.727@gmail.com>
Other (squashed) commits:
fix(tui): emit ui_send output atomically with the frame
Problem:
tui_ui_send() writes directly to the TTY, bypassing the output buffer.
Sequences sent via nvim_ui_send() (e.g. kitty multiple-cursors, or
visual-dot-repeat) always arrive in a separate TTY write from the frame
they were computed for. This manifests as "tearing", or e.g. in the case
of multicursor the terminal renders text with stale cursor overlays.
Solution:
- tui_ui_send(): while a frame is being assembled (pending invalid
regions or buffered output), buffer instead of writing directly.
- Out-of-frame sends (tty queries, clear-on-disable) still write
immediately.
- mcursor.lua: emit the terminal-cursor update at the end of the redraw
cycle (`on_end`, when screen positions are final) instead of
vim.schedule().
Problem:
Literal path comparison ignored one trailing slash for every buffer name,
including URIs. Generic URI syntax does not make a non-empty path
equivalent to the same path with a trailing slash, so distinct URI
buffers collapsed into one.
Solution:
Require equal lengths when comparing URI buffer names, while retaining
trailing-separator normalization for filesystem paths.
AI-assisted
Problem:
`shorten_fnames()` always redraws the statusline/tabline, even if no
buffer name changed. Since b296666e a temp context-switch
(`win_execute()`, `vim._with{win=}`) restores the CWD, so every such
switch flickers the message area.
Solution:
Redraw only if `shorten_buf_fname()` actually changed a name.
Problem:
Folding range markers are overwritten while ranges are evaluated. A range ending on a row can hide another range starting there, and multiple nested ranges ending together emit only the innermost ending level.
Solution:
Track starts and the number of ends per row before emitting markers. Prefer starts on shared boundary rows and use the outermost level when nested ranges end together.
AI-assisted
Problem:
Confirming cmdwin with a UTF-8 character containing 0x80 does not complete the
command.
Solution:
Escape K_SPECIAL bytes while feeding the cmdwin input after confirmation.
Problem:
The "." example mapping at `:h edit-repeat` doesn't work well with
`nvim_feedkeys(…, 'mt', false)`.
Solution:
Use `vim.b[ev.buf].maxseq` instead of `undotree()`.
Problem:
Replaying a deleted Lua mapping, may call an arbitrary function.
RHS of a Lua mapping embeds its LuaRef (`<K_LUA><ref><CR>`). The raw
keys may outlive the ref (redobuff ".", CmdAtom `keys`). If the mapping
is deleted, replaying it either (1) dereferences a freed registry slot,
or (2) calls whatever callback reused the slot (autocmd, timer, other
mapping).
Solution:
Assign a monotonic (never recycled) id to Lua mappings and encode the
mapping keys as `<K_LUA><id><CR>`.
Note: in the case of Vimscript, a deleted function raises E117, but if
the function is redefined with the same name, the mapping will find it.
Alternatives?:
- Globally ensure `LuaRef` ids are not recycled.
- Problem: could exhaust `int` in a long-lived Nvim session? Also,
difficult to impl bc the "recycling" is done by `luaL_ref` itself.
ref: 5ac2e47acc
Problem:
An operator completed by a Lua `:omap` textobject emits
`CmdAtom.type="mapping"` (lhs-only, no keys) instead of `type="operator"`.
Analysis:
`atom_redo_set()` declined K_LUA, though the prepped redo
("op" + K_LUA + id + CR) is exactly what "." replays. A no-edit "g@"
emits nothing at all.
Solution:
- `atom_redo_set`: don't decline K_LUA; the redo route now captures the
operator atom.
- `atom_capture_cmd`: don't early-return if the frame has prepped redo.
- op_function(): save/restore redobuff when invoking 'operatorfunc',
like `call_user_func()` does for Vimscript. (Else the Lua callback
may clobber the prepped "g@" redo / dot-repeat.)
fix#41482
TODO:
- async Lua (timer/vim.schedule) can still clobber the pending dot-repeat...
Problem:
If a snippet does not have a placeholder, we use insert mode instead of
select mode. From here <Esc> leaves the session and highlight active.
Solution:
Cancel the session on <Esc>.
Problem:
0c091cedc2 fixed the "immediate exit" but the 2s `nvim_get_proc()`
check still fails on slow (ASAN) CI.
Solution:
Check the pid after the `screen:expect`. Anyway, `assert_nolog` is the
"meaningful" part of the test since 0c091cedc2.
Problem:
When 'cmdheight' is 0, the MsgSeparator can obscure the statusline when
a message is displayed, making its contents disappear until
the next redraw.
Solution:
Skip the separator when 'cmdheight=0' available.
Problem:
A blank line before the Content-Length header makes header parsing
fail with "Content-Length not found in header". An LF in the 'name'
state falls through to the 'invalid' state, which only return to
'name' at the *next* LF. That LF terminates the next line, so the
line with Content-Length is swallowed.
The same issue happens when a junk line is a partial match of the
header name (e.g. "Cont\nContent-Length: ...").
Solution:
When seeing an LF in the 'name' state, reset the cursor and stay in
'name' rather than entering 'invalid'.
Problem:
The Vim "stuff" concept breaks the ability to reason about the call
stack and thus the boundaries of a `CmdAtom`: a stuffed translation ("x"
=> "dl") defers to the main loop. This "continuation" must be modeled in
`CmdAtom`, by checking global flags at undefined times, during undefined
circumstances.
Solution:
- After a stuffed "translation", eagerly execute the stuff buffer
(`exec_stuffed()`).
- Delete the CmdAtom "continuation" junk.
Note:
- op_colon runs its cmdline "nested", but that's fine because operators
already nest interactive sessions there (op_change runs edit()), and
the cmdline is frameless so the operator's frame can own/capture it.
Problem:
If an autocommand handler focuses the tab page we're closing during a
`:tabonly` (with some conditions), we hit an assert failure in
`win_close_othertab()`.
For this to occur, we need:
- `nvim_buf_delete()` to trigger `close_windows()` (as is done in
`cmdwin.lua`'s `_cleanup()`)
- `close_windows()` then calls `win_close_othertab()`, removing the
window (`win_free_mem()`)
- then in the caller, `tabpage_close_other()`, the loop continues (we
don't detect `tp_lastwin == wp` since we've unlinked the window)
- the loop assumes that `curtab != tp`
but we've refocused `curtab` so the `ex_win_close()` call passes `tp` as
`curtab`, causing the assert to fail
Solution:
Detect the focus of `curtab` and abort closing the tab.
Problem:
Undo places the cursor wherever the cursor happened to sit at "save
time" (`uh_cursor` is sampled lazily on the first change).
Examples:
- `i` preserves, but `a` does not
- `diw`, `atest<Esc>`, `d^` abandon the original position
- `D`, `o` restore it (by accident).
Solution:
`composite` tracks the pending atom (and its `origin`) across frames.
A `stuffed` continuation frame inherits the `origin` + prepped redo.
Store `origin` info in the undo header, so undo can restore it.
- Not for a mid-command undo break (i_CTRL-G_u).
- Undoing a mapping restores where the mapping started (which
technically may be different than where the "edit" started).
Problem:
An insert-session entered by a scheduled `feedkeys('i','n')` is
classified on entry (as "not typed"), and not reevaluated after that,
thus user input following it is not captured.
Repro: an `:imap` that does `feedkeys('<esc>','n')` and schedules
re-entering insert, emits one CmdAtom for the first press, then nothing
else.
Solution:
Sample `maptick` (ticked by `gotchars()` on typed input, including
mappings) at session start; if it advanced by session end, the session
is user input.
Problem:
`nlua_pcall()` references `_G.debug.traceback`. If user code deletes it
or breaks it some other way, various Lua features are broken.
_G.debug = nil
vim.schedule(function() end)
vim.wait(100)
E5113: Lua chunk: attempt to index a nil value
stack traceback:
[C]: in function 'loop_poll'
[string "vim/_core/editor"]:176: in function 'wait'
crash.lua:3: in main chunk
PANIC: unprotected error in call to Lua API (attempt to index a nil value)
Solution:
Check `_G.debug.traceback` before using it as errfunc. If it's broken,
omit the traceback and say so in the error message.
Note: We could cache `_G.debug` in LUA_REGISTRYINDEX on startup, but
that would prevent plugins from providing custom functionality there
(and we happen to do so in `tui_spec.lua` for example).
libvterm reads the first three sub-parameters of an SGR 38:2 or 48:2 sequence
as R:G:B, so a colour space id shifts the channels: the empty slot becomes red
via CSI_ARG_MISSING truncating to 255, and green and blue move over one.
Skip the colour space id when the colon-separated group holds more than three
arguments. The group length comes from CSI_ARG_HAS_MORE rather than the raw
argument count, so a following semicolon-separated parameter is not consumed.
Problem:
With these two lines (`\b` marking a backspace):
f\bfoo
xb\bb
line 1's bold run ends at byte 1 and line 2's begins at byte 1, so line 1
renders "fo" in bold rather than "f", and line 2's "b" is not bold at all.
Solution:
Only grow a highlight group that is on the current row.
Problem: ins_compl_dict_alloc() builds the same dict as
fill_complete_info_dict().
Solution: Call fill_complete_info_dict() instead (glepnir).
closes: vim/vim#21140303a153694
Problem:
Test sockets live under `$TMPDIR`, which the harness points at the build
dir. On macOS/BSD `sockaddr_un.sun_path` is 104 bytes, and a CI build
path plus "nvim.<pid>.<n>" leaves little room:
/Users/runner/work/neovim/neovim/build/Xtest_tmpdir_terminal/nvim.runner/aBcDeF/nvim.12345.0
Solution:
Point XDG_RUNTIME_DIR (`stdpath('run')`) at "/tmp/nvim_<pid>". 28 bytes:
/tmp/nvim_19916/nvim.19919.1
TODO?:
- `TEMP_DIR_NAMES` prefers `$TMPDIR` over `/tmp`, so on macOS
`stdpath('run')` defaults to the long `/var/folders/<xx>/<…>/T/` path
instead of the short `/tmp` alias.
Note:
- The 104-byte limit applies to the `bind()` arg, not its "realpath",
to, so `/tmp/…` symlinks can be used to workaround the limit.
Problem:
Cannot retry a test with its full `after_each`/`before_each` lifecycle.
Solution:
- Overload `it()` to accept an `opts` param:
```
it('flaky', { retries = 2 }, function(ctx) end) -- 3 attempts.
```
- Pass `ctx` to test functions.
- Fix a bug in `t.read_file_list()`.
