When we convert a Lua table to an Object, we consider the table a
"dictionary" if it contains only string keys, and an array if it
contains all numeric indices with no gaps. While rare, Lua tables can
have both strictly numeric indices and gaps (e.g. { [2] = 2 }). These
currently cannot be serialized because it is not considered an array.
However, we know the maximum index of the table and as long as all of
the keys in the table are numeric, it is still possible to serialize
this table as an array. The missing indices will have nil values.
Problem:
Empty string is a valid JSON key, but json_decode() treats an object
with empty key as ":help msgpack-special-dict". #20757
:echo json_decode('{"": "1"}')
{'_TYPE': [], '_VAL': [['', '1']]}
Note: vim returns `{'': '1'}`.
Solution:
Allow empty string as an object key.
Note that we still (currently) disallow empty keys in object_to_vim() (since 7c01d5ff92):
f64e4b43e1/src/nvim/api/private/converter.c (L333-L334)Fix#20757
Co-authored-by: Justin M. Keyes <justinkz@gmail.com>
Problem:
With vim.treesitter.foldexpr, `o`-ing two lines above a folded region
opens the fold. This does not happen with legacy foldexprs. For example,
make a markdown file with the following text (without indentation),
enable treesitter fold, and follow the instruction in the text.
put cursor on this line and type zoo<Esc>
initially folded, revealed by zo
# then this fold will be opened
initially folded, revealed by o<Esc>
Analysis:
* `o` updates folds first (done in `changed_lines`), evaluating
foldexpr, and then invokes `on_bytes` (done in `extmark_splice`).
* Treesitter fold allocates the foldinfo for added lines (`add_range`)
on `on_bytes`.
* Therefore, when treesitter foldexpr is invoked while running `o`, it
sees outdated foldinfo.
Solution:
`extmark_splice`, and then `changed_lines`. This seems to be the
standard order in other places, e.g., `nvim_buf_set_lines`.
Quick fix as follow up to https://github.com/neovim/neovim/pull/26108
kqueue only reports events on a watched folder itself, not for files
created or deleted within. So the approach the PR took doesn't work on FreeBSD.
We'll either need to bring back polling for it, combine watching with manual
file tracking, or disable LSP file watching on FreeBSD
Should help with https://github.com/neovim/neovim/issues/23291
On linux `new_fs_event` doesn't support recursive watching, but we can
still use it to watch folders.
The downside of this approach is that we may end up sending some false
`Deleted` events. For example, if you save a file named `foo` there will
be a intermediate `foo~` due to the save mechanism of neovim.
The events we get from vim.uv in that case are:
- rename: foo~
- rename: foo~
- rename: foo
- rename: foo
- change: foo
- change: foo
The mechanism in this PR uses a debounce to reduce this to:
- deleted: foo~
- changed: foo
`foo~` will be the false positive.
I suspect that for the LSP case this is good enough. If not, we may need
to follow up on this and keep a table in memory that tracks available
files.
The prefix option of the diagnostic virtual text can be a function,
but previously it was only a function of diagnostic.
This function should also have additional parameters index and total,
more consistently and similarily as in the prefix function for
`vim.diagnostic.open_float()`.
These additional parameters will be useful when there are too many
number of diagnostics in a single line.
Problem: Peeking and flushing output slows down execution.
Solution: Do not update the mode message when global_busy is set. Do not
flush when only peeking for a character. (Ken Takata)
cb574f4154
When tabstop and shiftwidth are not equal, tabs are inserted as individual
spaces and then rewritten as tab characters in a second pass. That second pass
did not call changed_bytes which resulted in events being omitted.
Fixes#25092
Problem: sidescrolloff and scrolloff options work slightly
different than other global-local options
Solution: Make it behave consistent for all global-local options
It was noticed, that sidescrolloff and scrolloff options behave
differently in comparison to other global-local window options like
'listchars'
So make those two behave like other global-local options. Also add some
extra documentation for a few special local-window options.
Add a few tests to make sure all global-local window options behave
similar
closes: vim/vim#12956closes: vim/vim#126434a8eb6e7a9
Co-authored-by: Christian Brabandt <cb@256bit.org>
If an iterator pipeline stage returns nil as its first return value, the
other return values are ignored and it is treated as if that stage
returned only nil (the semantics of returning nil are different between
different stages). This is consistent with how for loops work in Lua
more generally, where the for loop breaks when the first return value
from the function iterator is nil (see :h for-in for details).
Currently (as of nvim 0.9), the behavior of boolean params in
vim.api lua wrappers is inconsistent for optional parameters
(part of an `opts` dict) compared to positional parameters.
This was inadvertently changed in #24524 . While cleaning up this
inconsistency is something we might want eventually, it needs
to be discussed separately and the impact of existing code considered.
Problem: cache paths are derived by replacing each reserved/filesystem-
path-sensitive char with a `%` char in the original path. With this
method, two different files at two different paths (each containing `%`
chars) can erroneously resolve to the very same cache path in certain
edge-cases.
Solution: derive cache paths by url-encoding the original (path) instead
using `vim.uri_encode()` with `"rfc2396"`. Increment `Loader.VERSION` to
denote this change.
Problem:
On Windows, `rundll32` exits zero (success) even when given
a non-existent file.
Solution:
Mock vim.system() on Windows to force a "failure" case.
Problem:
helpers.tmpname() may create a local file, depending on circumstances.
Solution:
Only use helpers.tmpname() for its parent directory (the "temp root").
Use fs_mkdtemp() to actually get a unique name.
* perf(rtp): reduce rtp scans
Problem:
Scanning the filesystem is expensive and particularly affects
startuptime.
Solution:
Reduce the amount of redundant directory scans by relying less on glob
patterns and handle vim and lua sourcing lower down.
Problem: Bashslashes added as regexp in runtime completion may be
treated as path separator with some 'isfname' value.
Solution: Make curly braces work for runtime completion and use it.
* feat(lua): allow vim.wo to be double indexed
Problem: `vim.wo` does not implement `setlocal`
Solution: Allow `vim.wo` to be double indexed
Co-authored-by: Christian Clason <c.clason@uni-graz.at>
Problem:
Showing an error via vim.notify() makes it awkward for callers such as
lsp/handlers.lua to avoid showing redundant errors.
Solution:
Return the message instead of showing it. Let the caller decide whether
and when to show the message.
---
Rejected experiment: move vim.ui.open() to vim.env.open()
Problem:
`vim.ui` is where user-interface "providers" live, which can be
overridden. It would also be useful to have a "providers" namespace for
platform-specific features such as "open", clipboard, python, and the other
providers listed in `:help providers`. We could overload `vim.ui` to
serve that purpose as the single "providers" namespace, but
`vim.ui.nodejs()` for example seems awkward.
Solution:
`vim.env` currently has too narrow of a purpose. Overload it to also be
a namespace for `vim.env.open`.
diff --git a/runtime/lua/vim/_meta.lua b/runtime/lua/vim/_meta.lua
index 913f1fe20348..17d05ff37595 100644
--- a/runtime/lua/vim/_meta.lua
+++ b/runtime/lua/vim/_meta.lua
@@ -37,8 +37,28 @@ local options_info = setmetatable({}, {
end,
})
-vim.env = setmetatable({}, {
- __index = function(_, k)
+vim.env = setmetatable({
+ open = setmetatable({}, {
+ __call = function(_, uri)
+ print('xxxxx'..uri)
+ return true
+ end,
+ __tostring = function()
+ local v = vim.fn.getenv('open')
+ if v == vim.NIL then
+ return nil
+ end
+ return v
+ end,
+ })
+ },
+ {
+ __index = function(t, k, ...)
+ if k == 'open' then
+ error()
+ -- vim.print({...})
+ -- return rawget(t, k)
+ end
local v = vim.fn.getenv(k)
if v == vim.NIL then
return nil
Enforce consistent terminology (defined in
`gen_help_html.lua:spell_dict`) for common misspellings.
This does not spellcheck English in general (perhaps a future TODO,
though it may be noisy).
Problem:
- `vim.json` exposes various global options which:
- affect all Nvim Lua plugins (especially the LSP client)
- are undocumented and untested
- can cause confusing problems such as: cc76ae3abe
- `vim.json` exposes redundant mechanisms:
- `vim.json.null` is redundant with `vim.NIL`.
- `array_mt` is redundant because Nvim uses a metatable
(`vim.empty_dict()`) for empty dict instead, which `vim.json` is
configured to use by default (see `as_empty_dict`).
Example:
```
:lua vim.print(vim.json.decode('{"bar":[],"foo":{}}'))
--> { bar = {}, foo = vim.empty_dict() }
```
Thus we don't need to also decorate empty arrays with `array_mt`.
Solution:
Remove the functions from the public vim.json interface.
Comment-out the implementation code to minimize drift from upstream.
TODO:
- Expose the options as arguments to `vim.json.new()`
Problem: Current implementation of "remove trailing /" doesn't
account for the case of literal '/' as path.
Solution: Remove trailing / only if it preceded by something else.
Co-authored by: notomo <notomo.motono@gmail.com>
feat(lua): add vim.system()
Problem:
Handling system commands in Lua is tedious and error-prone:
- vim.fn.jobstart() is vimscript and comes with all limitations attached to typval.
- vim.loop.spawn is too low level
Solution:
Add vim.system().
Partly inspired by Python's subprocess module
Does not expose any libuv objects.
vim.version.range() couldn't parse them correctly.
For example, vim.version.range('<0.9.0'):has('0.9.0') returned `true`.
fix: range:has() accepts vim.version()
So that it's possible to compare a range with:
vim.version.range(spec):has(vim.version())
This ensures that colorschemes in 'rtp' are tried before ones in 'pp',
because some colorschemes in 'pp' may not work if not added to 'rtp'.
This also match the current documentation better.
`nvim_(get|set)_option_value` pick the current buffer / window by default for buffer-local/window-local (but not global-local) options. So specifying `buf = 0` or `win = 0` in opts is unnecessary for those options. This PR removes those to reduce code clutter.
Uncrustify is sensitive to version changes, which causes friction for
contributors that doesn't have that exact version. It's also simpler to
download and install the correct version than to have bespoke version
checking.
Revert the change to do_cmdline_cmd() from #5226.
This function is used in many places, so making it different from Vim
leads to small differences from Vim in the behavior of some functions
like execute() and assert_fails(). If DOCMD_VERBOSE really needs to be
removed somewhere, a do_cmdline() call without DOCMD_VERBOSE is also
shorter than a do_cmdline() call with DOCMD_VERBOSE.
This is a more robust method for tagging a packed table as it completely
eliminates the possibility of mistaking an actual table key as the
packed table tag.
Problem:
`vim.split('a:::', ':', {trimempty=true})` trims inner empty items.
Regression from 9c49c10470
Solution:
Set `empty_start=false` when first non-empty item is found.
close#23212
This was originally meant as a convenience but prevents possible
functionality. For example:
-- Get the keys of the table with even values
local t = { a = 1, b = 2, c = 3, d = 4 }
vim.iter(t):map(function(k, v)
if v % 2 == 0 then return k end
end):totable()
The example above would not work, because the map() function returns
only a single value, and cannot be converted back into a table (there
are many such examples like this).
Instead, to convert an iterator into a map-like table, users can use
fold():
vim.iter(t):fold({}, function(t, k, v)
t[k] = v
return t
end)
If pack() is called with a single value, it does not create a table; it
simply returns the value it is passed. When unpack is called with a
table argument, it interprets that table as a list of values that were
packed together into a table.
This causes a problem when the single value being packed is _itself_ a
table. pack() will not place it into another table, but unpack() sees
the table argument and tries to unpack it.
To fix this, we add a simple "tag" to packed table values so that
unpack() only attempts to unpack tables that have this tag. Other tables
are left alone. The tag is simply the length of the table.
vim.iter wraps a table or iterator function into an `Iter` object with
methods such as `filter`, `map`, and `fold` which can be chained to
produce iterator pipelines that do not create new tables at each step.
- vim.diagnostic.config() now accepts a function for the virtual_text.prefix
option, which allows for rendering e.g., diagnostic severities differently.
Problem: Vim9: exception in ISN_INSTR caught at wrong level.
Solution: Set the starting trylevel in exec_instructions(). (closesvim/vim#8214)
ff65288aa8
Co-authored-by: Bram Moolenaar <Bram@vim.org>
Problem: Confusing error for using a variable as a function.
Solution: If a function is not found but there is a variable, give a more
useful error. (issue vim/vim#9310)
2ef9156b42
Co-authored-by: Bram Moolenaar <Bram@vim.org>
feat(lua)!: add stricter vim.tbl_islist(), rename vim.tbl_isarray()
Problem: `vim.tbl_islist` allows gaps in tables with integer keys
("arrays").
Solution: Rename `vim.tbl_islist` to `vim.tbl_isarray`, add new
`vim.tbl.islist` that checks for consecutive integer keys that start
from 1.
* feat(lua): vim.tbl_contains supports general tables and predicates
Problem: `vim.tbl_contains` only works for list-like tables (integer
keys without gaps) and primitive values (in particular, not for nested
tables).
Solution: Rename `vim.tbl_contains` to `vim.list_contains` and add new
`vim.tbl_contains` that works for general tables and optionally allows
`value` to be a predicate function that is checked for every key.
The first argument which is non-nil is returned. This is useful when
using nested default values (e.g. in the EditorConfig plugin).
Before:
local enable = vim.F.if_nil(vim.b.editorconfig, vim.F.if_nil(vim.g.editorconfig, true))
After:
local enable = vim.F.if_nil(vim.b.editorconfig, vim.g.editorconfig, true)
Problem:
Codebase inconsistently binds vim.api onto a or api.
Solution:
Use api everywhere. a as an identifier is too short to have at the
module level.
Problem:
semver specifies that digit sequences in a prerelease string should be
compared as numbers, not lexically: https://semver.org/#spec-item-11
> Precedence for two pre-release versions with the same major, minor,
> and patch version MUST be determined by comparing each dot separated
> identifier from left to right until a difference is found as follows:
> 1. Identifiers consisting of only digits are compared numerically.
> 2. Identifiers with letters or hyphens are compared lexically in ASCII sort order.
> 3. Numeric identifiers always have lower precedence than non-numeric identifiers.
> 4. A larger set of pre-release fields has a higher precedence than a smaller set, if all of the preceding identifiers are equal.
Example:
1.0.0-alpha < 1.0.0-alpha.1 < 1.0.0-alpha.beta < 1.0.0-beta < 1.0.0-beta.2 < 1.0.0-beta.11 < 1.0.0-rc.1 < 1.0.0.
Solution:
cmp_prerel() treats all digit sequences in a prerelease string as
numbers. This doesn't _exactly_ match the spec, which specifies that
only dot-delimited digit sequences should be treated as numbers...
Problem:
- vim.split has more features than vim.gsplit.
- Cannot inspect the "separator" segments of vim.split or vim.gsplit.
Solution:
- Move common implementation from vim.split into vim.gsplit.
- TODO: deprecate vim.split in favor of vim.totable(vim.gsplit())?
- Introduce `keepsep` parameter.
Related: 84f66909e4
existing behavior of
:=
and
:[range]=
are unchanged. `|` is still allowed with this usage.
However,
:=p
and similar are changed in a way which could be construed as a breaking
change. Allowing |ex-flags| for := in the first place was a mistake as
any form of := DOES NOT MOVE THE CURSOR. So it would print one line number
and then print a completely different line contents after that.
Problem:
"tmux 3.2a" (output from "tmux -V") is not parsed easily.
Solution:
With `strict=false`, discard everything before the first digit.
- rename Semver => Version
- rename vim.version.version() => vim.version._version()
- rename matches() => has()
- remove `opts` from cmp()
Problem:
vim.deprecate() shows ":help deprecated" for third-party plugins. ":help
deprecated" only describes deprecations in Nvim, and is unrelated to any
3rd party deprecations.
Solution:
If `plugin` is specified, don't show ":help deprecated".
fix#22235
Problem:
The function name `vim.pretty_print`:
1. is verbose, which partially defeats its purpose as sugar
2. does not draw from existing precedent or any sort of convention
(except external projects like penlight or python?), which reduces
discoverability, and degrades signaling about best practices.
Solution:
- Rename to `vim.print`.
- Change the behavior so that
1. strings are printed without quotes
2. each arg is printed on its own line
3. tables are indented with 2 instead of 4 spaces
- Example:
:lua ='a', 'b', 42, {a=3}
a
b
42
{
a = 3
}
Comparison of alternatives:
- `vim.print`:
- pro: consistent with Lua's `print()`
- pro: aligns with potential `nvim_print` API function which will
replace nvim_echo, nvim_notify, etc.
- con: behaves differently than Lua's `print()`, slightly misleading?
- `vim.echo`:
- pro: `:echo` has similar "pretty print" behavior.
- con: inconsistent with Lua idioms.
- `vim.p`:
- pro: very short, fits with `vim.o`, etc.
- con: not as discoverable as "echo"
- con: less opportunity for `local p = vim.p` because of potential shadowing.
When a buffer update callback is called, textlock is active so buffer
text cannot be changed, but cursor can still be moved. This can cause
problems when the buffer update is in the middle of an operator, like
the one mentioned in #16729. The solution is to save cursor position and
restore it afterwards, like how cursor is saved and restored when
evaluating an <expr> mapping.
also make implicit submodules "uri" and "_inspector" work with completion
this is needed for `:lua=vim.uri_<tab>` wildmenu completion
to work even before uri or _inspector functions are used.
Problem:
If major<major but minor>minor, cmp_version_core returns 1
Solution:
- Fix logic in cmp_version_core
- Delete most eq()/gt()/lt() tests, they are redundant.
- version.cmp(): assert valid version
- add test for loading vim.version (the other tests use shared.lua in
the test runner)
- reduce test scopes, reword test descriptions
Searching the entire repo for a directory named "contrib" causes failure
if there happens to be another subdirectory with the name "contrib".
Instead, point directly to the correct contrib directory.
Problem:
No easy way to find files under certain directories (ex: grab all files under
`test/`) or exclude the content of certain paths (ex. `build/`, `.git/`)
Solution:
Pass the full `path` as an arg to the predicate.
Problem:
The sleep before collecting the initial screen state is confusing and
may lead to unexpected success if it comes after a blocking RPC call.
Solution:
Remove that sleep and add an "intermediate" argument.
Problem:
- API validation involves too much boilerplate.
- API validation errors are not consistently worded.
Solution:
Introduce some macros. Currently these are clumsy, but they at least
help with consistency and avoid some nesting.
If nothing matched in match_from_hashbang, also check the file extension table.
For a hashbang like '#!/bin/env foo', this will set the filetype to 'fooscript'
assuming the filetype for the 'foo' extension is 'fooscript' in the extension
table.
Problem:
Build is not reproducible, because generated source files (.c/.h/) are not
deterministic, mostly because Lua pairs() is unordered by design (for security).
https://github.com/LuaJIT/LuaJIT/issues/626#issuecomment-707005671https://www.lua.org/manual/5.1/manual.html#pdf-next
> The order in which the indices are enumerated is not specified [...]
>
>> The hardening of the VM deliberately randomizes string hashes. This in
>> turn randomizes the iteration order of tables with string keys.
Solution:
- Update the code generation scripts to be deterministic.
- That is only a partial solution: the exported function
(funcs_metadata.generated.h) and ui event
(ui_events_metadata.generated.h) metadata have some mpack'ed
tables, which are not serialized deterministically.
- As a workaround, introduce `PRG_GEN_LUA` cmake setting, so you can
inject a modified build of luajit (with LUAJIT_SECURITY_PRN=0)
that preserves table order.
- Longer-term we should change the mpack'ed data structure so it no
longer uses tables keyed by strings.
Closes#20124
Co-Authored-By: dundargoc <gocdundar@gmail.com>
Co-Authored-By: Arnout Engelen <arnout@bzzt.net>
Problem:
Tests that _intentionally_ fail certain conditions cause noise in
$NVIM_LOG_FILE:
$NVIM_LOG_FILE: /home/runner/work/neovim/neovim/build/.nvimlog
(last 100 lines)
WRN 2023-01-16T18:26:27.673 T599.7799.0 unsubscribe:519: RPC: ch 1: tried to unsubscribe unknown event 'doesnotexist'
WRN 2023-01-16T18:29:00.557 ?.11151 server_start:163: Failed to start server: no such file or directory: /X/X/X/...
WRN 2023-01-16T18:33:07.269 127.0.0.1:12345 server_start:163: Failed to start server: address already in use: 127.0.0.1
...
-- Output to stderr:
module 'vim.shared' not found:
no field package.preload['vim.shared']
no file './vim/shared.lua'
no file '/home/runner/nvim-deps/usr/share/lua/5.1/vim/shared.lua'
no file '/home/runner/nvim-deps/usr/share/lua/5.1/vim/shared/init.lua'
no file '/home/runner/nvim-deps/usr/lib/lua/5.1/vim/shared.lua'
no file '/home/runner/nvim-deps/usr/lib/lua/5.1/vim/shared/init.lua'
no file './vim/shared.so'
...
E970: Failed to initialize builtin lua modules
Solution:
- Log to a private $NVIM_LOG_FILE in tests that intentionally fail and
cause ERR log messages.
- Assert that the expected messages are actually logged.
Problem: Unable to customize the column next to a window ('gutter').
Solution: Add 'statuscolumn' option that follows the 'statusline' syntax,
allowing to customize the status column. Also supporting the %@
click execute function label. Adds new items @C and @s which
will print the fold and sign columns. Line numbers and signs
can be clicked, highlighted, aligned, transformed, margined etc.
The existing groups, Error, Hint, Info, Warn cover many use cases, but
neglect the occasion where a diagnostic message should communicate a
non-informative (not a Hint or Info) event. DiagnosticOk covers this
with a generic green colorscheme.
The BufWipeout autocmd is not 100% reliable and may leave stale entries
in the cache. This is sort of a hack/workaround to ensure
`vim.diagnostic.reset` calls don't fail if there are stale cache entries
but instead clears them
Fixes errors like
Error executing vim.schedule lua callback: /usr/share/nvim/runtime/lua/vim/diagnostic.lua:1458: Invalid buffer id: 22
stack traceback:
[C]: in function 'nvim_exec_autocmds'
/usr/share/nvim/runtime/lua/vim/diagnostic.lua:1458: in function 'reset'
While `return` and `return nil` are for most intents and purposes
identical, there are situations where they're not. For example,
calculating the amount of values via the `select()` function will yield
varying results:
```lua
local function nothing() return end
local function null() return nil end
select('#', nothing()) -- 0
select('#', null()) -- 1
```
`vim.tbl_get` currently returns both nil and no results, which makes it
unreliable to use in certain situations without manually accounting for
these discrepancies.
- use pcall when calling vim.secure.read from C
- catch keyboard interrupts in vim.secure.read, rethrow other errors
- selecting "view" in prompt runs :view command
- simplify lua stack cleanup with lua_gettop and lua_settop
Co-authored-by: ii14 <ii14@users.noreply.github.com>
Introduce vim.secure.trust() to programmatically manage the trust
database. Use this function in a new :trust ex command which can
be used as a simple frontend.
Resolves: https://github.com/neovim/neovim/issues/21092
Co-authored-by: Gregory Anders <greg@gpanders.com>
Co-authored-by: ii14 <ii14@users.noreply.github.com>
Extend the capabilities of is_os to detect more platforms such as
freebsd and openbsd. Also remove `iswin()` helper function as it can be
replaced by `is_os("win")`.
This introduces a `suffix` option to the `virt_text` config in
`vim.diagnostic.config()`. The suffix can either be a string which is appended
to the diagnostic message or a function returning such. The function receives a
`diagnostic` argument, which is the diagnostic table of the last diagnostic (the
one whose message is rendered as virt text).
Closes#18687
This introduces a `suffix` option to `vim.diagnostic.open_float()` (and
consequently `vim.diagnostic.config()`) that appends some text to each
diagnostic in the float.
It accepts the same types as `prefix`. For multiline diagnostics, the suffix is
only appended to the last line. By default, the suffix will render the
diagnostic error code, if any.
This function accepts a path to a file and prompts the user if the file
is trusted. If the user confirms that the file is trusted, the contents
of the file are returned. The user's decision is stored in a trust
database at $XDG_STATE_HOME/nvim/trust. When this function is invoked
with a path that is already marked as trusted in the trust database, the
user is not prompted for a response.
This is essentially a convenience wrapper around the `pending()`
function, similar to `skip_fragile()` but more general-purpose.
Also remove `pending_win32` function as it can be replaced by
`skip(iswin())`.
Followup to #20883
Related: #18144
This patch changes the behavior of the default `vim.ui.input` when the user
aborts with `<C-c>`. Currently, it produces an error message + stack and causes
`on_confirm` to not be called. With this patch, `<C-c>` will cause `on_confirm`
to be called with `nil`, the same behavior as when the user aborts with `<Esc>`.
I can think of three good reasons why the behavior should be this way:
1. Easier for the user to understand** It's not intuitive for there to be two
ways to abort an input dialog that have _different_ outcomes. As a user,
I would expect any action that cancels the input to leave me in the same
state. As a plugin author, I see no value in having two possible outcomes for
aborting the input. I have to handle both cases, but I can't think of
a situation where I would want to treat one differently than the other.
2. Provides an API that can be overridden by other implementations** The current
contract of "throw an error upon `<C-c>`" cannot be replicated by async
implementations of `vim.ui.input`. If the callsite wants to handle the case
of the user hitting `<C-c>` they need to use `pcall(vim.ui.input, ...)`,
however an async implementation will instantly return and so there will be no
way for it to produce the same error-throwing behavior when the user inputs
`<C-c>`. This makes it impossible to be fully API-compatible with the
built-in `vim.ui.input`.
3. Provides a useful guarantee to the callsite** As a plugin author, I want the
guarantee that `on_confirm` will _always_ be called (only catastrophic errors
should prevent this). If I am in the middle of some async thread of logic,
I need some way to resume that logic after handing off control to
`vim.ui.input`. The only way to handle the `<C-c>` case is with `pcall`,
which as already mentioned, breaks down if you're using an alternative
implementation.
fix(vim.ui.input): return empty string when inputs nothing
The previous behavior of `vim.ui.input()` when typing <CR> with
no text input (with an intention of having the empty string as input)
was to execute `on_confirm(nil)`, conflicting with its documentation.
Inputting an empty string should now correctly execute `on_confirm('')`.
This should be clearly distinguished from cancelling or aborting the
input UI, in which case `on_confirm(nil)` is executed as before.
Problem: Handling 'statusline' errors is spread out.
Solution: Pass the option name to the lower levels so the option can be
reset there when an error is encountered. (Luuk van Baal,
closesvim/vim#11467)
7b224fdf4a
Problem:
- pesc() returns multiple results, it should return a single result.
- tbl_islist() returns non-boolean in some branches.
- Docstring: @generic must be declared first
Solution:
Constrain docstring annotations.
Fix return types.
Co-authored-by: Justin M. Keyes <justinkz@gmail.com>
Problem:
- Docs HTML: "foo ~" headings (column_heading) are not aligned with
their table columns/contents because the leading whitespace is not
emitted.
- taglinks starting with hyphen like |-x| were not recognized.
- keycodes like `<foo>` and `CTRL-x` were not recognized.
- ToC is not scrollable.
Solution:
- Add ws() to the column_heading case.
- Update help parser to latest version
- supports `keycode`
- fixes for taglink, argument
- Update .toc CSS. https://github.com/neovim/neovim.github.io/issues/297
fix https://github.com/neovim/neovim.github.io/issues/297
When :undo! was introduced to Nvim the implementation of 'inccommand'
preview callback hasn't been fully decided yet, so not notifying buffer
update callbacks made sense for 'inccommand' preview callback in case it
needs to undo the changes itself.
Now it turns out that the undo-and-forget is done automatically for
'inccommand', so it doesn't make sense for :undo! to avoid notifying
buffer update callbacks anymore.
BREAKING CHANGE: When using a Funcref converted from a Lua function as
a method in Vim script, the result of the base expression is now passed
as the first argument instead of being ignored.
vim-patch:8.2.5117: crash when calling a Lua callback from a :def function
Problem: Crash when calling a Lua callback from a :def function. (Bohdan
Makohin)
Solution: Handle FC_CFUNC in call_user_func_check(). (closesvim/vim#10587)
7d149f899d
Makes it possible to use `vim.fs.find` to find files where only a
substring is known.
This is useful for `vim.lsp.start` to get the `root_dir` for languages
where the project-file is only known by its extension, not by the full
name.
For example in .NET projects there is usually a `<projectname>.csproj`
file in the project root.
Example:
vim.fs.find(function(x) return vim.endswith(x, '.csproj') end, { upward = true })
Doing so on `BufDelete` has issues:
- `BufDelete` is also fired for listed buffers that are made unlisted.
- `BufDelete` is not fired for unlisted buffers that are deleted.
This means that diagnostics will be lost for a buffer that becomes unlisted.
It also means that if an entry exists for an unlisted buffer, deleting that
buffer later will not remove its entry from the cache (and you may see "Invalid
buffer id" errors when using diagnostic functions if it was wiped).
Instead, remove a buffer from the cache if it is wiped out.
This means simply `:bd`ing a buffer will not clear its diagnostics now.
These were just added to avoid churn when changing the default
of 'display'. To simplify message handling logic, we might want
to remove support for printing messages in default_grid later on.
This would allow things like printing error messages safely in the
middle of redraw, or a future graduation of the 'multigrid' feature.
This enables vim.filetype.match to match based on a buffer (most
accurate) or simply a filename or file contents, which are less accurate
but may still be useful for some scenarios.
When matching based on a buffer, the buffer's name and contents are both
used to do full filetype matching. When using a filename, if the file
exists the file is loaded into a buffer and full filetype detection is
performed. If the file does not exist then filetype matching is only
performed against the filename itself. Content-based matching does the
equivalent of scripts.vim, and matches solely based on file contents
without any information from the name of the file itself (e.g. for
shebangs).
BREAKING CHANGE: use `vim.filetype.match({buf = bufnr})` instead
of `vim.filetype.match(name, bufnr)`
global-local window options need to be handled specially. When `win` is
given but `scope` is not, then we want to set the local version of the
option but not the global one, therefore we need to force
`scope='local'`.
Note this does not apply to window-local only options (e.g. 'number')
Example:
nvim_set_option_value('scrolloff', 10, {}) -- global-local window option; set global value
nvim_set_option_value('scrolloff', 20, {win=0}) -- global-local window option; set local value
nvim_set_option_value('number', true, {}) -- local window option
is now equivalent to:
nvim_set_option_value('scrolloff', 10, {})
nvim_set_option_value('scrolloff', 20, {win=0, scope='local'}) -- changed from before
nvim_set_option_value('number', true, {win=0}) -- unchanged from before
Only the global-local option with a `win` provided gets forced to local
scope.
`nvim_get_option_value` and `nvim_set_option_value` better handle
unsetting local options. For instance, this is currently not possible:
vim.bo.tagfunc = nil
This does not work because 'tagfunc' is marked as "local to buffer" and
does not have a fallback global option. However, using :setlocal *does*
work as expected
:setlocal tagfunc=
`nvim_set_option_value` behaves more like :set and :setlocal (by
design), so using these as the underlying API functions beneath vim.bo
and vim.wo makes those two tables act more like :setlocal. Note that
vim.o *already* uses `nvim_set_option_value` under the hood, so that
vim.o behaves like :set.
Steps to reproduce:
1. setting `vim.highlight.on_yank`
```
vim.api.nvim_create_autocmd({ "TextYankPost" }, {
pattern = { "*" },
callback = function()
vim.highlight.on_yank({ timeout = 200 })
end,
})
```
2. repeat typing `yeye` ...
3. causes the following error.
```
Error executing vim.schedule lua callback: vim/_editor.lua:0: handle 0x01e96970 is already closing
stack traceback:
[C]: in function 'close'
vim/_editor.lua: in function ''
vim/_editor.lua: in function <vim/_editor.lua:0>
```
📝 Test result before fix:
[----------] Global test environment setup.
[----------] Running tests from test/functional/lua/highlight_spec.lua
[ RUN ] vim.highlight.on_yank does not show errors even if buffer is wiped before timeout: 15.07 ms OK
[ RUN ] vim.highlight.on_yank does not show errors even if executed between timeout and clearing highlight: 15.07 ms ERR
test/helpers.lua:73: Expected objects to be the same.
Passed in:
(string) 'Error executing vim.schedule lua callback: vim/_editor.lua:0: handle 0x02025260 is already closing
stack traceback:
[C]: in function 'close'
vim/_editor.lua: in function ''
vim/_editor.lua: in function <vim/_editor.lua:0>'
Expected:
(string) ''
There can be other places that access window buffer info (e.g.
`tabpagebuflist()`), so checking `w_closing` in `win_findbuf()` doesn't
solve the crash in all cases, and may also cause Nvim's behavior to
diverge from Vim.
Fix#14998
Many filetypes from filetype.vim set buffer-local variables, meaning
vim.filetype.match cannot be used without side effects. Instead of
setting these buffer-local variables in the filetype detection functions
themselves, have vim.filetype.match return an optional function value
that, when called, sets these variables. This allows vim.filetype.match
to work without side effects.
`vim.keymap.del` takes an `opts` parameter that lets caller refer to and
delete buffer-local mappings. For some reason the implementation of
`vim.keymap.del` mutates the table that is passed in, setting
`opts.buffer` to `nil`. This is wrong and also undocumented.
vim.tbl_get takes a table with subsequent string arguments (variadic) that
index into the table. If the value pointed to by the set of keys exists,
the function returns the value. If the set of keys does not exist, the
function returns nil.
Looks like I did an oopsie; although API strings carry a size field, they should
still be usable as C-strings! (even though they may contain embedded NULs)
