Problem: Strange error when heredoc marker starts with "trim".
Solution: Check for whitespace after "trim" or "eval" (zeertzjq)
For :python3 etc., a heredoc marker that starts with a lower-case letter
is valid, and when it starts with "trim" it works in a script but not in
a function, and this PR makes it works in a function.
For :let, a heredoc marker that starts with a lower-case letter is not
valid, but when it starts with "trim" or "eval" the error can be a bit
confusing in a function, and this PR make it less confusing.
closes: vim/vim#16574449c2e5454
Problem: tests: test_compiler fails on Windows without Maven.
Solution: Add Xspotbugs directory to $PATH when mvn is not available
(zeertzjq).
closes: vim/vim#1657623da16d3d0
Problem: SpotBugs compiler can be improved
Solution: runtime(compiler): Improve defaults and error handling for
SpotBugs; update test_compiler.vim (Aliaksei Budavei)
runtime(compiler): Improve defaults and error handling for SpotBugs
* Keep "spotbugs#DefaultPreCompilerTestAction()" defined but
do not assign its Funcref to the "PreCompilerTestAction"
key of "g:spotbugs_properties": there are no default and
there can only be introduced arbitrary "*sourceDirPath"
entries; therefore, this assignment is confusing at best,
given that the function's implementation delegates to
whatever "PreCompilerAction" is.
* Allow for the possibility of relative source pathnames
passed as arguments to Vim for the Javac default actions,
and the necessity to have them properly reconciled when
the current working directory is changed.
* Do not expect users to remember or know that new source
files ‘must be’ ":argadd"'d to be then known to the Javac
default actions; so collect the names of Java-file buffers
and Java-file Vim arguments; and let users providing the
"@sources" file-lists in the "g:javac_makeprg_params"
variable update these file-lists themselves.
* Strive to not leave behind a fire-once Syntax ":autocmd"
for a Java buffer whenever an arbitrary pre-compile action
errors out.
* Only attempt to run a post-compiler action in the absence
of failures for a pre-compiler action. Note that warnings
and failures are treated alike (?!) by the Javac compiler,
so when previews are tried out with "--enable-preview",
remember about passing "-Xlint:-preview" too to also let
SpotBugs have a go.
* Properly group conditional operators when testing for key
entries in a user-defined variable.
* Also test whether "javaExternal" is defined when choosing
an implementation for source-file parsing.
* Two commands are provided to toggle actions for buffer-local
autocommands:
- SpotBugsRemoveBufferAutocmd;
- SpotBugsDefineBufferAutocmd.
For example, try this from "~/.vim/after/ftplugin/java.vim":
------------------------------------------------------------
if exists(':SpotBugsDefineBufferAutocmd') == 2
SpotBugsDefineBufferAutocmd BufWritePost SigUSR1
endif
------------------------------------------------------------
And ":doautocmd java_spotbugs User" can be manually used at will.
closes: vim/vim#16140368ef5a48c
Co-authored-by: Aliaksei Budavei <0x000c70@gmail.com>
Problem: Arguments of a function are missing after failing to redefine
it (after 8.2.2505), and heap-use-after-free with script-local
function (after 9.1.1063).
Solution: Don't clear arguments or free uf_name_exp when failing to
redefine an existing function (zeertzjq)
closes: vim/vim#1656704d2a3fdc0
Problem: heap-use-after-free and stack-use-after-scope with :14verbose
when using :return and :try (after 9.1.1063).
Solution: Move back the vim_free(tofree) and the scope of numbuf[].
(zeertzjq)
closes: vim/vim#165632101230f40
Problem: Cannot control cursor positioning of getchar().
Solution: Add "cursor" flag to {opts}, with possible values "hide",
"keep" and "msg".
related: vim/vim#10603closes: vim/vim#16569edf0f7db28
Problem:
When an evaluation {-item calls `nvim_eval_statusline()`, that nested
call may overwrite the same memory used for `stl_items`.
Solution:
Make `curitem` static and use it to compute an offset to avoid
overwriting `stl_items` in nested calls to `build_stl_str_hl()`. Move
miscellaneous statusline tests into `describe()` block.
When given, only that range will be checked for validity rather than the
entire tree. This is used in the highlighter to save CPU cycles since we
only need to parse a certain region at a time anyway.
Problem: preinsert text completions not deleted with <C-W>/<C-U>
(ddad431, after v9.1.1059)
Solution: handle <C-W> or <C-U> specifically and clear the completion
(glepnir)
fixes: vim/vim#16557closes: vim/vim#16565001c26cd61
Problem: getchar() can't distinguish between C-I and Tab.
Solution: Add {opts} to pass extra flags to getchar() and getcharstr(),
with "number" and "simplify" keys.
related: vim/vim#10603closes: vim/vim#16554e0a2ab397f
Cherry-pick tv_dict_has_key() from patch 8.2.4683.
Problem:
Event handler declared with `once=true` can re-trigger itself (i.e. more
than once!) by calling `nvim_exec_autocmds` or `:doautocmd`.
Analysis:
This happens because the callback is executed before deletion/cleanup
(`aucmd_del`). And calling `aucmd_del` before `call_autocmd_callback`
breaks the autocmd execution...
Solution:
Set `ac->pat=NULL` to temporarily "delete" the autocmd, then restore it
after executing the callback.
Fix#25526
Co-authored-by: Justin M. Keyes <justinkz@gmail.com>
Problem: completion: input text deleted with preinsert when adding leader
Solution: remove compl_length and check the ptr for being equal
to pattern when preinsert is active (glepnir)
closes: vim/vim#16545bfb4eea786
Co-authored-by: glepnir <glephunter@gmail.com>
Problem: Vim doesn't highlight to be inserted text when completing
Solution: Add support for the "preinsert" 'completeopt' value
(glepnir)
Support automatically inserting the currently selected candidate word
that does not belong to the latter part of the leader.
fixes: vim/vim#3433closes: vim/vim#16403edd4ac3e89
Co-authored-by: glepnir <glephunter@gmail.com>
Problem: `cmdline_show` is emitted unnecessarily each event
loop iteration, because `cmdline_was_last_drawn` is never set.
Solution: Keep track of whether the cmdline was last drawn to avoid
unnecessarily emitting cmdline_show. Set `redraw_state` to
emit `cmdline_pos` when emitting `CursorMovedC`. Only emit
`cmdline_pos` when cmdline was last drawn.
Since these assertions all use a common function to perform the test
assertions, it's difficult to figure out which test failed:
ERROR test/unit/testutil.lua @ 785: vim_snprintf() positional arguments
test/unit/testutil.lua:757: test/unit/testutil.lua:741: (string) '
test/unit/strings_spec.lua:143: Expected objects to be the same.
Passed in:
(number) 6400
Expected:
(number) 6'
exit code: 256
Adding context to the assertion makes it clearer what the problem is:
ERROR test/unit/testutil.lua @ 785: vim_snprintf() positional arguments
test/unit/testutil.lua:757: test/unit/testutil.lua:741: (string) '
test/unit/strings_spec.lua:149: snprintf(buf, 0, "%1$0.*2$b", cdata<unsigned int>: 0xf78d0f38, cdata<int>: 0xf78dc4e0) = 001100
Expected objects to be the same.
Passed in:
(number) 6400
Expected:
(number) 6'
exit code: 256
Problem:
`nvim_win_set_config` cannot set the title and footer independently.
When only one is given, the other is reset to the default of "left".
Solution:
Reuse existing title/footer value if not provided.
Problem: tests: off-by-one error in CheckCWD in test_debugger.vim
Solution: Fix off-by-one in CheckCWD leading to local tests failure
(Yee Cheng Chin)
Vim's test_debugger's Test_debug_backtrace_level test will fail if you
happen to run it in a Vim repository with full path of directory being
exactly 29 characters (e.g. `/Users/bob/developing/src/vim`). The test
does term dump comparison and the printout will overflow if the CWD is
too long. It does have a function to skip to test if it detects that but
it's off by one leading to this one situation where it will fail.
The reason why the logic didn't account for this is that Vim's message
printing will overflow the text if it prints a message at exactly the
width of the terminal. This could be considered a bug / quirk but that
will be another issue.
closes: vim/vim#165173acfbb4b54
Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
Problem: "nosort" enables fuzzy filtering even if "fuzzy" isn't in
'completeopt' (after v9.1.1049)
Solution: Only enable fuzzy filtering when "fuzzy" is in 'completeopt'.
(zeertzjq)
closes: vim/vim#16510d65aa1bbdb
Problem: insert-completed items are always sorted, although the LSP
spec[1] standard defines sortText in the returned
completionitem list. This means that the server has sorted the
results. When fuzzy is enabled, this will break the server's
sorting results.
Solution: disable sorting of candidates when "nosort" is set in
'completeopt'
[1]
https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#completionItemcloses: vim/vim#16501f400a0cc41
Co-authored-by: glepnir <glephunter@gmail.com>
Problem: Right aligned virtual text can cover up buffer text if virtual
text is too long
Solution: An additional option for `virt_text_pos` called
`eol_right_align` has been added to truncate virtual text if it would
have otherwise covered up buffer text. This ensures the virtual text
extends no further left than EOL.
Problem: nlua_call_ref_ctx() does not pop the return value in fast
context that did not error.
Solution: Fall through to end; calling nlua_call_pop_retval().
Problem: A right-click on the 'statuscolumn' does not open the
popupmenu, even if a cell without a clickdef is clicked.
Clicking the %C fold item does not open/close the fold.
Solution: Open the popupmenu when there is no clickdef like right-clicking
the sign/numbercolumn does. Fill "linebuf_vcol" when drawing the
'statuscolumn' to handle foldcolumn item clicks.
Problem: Combined highlighting was not applied to nvim_eval_statusline(),
and 'statuscolumn' sign segment/numhl highlights.
Solution: Add an additional `groups` element to the return value of
`nvim_eval_statusline()->highlights`. This is an array of stacked
highlight groups (highest priority last). Also resolve combined
highlights for the 'statuscolumn' sign segment/numhl highlights.
Expose/synchronize some drawline.c logic that is now mimicked in
three different places.
Problem: filetype: just files are not recognized
Solution: adjust filetype detection pattern, detect just shebang line,
include just ftplugin, indent and syntax plugin
(Peter Benjamin)
closes: vim/vim#1646672755b3c8e
Co-authored-by: Peter Benjamin <petermbenjamin@gmail.com>
Problem: filetype: N-Tripels and TriG files are not recognized
Solution: detect '*.nt' files as ntriples filetype and '*.trig' files
as trig filetype (Gordian Dziwis)
closes: vim/vim#16493c04334c33f
Co-authored-by: Gordian Dziwis <gordian@dziw.is>
Problem: Crash after scrolling and pasting in silent Ex mode.
(fizz-is-on-the-way)
Solution: Don't move cursor to line 0 when scrolling.
(zeertzjq)
closes: vim/vim#16506df098fedbc
This has been possible in the "backend" for a while but
API was missing.
Followup: we will need a `details2=true` mode for `nvim_get_hl_id_by_name`
to return information in a way forward compatible with even further
enhancements.
Problem:
Incremental preview is not allowed on 'nomodifiable' buffers.
Solution:
- Allow preview on 'nomodifiable' buffers.
- Restore the 'modifiable' option in case the preview function changes it.
Problem: filetype: setting bash filetype is backwards incompatible
Solution: revert patch v9.1.0965, detect bash scripts again as sh
filetype
This reverts commit b9b762c21f2b61e0e7d8fee43d4d3dc8ecffd721.
related: vim/vim#16309727c567a09
Co-authored-by: Christian Brabandt <cb@256bit.org>
vim-patch:9.1.1033: tests: shaderslang was removed from test_filetype erroneously
Problem: tests: shaderslang was removed from test_filetype erroneously
(Christian Clason, after v9.1.1030)
Solution: restore the test
1d2867df0c
Co-authored-by: Christian Brabandt <cb@256bit.org>
Follow up to https://github.com/neovim/neovim/pull/32072
If there is no prefix (e.g. at the start of word boundary or a line), it
always used the `filterText` because the `match` function always
returned false.
Problem:
Calling `xstrdup` with a NULL pointer causes a SIGSEGV if `lua_tostring` returns
NULL in `nlua_luv_thread_common_cfpcall`.
Crash stack trace:
- `_platform_strlen` → `xstrdup` (memory.c:469)
- `nlua_luv_thread_common_cfpcall` (executor.c:281)
Solution:
Check if `lua_tostring` returns NULL and pass NULL to `event_create` to avoid the crash.
Making this opt-out (on by default) was the wrong choice from the
beginning. It is too visually noisy to be enabled by default.
BREAKING CHANGE: Users must opt-in to the diagnostic virtual text
handler by adding
vim.diagnostic.config({ virtual_text = true })
to their config.
Problem:
With language servers like lemminx, completing xml tags like `<mo` first
shows the right candidates (`modules`) but after typing `d` the
candidates disappear.
This is because the server returns:
[...]
filterText = "<module",
label = "module",
textEdit = {
newText = "<module>$1</module>$0",
Which resulted in `module` being used as `word`, and `module` doesn't
match the prefix `<mo`. Typing `d` causes the `complete()` filtering
mechanism to kick in and remove the entry.
Solution:
Use `<module` from the `filterText` as `word` if the textEdit/label
heuristic doesn't match.
Problem: no way to get current selected item in a async context
Solution: add completed flag to show the entries of currently selected
index item (glepnir)
closes: vim/vim#16451037b028a22
Co-authored-by: glepnir <glephunter@gmail.com>
This commit adds basic support for the kitty keyboard protocol to
Neovim's builtin terminal. For now only the first mode ("Disambiguate
escape codes") is supported.
Problem: Vim9: Patch 9.1.1013 causes a few problems
Solution: Translate the function name only when it is a string
(Yegappan Lakshmanan)
fixes: vim/vim#16453closes: vim/vim#164509904cbca41
Cherry-pick call() change from patch 9.0.0345.
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: Vim9: Regression caused by patch v9.1.0646
Solution: Translate the function name before invoking it in call()
(Yegappan Lakshmanan)
fixes: vim/vim#16430closes: vim/vim#164456289f91591
N/A patch:
vim-patch:8.2.4176: Vim9: cannot use imported function with call()
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: v9.1.0743 causes regression with diff mode
Solution: Fix the regression with overlapping regions
closes: vim/vim#1645401f6509fb2
Co-authored-by: Yukihiro Nakadaira <yukihiro.nakadaira@gmail.com>
Problem: If a (vim.ui_attach) cmdline_hide callback triggers a redraw,
it may cause cmdline_show events for an already exited cmdline.
Solution: Avoid emitting cmdline_show event when ccline.cmdbuff is
already NULL. Unset ccline.cmdbuff before emitting cmdline_hide.
When `root_dir` is a function it can (and often will) call the provided
callback function in a fast API context (e.g. in the `on_exit` handler
of `vim.system`). When the callback function is executed we should
ensure that it runs vim.lsp.start on the main event loop.
On master (and also before #31539) closing a window could cause the
used wininfo for a buffer to change. This is due to always removing the
previous NULL wininfo when deleting a window, even if that wininfo had
higher priority than the the deleted window's own wininfo.
Instead delete the wininfo with lowest priority. This retains the memory
saving efect while not affecting the effective value of window options
and so on.
Problem: Anchoring a floating window to the tabline and laststatus is
cumbersome; requiring autocommands and looping over all
windows/tabpages.
Solution: Add new "tabline" and "laststatus" options to the `relative`
field of nvim_open_win() to place a window relative to.
Problem:
The floating window for hover and signature help always cuts off a few lines,
because the `_make_floating_popup_size` function counts empty lines as having
zero height.
Solution:
Ensure the height is at least 1.
Problem:
Whether an option is allowed to be empty isn't well defined and
isn't properly checked.
Solution:
- For non-list string options, explicitly check the option value
if it is empty.
- Annotate non-list string options that can accept an empty value.
- Adjust command completion to ignore the empty value.
- Render values in Lua meta files
The code represents a useful pattern in normal mode where remapping
`<tab>` will implicitly also remap `<c-i>` unless you remap that
explicitly. This relies on the _unmapped_ behavior being identical which
is not true in terminal mode, as vterm can distinguish these keys.
Vim seems to entangle this with kitty keyboard mode detection which
is irrelevant for us. Conditional fallbacks depending on
keyboard mode could be done completely inside `vterm/` without getchar.c
getting involved, I would think.
Problem: Popup menu internal error with some abbr in completion item.
Solution: Don't compute attributes when there is no corresponding text.
Reduce indent in pum_redraw() while at it (zeertzjq).
fixes: vim/vim#16427closes: vim/vim#164353a0cc36c69
**Problem:** `vim.treesitter.get_parser()` and `vim.treesitter.start()`
both parse the tree before returning it. This is problematic because if
this is a sync parse, it will stall the editor on large files. If it is
an async parse, the functions return stale trees.
**Solution:** Remove this parsing side effect and leave it to the user
to parse the returned trees, either synchronously or asynchronously.
**Problem:** Parsing can be slow for large files, and it is a blocking
operation which can be disruptive and annoying.
**Solution:** Provide a function for asynchronous parsing, which accepts
a callback to be run after parsing completes.
Co-authored-by: Lewis Russell <lewis6991@gmail.com>
Co-authored-by: Luuk van Baal <luukvbaal@gmail.com>
Co-authored-by: VanaIgr <vanaigranov@gmail.com>
**Problem:** Query parsing uses a weak cache which is invalidated
frequently
**Solution:** Make the cache strong, and invalidate it manually when
necessary (that is, when `rtp` is changed or `query.set()` is called)
Co-authored-by: Christian Clason <c.clason@uni-graz.at>
Problem: filetype: various ignore are not recognized
Solution: detect rg/docker/npm/vvsce ignore files as 'gitgnore' filetype
(Wu, Zhenyu)
Not only prettier, but many programs also support ignore files (like rg,
docker, npm, vscode). So use the gitignore filetype for them due to same syntax
closes: vim/vim#164288cbe2e0a0a
Co-authored-by: Wu, Zhenyu <wuzhenyu@ustc.edu>
Problem: [security]: heap-buffer-overflow with visual mode when
using :all, causing Vim trying to access beyond end-of-line
(gandalf)
Solution: Reset visual mode on :all, validate position in gchar_pos()
and charwise_block_prep()
This fixes CVE-2025-22134
Github Advisory:
https://github.com/vim/vim/security/advisories/GHSA-5rgf-26wj-48v8c9a1e257f1
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem: completion text is highlighted even with no pattern found
Solution: use ins_compl_leader_len() instead of checking
compl_leader.length (glepnir)
closes: vim/vim#164229fddb8ae77
Co-authored-by: glepnir <glephunter@gmail.com>
Problem: ComplMatchIns may highlight wrong text
Solution: don't highlight in case of fuzzy match,
skip-highlight when not inserting anything
(glepnir)
closes: vim/vim#16404e890887b80
Problem:
The `nvim_notify` API (note: unrelated to `vim.notify()` Lua API) was
not given any real motivation in https://github.com/neovim/neovim/pull/13843
There are, and were, idiomatic and ergonomic alternatives already.
Solution:
Deprecate `nvim_notify`.
Problem: tests: ruby tests fail with Ruby 3.4
Solution: adjust expected output for Ruby 3.4
(Yee Cheng Chin)
Vim's Ruby tests relied on explicit matching of output texts which are
fragile in design. Ruby 3.4 has changed the output slightly (using
'name' instead of `name', and also using more spaces in dictionary
printouts). Modify the Vim tests to be less fragile to such changes.
closes: vim/vim#16411ebea31e454
Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
Problem: leaking finished exception
(after v9.1.0984)
Solution: use finish_exception to clean up caught exceptions
(Yee Cheng Chin)
In Vimscript, v:exception/throwpoint/stacktrace are supposed to reflect
the currently caught exception, and be popped after the exception is
finished (via endtry, finally, or a thrown exception inside catch).
Vim9script does not handle this properly, and leaks them instead. This
is clearly visible when launching GVim with menu enabled. A caught
exception inside the s:BMShow() in menu.vim would show up when querying
`v:stacktrace` even though the exception was already caught and handled.
To fix this, just use the same functionality as Vimscript by calling
`finish_exception` to properly restore the states. Note that this
assumes `current_exception` is always the same as `caught_stack` which
believe should be the case.
Added tests for this. Also fix up test_stacktrace to properly test the
stack restore behavior where we have nested exceptions in catch blocks
and to also test the vim9script functionality properly.
- Also, remove its dependency on explicitly checking a line number in
runtest.vim which is a very fragile way to write tests as any minor
change in runtest.vim (shared among all tests) would require changing
test_stacktrace.vim. We don't actually need such granularity in the
test.
closes: vim/vim#164132051af1642
Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
