Problem: Color completion items display as plain text without visual preview
Solution: Parse RGB/hex colors from documentation and render with colored symbol ■
Closing the quickfix window previously triggered a WinClosed autocmd
that deleted all difftool autocmds and pushed an empty quickfix list,
making the difftool non-functional. Users who close the quickfix window
to gain screen real estate for viewing diffs had no way to continue
navigating entries.
Remove the qf_win tracking and its associated WinClosed autocmd so that
closing the quickfix window no longer tears down the difftool state.
Closing either diff window still performs full cleanup as before.
The BufWinEnter handler no longer passes with_qf to diff_files, so
navigating entries while the quickfix window is closed reuses the
existing diff layout without forcing a layout rebuild.
Fixes#37388
Problem: Unable to configure message targets based on message kind.
Solution: Add cfg.msg.targets mapping message kinds to "cmd/msg/pager".
Check the configured target when writing the message.
cfg.msg = { target = 'cmd', targets = { progress = 'msg', list_cmd = 'pager' } }
will for example use the 'msg' target for progress messages,
immediately open the pager for 'list_cmd' and use the cmdline
for all other message kinds.
fix: allocate hidden console for detached server
Starting the server with UV_PROCESS_DETACHED results in DETACHED_PROCESS, leaving the child without a console. Without a console:
CONIN$ / CONOUT$ cannot resolve, causing channel_from_stdio to fail.
ConPTY cannot attach, breaking :terminal.
This patch allocates a hidden console via AllocConsole() when the server has none, restoring working stdio and enabling ConPTY.
Also updates os_set_cloexec to clear HANDLE_FLAG_INHERIT on the RPC pipe
handles, matching the Unix F_DUPFD_CLOEXEC behavior.
Problem: closing a minimal float saves its style-imposed options into
buffer's wininfo, which get picked up by normal windows on :bnext.
Solution: don't save window options to wininfo for style=minimal windows.
Problem: memline: a crafted swap files with bogus pe_page_count/pe_bnum
values could cause a multi-GB allocation via mf_get(), and
invalid pe_old_lnum/pe_line_count values could cause a SEGV
when passed to readfile() (ehdgks0627, un3xploitable)
Solution: Add bounds checks on pe_page_count and pe_bnum against
mf_blocknr_max before descending into the block tree, and
validate pe_old_lnum >= 1 and pe_line_count > 0 before calling
readfile().
Github Advisory:
https://github.com/vim/vim/security/advisories/GHSA-r2gw-2x48-jj5p65c1a143c3
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem: A stack-buffer-overflow occurs when rendering a statusline
with a multi-byte fill character on a very wide terminal.
The size check in build_stl_str_hl() uses the cell width
rather than the byte length, allowing the subsequent fill
loop to write beyond the 4096-byte MAXPATHL buffer
(ehdgks0627, un3xploitable).
Solution: Update the size check to account for the byte length of
the fill character (using MB_CHAR2LEN).
Github Advisory:
https://github.com/vim/vim/security/advisories/GHSA-gmqx-prf2-8mwf4e5b9e31cb
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem: When parsing a malformed Emacs-style tags file, a 1-byte
heap-buffer-underflow read occurs if the 0x7f delimiter
appears at the very beginning of a line. This happens
because the code attempts to scan backward for a tag
name from the delimiter without checking if space exists.
(ehdgks0627, un3xploitable)
Solution: Add a check to ensure the delimiter (p_7f) is not at the
start of the buffer (lbuf) before attempting to isolate
the tag name.
GitHub Advisory:
https://github.com/vim/vim/security/advisories/GHSA-xcc8-r6c5-hvwv9b7dfa2948
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem: Crash with overlong emacs tag file, because of an OOB buffer
read (ehdgks0627, un3xploitable)
Solution: Check for end of buffer and return early.
Github Advisory:
https://github.com/vim/vim/security/advisories/GHSA-h4mf-vg97-hj8jf6a7f469a9
Cherry-pick a change from patch 9.0.0767.
Add missing change from patch 9.2.0070.
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem: tests: various tests leave swapfiles around
Solution: close open buffers using :bw! instead of :close!
2fa34b6422
Co-authored-by: Christian Brabandt <cb@256bit.org>
Problem: newline escape wrong in ex mode (Konrad Schwarz)
Solution: partly revert patch 7.3.014, remove backslash in front of a
newline when not in prompt mode in ex line mode
(Mohamed Akram)
This fixes newline escaping to allow passing multiple commands to
":global", multiple lines to shell commands, and ending lines in append
mode with backslashes. This should fix a POSIX/(traditional) VI
incompatiblity.
This reverts a previous incorrect attempt at patch v7.3.014 to fix
append mode which removed half of trailing backslashes which lead to,
eg. the following two commands being parsed as having a different number
of backslashes:
```
!echo foo\\\
```
```
!echo foo\\ \
```
fixes: vim/vim#6135fixes: vim/vim#7244closes: vim/vim#15120f3daa4525b
Co-authored-by: Mohamed Akram <mohd.akram@outlook.com>
Problem:
vim.fs.joinpath treats empty string as a path segment
(it adds a path separator for each empty item):
print(vim.fs.joinpath('', 'after/lsp', '')) -- '/after/lsp/'
print(vim.fs.joinpath('', '')) -- '/'
Especially problematic if the empty segment is the first segment, as
that converts the path to an absolute path.
Solution:
Ignore empty (length of 0) path segments.
Benchmark:
local function test(func)
local t = vim.uv.hrtime()
for _ = 1, 100000, 1 do
func('', 'this/is', 'a/very/long/path', '', 'it', 'really', 'is')
end
print(math.floor((vim.uv.hrtime() - t) / 1e6), 'ms')
end
- with Iter():filter() --> 370 ms
- building new segments table --> 208 ms
- with vim.tbl_filter --> 232 ms
- Instead of gsub split on `/` in all parts --> 1870 ms
Problem:
libvterm doesn't support parsing the dim and overline attributes, so when a program running in the embedded terminal emits one of these escape codes, we ignore it and don't surface it to the outer terminal.
Solution: tweak libvterm to add support for both attributes.
Problem:
Automatic background detection sets the background option too late,
which loads colorschemes twice and causes problems when the user's
terminal background doesn't match the default (#32109, #36211, #36416).
Solution:
Use a DA1 query to determine whether the TTY supports OSC 11. Wait for
background detection and setting to complete before processing user
config.
Note: To preserve the existing behavior as much as possible, this triggers
OptionSet manually on VimEnter (since it won't trigger automatically if
we set bg during startup). However, I'm unsure if this behavior is
truly desired given that the documentation says OptionSet is triggered
"After setting an option (except during |startup|)."
Also fixes flickering issue #28667. To check for flickering:
nvim --clean --cmd "set termguicolors" --cmd "echo \"foo\"" --cmd "sleep 10"
On master, this gives me a black screen for 10 seconds, but on this
branch, the background is dark or light depending on the terminal
background (since the option is now set during startup rather than after
VimEnter).
Problem: When nvim is started with "-s -" (read from stdin), ":restart"
drops the "-" argument but keeps "-s", leaving an orphaned "-s" in the
restarted server's argv, causing a crash.
Solution: Drop "-s" and its scriptfile argument when copying v:argv for
the restarted server. Like "-- [files…]", the scriptfile is a one-shot
startup input that should not be replayed on restart
Problem:
Regression from b99cdd0:
Pull diagnostics (from `textDocument/diagnostic`) and push diagnostics
(from `textDocument/publishDiagnostics`) use the same namespace, which
is a problem when using language servers that publish two different sets
of diagnostics on push vs pull, like rust-analyzer (see
https://github.com/rust-lang/rust-analyzer/issues/18709#issuecomment-2551394047).
Solution:
Rename `is_pull` to `pull_id` which accepts a pull namespace instead of
just a boolean.
Problem: Not possible to know when a session will be loaded.
Solution: Add the SessionLoadPre autocommand (Colin Kennedy).
fixes: vim/vim#19084closes: vim/vim#193061c0d468d72
Co-authored-by: Colin Kennedy <colinvfx@gmail.com>
Problem: Leading message newlines (not emitted with ext_messages since
4260f73) were responsible for resetting the redirection message
column (while the newline itself is later pruned...).
Solution: Ensure the redirection column is reset at the start of a message.
(Instead of re-adjusting all the newline callsites which can
themselves hopefully be pruned if ext_messages is enabled by
default.)
fix#37867
This bug happens when only one end is moved between different nodes,
but the other end is also moved but within the same node.
When this happens we need the correct previous position even for the
internal move. This code shall be refactored to make the intent clearer,
(and avoid some unnecessary processing) but this is a fix for the observable
bug.
Thanks to KevinGoodsell for providing a deterministic
reproduce using fuzzing techniques.
Problem: Timer removing a message from the msg buffer does not remove
empty lines if window is closed (col([ui.wins.msg]) fails).
Solution: Use nvim_buf_get_text() to check if line is empty.
The wait added in #37853 doesn't seem to do anything as request is sent
immediately on InsertLeave, and the number 4 also seems wrong. Instead,
the actual cause for the flakiness that the feed() (and hence the buffer
change) may arrive before the scheduled initialization of capabilities,
causing there be only only one textDocument/semanticTokens/full request
instead of two.
fixes#37586
when doing `packadd mypackage` up to two exact paths are added
to &rtp. Instead of recalculating runtime_search_path from scratch,
we can "just" splice these two paths in
This is simple in theory, but get complicated in practice as
"after" dirs do exist and need some wrangling.
Echasnovski did some benchmarking, to show that this reduces overhead
of a init.lua configuration style where separate `packadd!` calls are
used spread out during the config. In addition, "batched" addition
(either using "start" packages or packadd! a lot of opt packages at
once) does not regress.
A theoretical simplification could be to NEVER explicitly add "after"
dirs to &rtp, but implicitly add all existing "after" dirs in reverse
order when calculating the effective run time path. This might be tricky
to do without breaking 12 tpope plugins again tho.
We might also instead consider solutions where &rtp remains fully expanded but no longer is the main source of truth. But this is all post 0.12 work. This PR is an alright stopgap to make 0.12 fully support intended use cases of vim.pack.add() .
Problem:
- `:args` and `argv()` can change after startup.
- `v:arg` includes options/commands, not just files.
- Plugins (e.g. Oil) may rewrite directory args.
Solution:
- New read-only var `v:argf`: snapshot of file/dir args at startup.
- Unaffected by `:args` or plugins.
- Unlike `v:argv`, excludes options/commands.
- Paths are resolved to absolute paths when possible
Example:
nvim file1.txt dir1 file2.txt
:echo v:argf
" ['/home/user/project/file1.txt', '/home/user/project/dir1', '/home/user/project/file2.txt']
Problem: - Unintentionally inserting lines for a replaced multiline
message that also has multiple highlights.
- Scheduled check to see if the expanded cmdline window was
entered makes it difficult to keep track of what happens when
the key pressed to dismiss it results in a message.
- Reading the first line of an error message should be enough
notice for something going wrong.
- "search_cmd" messages should not be shown with 0 'cmdheight'.
- Unable to configure dynamically changed pager height.
- Enabling UI2 doesn't make sense with no UIs attached.
Solution: - Only insert a line for the first chunk after a newline.
- Use getmousepos() to check if the expanded cmdline was
clicked to enter the pager.
entering the pager to serve as a configuration interface.
- Don't expand the cmdline for error messages; user can press g<.
- Don't show "search_cmd" messages with 'cmdheight' set to 0.
- Change 'eventignorewin' to ensure WinEnter is fired when
- Have enable() return early when no UIs are attached.
Problem:
:InspectTree don't show luadoc injection lua file. Since luadoc share
the same "root" with comment in their common primary (lua) tree.
Current logic simply show the largest (comment injection) and ignore all
smaller one (luadoc injection).
Solution:
Handle different lang injections separately. Then sort them by
byte_length to ensure the draw tree consistent.
Problem: Newlines emitted with ext_messages intended to position
the message/prompt on the message grid.
Solution: Don't emit these newlines with ext_messages, followup to 4260f73e.
Problem: after #33036, an error from evaluating 'statusline' clears it and
doesn't draw the statusline. (causing glitchy redraws)
Solution: use the default value instead. If 'stl' is somehow ever empty, still
call redraw_custom_statusline to at least draw an empty statusline.
Ideally our default 'stl' shouldn't itself error too! :-)
Also adjust some prior screen:expect()s to avoid immediate success warnings.
When using feed_data() with hidden buffer, terminal refresh may arrive
during may_restore_curbuf(), causing "last cursor" positions to change.
Get the two "last cursor" positions in the same RPC call.
If the last nvim_eval arrives on RPC channel before rpc_close_event() is
processed, it will be scheduled immediately after rpc_close_event() and
before free_channel_event(), causing the test to fail.
Adding input doesn't help here, as the important part here is processing
main_loop.events, not running the main loop. Instead, use a non-fast API
request to ensure that previously queued events are processed.
Problem:
TUI does not support several standard SGR text attributes:
- dim/faint (SGR 2)
- blink (SGR 5)
- conceal (SGR 8)
- overline (SGR 53)
This means that when a program running in the embedded terminal emits
one of these escape codes, we drop it and don't surface it to the
outer terminal.
Solution:
- Add support for those attributes.
- Also add corresponding flags to `nvim_set_hl` opts, so users can set
these attributes in highlight groups.
- refactor(highlight): widen `HlAttrFlags` from `int16_t` to `int32_t`
Widen the `rgb_ae_attr` and `cterm_ae_attr` fields in HlAttrs from
int16_t to int32_t to make room for new highlight attribute flags,
since there was only one spare bit left.
- The C flag is named HL_CONCEALED to avoid colliding with the
existing HL_CONCEAL in syntax.h (which is a syntax group flag, not
an SGR attribute).
- Also note that libvterm doesn't currently support the dim and overline
attributes, so e.g. `printf '\e[2mThis should be dim\n'` and `printf
'\e[53mThis should have an overline\n'` are still not rendered
correctly when run from the embedded terminal.
Problem: filetype: sh filetype used for env files
Solution: Detect *.env and .env.* files as env filetype,
detect .envrc and .envrc.* as sh filetype,
include a simple env syntax script (DuckAfire)
Previously, .env files were handled by the shell syntax. While
functional, this limited the ability to support specific .env
implementations, such as CodeIgniter4 which allows dots in keys
(e.g., "foo.bar=0").
The new dedicated 'env' filetype and syntax script improves legibility
and prevents highlighting from breaking when encountering spaces.
Currently, the syntax does not support indentation; fields, variables,
and comments must start at the beginning of the line.
closes: vim/vim#19260d0fa375629
Co-authored-by: DuckAfire <155199080+duckafire@users.noreply.github.com>
Problem: completion: hang with line completion and fuzzy (Jesse Pavel)
Solution: Only check the line number when wrapping around the file
(Hirohito Higashi).
fixes: vim/vim#19434closes: vim/vim#19443d8648f7279
Co-authored-by: Hirohito Higashi <h.east.727@gmail.com>
Problem: When 'autocomplete' fires before compl_leader is initialized,
the prefix filter is bypassed. This allows non-prefix matches
(e.g. from fuzzy omnifuncs) to be shown in the popup menu and
incorrectly preinserted.
Solution: In get_leader_for_startcol(), if compl_leader.string is NULL,
fall back to using compl_orig_text as a filter for matches
starting at or after the completion column (Hirohito Higashi).
When 'autocomplete' first fires, compl_leader is NULL because
ins_compl_start() has not set it yet. This caused the prefix filter in
ins_compl_build_pum(), find_next_completion_match() and
find_common_prefix() to be bypassed, allowing non-prefix fuzzy omnifunc
matches to appear in the PUM and be preinserted.
Extend get_leader_for_startcol() to fall back to compl_orig_text when
compl_leader.string is NULL: if the match's cpt source startcol is less
than compl_col the match includes pre-compl_col text, so return
&compl_leader (NULL string) to signal "pass through"; otherwise return
&compl_orig_text so callers filter by the original text. The compl_col
<= 0 guard is kept only for the prepend-text path to avoid it
interfering with the NULL-leader fallback when compl_col is zero.
With this change all callers of get_leader_for_startcol() automatically
receive the correct filter string without additional helpers.
Also update Test_autocomplete_trigger Test 9 to reflect the new
behavior: 'faberge' is no longer shown when completing 'foo' because
it does not start with the current prefix.
Add Test_autocomplete_preinsert_null_leader() to verify that only
prefix-matching candidates appear in the PUM and are preinserted.
fixes: vim/vim#19328closes: vim/vim#19447a04ae0210b
Co-authored-by: Hirohito Higashi <h.east.727@gmail.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
