Problem:
`:restart` does not preserve window layout, etc.
Solution:
- Change `:restart` to save/restore a session automatically.
- Introduce "bang" variant `:restart!` to restart *without* session
save/restore.
- Introduce `v:startreason`.
- `ZR` maps to `:restart!`.
Problem:
On Windows, channel jobs inherit Nvim's stdio, so a background job
writing to CON (e.g. gutentags) draws onto the TUI and stays until
redraw.
Solution:
Give Windows job stdin/stderr libuv-created pipes (UV_CREATE_PIPE)
instead of inherited fds, so libuv spawns the child with
CREATE_NO_WINDOW and CON writes no longer leak onto the TUI.
Problem: Attach-time terminal probes cannot distinguish responses from
different attached UIs.
Solution: Identify the UI by RPC channel id in `TermResponse` and make
`vim.tty.request()` filter responses by channel.
Problem: Terminal background and truecolor detection runs only at startup,
gated on a UI being attached. A headless server has no UI then, so
`'background'` and `'termguicolors'` are never detected and remote
UIs ignore the terminal's theme.
Solution: Also (re)detect on UIEnter. The most recently attached terminal
wins; an explicit user value is preserved.
This was initially added so that the __NVIM_DETACH environment variable
can be added to the jobs started in this test. That environment variable
is no longer needed, and there is also vim.tbl_extend() that can be used
to add an environment variable to a job anyway.
Also, make a shallow copy of opts.env in setup_child_nvim(), as mutating
the opts.env passed in may mask problems in other tests.
Problem: when the TUI output buffer is filled during a redraw, flushing
it to the terminal also causes `CSI ? 2026 l` to be emitted, which ends
synchronized output. This causes visible flickering with large redraws.
Solution: track whether a synchronized output is currently active, and
only emit the end sequence on the final flush.
Problem:
- Windows users can't use terminfo to configure their terminal
capabilities. #37274
- Terminfo definitions sometimes get out of date or are simply
inaccurate.
- Eventually, we may want to drop terminfo, relying primarily on
built-in definitions. Users will still need some flexibility.
Solution:
Support $NVIM_TERMDEFS environment variable, which is JSON data that
defines "terminfo" definitions that override our builtin terminfo.
Problem:
During startup, we manually trigger a useless and misleading `OptionSet`
event, which doesn't set `v:option_*` values (this is a limitation of
`nvim_exec_autocmds`).
ad4bc2d90c/runtime/lua/vim/_core/defaults.lua (L939).
Solution:
The `nvim_exec_autocmds('OptionSet',…)` call does not serve any purpose
since 5cbb9d613b, so just drop it.
Problem:
The Lua test harness still ran through standalone -ll mode, so tests
depended on the low-level Lua path instead of the regular Nvim Lua
environment. That also meant os.exit() coverage had to carry an ASAN
workaround because Lua's raw process exit skipped Nvim teardown and let
LeakSanitizer interfere with the observed exit code.
Solution:
Run the harness and related fixtures with nvim -l. Patch os.exit() in
the main Lua state to exit through getout(), so scripts observe normal
Nvim shutdown while standalone -ll remains available for generator-style
scripts. As a consequence, the startup test can assert os.exit() without
disabling leak detection.
AI-assisted: Codex
Problem:
BufModifiedSet autocmd only triggered for current buffer during
redraw, causing delayed events when :wa writes non-current buffers.
Solution:
- Use the aucmd_defer approach to implement `Optionset modified`.
- Drop BufModifiedSet.
Problem:
When running nvim on a remote machine over SSH, if there is high ping,
then bg detection may not complete in time. This results in a warning
every time nvim is started. #38648
Solution:
Restore 'ttyfast' option and allow it to control whether or not bg
detection is performed. Because this is during startup and before any
user config or commands, we use the environment variable
`NVIM_NOTTYFAST` to allow disabling `ttyfast` during initialization.
Problem:
On Windows, path separators may become inconsistent for various reasons,
which makes normalization quite painful.
Solution:
Normalize paths to `/` at the entry boundaries and always use it
internally, converting back only in rare cases where `\` is really
needed (e.g. cmd.exe/bat scripts?).
This is the first commit in a series of incremental steps.
Note:
* some funcs won't respect shellslash. e.g. `expand/fnamemodify`
* some funcs still respect shellslash, but will be updated in a follow
PR. e.g. `ex_pwd/f_chdir/f_getcwd`
* uv's built-in funcs always return `\`. e.g. `uv.cwd/uv.exepath`
Co-authored-by: Justin M. Keyes <justinkz@gmail.com>
Problem: Destroying a terminal with pending `TermRequest` events leaks
memory.
Solution: Make `emit_termrequest` the sole owner of its `pending_send`
allocation.
Problem:
- The `ZR` feature makes it more obvious that we need some sort of flag so that
an `ExitPre` / `QuitPre` / `VimLeave` handler can handle restarts differently
than a normal exit. For example, it's common that users want `:mksession` on
restart, but perhaps not on a normal exit.
- Nvim has no way to report its "uptime".
Solution:
- Introduce `v:starttime`
- Introduce `v:exitreason`
RUN T339 TUI :restart ZR: Uncaught Error: test/client/uv_stream.lua:111: ECONNRESET
stack traceback:
[C]: in function 'error'
test/client/uv_stream.lua:111: in function <test/client/uv_stream.lua:109>
[C]: in function 'run'
test/client/session.lua:240: in function '_run'
test/client/session.lua:216: in function '_blocking_request'
test/client/session.lua:117: in function 'request'
...t_xdg_terminal/test/functional/terminal/tui_spec.lua:223: in function <...t_xdg_terminal/test/functional/terminal/tui_spec.lua:215>
[C]: in function 'pcall'
test/testutil.lua:82: in function 'retry'
...t_xdg_terminal/test/functional/terminal/tui_spec.lua:215: in function 'assert_restarted'
...t_xdg_terminal/test/functional/terminal/tui_spec.lua:275: in function <...t_xdg_terminal/test/functional/terminal/tui_spec.lua:232>
[C]: in function 'xpcall'
/home/runner/work/neovim/neovim/test/harness.lua:693: in function 'run_callable'
/home/runner/work/neovim/neovim/test/harness.lua:1008: in function 'run_test'
/home/runner/work/neovim/neovim/test/harness.lua:1083: in function 'run_suite'
/home/runner/work/neovim/neovim/test/harness.lua:1081: in function 'run_suite'
/home/runner/work/neovim/neovim/test/harness.lua:1081: in function 'run_suite'
/home/runner/work/neovim/neovim/test/harness.lua:1507: in function 'run_test_file'
/home/runner/work/neovim/neovim/test/harness.lua:1577: in function 'run_iteration'
/home/runner/work/neovim/neovim/test/harness.lua:1665: in function 'main'
/home/runner/work/neovim/neovim/test/runner.lua:30: in main chunk
-- Tests exited non-zero: 255
CMake Error at /home/runner/work/neovim/neovim/cmake/RunTests.cmake:135 (message):
functional tests failed with error: 255
Problem:
On Windows, :restart cannot immediately reuse the canonical --listen
address because named pipe release is asynchronous.
Solution:
Start the new Nvim server on a temporary address; in the new Nvim,
retry serverstart() with the original ("canonical") address until it
succeeds.
Replace the busted-based Lua test runner with a repo-local harness.
The new harness runs spec files directly under `nvim -ll`, ships its own
reporter and lightweight `luassert` shim, and keeps the helper/preload
flow used by the functional and unit test suites.
Keep the file boundary model shallow and busted-like by restoring `_G`,
`package.loaded`, `package.preload`, `arg`, and the process environment
between files, without carrying extra reset APIs or custom assertion
machinery.
Update the build and test entrypoints to use the new runner, add
black-box coverage for the harness itself, and drop the bundled
busted/luacheck dependency path.
AI-assisted: Codex
Problem:
The retry on Windows doesn't seem to actually work, as the test still
occasionally fails on Windows. Meanwhile the test can fail on Linux as
well:
FAILED test/functional/terminal/channel_spec.lua @ 123: chansend sends lines to terminal channel in proper order
test/functional/terminal/channel_spec.lua:131: retry() attempts: 1
test/functional/terminal/channel_spec.lua:134: Failed to match any screen lines.
Expected (anywhere): "echo "hello".*echo "world""
Actual:
|^ech$ o "hello" |
|echo "world" |
|hello |
|$ world |
|$ |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
|term://~/work/neovim/neovim/build/Xtest_xdg_terminal//32516:sh [-] |
| |
Solution:
Use a wait before the chansend() instead.
FAILED test/functional/terminal/cursor_spec.lua @ 419: :terminal cursor uses the correct attributes
test/functional/terminal/cursor_spec.lua:448: Expected objects to be the same.
Passed in:
(string) 'block'
Expected:
(string) 'vertical'
stack traceback:
test/functional/terminal/cursor_spec.lua:448: in function <test/functional/terminal/cursor_spec.lua:419>
Problem:
failures in s390x CI.
Solution:
- runtime/lua/man.lua: parse_path() can return nil but 3 callers didn't handle it.
- skip some tests on s390x.
TODO:
- TODO: why "build/bin/xxd is not executable" on s390x?
- TODO: other failures, not addressed (see below).
OTHER FAILURES:
FAILED test/functional/treesitter/fold_spec.lua @ 87: treesitter foldexpr recomputes fold levels after lines are added/removed
test/functional/treesitter/fold_spec.lua:95: Expected objects to be the same.
Passed in:
(table: 0x4013c18940) {
[1] = '0'
[2] = '0'
[3] = '0'
*[4] = '0'
[5] = '0'
...
Expected:
(table: 0x4005acf900) {
[1] = '0'
[2] = '0'
[3] = '>1'
*[4] = '1'
[5] = '1'
...
stack traceback:
(tail call): ?
test/functional/treesitter/fold_spec.lua:95: in function <test/functional/treesitter/fold_spec.lua:87>
FAILED test/functional/treesitter/select_spec.lua @ 52: treesitter incremental-selection works
test/functional/treesitter/select_spec.lua:63: Expected objects to be the same.
Passed in:
(string) 'bar(2)'
Expected:
(string) 'foo(1)'
stack traceback:
(tail call): ?
test/functional/treesitter/select_spec.lua:63: in function <test/functional/treesitter/select_spec.lua:52>
FAILED test/functional/treesitter/select_spec.lua @ 69: treesitter incremental-selection repeat
test/functional/treesitter/select_spec.lua:82: Expected objects to be the same.
Passed in:
(string) '2'
Expected:
(string) '4'
stack traceback:
(tail call): ?
test/functional/treesitter/select_spec.lua:82: in function <test/functional/treesitter/select_spec.lua:69>
FAILED test/functional/treesitter/select_spec.lua @ 98: treesitter incremental-selection history
test/functional/treesitter/select_spec.lua:111: Expected objects to be the same.
Passed in:
(string) 'bar(2)'
Expected:
(string) 'foo(1)'
stack traceback:
(tail call): ?
test/functional/treesitter/select_spec.lua:111: in function <test/functional/treesitter/select_spec.lua:98>
FAILED test/functional/treesitter/select_spec.lua @ 186: treesitter incremental-selection with injections works
test/functional/treesitter/select_spec.lua:201: Expected objects to be the same.
Passed in:
(string) 'lua'
Expected:
(string) 'foo'
stack traceback:
(tail call): ?
test/functional/treesitter/select_spec.lua:201: in function <test/functional/treesitter/select_spec.lua:186>
FAILED test/functional/treesitter/select_spec.lua @ 216: treesitter incremental-selection with injections ignores overlapping nodes
test/functional/treesitter/select_spec.lua:231: Expected objects to be the same.
Passed in:
(string) ' )'
Expected:
(string) ' foo('
stack traceback:
(tail call): ?
test/functional/treesitter/select_spec.lua:231: in function <test/functional/treesitter/select_spec.lua:216>
FAILED test/functional/treesitter/select_spec.lua @ 307: treesitter incremental-selection with injections handles disjointed trees
test/functional/treesitter/select_spec.lua:337: Expected objects to be the same.
Passed in:
(string) 'int'
Expected:
(string) '1}'
stack traceback:
(tail call): ?
test/functional/treesitter/select_spec.lua:337: in function <test/functional/treesitter/select_spec.lua:307>
ERROR test/functional/treesitter/parser_spec.lua @ 562: treesitter parser API can run async parses with string parsers
test/functional/treesitter/parser_spec.lua:565: attempt to index a nil value
stack traceback:
test/functional/testnvim/exec_lua.lua:124: in function <test/functional/testnvim/exec_lua.lua:105>
(tail call): ?
(tail call): ?
test/functional/treesitter/parser_spec.lua:563: in function <test/functional/treesitter/parser_spec.lua:562>
FAILED test/functional/core/job_spec.lua @ 1157: jobs jobstop() kills entire process tree #6530
test/functional/core/job_spec.lua:1244: retry() attempts: 94
test/functional/core/job_spec.lua:1246: Expected objects to be the same.
Passed in:
(table: 0x401dd74b30) {
[name] = 'sleep <defunct>'
[pid] = 33579
[ppid] = 1 }
Expected:
(userdata) 'vim.NIL'
stack traceback:
test/testutil.lua:89: in function 'retry'
test/functional/core/job_spec.lua:1244: in function <test/functional/core/job_spec.lua:1157>
Problem:
The "restart" event has some problems:
- all UI clients must implement a somewhat complex set of setups
- UI must be on the same machine as the server
- only works for the "current" UI
- race/edge case: If the user config has errors / waiting for input, are
all UIs able to attach while Nvim is waiting for input?
Solution:
- Perform the restart on the server, not the client.
- Pass listen address (instead of CLI args) in the UI event.
- Simplifies UI logic: they only need to attach to new address.
- Opens the door for more enhancements in the future, such as allowing
all UIs to reattach instead of only the "current" UI.
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Co-authored-by: Justin M. Keyes <justinkz@gmail.com>
Problem:
If buffer update callbacks poll for uv events during terminal scrollback
refresh, new output from PTY process may lead to incorrect scrollback.
Solution:
Don't poll for output to the same terminal as the one being refreshed.
Problem:
"[Process exited]" is implemented in C with anonymous namespace
and users have no way to hide it.
Solution:
- Handle "TermClose" event in Lua.
- User can delete the "nvim.terminal" augroup to avoid "[Process exited]".
Problem:
Applications running inside :terminal that use DEC private mode 2026
(synchronized output) to batch screen updates get garbled rendering.
Neovim's embedded libvterm does not handle mode 2026, so the
synchronization sequences are ignored and intermediate screen states
leak through as visual corruption.
Solution:
Add mode 2026 support to libvterm's state machine and wire it through
to terminal.c. When an application enables mode 2026, invalidation of
the terminal buffer is deferred until the application disables it,
causing all accumulated screen updates to flush as a single
atomic refresh.
* fix(terminal): harden sync output redraw gating
Problem:
The initial mode 2026 implementation gated invalidate_terminal()
but missed three other redraw paths: term_sb_push/term_sb_pop
bypassed the gate by directly adding to invalidated_terminals,
refresh_timer_cb could fire mid-sync flushing partial state, and
the 10ms timer delay after sync-end left a window for stale
repaints.
Solution:
- Gate term_sb_push/term_sb_pop during synchronized output
- Skip syncing terminals in refresh_timer_cb
- On sync end, schedule a zero-delay full-screen refresh via
sync_flush_pending flag in terminal_receive()
- Add news.txt entry for mode 2026 support
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* test(terminal): add vterm unit tests for mode 2026
Add unit-level tests for synchronized output (mode 2026) to
vterm_spec.lua, covering settermprop callbacks and DECRQM
query/response.
Suggested-by: justinmk
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(terminal): address review feedback for mode 2026
- Use multiqueue_put(main_loop.events) instead of restarting the
global refresh timer on sync end, to avoid affecting other
invalidated terminals.
- Add screen:expect_unchanged() to verify screen doesn't update
during sync mode.
- Merge buffer-lines test into existing test.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Problem: #38316 is a bit aggressive; we need not always leave Terminal mode if
autocmds put us in a different terminal.
Solution: don't skip entering; let terminal_check_focus handle whether we should
immediately leave.
Problem: heap-use-after-free possible when entering Terminal mode if
autocommands close the terminal.
Solution: set the refcount. Skip to the end if we must close the terminal.
Problem:
In autocmd examples, using "args" as the event-object name is vague and
may be confused with a user-command.
Solution:
Use "ev" as the conventional event-object name.
Problem: Since the "[Process exited]" msg is no longer part of buffer
contents, `jobstart`'s reuse of unmodified finished terminal buffers
does not clear the msg.
Solution: Delete the extmark if `term` is already closed.
Problem:
When a terminal process exits, "[Process Exited]" text is added
to the buffer contents.
Solution:
- Return `exitcode` field from `nvim_get_chan_info`.
- Show it in the default 'statusline'.
- Show exitcode as virtual text in the terminal buffer.
