Starting

Nvim :help pages, generated from source using the tree-sitter-vimdoc parser.


Starting Vim

Nvim arguments cli-arguments

Most often, Nvim is started to edit a single file with the command:
nvim filename
More generally, Nvim is started with:
nvim [option | filename] ..
Option arguments and file name arguments can be mixed, and any number of them can be given. However, watch out for options that take an argument.
The following items decide how to start editing:
-file --- filename One or more file names. The first one will be the current file and read into the buffer. The cursor will be positioned on the first line of the buffer. To avoid a file name starting with a '-' being interpreted as an option, precede the arglist with "--", e.g.:
nvim -- -filename
All arguments after "--" are interpreted as file names, no other options or "+command" arguments can follow.
-- - Alias for stdin (standard input). Example:
echo text | nvim - file
"text" is read into buffer 1, "file" is opened as buffer 2. In most cases (except -s, -es, --embed, --headless) if stdin is not a TTY then it is read as text, so "-" is implied:
echo text | nvim file
The buffer will be marked as modified, because it contains text that needs to be saved (except for readonly -R mode). If you don't like that, put these lines in your init.vim:
" Don't set 'modified' when reading from stdin
au StdinReadPost * set nomodified
To read stdin as Normal commands use -s with "-":
echo "ifoo" | nvim -s -
To read stdin as Ex commands use -es or -e:
echo "echo getpid()" | nvim -e - -V1
To open a file literally named "-", put it after "--":
echo foo | nvim -- -
To read stdin as text with --headless use "-".
-t -tag -t {tag} A tag. "tag" is looked up in the tags file, the associated file becomes the current file, and the associated command is executed. Mostly this is used for C programs, in which case "tag" often is a function name. The effect is that the file containing that function becomes the current file and the cursor is positioned on the start of the function (see tags).
-q -qf -q [errorfile] QuickFix mode. The file with the name [errorfile] is read and the first error is displayed. See quickfix. If [errorfile] is not given, the 'errorfile' option is used for the file name. See 'errorfile' for the default value.
(nothing) Without one of the four items above, Vim will start editing a new buffer. It's empty and doesn't have a file name.
startup-options The option arguments may be given in any order. Single-letter options can be combined after one dash. There can be no option arguments after the "--" argument.
--help -h --help -? -? -h Give usage (help) message and exit.
--version -v --version -v Print version information and exit. Same output as for :version command.
--clean --clean Mimics a fresh install of Nvim:
Skips initializations from files and environment variables.
No 'shada' file is read or written.
Excludes user directories from 'runtimepath'
Loads builtin plugins, unlike "-u NONE -i NONE".
--noplugin --noplugin Skip loading plugins. Resets the 'loadplugins' option. Note that the -u argument may also disable loading plugins:
argument load vimrc files load plugins
(nothing) yes yes -u NONE no no -u NORC no yes --noplugin yes no
--startuptime {fname} --startuptime During startup write timing messages to the file {fname}. This can be used to find out where time is spent while loading your config, plugins and opening the first file. When {fname} already exists new messages are appended.
-+ +[num] The cursor will be positioned on line "num" for the first file being edited. If "num" is missing, the cursor will be positioned on the last line.
-+/ +/{pat} The cursor will be positioned on the first line containing "pat" in the first file being edited (see pattern for the available search patterns). The search starts at the cursor position, which can be the first line or the cursor position last used from shada. To force a search from the first line use "+1 +/pat".
+{command} -+c -c -c {command} {command} will be executed after the first file has been read (and after autocommands and modelines for that file have been processed). "command" is interpreted as an Ex command. If the "command" contains spaces, it must be enclosed in double quotes (this depends on the shell that is used). Example:
vim  "+set si"  main.c
vim  "+find stdio.h"
vim  -c "set ff=dos"  -c wq  mine.mak
Note: You can use up to 10 "+" or "-c" arguments in a Vim command. They are executed in the order given. A "-S" argument counts as a "-c" argument as well.
--cmd {command} --cmd {command} will be executed before processing any vimrc file. Otherwise it acts like -c {command}. You can use up to 10 of these commands, independently from "-c" commands.
-S -S [file] Executes Vimscript or Lua (".lua") [file] after the first file has been read. See also :source. If [file] is not given, defaults to "Session.vim". Equivalent to:
-c "source {file}"
Can be repeated like "-c", subject to the same limit of 10 "-c" arguments. {file} cannot start with a "-".
-L -L -r -r Recovery mode. Without a file name argument, a list of existing swap files is given. With a file name, a swap file is read to recover a crashed editing session. See crash-recovery.
-R -R Readonly mode. The 'readonly' option will be set for all the files being edited. You can still edit the buffer, but will be prevented from accidentally overwriting a file. If you forgot that you are in View mode and did make some changes, you can overwrite a file by adding an exclamation mark to the Ex command, as in ":w!". The 'readonly' option can be reset with ":set noro" (see the options chapter, options). Subsequent edits will not be done in readonly mode. Calling the executable "view" has the same effect as the -R argument. The 'updatecount' option will be set to 10000, meaning that the swap file will not be updated automatically very often. See -M for disallowing modifications.
-m -m Modifications not allowed to be written. The 'write' option will be reset, so that writing files is disabled. However, the 'write' option can be set to enable writing again.
-M -M Modifications not allowed. The 'modifiable' option will be reset, so that changes are not allowed. The 'write' option will be reset, so that writing files is disabled. However, the 'modifiable' and 'write' options can be set to enable changes and writing.
-e -e -E -E Start Nvim in Ex mode gQ, see Ex-mode.
If stdin is not a TTY: -e reads/executes stdin as Ex commands. -E reads stdin as text (into buffer 1).
-es -es -Es -s-ex silent-mode -Es Script mode, aka "silent mode", aka "batch mode". No UI, disables most prompts and messages. Unrelated to -s. See also -S to run script files.
-es reads/executes stdin as Ex commands.
printf "put ='foo'\n%%print\n" | nvim -es
-Es reads stdin as text (into buffer 1). Use -c or "+" to send commands.
printf "foo\n" | nvim -Es +"%print"
These commands display on stdout: :list :number :print :set With :verbose or 'verbose', other commands display on stderr:
nvim -es +":verbose echo 'foo'"
nvim -V1 -es +foo
User config is skipped unless -u was given. Swap file is skipped (like -n). User shada is loaded (unless "-i NONE" is given).
-l -l {script} [args] Executes Lua {script} non-interactively (no UI) with optional [args] after processing any preceding Nvim cli-arguments, then exits. Exits 1 on Lua error. See -S to run multiple Lua scripts without args, with a UI. lua-args All [args] are treated as {script} arguments and stored in the Lua _G.arg global table, thus "-l" ends processing of Nvim arguments. The {script} name is stored at _G.arg[0].
Sets 'verbose' to 1 (like "-V1"), so Lua print() writes to output.
Arguments before "-l" are processed before executing {script}. This example quits before executing "foo.lua":
nvim +q -l foo.lua
This loads Lua module "bar" before executing "foo.lua":
nvim +"lua require('bar')" -l foo.lua
Skips user config unless -u was given. Disables plugins unless 'loadplugins' was set. Disables shada unless -i was given. Disables swapfile (like -n).
-b -b Binary mode. File I/O will only recognize <NL> to separate lines. The 'expandtab' option will be reset. The 'textwidth' option is set to 0. 'modeline' is reset. The 'binary' option is set. This is done after reading the vimrc but before reading any file in the arglist. See also edit-binary.
-A -A Arabic mode. Sets the 'arabic' option on.
-H -H Hebrew mode. Sets the 'hkmap' and 'rightleft' options on.
-V verbose -V[N] Verbose. Sets the 'verbose' option to [N] (default: 10). Messages will be given for each file that is ":source"d and for reading or writing a ShaDa file. Can be used to find out what is happening upon startup and exit. Example:
nvim -V8
-V[N]{file} Like -V and sets 'verbosefile' to {file} (must not start with a digit). Messages are not displayed, instead they are written to {file}. Example:
nvim -V20vimlog
-D -D Debugging. Go to debugging mode when executing the first command from a script. debug-mode
-n -n No swap-file will be used. Recovery after a crash will be impossible. Handy if you want to view or edit a file on a very slow medium (e.g., a floppy). Can also be done with ":set updatecount=0". You can switch it on again by setting the 'updatecount' option to some value, e.g., ":set uc=100". 'updatecount' is set to 0 AFTER executing commands from a vimrc file, but before the GUI initializations. Thus it overrides a setting for 'updatecount' in a vimrc file, but not in a gvimrc file. See startup. When you want to reduce accesses to the disk (e.g., for a laptop), don't use "-n", but set 'updatetime' and 'updatecount' to very big numbers, and type ":preserve" when you want to save your work. This way you keep the possibility for crash recovery.
-o -o[N] Open N windows, split horizontally. If [N] is not given, one window is opened for every file given as argument. If there is not enough room, only the first few files get a window. If there are more windows than arguments, the last few windows will be editing an empty file.
-O -O[N] Open N windows, split vertically. Otherwise it's like -o. If both the -o and the -O option are given, the last one on the command line determines how the windows will be split.
-p -p[N] Open N tab pages. If [N] is not given, one tab page is opened for every file given as argument. The maximum is set with 'tabpagemax' pages (default 50). If there are more tab pages than arguments, the last few tab pages will be editing an empty file. Also see tabpage. -d -d Start in diff-mode.
-u E282 -u {vimrc} The file {vimrc} is read for initializations. Most other initializations are skipped; see initialization.
This can be used to start Vim in a special mode, with special mappings and settings. A shell alias can be used to make this easy to use. For example:
alias vimc vim -u ~/.config/nvim/c_init.vim !*
Also consider using autocommands; see autocommand.
When {vimrc} is "NONE" (all uppercase), all initializations from files and environment variables are skipped. Plugins and syntax highlighting are also skipped.
When {vimrc} is "NORC" (all uppercase), this has the same effect as "NONE", but plugins and syntax highlighting are not skipped.
-i -i {shada} The file {shada} is used instead of the default ShaDa file. If the name "NONE" is used (all uppercase), no ShaDa file is read or written, even if 'shada' is set or when ":rsh" or ":wsh" are used. See also shada-file.
-s -s {scriptin} Read script file {scriptin}, interpreting characters as Normal-mode input. The same can be done with ":source!":
:source! {scriptin}
Reads from stdin if {scriptin} is "-":
echo "ifoo" | nvim -s -
If the end of the file is reached before Nvim exits, further characters are read from the keyboard.
Does not work with -es. See also complex-repeat.
-w_nr -w {number} -w{number} Set the 'window' option to {number}.
-w -w {scriptout} All keys that you type are recorded in the file "scriptout", until you exit Vim. Useful to create a script file to be used with "vim -s" or ":source!". Appends to the "scriptout" file if it already exists. {scriptout} cannot start with a digit. See also vim.on_key(). See also complex-repeat.
-W -W {scriptout} Like -w, but do not append, overwrite an existing file.
--api-info --api-info Print msgpack-encoded api-metadata and exit.
--embed --embed Use stdin/stdout as a msgpack-RPC channel, so applications can embed and control Nvim via the RPC API.
Waits for the client ("embedder") to call nvim_ui_attach() before sourcing startup files and reading buffers, so that UIs can deterministically handle (display) early messages, dialogs, etc. The client can do other requests before nvim_ui_attach (e.g. nvim_get_api_info for feature-detection). During this pre-startup phase the user config is of course not available (similar to --cmd).
Embedders _not_ using the UI protocol must pass --headless:
nvim --embed --headless
Then startup will continue without waiting for nvim_ui_attach. This is equivalent to:
nvim --headless --cmd "call stdioopen({'rpc': v:true})"
--headless --headless Start without UI, and do not wait for nvim_ui_attach. The builtin TUI is not used, so stdio works as an arbitrary communication channel. channel-stdio
Also useful for scripting (tests) to see messages that would not be printed by -es.
To detect if a UI is available, check if nvim_list_uis() is empty during or after VimEnter.
To read stdin as text, "-" must be given explicitly: --headless cannot assume that stdin is just text.
echo foo | nvim --headless +"%print" +"q!" -
See also --embed. See also -es, which also disables most messages.
--listen {addr} --listen Start RPC server on pipe or TCP address {addr}. Sets the primary listen address v:servername to {addr}. serverstart()

Initialization initialization startup

At startup, Nvim checks environment variables and files and sets values accordingly, proceeding as follows:
1. Set the 'shell' option SHELL COMSPEC The environment variable SHELL, if it exists, is used to set the 'shell' option. On Win32, the COMSPEC variable is used if SHELL is not set.
2. Process the arguments The options and file names from the command that start Vim are inspected. Buffers are created for all files (but not loaded yet). The -V argument can be used to display or log what happens next, useful for debugging the initializations.
3. Start a server (unless --listen was given) and set v:servername.
4. Wait for UI to connect. Nvim started with --embed waits for the UI to connect before proceeding to load user configuration.
6. Enable filetype and indent plugins. This does the same as the command:
:runtime! ftplugin.vim indent.vim
Skipped if the "-u NONE" command line argument was given.
7. Load user config (execute Ex commands from files, environment, …). $VIMINIT environment variable is read as one Ex command line (separate multiple commands with '|' or <NL>). config init.vim init.lua vimrc exrc A file containing initialization commands is generically called a "vimrc" or config file. It can be either Vimscript ("init.vim") or Lua ("init.lua"), but not both. E5422 See also vimrc-intro and base-directories.
The config file is located at: Unix ~/.config/nvim/init.vim (or init.lua) Windows ~/AppData/Local/nvim/init.vim (or init.lua) $XDG_CONFIG_HOME $XDG_CONFIG_HOME/nvim/init.vim (or init.lua)
If Nvim was started with "-u {file}" then {file} is used as the config and all initializations until 5. are skipped. $MYVIMRC is not set. "nvim -u NORC" can be used to skip these initializations without reading a file. "nvim -u NONE" also skips plugins and syntax highlighting. -u
If Nvim was started with -es all initializations until 5. are skipped. system-vimrc sysinit.vim a. The system vimrc file is read for initializations. If nvim/sysinit.vim file exists in one of $XDG_CONFIG_DIRS, it will be used. Otherwise the system vimrc file is used. The path of this file is given by the :version command. Usually it's "$VIM/sysinit.vim".
VIMINIT EXINIT $MYVIMRC b. Locations searched for initializations, in order of preference:
$VIMINIT environment variable (Ex command line).
User config: $XDG_CONFIG_HOME/nvim/init.vim (or init.lua).
Other config: {dir}/nvim/init.vim (or init.lua) where {dir} is any directory in $XDG_CONFIG_DIRS.
$EXINIT environment variable (Ex command line). $MYVIMRC is set to the first valid location unless it was already set or when using $VIMINIT.
c. If the 'exrc' option is on (which is NOT the default), the current directory is searched for the following files, in order of precedence:
".nvim.lua"
".nvimrc"
".exrc" The first that exists is used, the others are ignored.
8. Enable filetype detection. This does the same as the command:
:runtime! filetype.lua
Skipped if ":filetype off" was called or if the "-u NONE" command line argument was given.
9. Enable syntax highlighting. This does the same as the command:
:runtime! syntax/syntax.vim
Skipped if ":syntax off" was called or if the "-u NONE" command line argument was given.
10. Load the plugin scripts. load-plugins This does the same as the command:
:runtime! plugin/**/*.vim
:runtime! plugin/**/*.lua
The result is that all directories in 'runtimepath' will be searched for the "plugin" sub-directory and all files ending in ".vim" or ".lua" will be sourced (in alphabetical order per directory), also in subdirectories. First "*.vim" are sourced, then "*.lua" files.
However, directories in 'runtimepath' ending in "after" are skipped here and only loaded after packages, see below. Loading plugins won't be done when:
The 'loadplugins' option was reset in a vimrc file.
The --noplugin command line argument is used.
The --clean command line argument is used.
The "-u NONE" command line argument is used -u. Note that using -c 'set noloadplugins' doesn't work, because the commands from the command line have not been executed yet. You can use --cmd 'set noloadplugins' or --cmd 'set loadplugins' --cmd.
Packages are loaded. These are plugins, as above, but found in the "start" directory of each entry in 'packpath'. Every plugin directory found is added in 'runtimepath' and then the plugins are sourced. See packages.
The plugins scripts are loaded, as above, but now only the directories ending in "after" are used. Note that 'runtimepath' will have changed if packages have been found, but that should not add a directory ending in "after".
11. Set 'shellpipe' and 'shellredir' The 'shellpipe' and 'shellredir' options are set according to the value of the 'shell' option, unless they have been set before. This means that Nvim will figure out the values of 'shellpipe' and 'shellredir' for you, unless you have set them yourself.
12. Set 'updatecount' to zero, if "-n" command argument used
13. Set binary options if the -b flag was given.
14. Read the shada-file.
15. Read the quickfix file if the -q flag was given, or exit on failure.
16. Open all windows When the -o flag was given, windows will be opened (but not displayed yet). When the -p flag was given, tab pages will be created (but not displayed yet). When switching screens, it happens now. Redrawing starts. If the -q flag was given, the first error is jumped to. Buffers for all windows will be loaded, without triggering BufAdd autocommands.
17. Execute startup commands If a -t flag was given, the tag is jumped to. Commands given with -c and +cmd are executed. The starting flag is reset, has("vim_starting") will now return zero. The v:vim_did_enter variable is set to 1. The VimEnter autocommands are executed.
Saving the current state of Vim to a file
Whenever you have changed values of options or when you have created a mapping, then you may want to save them in a vimrc file for later use. See save-settings about saving the current state of settings to a file.
Avoiding trojan horses
trojan-horse While reading the "vimrc" or the "exrc" file in the current directory, some commands can be disabled for security reasons by setting the 'secure' option. This is always done when executing the command from a tags file. Otherwise it would be possible that you accidentally use a vimrc or tags file that somebody else created and contains nasty commands. The disabled commands are the ones that start a shell, the ones that write to a file, and ":autocmd". The ":map" commands are echoed, so you can see which keys are being mapped. If you want Vim to execute all commands in a local vimrc file, you can reset the 'secure' option in the EXINIT or VIMINIT environment variable or in the global exrc or vimrc file. This is not possible in vimrc or exrc in the current directory, for obvious reasons. On Unix systems, this only happens if you are not the owner of the vimrc file. Warning: If you unpack an archive that contains a vimrc or exrc file, it will be owned by you. You won't have the security protection. Check the vimrc file before you start Vim in that directory, or reset the 'exrc' option. Some Unix systems allow a user to do "chown" on a file. This makes it possible for another user to create a nasty vimrc and make you the owner. Be careful! When using tag search commands, executing the search command (the last part of the line in the tags file) is always done in secure mode. This works just like executing a command from a vimrc in the current directory.
If Vim startup is slow
slow-start If Vim takes a long time to start up, use the --startuptime argument to find out what happens.
If you have 'shada' enabled, the loading of the ShaDa file may take a while. You can find out if this is the problem by disabling ShaDa for a moment (use the Vim argument "-i NONE", -i). Try reducing the number of lines stored in a register with ":set shada='20,<50,s10". shada-file.
Troubleshooting broken configurations
bisect The extreme flexibility of editors like Vim and Emacs means that any plugin or setting can affect the entire editor in ways that are not initially obvious.
To find the cause of a problem in your config, you must "bisect" it: 1. Remove or disable half of your config. 2. Restart Nvim. 3. If the problem still occurs, goto 1. 4. If the problem is gone, restore half of the removed lines. 5. Continue narrowing your config in this way, until you find the setting or plugin causing the issue.
Intro message
:intro When Vim starts without a file name, an introductory message is displayed. It is removed as soon as the display is redrawn. To see the message again, use the ":intro" command. To avoid the intro message on startup, add the "I" flag to 'shortmess'.

$VIM and $VIMRUNTIME

$VIM The environment variable "$VIM" is used to locate various user files for Nvim, such as the user config. This depends on the system, see startup.
Nvim will try to get the value for $VIM in this order:
1. Environment variable $VIM, if it is set. 2. Path derived from the 'helpfile' option, unless it contains some environment variable too (default is "$VIMRUNTIME/doc/help.txt"). File name ("help.txt", etc.) is removed. Trailing directory names are removed, in this order: "doc", "runtime". 3. Path derived from the location of the nvim executable. 4. Compile-time defined installation directory (see output of ":version").
After doing this once, Nvim sets the $VIM environment variable.
$VIMRUNTIME The environment variable "$VIMRUNTIME" is used to locate various support files, such as the documentation and syntax-highlighting files. For example, the main help file is normally "$VIMRUNTIME/doc/help.txt".
Nvim will try to get the value for $VIMRUNTIME in this order:
1. Environment variable $VIMRUNTIME, if it is set. 2. Directory path "$VIM/vim{version}", if it exists, where {version} is the Vim version number without '-' or '.'. For example: "$VIM/vim82". 3. Directory path "$VIM/runtime", if it exists. 4. Value of $VIM environment variable. This is for backwards compatibility with older Vim versions. 5. If "../share/nvim/runtime" exists relative to v:progpath, it is used. 6. Path derived from the 'helpfile' option (if it doesn't contain '$') with "doc/help.txt" removed from the end.
After doing this once, Nvim sets the $VIMRUNTIME environment variable.
In case you need the value of $VIMRUNTIME in a shell (e.g., for a script that greps in the help files) you might be able to use this:
VIMRUNTIME="$(nvim --clean --headless --cmd 'echo $VIMRUNTIME|q')"

Suspending suspend

CTRL-Z v_CTRL-Z CTRL-Z Suspend Nvim, like ":stop". Works in Normal and in Visual mode. In Insert and Command-line mode, the CTRL-Z is inserted as a normal character. In Visual mode Nvim goes back to Normal mode.
:sus[pend][!] or :sus :suspend :st :stop :st[op][!] Suspend Nvim using OS "job control"; it will continue if you make it the foreground job again. Triggers VimSuspend before suspending and VimResume when resumed. If "!" is not given and 'autowrite' is set, every buffer with changes and a file name is written out. If "!" is given or 'autowrite' is not set, changed buffers are not written, don't forget to bring Nvim back to the foreground later!
In the GUI, suspending is implementation-defined.

Exiting exiting

There are several ways to exit Vim:
Close the last window with :quit. Only when there are no changes.
Close the last window with :quit!. Also when there are changes.
Close all windows with :qall. Only when there are no changes.
Close all windows with :qall!. Also when there are changes.
Use :cquit. Also when there are changes.
When using :cquit or when there was an error message Vim exits with exit code 1. Errors can be avoided by using :silent! or with :catch.

Saving settings save-settings

Mostly you will edit your vimrc files manually. This gives you the greatest flexibility. There are a few commands to generate a vimrc file automatically. You can use these files as they are, or copy/paste lines to include in another vimrc file.
:mk :mkexrc :mk[exrc] [file] Write current key mappings and changed options to [file] (default ".exrc" in the current directory), unless it already exists.
:mk[exrc]! [file] Always write current key mappings and changed options to [file] (default ".exrc" in the current directory).
:mkv :mkvi :mkvimrc :mkv[imrc][!] [file] Like ":mkexrc", but the default is ".nvimrc" in the current directory. The ":version" command is also written to the file.
These commands will write ":map" and ":set" commands to a file, in such a way that when these commands are executed, the current key mappings and options will be set to the same values. The options 'columns', 'endofline', 'fileformat', 'lines', 'modified', and 'scroll' are not included, because these are terminal or file dependent. Note that the options 'binary', 'paste' and 'readonly' are included, this might not always be what you want.
When special keys are used in mappings, The 'cpoptions' option will be temporarily set to its Vim default, to avoid the mappings to be misinterpreted. This makes the file incompatible with Vi, but makes sure it can be used with different terminals.
Only global mappings are stored, not mappings local to a buffer.
A common method is to use a default config file, make some modifications with ":map" and ":set" commands and write the modified file. First read the default vimrc in with a command like ":source ~piet/.vimrc.Cprogs", change the settings and then save them in the current directory with ":mkvimrc!". If you want to make this file your default config, move it to $XDG_CONFIG_HOME/nvim. You could also use autocommands autocommand and/or modelines modeline.
vimrc-option-example If you only want to add a single option setting to your vimrc, you can use these steps: 1. Edit your vimrc file with Vim. 2. Play with the option until it's right. E.g., try out different values for 'guifont'. 3. Append a line to set the value of the option, using the expression register '=' to enter the value. E.g., for the 'guifont' option:
o:set guifont=<C-R>=&guifont<CR><Esc>
[<C-R> is a CTRL-R, <CR> is a return, <Esc> is the escape key] You need to escape special characters, esp. spaces.

Views and Sessions views-sessions

This is introduced in sections 21.4 and 21.5 of the user manual.
View view-file A View is a collection of settings that apply to one window. You can save a View and when you restore it later, the text is displayed in the same way. The options and mappings in this window will also be restored, so that you can continue editing like when the View was saved.
Session session-file A Session keeps the Views for all windows, plus the global settings. You can save a Session and when you restore it later the window layout looks the same. You can use a Session to quickly switch between different projects, automatically loading the files you were last working on in that project.
Views and Sessions are a nice addition to ShaDa files, which are used to remember information for all Views and Sessions together shada-file.
You can quickly start editing with a previously saved View or Session with the -S argument:
vim -S Session.vim
:mks :mksession :mks[ession][!] [file] Write a Vim script that restores the current editing session. When [!] is included an existing file is overwritten. When [file] is omitted "Session.vim" is used.
The output of ":mksession" is like ":mkvimrc", but additional commands are added to the file. Which ones depends on the 'sessionoptions' option. The resulting file, when executed with a ":source" command: 1. Restores global mappings and options, if 'sessionoptions' contains "options". Script-local mappings will not be written. 2. Restores global variables that start with an uppercase letter and contain at least one lowercase letter, if 'sessionoptions' contains "globals". 3. Closes all windows in the current tab page, except the current one; closes all tab pages except the current one (this results in currently loaded buffers to be unloaded, some may become hidden if 'hidden' is set or otherwise specified); wipes out the current buffer, if it is empty and unnamed. 4. Restores the current directory if 'sessionoptions' contains "curdir", or sets the current directory to where the Session file is if 'sessionoptions' contains "sesdir". 5. Restores GUI Vim window position, if 'sessionoptions' contains "winpos". 6. Restores screen size, if 'sessionoptions' contains "resize". 7. Reloads the buffer list, with the last cursor positions. If 'sessionoptions' contains "buffers" then all buffers are restored, including hidden and unloaded buffers. Otherwise only buffers in windows are restored. 8. Restores all windows with the same layout. If 'sessionoptions' contains "help", help windows are restored. If 'sessionoptions' contains "blank", windows editing a buffer without a name will be restored. If 'sessionoptions' contains "winsize" and no (help/blank) windows were left out, the window sizes are restored (relative to the screen size). Otherwise, the windows are just given sensible sizes. 9. Restores the Views for all the windows, as with :mkview. But 'sessionoptions' is used instead of 'viewoptions'. 10. If a file exists with the same name as the Session file, but ending in "x.vim" (for eXtra), executes that as well. You can usex.vim files to specify additional settings and actions associated with a given Session, such as creating menu items in the GUI version.
After restoring the Session, the full filename of your current Session is available in the internal variable v:this_session. An example mapping:
:nmap <F2> :wa<Bar>exe "mksession! " .. v:this_session<CR>:so ~/sessions/
This saves the current Session, and starts off the command to load another.
A session includes all tab pages, unless "tabpages" was removed from 'sessionoptions'. tab-page
The SessionLoadPost autocmd event is triggered after a session file is loaded/sourced. SessionLoad-variable While the session file is loading the SessionLoad global variable is set to 1. Plugins can use this to postpone some work until the SessionLoadPost event is triggered.
:mkvie :mkview :mkvie[w][!] [file] Write a Vim script that restores the contents of the current window. When [!] is included an existing file is overwritten. When [file] is omitted or is a number from 1 to 9, a name is generated and 'viewdir' prepended. When the last path part of 'viewdir' does not exist, this directory is created. E.g., when 'viewdir' is "$VIM/vimfiles/view" then "view" is created in "$VIM/vimfiles". An existing file is always overwritten then. Use :loadview to load this view again. When [file] is the name of a file ('viewdir' is not used), a command to edit the file is added to the generated file.
The output of ":mkview" contains these items: 1. The argument list used in the window. When the global argument list is used it is reset to the global list. The index in the argument list is also restored. 2. The file being edited in the window. If there is no file, the window is made empty. 3. Restore mappings, abbreviations and options local to the window if 'viewoptions' contains "options" or "localoptions". For the options it restores only values that are local to the current buffer and values local to the window. When storing the view as part of a session and "options" is in 'sessionoptions', global values for local options will be stored too. 4. Restore folds when using manual folding and 'viewoptions' contains "folds". Restore manually opened and closed folds. 5. The scroll position and the cursor position in the file. Doesn't work very well when there are closed folds. 6. The local current directory, if it is different from the global current directory and 'viewoptions' contains "curdir".
Note that Views and Sessions are not perfect:
They don't restore everything. For example, defined functions, autocommands and ":syntax on" are not included. Things like register contents and command line history are in ShaDa, not in Sessions or Views.
Global option values are only set when they differ from the default value. When the current value is not the default value, loading a Session will not set it back to the default value. Local options will be set back to the default value though.
Existing mappings will be overwritten without warning. An existing mapping may cause an error for ambiguity.
When storing manual folds and when storing manually opened/closed folds, changes in the file between saving and loading the view will mess it up.
The Vim script is not very efficient. But still faster than typing the commands yourself!
:lo :loadview :lo[adview] [nr] Load the view for the current file. When [nr] is omitted, the view stored with ":mkview" is loaded. When [nr] is specified, the view stored with ":mkview [nr]" is loaded.
The combination of ":mkview" and ":loadview" can be used to store up to ten different views of a file. These are remembered in the directory specified with the 'viewdir' option. The views are stored using the file name. If a file is renamed or accessed through a (symbolic) link the view will not be found.
You might want to clean up your 'viewdir' directory now and then.
To automatically save and restore views for.c files:
au BufWinLeave *.c mkview
au BufWinEnter *.c silent! loadview

Shada ("shared data") file shada shada-file

If you exit Vim and later start it again, you would normally lose a lot of information. The ShaDa file can be used to remember that information, which enables you to continue where you left off. Its name is the abbreviation of SHAred DAta because it is used for sharing data between Neovim sessions.
This is introduced in section 21.3 of the user manual.
The ShaDa file is used to store:
The command line history.
The search string history.
The input-line history.
Contents of non-empty registers.
Marks for several files.
File marks, pointing to locations in files.
Last search/substitute pattern (for 'n' and '&').
The buffer list.
Global variables.
You could also use a Session file. The difference is that the ShaDa file does not depend on what you are working on. There normally is only one ShaDa file. Session files are used to save the state of a specific editing Session. You could have several Session files, one for each project you are working on. ShaDa and Session files together can be used to effectively enter Vim and directly start working in your desired setup. session-file
shada-read When Vim is started and the 'shada' option is non-empty, the contents of the ShaDa file are read and the info can be used in the appropriate places. The v:oldfiles variable is filled. The marks are not read in at startup (but file marks are). See initialization for how to set the 'shada' option upon startup.
shada-write When Vim exits and 'shada' is non-empty, the info is stored in the ShaDa file (it's actually merged with the existing one, if one exists shada-merging). The 'shada' option is a string containing information about what info should be stored, and contains limits on how much should be stored (see 'shada').
Notes for Unix:
The file protection for the ShaDa file will be set to prevent other users from being able to read it, because it may contain any text or commands that you have worked with.
If you want to share the ShaDa file with other users (e.g. when you "su" to another user), you can make the file writable for the group or everybody. Vim will preserve this when writing new ShaDa files. Be careful, don't allow just anybody to read and write your ShaDa file!
Vim will not overwrite a ShaDa file that is not writable by the current "real" user. This helps for when you did "su" to become root, but your $HOME is still set to a normal user's home directory. Otherwise Vim would create a ShaDa file owned by root that nobody else can read.
The ShaDa file cannot be a symbolic link. This is to avoid security issues.
Marks are stored for each file separately. When a file is read and 'shada' is non-empty, the marks for that file are read from the ShaDa file. NOTE: The marks are only written when exiting Vim, which is fine because marks are remembered for all the files you have opened in the current editing session, unless ":bdel" is used. If you want to save the marks for a file that you are about to abandon with ":bdel", use ":wsh". The '[' and ']' marks are not stored, but the '"' mark is. The '"' mark is very useful for jumping to the cursor position when the file was last exited. No marks are saved for files that start with any string given with the "r" flag in 'shada'. This can be used to avoid saving marks for files on removable media (for MS-Windows you would use "ra:,rb:"). The v:oldfiles variable is filled with the file names that the ShaDa file has marks for.
shada-file-marks Uppercase marks ('A to 'Z) are stored when writing the ShaDa file. The numbered marks ('0 to '9) are a bit special. When the ShaDa file is written (when exiting or with the :wshada command), '0 is set to the current cursor position and file. The old '0 is moved to '1, '1 to '2, etc. This resembles what happens with the "1 to "9 delete registers. If the current cursor position is already present in '0 to '9, it is moved to '0, to avoid having the same position twice. The result is that with "'0", you can jump back to the file and line where you exited Vim. To do that right away, try using this command:
vim -c "normal '0"
In a csh compatible shell you could make an alias for it:
alias lvim vim -c '"'normal "'"0'"'
For a bash-like shell:
alias lvim='vim -c "normal '\''0"'
Use the "r" flag in 'shada' to specify for which files no marks should be remembered.

MERGING shada-merging

When writing ShaDa files with :wshada without bang or at regular exit information in the existing ShaDa file is merged with information from current Neovim instance. For this purpose ShaDa files store timestamps associated with ShaDa entries. Specifically the following is being done:
1. History lines are merged, ordered by timestamp. Maximum amount of items in ShaDa file is defined by 'shada' option (shada-/, shada-:, [email protected], etc: one suboption for each character that represents history name (:history)). 2. Local marks and changes for files that were not opened by Neovim are copied to new ShaDa file. Marks for files that were opened by Neovim are merged, changes to files opened by Neovim are ignored. shada-' 3. Jump list is merged: jumps are ordered by timestamp, identical jumps (identical position AND timestamp) are squashed. 4. Search patterns and substitute strings are not merged: search pattern or substitute string which has greatest timestamp will be the only one copied to ShaDa file. 5. For each register entity with greatest timestamp is the only saved. shada-< 6. All saved variables are saved from current Neovim instance. Additionally existing variable values are copied, meaning that the only way to remove variable from a ShaDa file is either removing it by hand or disabling writing variables completely. shada-! 7. For each global mark entity with greatest timestamp is the only saved. 8. Buffer list and header are the only entries which are not merged in any fashion: the only header and buffer list present are the ones from the Neovim instance which was last writing the file. shada-%

COMPATIBILITY shada-compatibility

ShaDa files are forward and backward compatible. This means that
1. Entries which have unknown type (i.e. that hold unidentified data) are ignored when reading and blindly copied when writing. 2. Register entries with unknown register name are ignored when reading and blindly copied when writing. Limitation: only registers that use name with code in interval [1, 255] are supported. registers 3. Register entries with unknown register type are ignored when reading and merged as usual when writing. getregtype() 4. Local and global mark entries with unknown mark names are ignored when reading. When writing global mark entries are blindly copied and local mark entries are also blindly copied, but only if file they are attached to fits in the shada-' limit. Unknown local mark entry's timestamp is also taken into account when calculating which files exactly should fit into this limit. Limitation: only marks that use name with code in interval [1, 255] are supported. mark-motions 5. History entries with unknown history type are ignored when reading and blindly copied when writing. Limitation: there can be only up to 256 history types. history 6. Unknown keys found in register, local mark, global mark, change, jump and search pattern entries are saved internally and dumped when writing. Entries created during Neovim session never have such additions. 7. Additional elements found in replacement string and history entries are saved internally and dumped. Entries created during Neovim session never have such additions. 8. Additional elements found in variable entries are simply ignored when reading. When writing new variables they will be preserved during merging, but that's all. Variable values dumped from current Neovim session never have additional elements, even if variables themselves were obtained by reading ShaDa files.
"Blindly" here means that there will be no attempts to somehow merge them, even if other entries (with known name/type/etc) are merged. shada-merging

SHADA FILE NAME shada-file-name

Default name of the shada file is: Unix: "$XDG_STATE_HOME/nvim/shada/main.shada" Windows: "$XDG_STATE_HOME/nvim-data/shada/main.shada" See also base-directories.
To choose a different file name you can use:
The "n" flag in the 'shada' option.
The -i startup argument. "NONE" means no shada file is ever read or written. Also not for the commands below!
The 'shadafile' option. The value from the "-i" argument (if any) is stored in the 'shadafile' option.
For the commands below, another file name can be given, overriding the default and the name given with 'shada' or "-i" (unless it's NONE).

MANUALLY READING AND WRITING shada-read-write

Two commands can be used to read and write the ShaDa file manually. This can be used to exchange registers between two running Vim programs: First type ":wsh" in one and then ":rsh" in the other. Note that if the register already contained something, then ":rsh!" would be required. Also note however that this means everything will be overwritten with information from the first Vim, including the command line history, etc.
The ShaDa file itself can be edited by hand too, although we suggest you start with an existing one to get the format right. You need to understand MessagePack (or, more likely, find software that is able to use it) format to do this. This can be useful in order to create a second file, say "~/.my.shada" which could contain certain settings that you always want when you first start Neovim. For example, you can preload registers with particular data, or put certain commands in the command line history. A line in your config file like
:rshada! ~/.my.shada
can be used to load this information. You could even have different ShaDa files for different types of files (e.g., C code) and load them based on the file name, using the ":autocmd" command (see :autocmd). More information on ShaDa file format is contained in shada-format section.
E136 E929 shada-error-handling Some errors make Neovim leave temporary file named {basename}.tmp.X (X is any free letter from a to z) while normally it will create this file, write to it and then rename {basename}.tmp.X to {basename}. Such errors include:
Errors which make Neovim think that read file is not a ShaDa file at all: non-ShaDa files are not overwritten for safety reasons to avoid accidentally destroying an unrelated file. This could happen e.g. when typing "nvim -i file" in place of "nvim -R file" (yes, somebody did that at least with Vim). Such errors are listed at shada-critical-contents-errors.
If writing to the temporary file failed: e.g. because of the insufficient space left.
If renaming file failed: e.g. because of insufficient permissions.
If target ShaDa file has different from the Neovim instance's owners (user and group) and changing them failed. Unix-specific, applies only when Neovim was launched from root.
Do not forget to remove the temporary file or replace the target file with temporary one after getting one of the above errors or all attempts to create a ShaDa file may fail with E929. If you got one of them when using :wshada (and not when exiting Neovim: i.e. when you have Neovim session running) you have additional options:
First thing which you should consider if you got any error, except failure to write to the temporary file: remove existing file and replace it with the temporary file. Do it even if you have running Neovim instance.
Fix the permissions and/or file ownership, free some space and attempt to write again. Do not remove the existing file.
Use :wshada with bang. Does not help in case of permission error. If target file was actually the ShaDa file some information may be lost in this case. To make the matters slightly better use :rshada prior to writing, but this still will loose buffer-local marks and change list entries for any file which is not opened in the current Neovim instance.
Remove the target file from shell and use :wshada. Consequences are not different from using :wshada with bang, but "rm -f" works in some cases when you don't have write permissions.
:rsh :rshada E886 :rsh[ada][!] [file] Read from ShaDa file [file] (default: see above). If [!] is given, then any information that is already set (registers, marks, v:oldfiles, etc.) will be overwritten.
:wsh :wshada E137 :wsh[ada][!] [file] Write to ShaDa file [file] (default: see above). The information in the file is first read in to make a merge between old and new info. When [!] is used, the old information is not read first, only the internal info is written (also disables safety checks described in shada-error-handling). If 'shada' is empty, marks for up to 100 files will be written. When you get error "E929: All .tmp.X files exist, cannot write ShaDa file!" check that no old temp files were left behind (e.g. ~/.local/state/nvim/shada/main.shada.tmp*).
Note: Executing :wshada will reset all 'quote marks.
:o :ol :oldfiles :o[ldfiles] List the files that have marks stored in the ShaDa file. This list is read on startup and only changes afterwards with :rshada!. Also see v:oldfiles. The number can be used with c_#<. The output can be filtered with :filter, e.g.:
filter /\.vim/ oldfiles
The filtering happens on the file name.
:bro[wse] o[ldfiles][!] List file names as with :oldfiles, and then prompt for a number. When the number is valid that file from the list is edited. If you get the press-enter prompt you can press "q" and still get the prompt to enter a file number. Use ! to abandon a modified buffer. abandon

SHADA FILE FORMAT shada-format

ShaDa files are concats of MessagePack entries. Each entry is a concat of exactly four MessagePack objects:
1. First goes type of the entry. Object type must be an unsigned integer. Object type must not be equal to zero. 2. Second goes entry timestamp. It must also be an unsigned integer. 3. Third goes the length of the fourth entry. Unsigned integer as well, used for fast skipping without parsing. 4. Fourth is actual entry data. All currently used ShaDa entries use containers to hold data: either map or array. All string values in those containers are either binary (applies to filenames) or UTF-8, yet parser needs to expect that invalid bytes may be present in a UTF-8 string.
Exact format depends on the entry type:
Entry type (name) Entry data
1 (Header) Map containing data that describes the generator instance that wrote this ShaDa file. It is ignored when reading ShaDa files. Contains the following data:
Key Data
generator Binary, software used to generate ShaDa file. Is equal to "nvim" when ShaDa file was written by Neovim. version Binary, generator version. encoding Binary, effective 'encoding' value. max_kbyte Integer, effective shada-s limit value. pid Integer, instance process ID. * It is allowed to have any number of additional keys with any data. 2 (SearchPattern) Map containing data describing last used search or substitute pattern. Normally ShaDa file contains two such entries: one with "ss" key set to true (describes substitute pattern, see :substitute), and one set to false (describes search pattern, see search-commands). "su" key should be true on one of the entries. If key value is equal to default then it is normally not present. Keys:
Key Type Default Description
sm Boolean true Effective 'magic' value. sc Boolean false Effective 'smartcase' value. sl Boolean true True if search pattern comes with a line offset. See search-offset. se Boolean false True if search-offset requested to place cursor at (relative to) the end of the pattern. so Integer 0 Offset value. search-offset su Boolean false True if current entry was the last used search pattern. ss Boolean false True if current entry describes :substitute pattern. sh Boolean false True if v:hlsearch is on. With shada-h or 'nohlsearch' this key is always false. sp Binary N/A Actual pattern. Required. sb Boolean false True if search direction is backward. * any none Other keys are allowed for compatibility reasons, see shada-compatibility. 3 (SubString) Array containing last :substitute replacement string. Contains single entry: binary, replacement string used. More entries are allowed for compatibility reasons, see shada-compatibility. 4 (HistoryEntry) Array containing one entry from history. Should have two or three entries. First one is history type (unsigned integer), second is history line (binary), third is the separator character (unsigned integer, must be in interval [0, 255]). Third item is only valid for search history. Possible history types are listed in hist-names, here are the corresponding numbers: 0 - cmd, 1 - search, 2 - expr, 3 - input, 4 - debug. 5 (Register) Map describing one register (registers). If key value is equal to default then it is normally not present. Keys:
Key Type Def Description
rt UInteger 0 Register type:
No Description
0 charwise-register 1 linewise-register 2 blockwise-register rw UInteger 0 Register width. Only valid for blockwise-registers. rc Array of binary N/A Register contents. Each entry in the array represents its own line. NUL characters inside the line should be represented as NL according to NL-used-for-Nul. ru Boolean false Unnamed register. Whether the unnamed register had pointed to this register. n UInteger N/A Register name: character code in range [1, 255]. Example: quote0 register has name 48 (ASCII code for zero character).
any none Other keys are allowed for compatibility reasons, see shada-compatibility. 6 (Variable) Array containing two items: variable name (binary) and variable value (any object). Values are converted using the same code msgpackparse() uses when reading, msgpackdump() when writing, so there may appear msgpack-special-dicts. If there are more then two entries then the rest are ignored (shada-compatibility). 7 (GlobalMark) 8 (Jump) 10 (LocalMark) 11 (Change) Map containing some position description: Entry Position ~ GlobaMark Global mark position. 'A LocalMark Local mark position. 'a Jump One position from the jumplist. Change One position from the changelist.
Data contained in the map:
Key Type Default Description
l UInteger 1 Position line number. Must be greater then zero. c UInteger 0 Position column number. n UInteger 34 ('"') Mark name. Only valid for GlobalMark and LocalMark entries. f Binary N/A File name. Required. * any none Other keys are allowed for compatibility reasons, see shada-compatibility. 9 (BufferList) Array containing maps. Each map in the array represents one buffer. Possible keys:
Key Type Default Description
l UInteger 1 Position line number. Must be greater then zero. c UInteger 0 Position column number. f Binary N/A File name. Required. * any none Other keys are allowed for compatibility reasons, see shada-compatibility. * (Unknown) Any other entry type is allowed for compatibility reasons, see shada-compatibility.
E575 E576 Errors in ShaDa file may have two types: E575 used for all “logical” errors and E576 used for all “critical” errors. Critical errors trigger behaviour described in shada-error-handling when writing and skipping the rest of the file when reading and include: shada-critical-contents-errors
Any of first three MessagePack objects being not an unsigned integer.
Third object requesting amount of bytes greater then bytes left in the ShaDa file.
Entry with zero type. I.e. first object being equal to zero.
MessagePack parser failing to parse the entry data.
MessagePack parser consuming less or requesting greater bytes then described in the third object for parsing fourth object. I.e. when fourth object either contains more then one MessagePack object or it does not contain complete MessagePack object.

Standard Paths standard-path

Nvim stores configuration, data, and logs in standard locations. Plugins are strongly encouraged to follow this pattern also. Use stdpath() to get the paths.
base-directories xdg The "base" (root) directories conform to the XDG Base Directory Specification. https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html The $XDG_CONFIG_HOME, $XDG_DATA_HOME, $XDG_RUNTIME_DIR, and $XDG_STATE_HOME environment variables are used if defined, else default values (listed below) are used.
CONFIG DIRECTORY (DEFAULT)
$XDG_CONFIG_HOME Nvim: stdpath("config") Unix: ~/.config ~/.config/nvim Windows: ~/AppData/Local ~/AppData/Local/nvim
DATA DIRECTORY (DEFAULT)
$XDG_DATA_HOME Nvim: stdpath("data") Unix: ~/.local/share ~/.local/share/nvim Windows: ~/AppData/Local ~/AppData/Local/nvim-data
RUN DIRECTORY (DEFAULT)
$XDG_RUNTIME_DIR Nvim: stdpath("run") Unix: /tmp/nvim.user/xxx /tmp/nvim.user/xxx Windows: $TMP/nvim.user/xxx $TMP/nvim.user/xxx
STATE DIRECTORY (DEFAULT)
$XDG_STATE_HOME Nvim: stdpath("state") Unix: ~/.local/state ~/.local/state/nvim Windows: ~/AppData/Local ~/AppData/Local/nvim-data
Note: Throughout the user manual these defaults are used as placeholders, e.g. "~/.config" is understood to mean "$XDG_CONFIG_HOME or ~/.config".

LOG FILE $NVIM_LOG_FILE E5430

Besides 'debug' and 'verbose', Nvim keeps a general log file for internal debugging, plugins and RPC clients.
:echo $NVIM_LOG_FILE
By default, the file is located at stdpath("log")/log unless that path is inaccessible or if $NVIM_LOG_FILE was set before startup.
Main
Commands index
Quick reference