vis - Man Page

vis uses selections as core editing primitives. A selection is a non-empty, directed range with two endpoints called cursor and anchor. A selection can be anchored in which case the anchor remains fixed while only the position of the cursor is adjusted. For non-anchored selections both endpoints are updated. A singleton selection covers one character on which both cursor and anchor reside. There always exists a primary selection which remains visible (i.e. changes to its position will adjust the viewport).

vis employs the same modal editing approach as vi. It supports a ‘normal’, ‘operator pending’, ‘insert’, ‘replace’ and ‘visual’ (in both line and character wise variants) mode. The visual block and ex modes are deliberately not implemented, instead vis has built in support for multiple selections and an interactive variant of the structural regular expression based command language of sam.

In normal mode all selections are non-anchored and reduced to a single character.

vis uses an undo tree to keep track of text revisions. The u (undo) and ⟨C-r⟩ (redo) commands can be used to traverse the tree along the main branch. g+ and g- traverse the history in chronological order. The :earlier and :later commands provide means to restore the text to an arbitrary state.

A mark associates a symbolic name to a set of selections. A stored selection becomes invalid when its delimiting boundaries change in the underlying buffer. If said changes are later undone the mark becomes valid again. m sets a mark, M restores it. For example, 'am sets the mark a while 'aM restores it. Use of m or M without specifying a mark uses the default mark.

Available marks are:

''

default mark

'^

active selections when leaving visual mode

'a–'z

general purpose marks

No marks across files are supported. Marks are not preserved over editing sessions.

A per window, fixed sized file local jump list exists which stores marks (i.e. set of selections).

g<

jump backward

g>

jump forward

gs

save currently active selections

Registers are named lists of text. Uninitialized register slots default to the empty string. Available registers are:

""

default register

"a–"z

general purpose registers

"A–"Z

append to corresponding general purpose register

"*, "+

system clipboard integration via shell script vis-clipboard(1)

"0

yank register, most recently yanked range

"1–"9

"&

sub expression matches of most recent x or y command

"/

search register, most recently used search pattern

":

command register, most recently executed command

"_

black hole (/dev/null) register, ignore content is always empty

"#

selection number (readonly)

If no explicit register is specified the default register is used.

The general purpose registers "a–"z can be used to record macros. Use one of "A–"Z to append to an existing macro. q starts a recording, @ plays it back. @@ refers to the most recently recorded macro. @: repeats the last :-command. @/ is equivalent to n in normal mode. These operations always use the first register slot.

vis always assumes the input file to be UTF-8 encoded with \n line endings. If you wish to edit files with legacy encodings or non-Unix line endings, use iconv(1) and dos2unix(1) to convert them as needed. ⟨Tab⟩ can optionally be expanded to a configurable number of spaces (see Set Options).

The mouse is currently not used at all.

vis currently defers regular expression matching to the underlying C library. It uses what POSIX refers to as “Extended Regular Expressions” as described in regex(7). The anchors ^ and $ match the beginning / end of the range they are applied to. Additionally \n and \t may be used to refer to newlines and tabs, respectively. The . atom matches any character except newline. The empty regular expression stands for the last complete expression encountered.

An address identifies a substring (or range) in a file. In the following “character n” means the null string after the n-th character in the file, with 1 the first character in the file. “Line n” means the n-th match, starting at the beginning of the file, of the regular expression “.*\n?”.

All windows always have at least one current substring which is the default address. In sam this is referred to as dot. In vis multiple “dots” (or selections) can exist at the same time.

#n

The empty string after character n; #0 is the beginning of the file.

n

Line n.

/regexp/

?regexp?

The substring that matches the regular expression, found by looking towards the end (/) or beginning (?) of the file. The search does not wrap around when hitting the end (start) of the file.

0

The string before the first full line. This is not necessarily the null string; see + and - below.

$

The null string at the end of the file.

.

Dot, the current range.

'm

The mark m in the file.

In the following, a1 and a2 are addresses.

a1+a2

The address a2 evaluated starting at the end of a1.

a1-a2

The address a2 evaluated looking the reverse direction starting at the beginning of a1.

a1,a2

The substring from the beginning of a1 to the end of a2. If a1 is missing, 0 is substituted. If a2 is missing, $ is substituted.

a1;a2

Like a1,a2 but with a2 evaluated at the end of, and range set to, a1.

The operators + and - are high precedence, while , and ; are low precedence.

In both + and - forms, if a2 is a line or character address with a missing number, the number defaults to 1. If a1 is missing, . is substituted. If both a1 and a2 are present and distinguishable, + may be elided. a2 may be a regular expression; if it is delimited by ? characters, the effect of the + or - is reversed. The % sign is an alias for , and hence 0,$. It is an error for a compound address to represent a malformed substring.

In the following, text demarcated by slashes represents text delimited by any printable ASCII character except alphanumerics. Any number of trailing delimiters may be elided, with multiple elisions then representing null strings, but the first delimiter must always be present. In any delimited text, newline may not appear literally; \n and \t may be typed for newline and tab; \/ quotes the delimiter, here /. An ampersand & and \n, where n is a digit (1–9) are replaced by the corresponding register. Backslash is otherwise interpreted literally.

Most commands may be prefixed with an address to indicate their range of operation. If a command takes an address and none is supplied, a default address is used. In normal mode this equates to the character the selection is currently over. If only one selection exists x and y default to the whole file 0,$. In normal mode the write commands w and wq always apply to the whole file. Commands are executed once for every selection. In visual mode the commands are applied to every selection as if an implicit x command, matching the existing selections, was present.

In the description, “range” is used to represent whatever address is supplied.

Many commands create new selections as a side effect when issued from a visual mode. If so, it is always to the “result” of the change: the new text for an insertion, the empty string for a deletion, the command output of a filter etc. If after a successful command execution no selections remain, the editor will switch to normal mode, otherwise it remains in visual mode. This allows interactive refinements of ranges.

a[count]/text/

Insert the text count times into the file after the range.

May also be written as

 a
 lines
 of
 text
 .
    

c or i

Same as a, but c replaces the text, while i inserts before the range.

d

Delete the text in range.

p

Create a new selection for the range.

e[!] [file name]

Replace the file by the contents of the named external file. If no file name is given, reload file from disk.

r file name

Replace the text in the range by the contents of the named external file.

w[!] [file name]

Write the range (default 0,$) to the named external file.

wq[!] [file name]

Same as w, but close file afterwards.

If the file name argument is absent from any of these, the current file name is used. e always sets the file name, w will do so if the file has no name. Forcing the e command with ! will discard any unsaved changes. Forcing w will overwrite the file on disk even if it has been externally modified since loading it. Write commands with a non-default addresses and no file name are destructive and need always to be forced.

< shell command

Replace the range by the standard output of the shell command.

> shell command

Sends the range to the standard input of the shell command.

| shell command

Send the range to the standard input, and replace it by the standard output, of the shell command.

! shell command

Run interactive shell command, redirect keyboard input to it.

cd directory

Change working directory. If no directory is specified, $HOME is used.

In any of <, >, |, or !, if the shell command is omitted, the last shell command (of any type) is substituted. Unless the file being edited is unnamed, all these external commands can refer to its absolute path and file name through the vis_filepath and vis_filename environment variables.

x/regexp/ [command]

For each match of the regular expression in the range, run the command with range set to the match. If the regular expression and its slashes are omitted, /.*\n/ is assumed. Null string matches potentially occur before every character of the range and at the end of the range.

The "1–"9 and "& registers are updated with the (sub) expression matches of the pattern.

y/regexp/ [command]

Like x, but run the command for each substring that lies before, between, or after the matches that would be generated by x. There is no default behavior. Null substrings potentially occur before every character in the range.

X/regexp/ command

For each file whose file name matches the regular expression, make that the current file and run the command. If the expression is omitted, the command is run in every file.

Y/regexp/ command

Same as X, but for files that do not match the regular expression, and the expression is required.

g[count][/regexp/] command

v[count][/regexp/] command

If the count range contains (g) or does not contain (v) a match for the expression, run command on the range.

The count specifier has the following format, where n and m are integers denoting the ranges.

n,m

The closed interval from n to m. If n is missing, 1 is substituted. If m is missing, ∞ is substituted. Negative values are interpreted relative to the last range.

%n

Matches every n-th range.

These may be nested arbitrarily deeply. An empty command in an x or y defaults to p. X, Y, g and v do not have defaults.

Commands may be grouped by enclosing them in curly braces. Semantically, the opening brace is like a command: it takes an (optional) address and runs each sub-command on the range. Commands within the braces are executed sequentially, but changes made by one command are not visible to other commands.

When a command makes a number of changes to a file, as in x/re/ c/text/, the addresses of all changes are computed based on the initial state. If the changes are non-overlapping, they are applied in the specified order. Conflicting changes are rejected.

Braces may be nested arbitrarily.

Operators perform a certain operation on a text range indicated by either a motion, a text object or an existing selection.

When used in normal mode, the following operators wait for a motion, putting vis into operator pending mode.

c

change, delete range and enter insert mode

d

delete, cut range to register

<

shift-left, decrease indent

>

shift-right, increase indent

y

yank, copy range to register

When used in normal mode, the following actions take effect immediately.

=

format, filter range through fmt(1)

gu

make lowercase

gU

make uppercase

g~

swap case

J

join lines, insert spaces in between

gJ

join lines remove any delimiting white spaces

p

put register content after cursor

P

put register content before cursor

Motions take an initial file position and transform it to a destination file position, thereby defining a range.

0

start of line

b

previous start of a word

B

previous start of a WORD

$

end of line

e

next end of a word

E

next end of a WORD

F⟨char⟩

to next occurrence of ⟨char⟩ to the left

f⟨char⟩

to next occurrence of ⟨char⟩ to the right

^

first non-blank of line

g0

begin of display line

g$

end of display line

ge

previous end of a word

gE

previous end of a WORD

gg

begin of file

G

goto line or end of file

gj

display line down

gk

display line up

gh

codepoint left

gl

codepoint right

gH

byte left

gL

byte right

g_

last non-blank of line

gm

middle of display line

g|

goto column

h

char left

H

goto top/home line of window

j

line down

k

line up

l

char right

L

goto bottom/last line of window

%

match bracket, quote or backtick

}

next paragraph

)

next sentence

N

repeat last search backwards

n

repeat last search forward

[{

previous start of block

]}

next start of block

[(

previous start of parentheses pair

])

next start of parentheses pair

{

previous paragraph

(

previous sentence

;

repeat last to/till movement

,

repeat last to/till movement but in opposite direction

#

search word under selection backwards

*

search word under selection forwards

T⟨char⟩

till before next occurrence of ⟨char⟩ to the left

t⟨char⟩

till before next occurrence of ⟨char⟩ to the right

?pattern

to next match of pattern in backward direction

/pattern

to next match of pattern in forward direction

w

next start of a word

W

next start of a WORD

Text objects take an initial file position and transform it to a range where the former does not necessarily have to be contained in the latter. All of the following text objects are implemented in an inner variant (prefixed with i) where the surrounding white space or delimiting characters are not part of the resulting range and a normal variant (prefixed with a) where they are.

w

word

W

WORD

s

sentence

p

paragraph

[, ], (, ), {, }, <, >, ", ', `

block enclosed by these symbols

Further available text objects include:

gn

matches the last used search term in forward direction

gN

matches the last used search term in backward direction

al

current line

il

current line without leading and trailing white spaces

vis supports multiple selections with immediate visual feedback. There always exists one primary selection located within the current view port. Additional selections can be created as needed. If more than one selection exists, the primary one is styled differently.

To manipulate selections use in normal mode:

⟨C-k⟩

create count new selections on the lines above

⟨C-M-k⟩

create count new selections on the lines above the first selection

⟨C-j⟩

create count new selections on the lines below

⟨C-M-j⟩

create count new selections on the lines below the last selection

⟨C-p⟩

remove primary selection

⟨C-n⟩

select word the selection is currently over, switch to visual mode

⟨C-u⟩

make the count previous selection primary

⟨C-d⟩

make the count next selection primary

⟨C-c⟩

remove the count selection column

⟨C-l⟩

remove all but the count selection column

⟨Tab⟩

try to align all selections on the same column

⟨Escape⟩

dispose all but the primary selection

The visual modes were enhanced to recognize:

I

create a selection at the start of every selected line

A

create a selection at the end of every selected line

⟨Tab⟩

left align selections by inserting spaces

⟨S-Tab⟩

right align selections by inserting spaces

⟨C-a⟩

create new selections everywhere matching current word or selection

⟨C-n⟩

create new selection and select next word matching current selection

⟨C-x⟩

clear (skip) current selection, but select next matching word

⟨C-p⟩

remove primary selection

⟨C-u⟩ ⟨C-k⟩

make the count previous selection primary

⟨C-d⟩ ⟨C-j⟩

make the count next selection primary

⟨C-c⟩

remove the count selection column

⟨C-l⟩

remove all but the count selection column

+

rotate selections rightwards count times

-

rotate selections leftwards count times

_

trim selections, remove leading and trailing white space

o

flip selection direction, swap cursor and anchor

⟨Escape⟩

clear all selections, switch to normal mode

In insert and replace mode:

⟨S-Tab⟩

align all selections by inserting spaces

Selections can be manipulated using set operations. The first operand is the currently active selections while the second can be specified as a mark. For example, 'a| produces the union of the mark a and the current selection. Use of set operations without specifying a mark use the default mark as the first operand.

|

set union

&

set intersection

\

set minus

~

set complement

A file must be opened in at least one window. If the last window displaying a certain file is closed all unsaved changes are discarded. Windows are equally sized and can be displayed in rows (horizontally) or in columns (vertically). The ⟨C-w⟩ h, ⟨C-w⟩ j, ⟨C-w⟩ k and ⟨C-w⟩ l key mappings can be used to switch between windows.

:new

open an empty window, arranged as a new row (horizontally)

:vnew

open an empty window, arranged as a new column (vertically)

:open[!] [file name]

open a new window, displaying file name if given

:split [file name]

split window into rows (horizontally)

:vsplit [file name]

split window into columns (vertically)

:q[!] [exit code]

close currently focused window

:qall[!] [exit code]

close all windows, terminate editor with exit code (defaults to 0)

Commands taking a file name will invoke the vis-open(1) utility, if given a file pattern or directory.

vis supports global as well as window local run time key mappings which are always evaluated recursively.

:map[!] mode lhs rhs

add a global key mapping

:map-window[!] mode lhs rhs

add a window local key mapping

:unmap mode lhs

remove a global key mapping

:unmap-window mode lhs

remove a window local key mapping

In the above mode refers to one of ‘normal’, ‘insert’, ‘replace’, ‘visual’, ‘visual-line’ or ‘operator-pending’; lhs refers to the key to map and rhs is a key action or alias. An existing mapping may be overridden by forcing the map command by specifying !.

Because key mappings are always recursive, doing something like:

:map! normal j 2j

will not work because it would enter an endless loop. Instead, vis uses pseudo keys referred to as key actions which can be used to invoke a set of available editor functions. :help lists all currently active key bindings as well as all available symbolic keys.

In order to facilitate usage of non-latin keyboard layouts, vis allows one to map locale specific keys to their latin equivalents by means of the

:langmap locale-keys latin-keys

command. As an example, the following maps the movement keys in Russian layout:

:langmap ролд hjkl

If the key sequences have not the same length, the remainder of the longer sequence will be discarded.

The defined mappings take effect in all non-input modes, i.e. everywhere except in insert and replace mode.

:earlier [count]

revert to older text state

:later [count]

revert to newer text state

If count is suffixed by either of d (days), h (hours), m (minutes) or s (seconds) it is interpreted as an offset from the current system time and the closest available text state is restored.