Go to the previous, next chapter.
This chapter describes Vi commands other than Ex commands implemented in
VIP. Except for the last section which discusses insert mode, all the
commands described in this chapter are to be used in vi mode.
Most Vi commands accept a numeric argument which can be supplied as
a prefix to the commands. A numeric argument is also called a count.
In many cases, if a count is given, the command is executed that many times.
For instance, 5 d d deletes 5 lines while simple d d deletes a
line. In this manual the metavariable n will denote a count.
The keys C-g and C-l are unique in that their associated
functions are the same in any of emacs, vi and insert mode.
- C-g
-
Quit. Cancel running or partially typed command (
keyboard-quit
).
- C-l
-
Clear the screen and reprint everything (
recenter
).
In Emacs many commands are bound to the key strokes that start with
C-x, C-c and ESC. These commands can be
accessed from vi mode as easily as from emacs mode.
- C-x
- C-c
- ESC
-
Typing one of these keys have the same effect as typing it in emacs mode.
Appropriate command will be executed according as the keys you type after
it. You will be in vi mode again after the execution of the command.
For instance, if you type ESC < (in vi mode) then the cursor will
move to the beginning of the buffer and you will still be in vi mode.
- C
- X
-
Typing one of these keys have the effect of typing the corresponding
control character in emacs mode. Moreover, if you type an upper case
character following it, that character will also be translated to the
corresponding control character. Thus typing X W in vi mode is the
same as typing C-x C-w in emacs mode. You will be in vi mode again
after the execution of a command.
- \
-
Escape to emacs mode. Hitting the \ key will take you to emacs mode,
and you can execute a single Emacs command. After executing the
Emacs command you will be in vi mode again. You can give a count before
typing \. Thus 5 \ +, as well as \ C-u 5 +, will insert
`+++++' before point.
In Emacs the text you edit is stored in a buffer.
See GNU Emacs Manual, for details. There is always one selected
buffer which is called the current buffer.
You can see the contents of buffers through windows created by Emacs.
When you have multiple windows on the screen only one of them is selected.
Each buffer has a unique name, and each window has a mode line which shows
the name of the buffer associated with the window and other information
about the status of the buffer. You can change the format of the mode
line, but normally if you see `**' at the beginning of a mode line it
means that the buffer is modified. If you write out the content of
the buffer to a file, then the buffer will become not modified. Also if
you see `%%' at the beginning of the mode line, it means that the file
associated with the buffer is write protected.
We have the following commands related to windows and buffers.
- C-n
-
Move cursor to the next-window (
vip-next-window
).
- X 1
-
Delete other windows and make the selected window fill the screen
(
delete-other-windows
).
- X 2
-
Split current window into two windows (
split-window-vertically
).
- X 3
-
Show current buffer in two windows.
- s buffer RET
-
Select or create a buffer named buffer (
vip-switch-to-buffer
).
- S buffer RET
-
Similar but select a buffer named buffer in another window
(
vip-switch-to-buffer-other-window
).
- K
-
Kill the current buffer if it is not modified or if it is not associated
with a file (
vip-kill-buffer
).
- X B
-
List the existing buffers (
list-buffers
).
As buffer name completion is provided, you have only to type in
initial substring of the buffer name which is sufficient to identify it
among names of existing buffers. After that, if you hit TAB the rest
of the buffer name will be supplied by the system, and you can confirm it
by RET. The default buffer name to switch to will also be prompted,
and you can select it by giving a simple RET. See GNU Emacs Manual
for details of completion.
We have the following commands related to files. They are used to visit,
save and insert files.
- v file RET
-
Visit specified file in the current window (
vip-find-file
).
- V file RET
-
Visit specified file in another window (
vip-find-file-other-window
).
- X S
-
Save current buffer to the file associated with the buffer. If no file is
associated with the buffer, the name of the file to write out the content
of the buffer will be asked in the minibuffer.
- X W file RET
-
Write current buffer into a specified file.
- X I file RET
-
Insert a specified file at point.
- g
-
Give information on the file associated with the current buffer. Tell you
the name of the file associated with the buffer, the line number of the
current point and total line numbers in the buffer. If no file is
associated with the buffer, this fact will be indicated by the null file
name `""'.
In Emacs, you can edit a file by visiting it. If you wish to visit a
file in the current window, you can just type v. Emacs maintains the
default directory which is specific to each buffer. Suppose, for
instance, that the default directory of the current buffer is
`/usr/masahiko/lisp/'. Then you will get the following prompt in the
minibuffer.
visit file: /usr/masahiko/lisp/
If you wish to visit, say, `vip.el' in this directory, then you can
just type `vip.el' followed by RET. If the file `vip.el'
already exists in the directory, Emacs will visit that file, and if not,
the file will be created. Emacs will use the file name (`vip.el', in
this case) as the name of the buffer visiting the file. In order to make
the buffer name unique, Emacs may append `<2>', `<3>' etc., to
the buffer name. As the file name completion is provided here, you
can sometime save typing. For instance, suppose there is only one file in the
default directory whose name starts with `v', that is `vip.el'.
Then if you just type v TAB then it will be completed to
`vip.el'. Thus, in this case, you just have to type v v TAB
RET to visit `/usr/masahiko/lisp/vip.el'. Continuing the
example, let us now suppose that you wished to visit the file
`/usr/masahiko/man/vip.texinfo'. Then to the same prompt which you get
after you typed v, you can enter `/usr/masahiko/man/vip.texinfo' or
`../man/vip.texinfo' followed by RET.
Use V instead of v, if you wish to visit a file in another
window.
You can verify which file you are editing by typing g. (You can also
type X B to get nformation on other buffers too.) If you type
g you will get an information like below in the echo area:
"/usr/masahiko/man/vip.texinfo" line 921 of 1949
After you edited the buffer (`vip.texinfo', in our example) for a while,
you may wish to save it in a file. If you wish to save it in the file
associated with the buffer (`/usr/masahiko/man/vip.texinfo', in this
case), you can just say X S. If you wish to save it in another file,
you can type X W. You will then get a similar prompt as you get for
v, to which you can enter the file name.
In this and next section we discuss commands for moving around in the
buffer. These command do not change the content of the buffer. The
following commands are useful for viewing the content of the current
buffer.
- SPC
- C-f
-
Scroll text of current window upward almost full screen. You can go
forward in the buffer by this command (
vip-scroll
).
- RET
- C-b
-
Scroll text of current window downward almost full screen. You can go
backward in the buffer by this command (
vip-scroll-back
).
- C-d
-
Scroll text of current window upward half screen. You can go
down in the buffer by this command (
vip-scroll-down
).
- C-u
-
Scroll text of current window downward half screen. You can go
up in the buffer by this command (
vip-scroll-up
).
- C-y
-
Scroll text of current window upward by one line (
vip-scroll-down-one
).
- C-e
-
Scroll text of current window downward by one line (
vip-scroll-up-one
).
You can repeat these commands by giving a count. Thus, 2 SPC
has the same effect as SPC SPC.
The following commands reposition point in the window.
- z H
- z RET
-
Put point on the top (home) line in the window. So the current line
becomes the top line in the window. Given a count n, point will be
placed in the n-th line from top (
vip-line-to-top
).
- z M
- z .
-
Put point on the middle line in the window. Given a count n,
point will be placed in the n-th line from the middle line
(
vip-line-to-middle
).
- z L
- z -
-
Put point on the bottom line in the window. Given a count n,
point will be placed in the n-th line from bottom
(
vip-line-to-bottom
).
- C-l
- Center point in window and redisplay screen (
recenter
).
The following commands are used to mark positions in the buffer.
- m ch
-
Store current point in the register ch. ch must be a lower
case character between a and z.
- m <
- Set mark at the beginning of current buffer.
- m >
- Set mark at the end of current buffer.
- m .
- Set mark at point.
- m ,
- Jump to mark (and pop mark off the mark ring).
Emacs uses the mark ring to store marked positions. The commands
m <, m > and m . not only set mark but also add it as the
latest element of the mark ring (replacing the oldest one). By repeating
the command `m ,' you can visit older and older marked positions. You
will eventually be in a loop as the mark ring is a ring.
Commands for moving around in the current buffer are collected here. These
commands are used as an `argument' for the delete, change and yank commands
to be described in the next section.
- h
-
Move point backward by one character. Signal error if point is at the
beginning of buffer, but (unlike Vi) do not complain otherwise
(
vip-backward-char
).
- l
-
Move point backward by one character. Signal error if point is at the
end of buffer, but (unlike Vi) do not complain otherwise
(
vip-forward-char
).
- j
-
Move point to the next line keeping the current column. If point is on the
last line of the buffer, a new line will be created and point will move to
that line (
vip-next-line
).
- k
-
Move point to the previous line keeping the current column
(
vip-next-line
).
- +
-
Move point to the next line at the first non-white character. If point is
on the last line of the buffer, a new line will be created and point will
move to the beginning of that line (
vip-next-line-at-bol
).
- -
-
Move point to the previous line at the first non-white character
(
vip-previous-line-at-bol
).
If a count is given to these commands, the commands will be repeated that
many times.
- 0
-
Move point to the beginning of line (
vip-beginning-of-line
).
- ^
-
Move point to the first non-white character on the line
(
vip-bol-and-skip-white
).
- $
-
Move point to the end of line (
vip-goto-eol
).
- n |
-
Move point to the n-th column on the line (
vip-goto-col
).
Except for the | command, these commands neglect a count.
- w
-
Move point forward to the beginning of the next word
(
vip-forward-word
).
- W
-
Move point forward to the beginning of the next word, where a word is
considered as a sequence of non-white characters (
vip-forward-Word
).
- b
-
Move point backward to the beginning of a word (
vip-backward-word
).
- B
-
Move point backward to the beginning of a word, where a word is
considered as a sequence of non-white characters (
vip-forward-Word
).
- e
-
Move point forward to the end of a word (
vip-end-of-word
).
- E
-
Move point forward to the end of a word, where a word is
considered as a sequence of non-white characters (
vip-end-of-Word
).
Here the meaning of the word `word' for the w, b and e
commands is determined by the syntax table effective in the current
buffer. Each major mode has its syntax mode, and therefore the meaning of
a word also changes as the major mode changes. See GNU Emacs Manual for
details of syntax table.
- H
-
Move point to the beginning of the home (top) line of the window.
Given a count n, go to the n-th line from top
(
vip-window-top
).
- M
-
Move point to the beginning of the middle line of the window. Given
a count n, go to the n-th line from the middle line
(
vip-window-middle
).
- L
-
Move point to the beginning of the lowest (bottom) line of the
window. Given count, go to the n-th line from bottom
(
vip-window-bottom
).
These commands can be used to go to the desired line visible on the screen.
- (
-
Move point backward to the beginning of the sentence
(
vip-backward-sentence
).
- )
-
Move point forward to the end of the sentence
(
vip-forward-sentence
).
- {
-
Move point backward to the beginning of the paragraph
(
vip-backward-paragraph
).
- }
-
Move point forward to the end of the paragraph
(
vip-forward-paragraph
).
A count repeats the effect for these commands.
- G
-
Given a count n, move point to the n-th line in the buffer on
the first non-white character. Without a count, go to the end of the buffer
(
vip-goto-line
).
- ` `
-
Exchange point and mark (
vip-goto-mark
).
- ` ch
- Move point to the position stored in the register ch. ch must
be a lower case letter.
- ' '
-
Exchange point and mark, and then move point to the first non-white
character on the line (
vip-goto-mark-and-skip-white
).
- ' ch
- Move point to the position stored in the register ch and skip to the
first non-white character on the line. ch must be a lower case letter.
- %
-
Move point to the matching parenthesis if point is looking at (,
), {, 2, [ or ]
(
vip-paren-match
).
The command G mark point before move, so that you can return to the
original point by ` `. The original point will also be stored in
the mark ring.
The following commands are useful for moving points on the line. A count
will repeat the effect.
- f ch
-
Move point forward to the character ch on the line. Signal error if
ch could not be found (
vip-find-char-forward
).
- F ch
-
Move point backward to the character ch on the line. Signal error if
ch could not be found (
vip-find-char-backward
).
- t ch
-
Move point forward upto the character ch on the line. Signal error if
ch could not be found (
vip-goto-char-forward
).
- T ch
-
Move point backward upto the character ch on the line. Signal error if
ch could not be found (
vip-goto-char-backward
).
- ;
-
Repeat previous f, t, F or T command
(
vip-repeat-find
).
- ,
-
Repeat previous f, t, F or T command, in the
opposite direction (
vip-repeat-find-opposite
).
Following commands are available for searching and replacing.
- / string RET
-
Search the first occurrence of the string string forward starting
from point. Given a count n, the n-th occurrence of
string will be searched. If the variable
vip-re-search
has value
t
then regular expression search is done and the string
matching the regular expression string is found. If you give an
empty string as string then the search mode will change from vanilla
search to regular expression search and vice versa
(vip-search-forward
).
- ? string RET
-
Same as /, except that search is done backward
(
vip-search-backward
).
- n
-
Search the previous search pattern in the same direction as before
(
vip-search-next
).
- N
-
Search the previous search pattern in the opposite direction
(
vip-search-Next
).
- C-s
-
Search forward incrementally. See GNU Emacs Manual for details
(
isearch-forward
).
- C-r
-
Search backward incrementally (
isearch-backward
).
- R string RET newstring
-
There are two modes of replacement, vanilla and regular expression.
If the mode is vanilla you will get a prompt `Replace string:',
and if the mode is regular expression you will ge a prompt
`Replace regexp:'. The mode is initially vanilla, but you can
toggle these modes by giving a null string as string. If the mode is
vanilla, this command replaces every occurrence of string with
newstring. If the mode is regular expression, string is
treated as a regular expression and every string matching the regular
expression is replaced with newstring (
vip-replace-string
).
- Q string RET newstring
-
Same as R except that you will be asked form confirmation before each
replacement
(
vip-query-replace
).
- r ch
-
Replace the character point is looking at by the character ch. Give
count, replace that many characters by ch (
vip-replace-char
).
The commands / and ? mark point before move, so that you can
return to the original point by ` `.
In this section, commands for modifying the content of a buffer are
described. These commands affect the region determined by a motion command
which is given to the commands as their argument.
We classify motion commands into point commands and
line commands. The point commands are as follows:
h, l, 0, ^, $, w, W, b, B, e, E, (, ), /, ?, `, f, F, t, T, %, ;, ,
The line commands are as follows:
j, k, +, -, H, M, L, {, 2, G, '
If a point command is given as an argument to a modifying command, the
region determined by the point command will be affected by the modifying
command. On the other hand, if a line command is given as an argument to a
modifying command, the region determined by the line command will be
enlarged so that it will become the smallest region properly containing the
region and consisting of whole lines (we call this process expanding
the region), and then the enlarged region will be affected by the modifying
command.
- d motion-command
-
Delete the region determined by the motion command motion-command.
For example, d $ will delete the region between point and end of
current line since $ is a point command that moves point to end of line.
d G will delete the region between the beginning of current line and
end of the buffer, since G is a line command. A count given to the
command above will become the count for the associated motion command.
Thus, 3 d w will delete three words.
It is also possible to save the deleted text into a register you specify.
For example, you can say " t 3 d w to delete three words and save it
to register t. The name of a register is a lower case letter between
a and z. If you give an upper case letter as an argument to
a delete command, then the deleted text will be appended to the content of
the register having the corresponding lower case letter as its name. So,
" T d w will delete a word and append it to register t. Other
modifying commands also accept a register name as their argument, and we
will not repeat similar explanations.
We have more delete commands as below.
- d d
-
Delete a line. Given a count n, delete n lines.
- d r
-
Delete current region.
- d R
-
Expand current region and delete it.
- D
-
Delete to the end of a line (
vip-kill-line
).
- x
-
Delete a character after point. Given n, delete n characters
(
vip-delete-char
).
- DEL
-
Delete a character before point. Given n, delete n characters
(
vip-delete-backward-char
).
Yank commands yank a text of buffer into a (usually anonymous) register.
Here the word `yank' is used in Vi's sense. Thus yank commands do not
alter the content of the buffer, and useful only in combination with
commands that put back the yanked text into the buffer.
- y motion-command
-
Yank the region determined by the motion command motion-command.
For example, y $ will yank the text between point and the end of line
into an anonymous register, while "c y $ will yank the same text into
register c.
Use the following command to yank consecutive lines of text.
- y y
- Y
-
Yank a line. Given n, yank n lines (
vip-yank-line
).
- y r
-
Yank current region.
- y R
-
Expand current region and yank it.
Deleted or yanked texts can be put back into the buffer by the command
below.
- p
-
Insert, after the character point is looking at, most recently
deleted/yanked text from anonymous register. Given a register name
argument, the content of the named register will be put back. Given a
count, the command will be repeated that many times. This command also
checks if the text to put back ends with a new line character, and if so
the text will be put below the current line (
vip-put-back
).
- P
-
Insert at point most recently deleted/yanked text from anonymous register.
Given a register name argument, the content of the named register will
be put back. Given a count, the command will be repeated that many times.
This command also checks if the text to put back ends with a new line
character, and if so the text will be put above the current line rather
than at point (
vip-Put-back
).
Thus, " c p will put back the content of the register c into the
buffer. It is also possible to specify number register which is a
numeral between 1 and 9. If the number register n is
specified, n-th previously deleted/yanked text will be put back. It
is an error to specify a number register for the delete/yank commands.
Most commonly used change command takes the following form.
- c motion-command
-
Replace the content of the region determined by the motion command
motion-command by the text you type. If the motion command is a
point command then you will type the text into minibuffer, and if the
motion command is a line command then the region will be deleted first and
you can insert the text in insert mode.
For example, if point is at the beginning of a word `foo' and you
wish to change it to `bar', you can type c w. Then, as w
is a point command, you will get the prompt `foo =>' in the
minibuffer, for which you can type b a r RET to complete the change
command.
- c c
-
Change a line. Given a count, that many lines are changed.
- c r
-
Change current region.
- c R
-
Expand current region and change it.
VIP records the previous modifying command, so that it is easy to repeat
it. It is also very easy to undo changes made by modifying commands.
- u
-
Undo the last change. You can undo more by repeating undo by the repeat
command `.'. For example, you can undo 5 previous changes by typing
`u....'. If you type `uu', then the second `u' undoes the
first undo command (
vip-undo
).
- .
-
Repeat the last modifying command. Given count n it becomes the new
count for the repeated command. Otherwise, the count for the last
modifying command is used again (
vip-repeat
).
Miscellaneous Vi commands are collected here.
- Z Z
-
Exit Emacs. If modified buffers exist, you will be asked whether you wish
to save them or not (
save-buffers-kill-emacs
).
- ! motion-command format-command
- n ! ! format-command
-
The region determined by the motion command motion-command will be
given to the shell command format-command and the region will be
replaced by its output. If a count is given, it will be passed to
motion-command. For example, `3!Gsort' will sort the region
between point and the 3rd line. If ! is used instead of
motion-command then n lines will be processed by
format-command (
vip-command-argument
).
- J
-
Join two lines. Given count, join that many lines. A space will be
inserted at each junction (
vip-join-lines
).
- < motion-command
- n < <
-
Shift region determined by the motion command motion-command to
left by shift-width (default is 8). If < is used instead of
motion-command then shift n lines
(
vip-command-argument
).
- > motion-command
- n > >
-
Shift region determined by the motion command motion-command to
right by shift-width (default is 8). If < is used instead of
motion-command then shift n lines
(
vip-command-argument
).
- = motion-command
-
Indent region determined by the motion command motion-command. If
= is used instead of motion-command then indent n lines
(
vip-command-argument
).
- *
-
Call last remembered keyboard macro.
- #
- A new vi operator. See section Some New Commands, for more details.
The following keys are reserved for future extensions, and currently
assigned to a function that just beeps (vip-nil
).
&, @, U, [, ], _, q, ~
VIP uses a special local keymap to interpret key strokes you enter in vi
mode. The following keys are bound to nil in the keymap. Therefore,
these keys are interpreted by the global keymap of Emacs. We give below a
short description of the functions bound to these keys in the global
keymap. See GNU Emacs Manual for details.
- C-@
-
Set mark and push previous mark on mark ring (
set-mark-command
).
- TAB
-
Indent line for current major mode (
indent-for-tab-command
).
- LFD
-
Insert a newline, then indent according to mode (
newline-and-indent
).
- C-k
-
Kill the rest of the current line; before a newline, kill the newline.
With prefix argument, kill that many lines from point. Negative arguments
kill lines backward (
kill-line
).
- C-l
-
Clear the screen and reprint everything (
recenter
).
- n C-p
-
Move cursor vertically up n lines (
previous-line
).
- C-q
-
Read next input character and insert it. Useful for inserting control
characters
(
quoted-insert
).
- C-r
-
Search backward incrementally (
isearch-backward
).
- C-s
-
Search forward incrementally (
isearch-forward
).
- n C-t
-
Interchange characters around point, moving forward one character. With
count n, take character before point and drag it forward past n
other characters. If no argument and at end of line, the previous two
characters are exchanged (
transpose-chars
).
- n C-v
-
Scroll text upward n lines. If n is not given, scroll near
full screen (
scroll-up
).
- C-w
-
Kill between point and mark. The text is save in the kill ring. The
command P or p can retrieve it from kill ring
(
kill-region
).
You can enter insert mode by one of the following commands. In addition to
these, you will enter insert mode if you give a change command with a line
command as the motion command. Insert commands are also modifying commands
and you can repeat them by the repeat command . (vip-repeat
).
- i
-
Enter insert mode at point (
vip-insert
).
- I
-
Enter insert mode at the first non white character on the line
(
vip-Insert
).
- a
-
Move point forward by one character and then enter insert mode
(
vip-append
).
- A
-
Enter insert mode at end of line (
vip-Append
).
- o
-
Open a new line below the current line and enter insert mode
(
vip-open-line
).
- O
-
Open a new line above the current line and enter insert mode
(
vip-Open-line
).
- C-o
-
Insert a newline and leave point before it, and then enter insert mode
(
vip-open-line-at-point
).
Insert mode is almost like emacs mode. Only the following 4 keys behave
differently from emacs mode.
- ESC
-
This key will take you back to vi mode (
vip-change-mode-to-vi
).
- C-h
-
Delete previous character (
delete-backward-char
).
- C-w
-
Delete previous word (
vip-delete-backward-word
).
- C-z
-
This key simulates ESC key in emacs mode. For instance, typing
C-z x in insert mode iw the same as typing ESC x in emacs mode
(
vip-ESC
).
You can also bind C-h to help-command
if you like.
(See section Customizing Key Bindings, for details.) Binding C-h to
help-command
has the effect of making the meaning of C-h
uniform among emacs, vi and insert modes.
When you enter insert mode, VIP records point as the start point of
insertion, and when you leave insert mode the region between point and
start point is saved for later use by repeat command etc. Therefore, repeat
command will not really repeat insertion if you move point by emacs
commands while in insert mode.
Go to the previous, next chapter.