new copyright; att/bsd/shared

[unix-history] / usr / src / contrib / jove-4.14.6 / USD.doc / jove.4

.\" Copyright (c) 1986 The Regents of the University of California.
.\" All rights reserved.
.\"
.\" %sccs.include.redist.roff%
.\"
.\"	@(#)jove.4	6.2 (Berkeley) %G%
.\"
.bp
.NH 1
Alphabetical List of Commands and Variables
.dc "Prefix-1" "Escape"
This reads the next character and runs a command based on the character
typed.  If you wait for more than a second or so before typing the next
character, the message "ESC" will be printed on the message line to
remind you that \s-2JOVE\s0 is waiting for another character.
.dc "Prefix-2" "C-X"
This reads the next character and runs a command based on the character
typed.  If you wait for more than a second or so before typing another
character, the message "C-X" will be printed on the message line to
remind you that \s-2JOVE\s0 is waiting for another character.
.dc "Prefix-3" "Not Bound"
This reads the next character and runs a command based on the character
typed.  If you wait for more than a second or so before typing the next
character, the character that invoked Prefix-3 will be printed on the
message line to remind you that \s-2JOVE\s0 is waiting for another one.
.dc "allow-^S-and-^Q" "(variable)"
This variable, when set, tells \s-2JOVE\s0 that your terminal does not need
to use the characters C-S and C-Q for flow control, and that it is
okay to bind things to them.  This variable should be set depending
upon what kind of terminal you have.
.dc "allow-bad-filenames" "(variable)"
If set, this variable permits filenames to contain "bad" characters
such as those from the set *&%!"`[]{}.  These files are harder to deal
with, because the characters mean something to the shell.  The default
value is "off".
.dc "append-region" "Not Bound"
This appends the region to a specified file.  If the file does not
already exist it is created.
.dc "apropos" "Not Bound"
This types out all the commands, variables and macros with the specific
keyword in their names.  For each command and macro that contains the
string, the key sequence that can be used to execute the command or macro is
printed; with variables, the current value is printed.  So, to find all the
commands that are related to windows, you type
.DS
ESC X apropos window<Return>
.DE
.dc "auto-case-abbrev" "(variable)"
When this variable is on (the default), word abbreviations are adjusted for
case automatically.  For example, if "jove" were the abbreviation for
"jonathan's own version of emacs", then typing "jove" would give you
"jonathan's own version of emacs", typing "Jove" would give you "Jonathan's
own version of emacs", and typing "JOVE" would give you "Jonathan's Own
Version of Emacs".  When this variable is "off", upper and lower case are
distinguished when looking for the abbreviation, i.e., in the example above,
"JOVE" and "Jove" would not be expanded unless they were defined separately.
.dc "auto-execute-command" "Not Bound"
This tells \s-2JOVE\s0 to execute a command automatically when a file whose
name matches a specified pattern is visited.  The first argument is the
command you want executed and the second is a regular expression
pattern that specifies the files that apply.  For example, if you want
to be in show-match-mode when you edit C source files (that is, files
that end with ".c" or ".h") you can type
.ID
ESC X auto-execute-command show-match-mode .*\.[ch]$
.DE
.dc "auto-execute-macro" "Not Bound"
This is like
.IQ auto-execute-command
except you use it to execute macros
automatically instead of built-in commands.
.dc "auto-fill-mode" "Not Bound"
This turns on Auto Fill mode (or off if it's currently on) in the
selected buffer.  When \s-2JOVE\s0 is in Auto Fill mode it automatically
breaks lines for you when you reach the right margin so you don't have
to remember to hit Return.  \s-2JOVE\s0 uses 78 as the right margin but you
can change that by setting the variable
.IQ right-margin
to another
value.  See the
.IQ set
command to learn how to do this.
.dc "auto-indent-mode" "Not Bound"
This turns on Auto Indent mode (or off if it's currently on) in the
selected buffer.  When \s-2JOVE\s0 is in Auto Indent mode, Return indents the
new line to the same position as the line you were just on.  This is
useful for lining up C code (or any other language (but what else is
there besides C?)).  This is out of date because of the new command
called
.IQ newline-and-indent
but it remains because of several
"requests" on the part of, uh, enthusiastic and excitable users, that
it be left as it is.
.dc "backward-character" "C-B"
This moves point backward over a single character.  If point is at the
beginning of the line it moves to the end of the previous line.
.dc "backward-paragraph" "ESC ["
This moves point backward to the beginning of the current or previous
paragraph.  Paragraphs are bounded by lines that begin with a Period or
Tab, or by blank lines; a change in indentation may also signal a break
between paragraphs, except that \s-2JOVE\s0 allows the first line of a paragraph
to be indented differently from the other lines.
.dc "backward-s-expression" "ESC C-B"
This moves point backward over a s-expression.  It is just like
.IQ forward-s-expression
with a negative argument.
.dc "backward-sentence" "ESC A"
This moves point backward to the beginning of the current or previous
sentence.  \s-2JOVE\s0 considers the end of a sentence to be the characters
".", "!" or "?" followed by a Return or by one or more spaces.
.dc "backward-word" "ESC B"
This moves point backward to the beginning of the current or previous
word.
.dc "bad-filename-extensions" "(variable)"
This contains a list of words separated by spaces which are to be
considered bad filename extensions, and so will not be counted in
filename completion.  The default is ".o" so if you have jove.c and
jove.o in the same directory, the filename completion will
.IQ not
complain
of an ambiguity because it will ignore jove.o.
.dc "beginning-of-file" "ESC <"
This moves point backward to the beginning of the buffer.  This
sometimes prints the "Point Pushed" message.  If the top of the buffer
isn't on the screen \s-2JOVE\s0 will set the mark so you can go back to where
you were if you want.
.dc "beginning-of-line" "C-A"
This moves point to the beginning of the current line.
.dc "beginning-of-window" "ESC ,"
This moves point to the beginning of the current window.  The sequence
"ESC ," is the same as "ESC <" (beginning of file) except without the shift
key on the "<", and can thus can easily be remembered.
.dc "bind-to-key" "Not Bound"
This attaches a key to an internal \s-2JOVE\s0 command so that future hits on
that key invoke that command.  For example, to make "C-W" erase the
previous word, you type "ESC X bind-to-key kill-previous-word C-W".
.dc "bind-macro-to-key" "Not Bound"
This is like
.IQ bind-to-key
except you use it to attach keys to
named macros.
.dc "bind-macro-to-word-abbrev" "Not Bound"
This command allows you to bind a macro to a previously defined word
abbreviation.  Whenever you type the abbreviation, it will first be expanded
as an abbreviation, and then the macro will be executed.  Note that if the
macro moves around, you should set the mark first (C-@) and then exchange
the point and mark last (C-X C-X).
.dc "buffer-position" "Not Bound"
This displays the current file name, current line number, total number
of lines, percentage of the way through the file, and the position of
the cursor in the current line.
.dc "c-mode" "Not Bound"
This turns on C mode in the currently selected buffer.  This is one of
currently four possible major modes:  Fundamental, Text, C, Lisp.
When in C or Lisp mode, Tab, "}", and ")" behave a little differently
from usual: They are indented to the "right" place for C (or Lisp)
programs.  In \s-2JOVE\s0, the "right" place is simply the way the author
likes it (but I've got good taste).
.dc "case-character-capitalize" "Not Bound"
This capitalizes the character after point, i.e., the character undo
the cursor.  If a negative argument is supplied that many characters
.IQ before
point are upper cased.
.dc "case-ignore-search" "(variable)"
This variable, when set, tells \s-2JOVE\s0 to treat upper and lower case as
the same when searching.  Thus "jove" and "JOVE" would match, and
"JoVe" would match either.  The default value of this variable is "off".
.dc "case-region-lower" "Not Bound"
This changes all the upper case letters in the region to their lower
case equivalent.
.dc "case-region-upper" "Not Bound"
This changes all the lower case letters in the region to their upper
case equivalent.
.dc "case-word-capitalize" "ESC C"
This capitalizes the current word by making the current letter upper
case and making the rest of the word lower case.  Point is moved to
the end of the word.  If point is not positioned on a word it is first
moved forward to the beginning of the next word.  If a negative
argument is supplied that many words
.IQ before
point are capitalized.
This is useful for correcting the word just typed without having to
move point to the beginning of the word yourself.
.dc "case-word-lower" "ESC L"
This lower-cases the current word and leaves point at the end of it.
If point is in the middle of a word the rest of the word is
converted.  If point is not in a word it is first moved forward to the
beginning of the next word.  If a negative argument is supplied that
many words
.IQ before
point are converted to lower case.  This is useful
for correcting the word just typed without having to move point to the
beginning of the word yourself.
.dc "case-word-upper" "ESC U"
This upper-cases the current word and leaves point at the end of it.
If point is in the middle of a word the rest of the word is
converted.  If point is not in a word it is first moved forward to the
beginning of the next word.  If a negative argument is supplied that
many words
.IQ before
point are converted to upper case.  This is useful
for correcting the word just typed without having to move point to the
beginning of the word yourself.
.dc "character-to-octal-insert" "Not Bound"
This inserts a Back-slash followed by the ascii value of the next
character typed.  For example, "C-G" inserts the string "\\007".
.dc "cd" "Not Bound"
This changes the current directory.
.dc "clear-and-redraw" "ESC C-L"
This clears the entire screen and redraws all the windows.  Use this
when \s-2JOVE\s0 gets confused about what's on the screen, or when the screen
gets filled with garbage characters or output from another program.
.dc "comment-format" "(variable)"
This variable tells \s-2JOVE\s0 how to format your comments when you run the
command
.IQ fill-comment.
Its format is this:
.ID
<open pattern>%!<line header>%c<line trailer>%!<close pattern>
.DE
The %!, %c, and %! must appear in the format; everything else is optional.
A newline (represented by %n) may appear in the open or close patterns.  %%
is the representation for %.  The default comment format is for C comments.
See
.IQ fill-comment
for more.
.dc "compile-it" "C-X C-E"
This compiles your program by running the UNIX command "make" into a buffer,
and automatically parsing the error messages that are created (if any).  See
the
.IQ parse-errors
and
.IQ parse-special-errors
commands.  To compile
a C program without "make", use "C-U C-X C-E" and \s-2JOVE\s0 will prompt
for a command to run instead of make.  (And then the command you type will
become the default command.)  You can use this to parse the output from the
C compiler or the "grep" or "lint" programs.
.dc "continue-process" "Not Bound"
This sends SIGCONT to the current interactive process,
.IQ if
the process
is currently stopped.
.dc "copy-region" "ESC W"
This takes all the text in the region and copies it onto the kill ring
buffer.  This is just like running
.IQ kill-region
followed by the
.IQ yank
command.  See the
.IQ kill-region
and
.IQ yank
commands.
.dc "current-error" "Not Bound"
This moves to the current error in the list of parsed errors.  See the
.IQ next-error
and
.IQ previous-error
commands for more detailed
information.
.dc "date" "Not Bound"
This prints the date on the message line.
.dc "define-mode-word-abbrev" "Not Bound"
This defines a mode-specific abbreviation.
.dc "define-global-word-abbrev" "Not Bound"
This defines a global abbreviation.
.dc "delete-blank-lines" "C-X C-O"
This deletes all the blank lines around point.  This is useful when you
previously opened many lines with "C-O" and now wish to delete the
unused ones.
.dc "delete-buffer" "C-X K"
This deletes a buffer and frees up all the memory associated with it.
Be careful!  Once a buffer has been deleted it is gone forever.  \s-2JOVE\s0
will ask you to confirm if you try to delete a buffer that needs
saving.  This command is useful for when \s-2JOVE\s0 runs out of space to
store new buffers.
.dc "delete-macro" "Not Bound"
This deletes a macro from the list of named macros.  It is an error to
delete the keyboard-macro.  Once the macro is deleted it is gone forever.
If you are about to save macros to a file and decide you don't want to save
a particular one, delete it.
.dc "delete-next-character" "C-D"
This deletes the character that's just after point (that is, the
character under the cursor).  If point is at the end of a line, the
line separator is deleted and the next line is joined with the current
one.
.dc "delete-other-windows" "C-X 1"
This deletes all the other windows except the current one.  This can be
thought of as going back into One Window mode.
.dc "delete-previous-character" "Rubout"
This deletes the character that's just before point (that is, the
character before the cursor).  If point is at the beginning of the
line, the line separator is deleted and that line is joined with the
previous one.
.dc "delete-white-space" "ESC \\\\"
This deletes all the Tabs and Spaces around point.
.dc "delete-current-window" "C-X D"
This deletes the current window and moves point into one of the
remaining ones.  It is an error to try to delete the only remaining
window.
.dc "describe-bindings" "Not Bound"
This types out a list containing each bound key and the command that gets
invoked every time that key is typed.  To make a wall chart of \s-2JOVE\s0
commands, set
.IQ send-typeout-to-buffer
to "on" and \s-2JOVE\s0 will
store the key bindings in a buffer which you can save to a file and then
print.
.dc "describe-command" "Not Bound"
This prints some info on a specified command.
.dc "describe-key" "Not Bound"
This waits for you to type a key and then tells the name of the
command that gets invoked every time that key is hit.  Once you have
the name of the command you can use the
.IQ describe-command
command
to find out exactly what it does.
.dc "describe-variable" "Not Bound"
This prints some info on a specified variable.
.dc "digit" "ESC [0-9]"
This reads a numeric argument.  When you type "ESC" followed by a
number, "digit" keeps reading numbers until you type some other
command.  Then that command is executes with the numeric argument you
specified.
.dc "digit-1" "Not Bound"
This pretends you typed "ESC 1".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "digit-2" "Not Bound"
This pretends you typed "ESC 2".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "digit-3" "Not Bound"
This pretends you typed "ESC 3".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "digit-4" "Not Bound"
This pretends you typed "ESC 4".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "digit-5" "Not Bound"
This pretends you typed "ESC 5".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "digit-6" "Not Bound"
This pretends you typed "ESC 6".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "digit-7" "Not Bound"
This pretends you typed "ESC 7".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "digit-8" "Not Bound"
This pretends you typed "ESC 8".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "digit-9" "Not Bound"
This pretends you typed "ESC 9".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "digit-0" "Not Bound"
This pretends you typed "ESC 0".  This is useful for terminals that
have keypads that send special sequences for numbers typed on the
keypad as opposed to numbers typed from the keyboard.  This can save
having type "ESC" when you want to specify an argument.
.dc "dirs" "Not Bound"
This prints out the directory stack.  See the "cd", "pushd", "popd"
commands for more info.
.dc "disable-biff" "(variable)"
When this is set, \s-2JOVE\s0 disables biff when you're editing and enables it
again when you get out of \s-2JOVE\s0, or when you pause to the parent shell
or push to a new shell. (This means arrival of new mail will not be
immediately apparent but will not cause indiscriminate writing on the
display). The default is "off".
.dc "dstop-process" "Not Bound"
Send the "dsusp" character to the current process.  This is the
character that suspends a process on the next read from the
terminal.  Most people have it set to C-Y.  This only works if
you have the interactive process feature, and if you are in a
buffer bound to a process.
.dc "edit-word-abbrevs" "Not Bound"
This creates a buffer with a list of each abbreviation and the phrase
it expands into, and enters a recursive edit to let you change the
abbreviations or add some more.  The format of this list is
"abbreviation:phrase" so if you add some more you should follow that
format.  It's probably simplest just to copy some already existing
abbreviations and edit them.  When you are done you type "C-X C-C" to
exit the recursive edit.
.dc "end-of-file" "ESC >"
This moves point forward to the end of the buffer.  This sometimes
prints the "Point Pushed" message.  If the end of the buffer isn't on
the screen \s-2JOVE\s0 will set the mark so you can go back to where you were
if you want.
.dc "end-of-line" "C-E"
This moves point to the end of the current line.  If the line is too
long to fit on the screen \s-2JOVE\s0 will scroll the line to the left to
make the end of the line visible.  The line will slide back to its
normal position when you move backward past the leftmost visible character
or when you move off the line altogether.
.dc "end-of-window" "ESC ."
This moves point to the last character in the window.
.dc "eof-process" "Not Bound"
Sends EOF to the current interactive process.  This only works on
versions of \s-2JOVE\s0 which run under 4.2-3 BSD VAX UNIX.  You can't send
EOF to processes on the 2.9 BSD PDP-11 UNIX.
.dc "erase-buffer" "Not Bound"
This erases the contents of the specified buffer.  This is like
.IQ delete-buffer
except it only erases the contents of the buffer, not
the buffer itself.  If you try to erase a buffer that needs saving you
will be asked to confirm it.
.dc "error-window-size" "(variable)"
This is the percentage of the screen to use for the error-window on the
screen.  When you execute
.IQ compile-it,
.IQ error-window-size
percent of the screen will go to the error window.  If the window already
exists and is a different size, it is made to be this size.  The default
value is 20%.
.dc "exchange-point-and-mark" "C-X C-X"
This moves point to mark and makes mark the old point.  This is for
quickly moving from one end of the region to another.
.dc "execute-named-command" "ESC X"
This is the way to execute a command that isn't bound to any key.
When you are prompted with ": " you can type the name of the
command.  You don't have to type the entire name.  Once the command
is unambiguous you can type Space and \s-2JOVE\s0 will fill in the rest for
you.
If you are not sure of the name of the command, type "?" and \s-2JOVE\s0
will print a list of all the commands that you could possibly match
given what you've already typed.  If you don't have any idea what the
command's name is but you know it has something to do with windows
(for example), you can do "ESC X apropos window" and \s-2JOVE\s0 will print a
list of all the commands that are related to windows.
If you find yourself constantly executing the same commands this way
you probably want to bind them to keys so that you can execute them
more quickly.  See the
.IQ bind-to-key
command.
.dc "execute-keyboard-macro" "C-X E"
This executes the keyboard macro.  If you supply a numeric argument the
macro is executed that many times.
.dc "execute-macro" "Not Bound"
This executes a specified macro.  If you supply a numeric argument the
macro is executed that many times.
.dc "exit-jove" "C-X C-C"
This exits \s-2JOVE\s0.  If any buffers need saving \s-2JOVE\s0 will print a warning
message and ask for confirmation.  If you leave without saving your
buffers all your work will be lost.  If you made a mistake and really
do want to exit then you can.  If you are in a recursive editing level
.IQ exit-jove
will return you from that.
.dc "file-creation-mode" "(variable)"
This variable has an octal value.
It contains the mode (see
.IQ chmod(1)
) with which files should be created.  This mode gets modified by your
current umask setting (see
.IQ umask(1)
).  The default value is usually
.IQ 0666
or
.IQ 0644.
.dc "files-should-end-with-newline" "(variable)"
This variable indicates that all files should always have a newline
at the end.  This is often necessary for line printers and the like.
When set, if \s-2JOVE\s0 is writing a file whose last character is not a
newline, it will add one automatically.
.dc "fill-comment" "Not Bound"
This command fills in your C comments to make them pretty and readable.
This filling is done according the variable
.IQ comment-format.
.DS L
/*
 * the default format makes comments like this.
 */
.DE
This can be changed by changing the format variable.  Other languages
may be supported by changing the format variable appropriately.  The
formatter looks backwards from dot for an open comment symbol.  If 
found, all indentation is done relative the position of the first character
of the open symbol.  If there is a matching close symbol, the entire 
comment is formatted.  If not, the region between dot and the open symbol
is reformatted.
.dc "fill-paragraph" "ESC J"
This rearranges words between lines so that all the lines in the current
paragraph extend as close to the right margin as possible, ensuring that
none of the lines will be greater than the right margin.  The default value
for
.IQ right-margin
is 78, but can be changed with the
.IQ set
and
.IQ right-margin-here
commands.  \s-2JOVE\s0 has a complicated algorithm
for determining the beginning and end of the paragraph.  In the normal case
\s-2JOVE\s0 will give all the lines the same indent as they currently have,
but if you wish to force a new indent you can supply a numeric argument to
.IQ fill-paragraph
(e.g., by typing C-U ESC J)
and \s-2JOVE\s0 will indent each line to the column
specified by the
.IQ left-margin
variable.  See also the
.IQ left-margin
variable and
.IQ left-margin-here
command.
.dc "fill-region" "Not Bound"
This is like
.IQ fill-paragraph,
except it operates on a region instead of
just a paragraph.
.dc "filter-region" "Not Bound"
This sends the text in the region to a UNIX command, and replaces the
region with the output from that command.  For example, if you are
lazy and don't like to take the time to write properly indented C
code, you can put the region around your C file and
.IQ filter-region
it
through
.IQ cb,
the UNIX C beautifier.  If you have a file that contains
a bunch of lines that need to be sorted you can do that from inside
\s-2JOVE\s0 too, by filtering the region through the
.IQ sort
UNIX command.
Before output from the command replaces the region \s-2JOVE\s0 stores the old
text in the kill ring, so if you are unhappy with the results you can
easily get back the old text with "C-Y".
.dc "find-file" "C-X C-F"
This visits a file into its own buffer and then selects that buffer.
If you've already visited this file in another buffer, that buffer is
selected.  If the file doesn't yet exist, \s-2JOVE\s0 will print "(New file)"
so that you know.
.dc "find-tag" "C-X T"
This finds the file that contains the specified tag.  \s-2JOVE\s0 looks up
tags by default in the "tags" file in the current directory.  You can change
the default tag name by setting the
.IQ tag-file
variable to another
name.  If you specify a numeric argument to this command, you will be
prompted for a tag file.  This is a good way to specify another tag file
without changing the default.  If the tag cannot be found the error is
reported and point stays where it is.
.dc "find-tag-at-point" "Not Bound"
This finds the file that contains the tag that point is currently on.
See
.IQ find-tag.
.dc "first-non-blank" "ESC M"
This moves point back to the indent of the current line.
.dc "forward-character" "C-F"
This moves forward over a single character.  If point is at the end of
the line it moves to the beginning of the next one.
.dc "forward-paragraph" "ESC ]"
This moves point forward to the end of the current or next paragraph.
Paragraphs are bounded by lines that begin with a Period or Tab, or by blank
lines; a change in indentation may also signal a break between paragraphs,
except that \s-2JOVE\s0 allows the first line of a paragraph to be indented
differently from the other lines.
.dc "forward-s-expression" "ESC C-F"
This moves point forward over a s-expression.  If the first significant
character after point is "(", this moves past the matching ")".  If the
character begins an identifier, this moves just past it.  This is mode
dependent, so this will move over atoms in LISP mode and C identifiers in C
mode.  \s-2JOVE\s0 also matches "{".
.dc "forward-sentence" "ESC E"
This moves point forward to the end of the current or next sentence.
\s-2JOVE\s0 considers the end of a sentence to be the characters ".", "!" or
"?" followed by a Return, or one or more spaces.
.dc "forward-word" "ESC F"
This moves point forward to the end of the current or next word.
.dc "fundamental-mode" "Not Bound"
This sets the major mode to Fundamental.  This affects what \s-2JOVE\s0
considers as characters that make up words.  For instance,
Single-quote is not part of a word in Fundamental mode, but is in Text
mode.
.dc "goto-line" "ESC G"
If a numeric argument is supplied point moves to the beginning of that
line.  If no argument is supplied, point remains where it is.  This is
so you don't lose your place unintentionally, by accidentally hitting
the "G" instead of "F".
.dc "grind-s-expr" "Not Bound"
When point is positioned on a "(", this re-indents that LISP expression.
.dc "grow-window" "C-X ^"
This makes the current window one line bigger.  This only works when
there is more than one window and provided there is room to change the
size.
.dc "paren-flash" ") } ]"
This handles the C mode curly brace indentation, the Lisp mode paren
indentation, and the Show Match mode paren/curly brace/square bracket
flashing.
.dc "handle-tab" "Tab"
This handles indenting to the "right" place in C and Lisp mode, and
just inserts itself in Text mode.
.dc "i-search-forward" "Not Bound"
Incremental search.  Like search-forward except that instead of prompting
for a string and searching for that string all at once, it accepts the string
one character at a time.  After each character you type as part of the search
string, it searches for the entire string so far.  When you like what it
found, type the Return key to finish the search.  You can take back a
character with Rubout and the search will back up to the position before
that character was typed.  C-G aborts the search.
.dc "i-search-reverse" "Not Bound"
Incremental search.  Like search-reverse except that instead of prompting
for a string and searching for that string all at once, it accepts the string
one character at a time.  After each character you type as part of the search
string, it searches for the entire string so far.  When you like what it
found, type the Return key to finish the search.  You can take back a
character with Rubout and the search will back up to the position before
that character was typed.  C-G aborts the search.
.dc "insert-file" "C-X C-I"
This inserts a specified file into the current buffer at point.  Point
is positioned at the beginning of the inserted file.
.dc "internal-tabstop" "(variable)"
The number of spaces \s-2JOVE\s0 should print when it displays a tab character.
The default value is 8.
.dc "interrupt-process" "Not Bound"
This sends the interrupt character (usually C-C) to the interactive process
in the current buffer.  This is only for versions of \s-2JOVE\s0 that have the
interactive processes feature.  This only works when you are inside a buffer
that's attached to a process.
.dc "i-shell" "Not Bound"
This starts up an interactive shell in a window.  \s-2JOVE\s0 uses "shell-1"
as the name of the buffer in which the interacting takes place.  See
the manual for information on how to use interactive processes.
.dc "i-shell-command" "Not Bound"
This is like
.IQ shell-command
except it lets you continue with your
editing while the command is running.  This is really useful for long
running commands with sporadic output.  See the manual for information
on how to use interactive processes.
.dc "kill-next-word" "ESC D"
This kills the text from point to the end of the current or next word.
.dc "kill-previous-word" "ESC Rubout"
This kills the text from point to the beginning of the current or
previous word.
.dc "kill-process" "Not Bound"
This command prompts for a buffer name or buffer number (just as
select-buffer does) and then sends the process in that buffer a
kill signal (9).
.dc "kill-region" "C-W"
This deletes the text in the region and saves it on the kill ring.
Commands that delete text but save it on the kill ring all have the
word "kill" in their names.  Type "C-Y" to yank back the most recent
kill.
.dc "kill-s-expression" "ESC C-K"
This kills the text from point to the end of the current or next
s-expression.
.dc "kill-some-buffers" "Not Bound"
This goes through all the existing buffers and asks whether or not to kill
them.  If you decide to kill a buffer, and it turns out that the buffer is
modified, \s-2JOVE\s0 will offer to save it first.  This is useful for when \s-2JOVE\s0
runs out of memory to store lines (this only happens on PDP-11's) and you
have lots of buffers that you are no longer using.
.dc "kill-to-beginning-of-sentence" "C-X Rubout"
This kills from point to the beginning of the current or previous
sentence.
.dc "kill-to-end-of-line" "C-K"
This kills from point to the end of the current line.  When point is
at the end of the line the line separator is deleted and the next line
is joined with current one.  If a numeric argument is supplied that
many lines are killed; if the argument is negative that many lines
.IQ before
point are killed; if the argument is zero the text from point
to the beginning of the line is killed.
.dc "kill-to-end-of-sentence" "ESC K"
This kills from point to the end of the current or next sentence.  If a
negative numeric argument is supplied it kills from point to the
beginning of the current or previous sentence.
.dc "left-margin" "(variable)"
This is how far lines should be indented when auto-indent mode is on,
or when the
.IQ newline-and-indent
command is run (usually by typing
LineFeed).  It is also used by fill-paragraph and auto-fill mode.
If the value is zero (the default) then the left margin is determined
from the surrounding lines.
.dc "left-margin-here" "Not Bound"
This sets the
.IQ left-margin
variable to the current position of
point.  This is an easy way to say, "Make the left margin begin here,"
without having to count the number of spaces over it actually is.
.dc "lisp-mode" "Not Bound"
This turns on Lisp mode.  Lisp mode is one of four mutually exclusive major
modes: Fundamental, Text, C, and Lisp.  In Lisp mode, the characters Tab
and ) are treated specially, similar to the way they are treated in C mode.
Also, Auto Indent mode is affected, and handled specially.
.dc "list-buffers" "C-X C-B"
This types out a list containing various information about each buffer.
Right now that list looks like this:
.DS
.ta \w'NO111'u +\w'Lines1'u +\w'Scratch111'u +\w'*1'u +\w'commands.doc111'u
\ (* means the buffer needs saving)
\ NO	Lines	Type		Name	File
\ --	-----	----		----	----
\ 1	1	File		Main	[No file]
\ 2	1	Scratch	*	Minibuf	[No file]
\ 3	519	File	*	commands.doc	commands.doc
.DE
The first column lists the buffer's number.  When \s-2JOVE\s0 prompts for a
buffer name you can either type in the full name, or you can simply
type the buffer's number.  The second column is the number of lines in
the buffer.  The third says what type of buffer.  There are four
types: "File", "Scratch", "Process", "I-Process".  "File" is simply a
buffer that holds a file; "Scratch" is for buffers that \s-2JOVE\s0 uses
internally; "Process" is one that holds the output from a UNIX
command; "I-Process" is one that has an interactive process attached to
it.  The next column contains the name of the buffer.  And the last
column is the name of the file that's attached to the buffer.  In this
case, both Minibuf and commands.doc have been changed but not yet
saved.  In fact Minibuf won't be saved since it's an internal \s-2JOVE\s0
buffer that I don't even care about.
.dc "list-processes" "Not Bound"
This makes a list somewhat like "list-buffers" does, except its
list consists of the current interactive processes.  Right now the list
looks like this:
.DS
.ta \w'shell-111111111111'u +\w'Running1111111111'u
\ Buffer	Status	Command name
\ ------	------	------- ----
\ shell-1	Running	i-shell
\ fgrep 	Done	fgrep -n Buffer *.c
.DE
The first column has the name of the buffer to which the process is
attached.  The second has the status of the process; if a process has
exited normally the status is "Done" as in fgrep; if the process
exited with an error the status is "Exit N" where N is the value of
the exit code; if the process was killed by some signal the status is
the name of the signal that was used; otherwise the process is
running.  The last column is the name of the command that is being run.
.dc "mailbox" "(variable)"
Set this to the full pathname of your mailbox.  \s-2JOVE\s0 will look here to
decide whether or not you have any unread mail.  This defaults to
/usr/spool/mail/$USER, where $USER is set to your login name.
.dc "mail-check-frequency" "(variable)"
This is how often (in seconds) \s-2JOVE\s0 should check your mailbox for
incoming mail.  See also the
.IQ mailbox
and
.IQ disable-biff
variables.
.dc "make-backup-files" "(variable)"
If this variable is set, then whenever \s-2JOVE\s0 writes out a file, it will
move the previous version of the file (if there was one) to "#filename".
This is often convenient if you save a file by accident.  The default
value of this variable is "off".
.IQ Note:
this is an optional part of
\s-2JOVE\s0, and your guru may not have it enabled, so it may not work.
.dc "make-buffer-unmodified" "ESC ~"
This makes \s-2JOVE\s0 think the selected buffer hasn't been changed even if
it has.  Use this when you accidentally change the buffer but don't
want it considered changed.  Watch the mode line to see the * disappear
when you use this command.  
.dc "make-macro-interactive" "Not Bound"
This command is meaningful only while you are defining a keyboard macro.
Ordinarily, when a command in a macro definition requires a trailing
text argument (file name, search string, etc.), the argument you
supply becomes part of the macro definition.  If you want to be able
to supply a different argument each time the macro is used, then while
you are defining it, you should give the make-macro-interactive
command just before typing the argument which will be used during the
definition process.  Note: you must bind this command to a key in
order to use it; you can't say ESC X make-macro-interactive.
.dc "mark-threshold" "(variable)"
This variable contains the number of lines point may move by before
the mark is set.  If, in a search or something, point moves by more
than this many lines, the mark is set so that you may return easily.
The default value of this variable is 22 (one screenful, on most
terminals).
.dc "marks-should-float" "(variable)"
When this variable is "off", the position of a mark is remembered as a line
number within the buffer and a character number within the line.  If you add
or delete text before the mark, it will no longer point to the text you
marked originally because that text is no longer at the same line and
character number.  When this variable is "on", the position of a mark is
adjusted to compensate for each insertion and deletion.  This makes marks
much more sensible to use, at the cost of slowing down insertion and
deletion somewhat.  The default value is "on".
.dc "match-regular-expressions" "(variable)"
When set, \s-2JOVE\s0 will match regular expressions in search patterns.
This makes special the characters ., *, [, ], ^, and $, and the two-character
sequences \e<, \e>, \e\|{, \e\|} and \e\||.
See the
.IQ ed(1)
manual page, the tutorial "Advanced Editing in 
.UX
", and the section above "Searching with Regular Expressions"
for more information.
.dc "meta-key" "(variable)"
You should set this variable to "on" if your terminal has a real Meta
key.
If your terminal has such a key, then a key sequence like ESC Y can
be entered by holding down Meta and typing Y.
.dc "mode-line" "(variable)"
The format of the mode line can be determined by setting this variable.
The items in the line are specified using a printf(3) format, with the
special things being marked as "%x".  Digits may be used between the
'%' and the 'x' to mean repeat that many times.
\&'x' may be:
.DS I
.ta .5i 1i 1.5i
	C	check for new mail, and displays "[New mail]" if there
		is any (see also the mail-check-interval and disable-biff
		variables)
	F	the current file name, with leading path stripped
	M	the current list of major and minor modes
	b	the current buffer name
	c	the fill character (-)
	d	the current directory
	e	end of string--this must be the last item in the string
	f	the current file name
	l	the current load average (updated automatically)
	m	the buffer-modified symbol (*)
	n	the current buffer number
	s	space, but only if previous character is not a space
	t	the current time (updated automatically)
	[ ]	the square brackets printed when in a recursive edit
	( )	items enclosed in %( ... %) will only be printed on
		the bottom mode line, rather than copied when the
		window is split
.DE
In addition, any other character is simply copied into the mode line.
Characters may be escaped with a backslash.  To get a feel for all
this, try typing "ESC X print mode-line" and compare the result with
your current mode line.
.dc "mode-line-should-standout" "(variable)"
If set, the mode line will be printed in reverse video, if your
terminal supports it.  The default for this variable is "off".
.dc "name-keyboard-macro" "Not Bound"
This copies the keyboard macro and gives it a name freeing up the
keyboard macro so you can define some more.  Keyboard macros with
their own names can be bound to keys just like built in commands
can.  See the
.IQ read-macros-file-file
and
.IQ write-macros-to-file
commands.
.dc "newline" "Return"
This divides the current line at point moving all the text to the
right of point down onto the newly created line.  Point moves down to
the beginning of the new line.
.dc "newline-and-backup" "C-O"
This divides the current line at point moving all the text to the
right of point down onto the newly created line.  The difference
between this and "newline" is that point does not move down to the
beginning of the new line.
.dc "newline-and-indent" "LineFeed"
This behaves the same was as Return does when in Auto Indent mode.
This makes Auto Indent mode obsolete but it remains in the name of
backward compatibility.
.dc "next-error" "C-X C-N"
This moves to the next error in the list of errors that were parsed
with
.IQ parse-errors
or
.IQ parse-special-errors.
In one window the list
of errors is shown with the current one always at the top.  In another
window is the file that contains the error.  Point is positioned in
this window on the line where the error occurred.
.dc "next-line" "C-N"
This moves down to the next line.
.dc "next-page" "C-V"
This displays the next page of the buffer by taking the bottom line of
the window and redrawing the window with it at the top.  If there isn't
another page in the buffer \s-2JOVE\s0 rings the bell.  If a numeric argument
is supplied the screen is scrolled up that many lines; if the argument
is negative the screen is scrolled down.
.dc "next-window" "C-X N"
This moves into the next window.  Windows live in a circular list so
when you're in the bottom window and you try to move to the next one
you are moved to the top window.  It is an error to use this command
with only one window.
.dc "number-lines-in-window" "Not Bound"
This displays the line numbers for each line in the buffer being
displayed.  The number isn't actually part of the text; it's just
printed before the actual buffer line is.  To turn this off you run
the command again; it toggles.
.dc "over-write-mode" "Not Bound"
This turns Over Write mode on (or off if it's currently on) in the selected
buffer.  When on, this mode changes the way the self-inserting characters
work.  Instead of inserting themselves and pushing the rest of the line over
to the right, they replace or over-write the existing character.  Also,
Rubout replaces the character before point with a space instead of deleting
it.  When Over Write mode is on "OvrWt" is displayed on the mode line.
.dc "page-next-window" "ESC C-V"
This displays the next page in the next window.  This is exactly the
same as "C-X N C-V C-X P".
.dc "paren-flash-delay" "(variable)"
How long, in tenths of seconds, \s-2JOVE\s0 should pause on a matching
parenthesis in
.IQ Show Match
mode.  The default is 5.
.dc "parse-errors" "Not Bound"
This takes the list of C compilation errors (or output from another program
in the same format) in the current buffer and parses them for use with the
.IQ next-error
and
.IQ previous-error
and
.IQ current-error
commands.
This is a very useful tool and helps with compiling C programs and when used
in conjunction with the "grep" UNIX command very helpful in making changes
to a bunch of files.  This command understands errors produced by cc, cpp,
and lint; plus any other program with the same format (e.g., "grep -n").
\s-2JOVE\s0 visits each file that has an error and remembers each line that
contains an error.  It doesn't matter if later you insert or delete
some lines in the buffers containing errors; \s-2JOVE\s0 remembers where
they are regardless.
.IQ next-error
is automatically executed after one
of the parse commands, so you end up at the first error.
.dc "parse-special-errors" "Not Bound"
This parses errors in an unknown format.  Error parsing works with
regular expression search strings with \\('s around the the file name
and the line number.  So, you can use
.IQ parse-special-errors
to parse
lines that are in a slightly different format by typing in your own
search string.  If you don't know how to use regular expressions you
can't use this command.
.dc "parse-spelling-errors-in-buffer" "Not Bound"
This parses a list of words in the current buffer and looks them up in
another buffer that you specify.  This will probably go away soon.
.dc "pause-jove" "ESC S"
This stops \s-2JOVE\s0 and returns control to the parent shell.  This
only works for users using the C-shell, and on systems that have the
job control facility.  To return to \s-2JOVE\s0 you type "fg" to the C-shell.
.dc "physical-tabstop" "(variable)"
How many spaces your terminal prints when it prints a tab character.
.dc "pop-mark" "Not Bound"
This gets executed when you run
.IQ set-mark
with a numeric argument.
\s-2JOVE\s0 remembers the last 16 marks and you use
.IQ pop-mark
to go
backward through the ring of marks.  If you execute "
.IQ pop-mark
enough
times you will eventually get back to where you started.
.dc "popd" "Not Bound"
This pops one entry off the directory stack.  Entries are pushed with
the
.IQ pushd
command.  The names were stolen from the C-shell and the
behavior is the same.
.dc "previous-error" "C-X C-P"
This is the same as
.IQ next-error
except it goes to the previous error.
See
.IQ next-error
for documentation.
.dc "previous-line" "C-P"
This moves up to the previous line.
.dc "previous-page" "ESC V"
This displays the previous page of the current buffer by taking the top
line and redrawing the window with it at the bottom.  If a numeric
argument is supplied the screen is scrolled down that many lines; if
the argument is negative the screen is scrolled up.
.dc "previous-window" "C-X P and C-X O"
This moves into the next window.  Windows live in a circular list so
when you're in the top window and you try to move to the previous one
you are moved to the bottom window.  It is an error to use this command
with only one window.
.dc "print" "Not Bound"
This prints the value of a \s-2JOVE\s0 variable.
.dc "print-message" "Not Bound"
This command prompts for a message, and then prints it on the bottom
line where \s-2JOVE\s0 messages are printed.
.dc "process-bind-to-key" "Not Bound"
This command is identical to bind-to-key, except that it only affects
your bindings when you are in a buffer attached to a process.  When
you enter the process buffer, any keys bound with this command will
automatically take their new values.  When you switch to a non-process
buffer, the old bindings for those keys will be restored.  For example,
you might want to execute
.DS I
process-bind-to-key stop-process ^Z
process-bind-to-key interrupt-process ^C
.DE
Then, when you start up an interactive process and switch into that
buffer, C-Z will execute stop-process and C-C will execute interrupt-
process.  When you switch back to a non-process buffer, C-Z will go
back to executing scroll-up (or whatever you have it bound to).
.dc "process-newline" "Return"
This this only gets executed when in a buffer that is attached to an
interactive-process.  \s-2JOVE\s0 does two different things depending on where
you are when you hit Return.  When you're at the end of the I-Process
buffer this does what Return normally does, except it also makes the
line available to the process.  When point is positioned at some other
position that line is copied to the end of the buffer (with the prompt
stripped) and point is moved there with it, so you can then edit that
line before sending it to the process.  This command
.IQ must
be bound
to the key you usually use to enter shell commands (Return), or else
you won't be able to enter any.
.dc "process-prompt" (variable)
What a prompt looks like from the i-shell and i-shell-command
processes.  The default is "% ", the default C-shell prompt.  This is
actually a regular expression search string.  So you can set it to be
more than one thing at once using the \\| operator.  For instance, for
LISP hackers, the prompt can be
.DS
"% \\|-> \\|<[0-9]>: ".
.DE
.dc "push-shell" "Not Bound"
This spawns a child shell and relinquishes control to it.  This works
on any version of UNIX, but this isn't as good as
.IQ pause-jove
because
it takes time to start up the new shell and you get a brand new
environment every time.  To return to \s-2JOVE\s0 you type "C-D".
.dc "pushd" "Not Bound"
This pushes a directory onto the directory stack and cd's into it.  It
asks for the directory name but if you don't specify one it switches
the top two entries no the stack.  It purposely behaves the same as
C-shell's
.IQ pushd.
.dc "pwd" "Not Bound"
This prints the working directory.
.dc "quadruple-numeric-argument" "C-U"
This multiplies the numeric argument by 4.  So, "C-U C-F" means
forward 4 characters and "C-U C-U C-N" means down 16 lines.
.dc "query-replace-string" "ESC Q"
This replaces the occurrences of a specified string with a specified
replacement string.  When an occurrence is found point is moved to it
and then \s-2JOVE\s0 asks what to do.  The options are:
.DS I
.ta \w'Rubout111'u
Space	to replace this occurrence and go on to the next one.
Period	to replace this occurrence and then stop.
Rubout	to skip this occurrence and go on to the next one.
C-R	to enter a recursive edit.  This lets you temporarily
	suspend the replace, do some editing, and then return
	to continue where you left off.  To continue with the
	Query Replace type "C-X C-C" as if you were trying to
	exit \s-2JOVE\s0.  Normally you would but when you are in a
	recursive edit all it does is exit that recursive
	editing level.
C-W	to delete the matched string and then enter a recursive
	edit.
U	to undo the last replacement.
P or !	to go ahead and replace the remaining occurrences without
	asking.
Return	to stop the Query Replace.
.DE
The search for occurrences starts at point and goes to the end of the
buffer, so to replace in the entire buffer you must first go to the
beginning.
.dc "quit-process" "Not Bound"
This is the same as typing "C-\\" (the Quit character) to a normal UNIX
process, except it sends it to the current process in \s-2JOVE\s0.  This is
only for versions of \s-2JOVE\s0 that have the interactive processes feature.
This only works when you are inside a buffer that's attached to a
process.
.dc "quoted-insert" "C-Q"
This lets you insert characters that normally would be executed as
other \s-2JOVE\s0 commands.  For example, to insert "C-F" you type "C-Q C-F".
.dc "read-word-abbrev-file" "Not Bound"
This reads a specified file that contains a bunch of abbreviation
definitions, and makes those abbreviations available.  If the selected
buffer is not already in Word Abbrev mode this command puts it in
that mode.
.dc "read-macros-from-file" "Not Bound"
This reads the specified file that contains a bunch of macro
definitions, and defines all the macros that were currently defined
when the file was created.  See
.IQ write-macros-to-file
to see how to
save macros.
.dc "redraw-display" "C-L"
This centers the line containing point in the window.  If that line is
already in the middle the window is first cleared and then redrawn.
If a numeric argument is supplied, the line is positioned at that
offset from the top of the window.  For example, "ESC 0 C-L" positions
the line containing point at the top of the window.
.dc "recursive-edit" "Not Bound"
This enters a recursive editing level.  This isn't really very
useful.  I don't know why it's available for public use.  I think I'll
delete it some day.
.dc "rename-buffer" "Not Bound"
This lets you rename the current buffer.
.dc "replace-in-region" "Not Bound"
This is the same as
.IQ replace-string
except that it is restricted
to occurrences between Point and Mark.
.dc "replace-string" "ESC R"
This replaces all occurrences of a specified string with a specified
replacement string.  This is just like
.IQ query-replace-string
except
it replaces without asking.
.dc "right-margin" "(variable)"
Where the right margin is for
.IQ "Auto Fill"
mode and the
.IQ justify-paragraph
and
.IQ justify-region
commands.  The default is 78.
.dc "right-margin-here" "Not Bound"
This sets the
.IQ right-margin
variable to the current position of
point.  This is an easy way to say, "Make the right margin begin here,"
without having to count the number of spaces over it actually is.
.dc "save-file" "C-X C-S"
This saves the current buffer to the associated file.  This makes your
changes permanent so you should be sure you really want to.  If the
buffer has not been modified
.IQ save-file
refuses to do the save.  If
you really do want to write the file you can use "C-X C-W" which
executes
.IQ write-file.
.dc "scroll-down" "ESC Z"
This scrolls the screen one line down.  If the line containing point
moves past the bottom of the window point is moved up to the center of
the window.  If a numeric argument is supplied that many lines are
scrolled; if the argument is negative the screen is scrolled up
instead.
.dc "scroll-step" "(variable)"
How many lines should be scrolled if the
.IQ previous-line
or
.IQ next-line
commands move you off the top or bottom of the screen.  You
may wish to decrease this variable if you are on a slow terminal.
.dc "scroll-up" "C-Z"
This scrolls the screen one line up.  If the line containing point
moves past the top of the window point is moved down to the center of
the window.  If a numeric argument is supplied that many lines are
scrolled; if the argument is negative the screen is scrolled down
instead.
.dc "search-exit-char" "(variable)"
Set this to the character you want to use to exit incremental search.
The default is Newline, which makes i-search compatible with normal
string search.
.dc "search-forward" "C-S"
This searches forward for a specified search string and positions
point at the end of the string if it's found.  If the string is not
found point remains unchanged.  This searches from point to the end of
the buffer, so any matches before point will be missed.
.dc "search-reverse" "C-R"
This searches backward for a specified search string and positions
point at the beginning if the string if it's found.  If the string is
not found point remains unchanged.  This searches from point to the
beginning of the buffer, so any matches after point will be missed.
.dc "select-buffer" "C-X B"
This selects a new or already existing buffer making it the current
one.  You can type either the buffer name or number.  If you type in
the name you need only type the name until it is unambiguous, at which
point typing Escape or Space will complete it for you.  If you want to
create a new buffer you can type Return instead of Space, and a new
empty buffer will be created.
.dc "self-insert" "Most Printing Characters"
This inserts the character that invoked it into the buffer at point.
Initially all but a few of the printing characters are bound to
.IQ self-insert.
.dc "send-typeout-to-buffer" "(variable)"
When this is set \s-2JOVE\s0 will send output that normally overwrites the
screen (temporarily) to a buffer instead.  This affects commands like
.IQ list-buffers,
.IQ list-processes,
and other commands that use command
completion.  The default value is "off".
.dc "set" "Not Bound"
This gives a specified variable a new value.  Occasionally you'll see
lines like "set this variable to that value to do this".  Well, you
use the
.IQ set
command to do that.
.dc "set-mark" "C-@"
This sets the mark at the current position in the buffer.  It prints
the message "Point pushed" on the message line.  It says that instead
of "Mark set" because when you set the mark the previous mark is still
remembered on a ring of 16 marks.  So "Point pushed" means point is
pushed onto the ring of marks and becomes the value of "the mark".
To go through the ring of marks you type "C-U C-@", or execute the
.IQ pop-mark
command.  If you type this enough times you will get back
to where you started.
.dc "shell" "(variable)"
The shell to be used with all the shell commands command.  If your SHELL
environment variable is set, it is used as the value of
.IQ shell;
otherwise "/bin/csh" is the default.
.dc "shell-command" "C-X !"
This runs a UNIX command and places the output from that command in a
buffer.  \s-2JOVE\s0 creates a buffer that matches the name of the command
you specify and then attaches that buffer to a window.  So, when you
have only one window running this command will cause \s-2JOVE\s0 to split the
window and attach the new buffer to that window.  Otherwise, \s-2JOVE\s0
finds the most convenient of the available windows and uses that one
instead.  If the buffer already exists it is first emptied, except that if
it's holding a file, not some output from a previous command, \s-2JOVE\s0
prints an error message and refuses to execute the command.  If you
really want to execute the command you should delete that buffer
(saving it first, if you like) or use
.IQ shell-command-to-buffer,
and
try again.
.dc "shell-command-to-buffer" "Not Bound"
This is just like
.IQ shell-command
except it lets you specify the
buffer to use instead of \s-2JOVE\s0.
.dc "shell-flags" "(variable)"
This defines the flags that are passed to shell commands.  The default is
"-c".  See the
.IQ shell
variable to change the default shell.
.dc "show-match-mode" "Not Bound"
This turns on Show Match mode (or off if it's currently on) in the
selected buffer.  This changes "}" and ")" so that when they are typed
the are inserted as usual, and then the cursor flashes back to the
matching "{" or "(" (depending on what was typed) for about half a
second, and then goes back to just after the "}" or ")" that invoked
the command.  This is useful for typing in complicated expressions in
a program.  You can change how long the cursor sits on the matching
paren by setting the "paren-flash-delay" variable in tenths of a
second.  If the matching "{" or "(" isn't visible nothing happens.
.dc "shrink-window" "Not Bound"
This makes the current window one line shorter, if possible.  Windows
must be at least 2 lines high, one for the text and the other for the
mode line.
.dc "source" "Not Bound"
This reads a bunch of \s-2JOVE\s0 commands from a file.  The format of the
file is the same as that in your initialization file (your ".joverc")
in your main directory.  There should be one command per line and it
should be as though you typed "ESC X" while in \s-2JOVE\s0.  For example,
here's part of my initialization file:
.DS I
bind-to-key i-search-reverse ^R
bind-to-key i-search-forward ^S
bind-to-key pause-jove ^[S
.DE
What they do is make "C-R" call the
.IQ i-search-reverse
command and
"C-S" call
.IQ i-search-forward
and "ESC S" call
.IQ pause-jove.
.dc "spell-buffer" "Not Bound"
This runs the current buffer through the UNIX
.IQ spell
program and places
the output in buffer "Spell".  Then \s-2JOVE\s0 lets you edit the list of
words, expecting you to delete the ones that you don't care about, i.e., the
ones you know are spelled correctly.  Then the
.IQ parse-spelling-errors-in-buffer
command comes along and finds all the
misspelled words and sets things up so the error commands work.
.dc "split-current-window" "C-X 2"
This splits the current window into two equal parts (providing the
resulting windows would be big enough) and displays the selected
buffer in both windows.  Use "C-X 1" to go back to 1 window mode.
.dc "start-remembering" "C-X ("
This starts remembering your key strokes in the Keyboard macro.  To
stop remembering you type "C-X )".  Because of a bug in \s-2JOVE\s0 you can't
stop remembering by typing "ESC X stop-remembering";
.IQ stop-remembering
must be bound to "C-X )" in order to make things work correctly.  To
execute the remembered key strokes you type "C-X E" which runs the
.IQ execute-keyboard-macro
command.
Sometimes you may want a macro to accept different input each time it runs.
To see how to do this, see the
.IQ make-macro-interactive
command.
.dc "stop-process" "Not Bound"
This sends a stop signal (C-Z, for most people) to the current process.
It only works if you have the interactive process feature, and you are
in a buffer attached to a process.
.dc "stop-remembering" "C-X )"
This stop the definition of the keyboard macro.  Because of a bug in
\s-2JOVE\s0, this must be bound to "C-X )".  Anything else will not work
properly.
.dc "string-length" "Not Bound"
This prints the number of characters in the string that point sits in.
Strings are surrounded by double quotes.  \s-2JOVE\s0 knows that "\\007" is
considered a single character, namely "C-G", and also knows about
other common ones, like "\\r" (Return) and "\\n" (LineFeed).  This is
mostly useful only for C programmers.
.dc "suspend-jove" "ESC S"
This is a synonym for
.IQ pause-jove.
.dc "sync-frequency" "(variable)"
The temporary files used by \s-2JOVE\s0 are forced out to disk every
.IQ sync-frequency
modifications.  The default is 50, which really makes
good sense.  Unless your system is very unstable, you probably
shouldn't fool with this.
.dc "tag-file" "(variable)"
This the name of the file in which \s-2JOVE\s0 should look up tag
definitions.  The default value is "./tags".
.dc "text-mode" "Not Bound"
This sets the major mode to Text.  Currently the other modes are
Fundamental, C and Lisp mode.
.dc "transpose-characters" "C-T"
This switches the character before point with the one after point, and
then moves forward one.  This doesn't work at the beginning of the
line, and at the end of the line it switches the two characters before
point.  Since point is moved forward, so that the character that was
before point is still before point, you can use "C-T" to drag a
character down the length of a line.  This command pretty quickly
becomes very useful.
.dc "transpose-lines" "C-X C-T"
This switches the current line with the one above it, and then moves
down one so that the line that was above point is still above point.
This, like
.IQ transpose-characters,
can be used to drag a line down a page.
.dc "unbind-key" "Not Bound"
Use this to unbind
.IQ any
key sequence.  You can use this to unbind even a
prefix command, since this command does not use "key-map completion".  For
example, "ESC X unbind-key ESC [" unbinds the sequence "ESC [".  This is
useful for "turning off" something set in the system-wide ".joverc" file.
.dc "update-time-frequency" "(variable)"
How often the mode line is updated (and thus the time and load
average, if you display them).  The default is 30 seconds.
.dc "use-i/d-char" "(variable)"
If your terminal has insert/delete character capability you can tell \s-2JOVE\s0
not to use it by setting this to "off".  In my opinion it is only worth using
insert/delete character at low baud rates.  WARNING: if you set this to
"on" when your terminal doesn't have insert/delete character capability,
you will get weird (perhaps fatal) results.
.dc "version" "Not Bound"
Displays the version number of this \s-2JOVE\s0.
.dc "visible-bell" "(variable)"
Use the terminal's visible bell instead of beeping.  This is set
automatically if your terminal has the capability.
.dc "visible-spaces-in-window" "Not Bound"
This displays an underscore character instead of each space in the
window and displays a greater-than followed by spaces for each tab
in the window.  The actual text in the buffer is not changed; only
the screen display is affected.  To turn this off you run the command
again; it toggles.
.dc "visit-file" "C-X C-V"
This reads a specified file into the current buffer replacing the old
text.  If the buffer needs saving \s-2JOVE\s0 will offer to save it for you.
Sometimes you use this to start over, say if you make lots of changes
and then change your mind.  If that's the case you don't want \s-2JOVE\s0 to
save your buffer and you answer "NO" to the question.
.dc "window-find" "C-X 4"
This lets you select another buffer in another window three
different ways.  This waits for another character which can be one of
the following:
.DS I
.ta .5i 1i 1.5i
T	Finds a tag in the other window.
F	Finds a file in the other window.
B	Selects a buffer in the other window.
.DE
This is just a convenient short hand for "C-X 2" (or "C-X O" if there are
already two windows) followed by the appropriate sequence for invoking each
command.  With this, though, there isn't the extra overhead of having to
redisplay.  In addition, you don't have to decide whether to type "C-X 2" or
"C-X O" since "C-X 4" does the right thing.
.dc "word-abbrev-mode" "Not Bound"
This turns on Word Abbrev mode (or off if it's currently on) in the
selected buffer.  Word Abbrev mode lets you specify a word (an
abbreviation) and a phrase with which \s-2JOVE\s0 should substitute the
abbreviation.  You can use this to define words to expand into long
phrases, e.g., "jove" can expand into "Jonathan's Own Version of
Emacs"; another common use is defining words that you often misspell
in the same way, e.g., "thier" => "their" or "teh" => "the".  See
the information on the
.IQ auto-case-abbrev
variable.
.sp 1
There are two kinds of abbreviations: mode specific and global.  If
you define a Mode specific abbreviation in C mode, it will expand only
in buffers that are in C mode.  This is so you can have the same
abbreviation expand to different things depending on your context.
Global abbreviations expand regardless of the major mode of the
buffer.  The way it works is this: \s-2JOVE\s0 looks first in the mode
specific table, and then in the global table.  Whichever it finds it
in first is the one that's used in the expansion.  If it doesn't find
the word it is left untouched.
\s-2JOVE\s0 tries to expand words as they are typed, when you type a
punctuation character or Space or Return.  If you are in Auto Fill
mode the expansion will be filled as if you typed it yourself.
.dc "wrap-search" "(variable)"
If set, searches will "wrap around" the ends of the buffer instead
of stopping at the bottom or top.  The default is "off".
.dc "write-files-on-make" "(variable)"
When set, all modified files will be written out before calling
make when the
.IQ compile-it
command is executed.  The default is "on".
.dc "write-word-abbrev-file" "Not Bound"
This writes the currently defined abbreviations to a specified file.
They can be read back in and automatically defined with
.IQ read-word-abbrev-file.
.dc "write-file" "C-X C-W"
This saves the current buffer to a specified file, and then makes that
file the default file name for this buffer.  If you specify a file
that already exists you are asked to confirm over-writing it.
.dc "write-macros-to-file" "Not Bound"
This writes the currently defined macros to a specified file.  The
macros can be read back in with
.IQ read-macros-from-file
so you can
define macros and still use them in other instantiations of \s-2JOVE\s0.
.dc "write-modified-files" "C-X C-M"
This saves all the buffers that need saving.  If you supply a numeric
argument it asks for each buffer whether you really want to save it.
.dc "write-region" "Not Bound"
This writes the text in the region to a specified file.  If the file
already exists you are asked to confirm over-writing it.
.dc "yank" "C-Y"
This undoes the last kill command.  That is, it inserts the killed
text at point.  When you do multiple kill commands in a row, they are
merged so that yanking them back with "C\-Y" yanks back all of them.
.dc "yank-pop" "ESC Y"
This yanks back previous killed text.  \s-2JOVE\s0 has a kill ring on which
the last 10 kills are stored.
.IQ Yank
yanks a copy of the text at the
front of the ring.  If you want one of the last ten kills you use "ESC
Y" which rotates the ring so another different entry is now at the
front.  You can use "ESC Y" only immediately following a "C-Y" or
another "ESC Y".  If you supply a negative numeric argument the ring
is rotated the other way.  If you use this command enough times in a
row you will eventually get back to where you started.  Experiment
with this.  It's extremely useful.