XEmacs 21.2-b1

[chise/xemacs-chise.git.1] / man / term.texi

@\input texinfo @c -*-texinfo-*-
settitle Notes about emacs Term mode
@setfilename ../info/term.info

@titlepage
@sp 6
@center @titlefont(Notes about Emacs TERM Mode)
@end titlepage

@ifinfo
@c @format
@c START-INFO-DIR-ENTRY
@c * term mode::                       Emacs terminal emulator mode.
@c END-INFO-DIR-ENTRY
@c @end format

@node Top, , (DIR)
@top Terminal emulator mode
@end ifinfo

This is some notes about the term Emacs mode.

@menu
* term mode::
@end menu

@node term mode
@chapter Term Mode

@menu
* Overview::
* Connecting to remote computers::
* Paging::
* Terminal escapes::
@end menu

The @code{term} package includes the major modes @code{term},
@code{shell}, and @code{gud} (for running gbd or another debugger).
It is a replacement for the comint mode of Emacs 19,
as well as shell, gdb, terminal, and telnet modes.
The package works best with recent releases of Emacs 19,
but will also work reasonably well with Emacs 18 as well as Lucid Emacs 19.

The file @code{nshell.el} is a wrapper to use unless term mode
is built into Emacs.  If works around some of the missing
in older Emacs versions.
To use it, edit the paths in @code{nshell.el}, appropriately,
and then @code{M-x load-file nshell.el RET}.
This will also load in replacement shell and gud modes.

@node Overview
@section Overview

The @code{term} mode is used to control a program (an "inferior process").
It sends most keyboard input characters to the program,
and displays output from the program in the buffer.
This is similar to the traditional comint mode, and
modes derived from it (such as shell and gdb modes).
You can do with the new term-based shell the same sort
of things you could do with the old shell mode,
using more or less the same interface.  However, the
new mode is more flexible, and works somewhat differently.

@menu
* Output from the inferior::
* subbuffer:: The sub-buffer
* altsubbuffer:: The alternate sub-buffer
* Input to the inferior::
@end menu

@node Output from the inferior
@subsection Output from the inferior

In typical usage, output from the inferior is
added to the end of the buffer.  If needed, the window
will be scrolled, just like a regular terminal.
(Only one line at a time will be scrolled, just like
regular terminals, and in contrast to the old shell mode.)
Thus the buffer becomes a log of your interaction with the
inferior, just like the old shell mode.

Like a real terminal, term maintains a "cursor position."
This is the @code{process-mark} of the inferior process.
If the process-mark is not at the end of the buffer, output from
the inferior will overwrite existing text in the buffer.
This is like a real terminal, but unlike the old shell mode
(which inserts the output, instead of overwriting).

Some programs (such as Emacs itself) need to control the
appearance on the screen in detail.  They do this by
sending special control codes.  The exact control
codes needed from terminal to terminal, but nowadays
most terminals and terminal emulators (including xterm)
understand the so-called "ANSI escape sequences" (first
popularized by the Digital's VT100 family of terminal).
The term mode also understands these escape sequences,
and for each control code does the appropriate thing
to change the buffer so that the appearance of the window
will match what it would be on a real terminal.
(In contrast, the old shell mode doesn't handle
terminal control codes at all.)

See <...> for the specific control codes.

@node subbuffer
@subsection The sub-buffer

A program that talks to terminal expects the terminal to have a fixed size. 
If the program is talking a terminal emulator program such as @code{xterm},
that size can be changed (if the xterm window is re-sized), but programs
still assume a logical terminal that has a fixed size independent
of the amount of output transmitted by the programs.

To programs that use it, the Emacs terminal emulator acts as if it
too has a fixed size.  The @dfn{sub-buffer} is the part of a @code{term}-mode
buffer that corresponds to a "normal" terminal.  Most of the time
(unless you explicitly scroll the window displaying the buffer),
the sub-buffer is the part of the buffer that is displayed in a window.

The sub-buffer is defined in terms of three buffer-local-variable:

@defvar term-height
The height of the sub-buffer, in screen lines.
@end defvar

@defvar term-width
The width of the sub-buffer, in screen columns.
@end defvar

@defvar term-home-marker
The "home" position, that is the top left corner of the sub-buffer.
@end defvar

The sub-buffer is assumed to be the end part of the buffer;
the @code{term-home-marker} should never be more than
@code{term-height} screen lines from the end of the buffer.

@node altsubbuffer
@subsection The alternate sub-buffer

When a "graphical" program finishes, it is nice to
restore the screen state to what it was before the program started.
Many people are used to this behavior from @code{xterm}, and
its also offered by the @code{term} emulator.

@defun term-switch-to-alternate-sub-buffer set
If @var{set} is true, and we're not already using the alternate sub-buffer,
switch to it.  What this means is that the @code{term-home-marker}
is saved (in the variable @code{term-saved-home-marker}), and the
@code{term-home-marker} is set to the end of the buffer.

If @var{set} is false and we're using the alternate sub-buffer,
switch back to the saved sub-buffer.  What this means is that the
(current, alternate) sub-buffer is deleted (using
@code{(delete-region term-home-marker (point-max))}), and then the
@code{term-home-marker} is restored (from @code{term-saved-home-marker}).
@end defun

@node Input to the inferior
@subsection Input to the inferior

Characters typed by the user are sent to the inferior.
How this is done depends on whether the @code{term} buffer
is in "character" mode or "line" mode.
(A @code{term} buffer can also be in "pager" mode.
This is discussed <later>.)
Which of these is currently active is specified in the mode line.
The difference between them is the key-bindings available.

In character mode, one character (by default @key{C-c}) is special,
and is a prefix for various commands.  All other characters are
sent directly to the inferior process, with no interpretation by Emacs.
Character mode looks and feels like a real terminal, or a conventional
terminal emulator such as xterm.

In line mode, key commands mostly have standard Emacs actions.
Regulars characters insert themselves into the buffer.
When return is typed, the entire current line of the buffer
(except possibly the prompt) is sent to the inferior process.
Line mode is basically the original shell mode from earlier Emacs versions.

To switch from line mode to character mode type @kbd{C-c c}.
To switch from character mode to line mode type @kbd{C-c l}.

In either mode, "echoing" of user input is handled by the inferior.
Therefor, in line mode after an input line at the end of the buffer
is sent to the inferior, it is deleted from the buffer.
This is so that the inferior can echo the input, if it wishes
(which it normally does).

@node Connecting to remote computers
@section Connecting to remote computers

If you want to login to a remove computer, you can do that just as
you would expect, using whatever commands you would normally use.

(This is worth emphasizing, because earlier versions of @code{shell}
mode would not work properly if you tried to log in to some other
computer, because of the way echoing was handled.  That is why
there was a separate @code{telnet} mode to partially compensate for
these problems.  The @code{telnet} mode is no longer needed, and
is basically obsolete.)

A program that asks you for a password will normally suppress
echoing of the password, so the password will not show up in the buffer.
This will happen just as if you were using a real terminal, if
the buffer is in char mode.  If it is in line mode, the password
will be temporarily visible, but will be erased when you hit return.
(This happens automatically; there is no special password processing.)

When you log in to a different machine, you need to specify the
type of terminal your using.   If you are talking to a Bourne-compatible
shell, and your system understands the @code{TERMCAP} variable,
you can use the command @kbd{M-x shell-send-termcap}, which
sends a string specifying the terminal type and size.
(This command is also useful after the window has changed size.)

If you need to specify the terminal type manually, you can try the
terminal types "ansi" or "vt100".

You can of course run gdb on that remote computer.  One useful
trick:  If you invoke gdb with the @code{--fullname} option,
it will send special commands to Emacs that will cause Emacs to
pop up the source files you're debugging.  This will work
whether or not gdb is running on a different computer than Emacs,
assuming can access the source files specified by gdb.

@node Paging
@section Paging

When the pager is enabled, Emacs will "pause" after each screenful
of output (since the last input sent to the inferior).
It will enter "pager" mode, which feels a lot like the "more"
program:  Typing a space requests another screenful of output.
Other commands request more or less output, or scroll backwards
in the @code{term} buffer.  In pager mode, type @kbd{h} or @kbd{?}
to display a help message listing all the available pager mode commands.

In either character or line mode, type @kbd{C-c p} to enable paging,
and @kbd{C-c D} to disable it.

@node Terminal escapes
@section Terminal Escape sequences

A program that does "graphics" on a terminal controls the
terminal by sending strings called @dfn{terminal escape sequences}
that the terminal (or terminal emulator) interprets as special commands.
The @code{term} mode includes a terminal emulator that understands
standard ANSI escape sequences, originally popularized by VT100 terminals,
and now used by the @code{xterm} program and most modern terminal
emulator software.

@menu
* Cursor motion:: Escape sequences to move the cursor
* Erasing:: Escape commands for erasing text
* Inserting and deleting:: Escape sequences to insert and delete text
* Scrolling:: Escape sequences to scroll part of the visible window
* Command hook::
* Miscellaneous escapes::
@end menu

printing chars

tab

LF

@node Cursor motion
@subsection Escape sequences to move the cursor

@table @kbd
@item RETURN
Moves to the beginning of the current screen line.

@item C-b
Moves backwards one column.  (Tabs are broken up if needed.)
@comment Line wrap FIXME

@item Esc [ R ; C H
Move to screen row R, screen column C, where (R=1) is the top row,
and (C=1) is the leftmost column.  Defaults are R=1 and C=1.

@item Esc [ N A
Move N (default 1) screen lines up.
@item Esc [ N B
Move N (default 1) screen lines down.
@item Esc [ N C
Move N (default 1) columns right.
@item Esc [ N D
Move N (default 1) columns left.
@end table

@node Erasing
@subsection Escape commands for erasing text

These commands "erase" part of the sub-buffer.
Erasing means replacing by white space; it is not the same as deleting.
The relative screen positions of things that are not erased remain
unchanged with each other, as does the relative cursor position.

@table @kbd
@item E [ J
Erase from cursor to end of screen.
@item E [ 0 J
Same as E [ J.
@item E [ 1 J
Erase from home position to point.
@item E [ 2 J
Erase whole sub-buffer.
@item E [ K
Erase from point to end of screen line.
@item E [ 0 K
Same as E [ K.
@item E [ 1 K
Erase from beginning of screen line to point.
@item E [ 2 K
Erase whole screen line.
@end table

@node Inserting and deleting
@subsection Escape sequences to insert and delete text

@table @kbd
@item Esc [ N L
Insert N (default 1) blank lines.
@item Esc [ N M
Delete N (default 1) lines.
@item Esc [ N P
Delete N (default 1) characters.
@item Esc [ N @@
Insert N (default 1) spaces.
@end table

@node Scrolling
@subsection Escape sequences to scroll part of the visible window

@table @kbd
@item Esc D
Scroll forward one screen line.

@item Esc M
Scroll backwards one screen line.

@item Esc [ T ; B r
Set the scrolling region to be from lines T down to line B inclusive,
where line 1 is the topmost line.
@end table

@node Command hook
@subsection Command hook

If @kbd{C-z} is seen, any text up to a following @key{LF} is scanned.
The text in between (not counting the initial C-z or the final LF)
is passed to the function that is the value of @code{term-command-hook}.

The default value of the @code{term-command-hook} variable
is the function @code{term-command-hook}, which handles the following:

@table @kbd
@item C-z C-z FILENAME:LINENUMBER:IGNORED LF
Set term-pending-frame to @code{(cons "FILENAME" LINENUMBER)}.
When the buffer is displayed in the current window, show
the FILENAME in the other window, and show an arrow at LINENUMBER.
Gdb emits these strings when invoked with the flag --fullname.
This is used by gdb mode; you can also invoke gdb with this flag
from shell mode.

@item C-z / DIRNAME LF
Set the directory of the term buffer to DIRNAME

@item C-z ! LEXPR LF
Read and evaluate LEXPR as a Lisp expression.
The result is ignored.
@end table

@node Miscellaneous escapes
@subsection Miscellaneous escapes

@table @kbd
@item C-g (Bell)
Calls @code{(beep t)}.

@item Esc 7
Save cursor.

@item Esc 8
Restore cursor.

@item Esc [ 47 h
Switch to the alternate sub-buffer,
@item Esc [ 47 l
Switch back to the regular sub-buffer,
@end table

@bye