Path: seismo!harvard!talcott!panda!decvax!minow (Martin Minow)
From: sources-request@panda.UUCP
Newsgroups: mod.sources
Subject: MicroEmacs (Part 6 of 6)
Message-ID: <1673@panda.UUCP>
Date: 14 Apr 86 22:33:43 GMT
Sender: jpn@panda.UUCP
Lines: 838
Approved: jpn@panda.UUCP

Mod.sources:  Volume 4, Issue 73
Submitted by: decvax!minow (Martin Minow)

#! /bin/sh
# This is a shell archive, meaning:
# 1. Remove everything above the #! /bin/sh line.
# 2. Save the resulting text in a file.
# 3. Execute the file with /bin/sh (not csh) to create the files:
#	uemacs.tex
# This archive created: Sun Apr 13 11:18:26 1986
export PATH; PATH=/bin:$PATH
echo shar: extracting "'uemacs.tex'" '(36749 characters)'
if test -f 'uemacs.tex'
then
	echo shar: will not over-write existing file "'uemacs.tex'"
else
cat << \SHAR_EOF > 'uemacs.tex'
%%
%%	W A R N I N G
%%
%% This file is quite out of date.  It is, however, the best we have.
%% Build the documentation using LaTeX.
%%
\newcounter{Version}
\setcounter{Version}{28}		% update at each new release
\newlength{\Cmdwidth}			% used to indent command lists
\setlength{\Cmdwidth}{8em}		% set by eyeball
\documentstyle[twoside,12pt]{article}
\pagestyle{headings}
%\makeindex				% someday
\load{\footnotesize}{\sf}		% Load smaller
\load{\small}{\sf}			% Sans-Serif fonts
\hyphenation{Mi-cro-EMACS}
\renewcommand{\labelitemii}{$\bullet$}	% Second-level \item
\parskip 3pt plus 2pt minus 1pt		% this looks better than
\parindent 20pt				% the standard style
%
% Define textual macros:
% \_			not correct in standard TeX for big fonts
%
% The following are used to switch fonts for special Uemacs terms.
%
% \Name{xxx}		Command name (changes font to Sans-Serif)
% \META{}		META in running text
% \CONTROL{}		CONTROL in running text
% \Text{xxx}		Display text (changes font to typewriter)
% \Nameitem{xxx}	\item[] if contents are command names.
%			(Will put name in index eventually)
%
\renewcommand{\_}{{\tt \char95}}
\newcommand{\Text}[1]{{\tt #1}}
\newcommand{\Name}[1]{{\sf #1}}
\newcommand{\META}{{\underline{\sc meta}}}
\newcommand{\CONTROL}{{\underline{\sc control}}}
\newcommand{\Nameitem}[1]{\item [\Name{#1}]}
%
% Environment for command lists.  Like \begin{description}, but
% with special indentation.  Usage:
% \begin{cmdlist}
%   \Nameitem{xxx} Explanation of item xxx.
%   \Nameitem{yyy} Explanation of item yyy.
% \end{cmdlist}
%
\newenvironment{cmdlist}{%
  \list{}{%					% homemade description envt.
    \setlength{\rightmargin}{0em}		% on either side.
    \setlength{\labelwidth}{\Cmdwidth}	% \item indentation
    \setlength{\labelsep}{1em}			% between item and text
    \setlength{\leftmargin}{\labelwidth}	% Indent by indentation
    \addtolength{\leftmargin}{\labelsep}	%   + a little space
    \renewcommand{\makelabel}[1]{##1\hfill}	% Add \index command someday
  }
}{\endlist}

\begin{document}
\author{\
David G. Conroy \\
Digital Equipment Corporation \\
146 Main St. \\
Maynard, MA 01754
}
\title{\
  MicroEmacs \\[3ex]
  {\large Reasonable Editing } \\
  {\large on } \\
  {\large Tiny Computers } \\
}
\maketitle
\abstract{
This document describes MicroEMACS,
a public-domain display editor, loosly base on the EMACS display editor
written by Richard Stallman at MIT. It is intended as a reference manual
for users already familiar with EMACS.

This is MicroEmacs Version \arabic{Version}.
}
\vfill
\vfill			% Set this 2/3 down on the page.
\begin{quotation}
This document was produced using the \LaTeX{} document
preparation system developed by Leslie
Lamport.  All page composition and make-up was performed
by \TeX,\footnote{\TeX{} is a registered trademark of the
American Mathematical Society.} the typesetting system developed by
Donald E. Knuth at Stanford University.
\end{quotation}
\vfill
\newpage
\tableofcontents
\newpage
\section{Introduction}
This document describes MicroEMACS, a
public-domain display editor, loosely based on EMACS, that runs on little
computers.

We call it Micro{\em EMACS\/} to emphasize the fact that
most of the commands are very similar to, if not identical to,
fundamental mode EMACS commands (actually, it implements a set of
rebindings used by a group of users at Digital).

We call it {\em Micro\/}EMACS to emphasize the fact that it is
but a shadow of full EMACS. No attempt has been made to make MicroEMACS
wildly
customizable (without writing code), or to have extensive online documentation.
All of this sophistication was thrown away right at the start, because
it was important that MicroEMACS run on little computers.
In all fairness, is should be stated here and now that the most popular
small computer these days is the MicroVAX!
\section{Some Basic Concepts}
This section describes a number of basic concepts,
which will be used again and again in the following descriptions.
Most of it should be familiar to EMACS users.
New users are advised not to get stuck reading this section,
because most of the information will become clear the first time you
play with the editor on a computer.
\subsection{Screen Layout}
MicroEMACS divides the screen into two areas. The very last
line on the screen, which is used to display messages and
ask one line questions, is called the echo line. Typing in the echo line
is like typing commands to the operating system. More details later.

The remainer of the screen is filled with text windows.
When MicroEMACS is first started, there is only one text window, and it
fills the screen. Windows can be created, deleted, and adjusted in
size using MicroEMACS commands.
\subsection{Mode Lines}
The last line of a window (the line, usually in reverse video,
that starts with \Text{MicroEMACS}) is
the mode line for the window.
It contains information about what is being displayed in the
window. Mode lines separate windows from each other and from the
echo line.

Three important bits of information are displayed in the mode line.
these are the {\em buffer name\/}, the {\em file name\/},
and the {\em buffer changed flag\/}.

A window is always displaying a buffer. The name of the buffer being
displayed in the window is shown in the mode line, right after the
\Text{MicroEMACS}.

A buffer may have a file associated with it.
Usually this is the last file read into or written from the
buffer. If a file is associated with the buffer, the name of the
file is displayed in the mode line, prefixed by \Text{File:}.
On operating systems in which file names are case insensitive
(for example, VMS) the file name will always be in lower case.

MicroEMACS knows if a buffer has been changed since it was last
written to its associated file. It informs you of this fact by displaying
a \verb+*+ in the mode line immediately to the left of the \Text{MicroEMACS}
string.
\subsection{Keyboard Conventions}
Internally, MicroEMACS represents keyboard characters in a special
10 bit code. This code consists of 8 bits of character data (ASCII plus
whatever characters your terminal puts in the right side graphics set), and
2 flags, the \CONTROL{} flag and the \META{} flag.
All 4 combinations of the two flags are possible.

This manual represents 10 bit characters in a standard way. If you
have a character \Text{A}, then that character with
the \CONTROL{} bit set is
written as \Name{C-A}. That character with the
\META{} bit set is written as \Name{M-A}.
If both bits are set, it is written as \Name{C-M-A}.

Terminals (usually) cannot deal directly deal with the 10 bit
character set.
Characters with the flag bits set are entered by prefixing the
desired character with another character.
To get a character with the \CONTROL{} bit set,
prefix the character with \Text{Control-}\verb+^+ (the
word \Text{Control} here, and in the next few examples, means the key marked
\Text{Control} or \Text{CTRL} on the ASCII keyboard).
To get a character with the
\META{} bit set, prefix the character
with \Text{Control-}\verb+[+, \Text{Escape}, or \Text{ESC}.
To get a character with the \CONTROL{}
and \META{} bits set, prefix the character
with \Text{Control-}\verb+\+.

Those of you who understand ASCII are probably asking what you get
if you prefix a character with \Text{Control-}\verb+]+.
This character is not a prefix;
it is reserved for future use as an ``exit from recursive editing level''
command.

CONTROL characters are very important, so there is a shorthand way
of typing them.
Characters between \Text{Control-@} and \Text{Control-Z} are mapped
to the characters \Text{@} through \Text{Z} with the \CONTROL{} bit set.
Characters between
\Text{Control-a} and \Text{Control-z} are mapped to
the characters \Text{A} through \Text{Z}
with the \CONTROL{} bit set.
Most of the time \Text{Control-X} and \Name{C-X} are the same.
\subsubsection{The VT100 Keyboard}
MicroEMACS understands the arrow keys and the keypad on VT100
style keyboards. Only the PF1 through PF4 keys on the keypad have any
special meaning; the digits,
the \Text{-} key, and the \Text{,} key are treated just
like the non keypad versions of those characters.
The \Text{Enter} key is treated
exactly like the \Text{Return} key.

The arrow keys may be used for moving around. The left arrow is the
same as  \Name{C-B},
the right arrow is the same as \Name{C-F}, the up arrow is the same
as \Name{C-P}, and the down arrow is the same as \Name{C-N}.

The four function keys are command keys.
The \Text{PF1} key is the same as \Name{M-X}
(the \Text{PF1} key is where the \Text{Gold} key normally
resides on DEC products, so
it is a good choice for the key that asks for extended commands).
The \Text{PF2} key is the same as \Name{C-Q} (quote character).
The \Text{PF3} key is the same as \Name{C-S} (search forward).
The \Text{PF4} key is the same as \Name{C-R} (search reverse). 
These assignments have not been proven optimal, so they may get changed
in the future.
\subsubsection{The LK201 Keyboard}
The escape key, used in all \META{} commands, is very poorly placed on
the LK201 keyboard. To make typing a bit easier, the grave accent is also
considered to be a \META{} prefix on the LK201;
the grave accent is right where you expect the escape key to be located.
A grave accent must be quoted to exter it into the text.

The arrow keys and all of the keys on the keypad
work just like they do on the VT100 keyboard.

The keys immediately above the arrow keys try to do what their name
implies. \Text{Next Screen} is the same as \Name{C-V}.
\Text{Prev Screen} is the same as
\Name{M-V}. \Text{Select} is the same as \Name{C-@} (set mark).
\Text{Remove} is the same as
\Name{C-W} (kill region). \Text{Insert here}
is the same as \Name{C-Y} (yank from killbuffer).
\Text{Find} is the same as \Name{C-S} (search forward).

The \Text{F6}, \Text{F7}, \Text{F8},
\Text{F9}, \Text{F10}, and \Text{F14} keys are unused.
The \Text{F11} is escape, which is a \META{} prefix key.
The \Text{F12} key is backspace. The \Text{F13} key is linefeed.

The \Text{Help} key is unused. The \Text{Do}
key is the same as \Name{C-X E} (execute keyboard macro).

The \Text{F17} key is the same as \Name{C-X P} (previous window).
The \Text{F18} key is the same as \Name{C-X N} (next window).
The \Text{F19} key is the same as \Name{C-X Z} (grow window).
The \Text{F20} key is the same as \Name{C-X C-Z} (shrink window).
\subsection{The Echo Line}
The echo line has two purposes; displaying messages and asking one
line questions.

Two types of messages are displayed in the echo line.
Informational messages tell you something useful, but do not imply in any
way that something is wrong. These messages are always enclosed in square
brackets (\verb+[+ and \verb+]+).
Error messages indicate that something has
prevented a command from being executed. They are never enclosed in square
brackets, and will be accompanied by a beep from the terminal's bell if
the sound of the bell is not too obnoxious.

The echo line is normally cleared by the next keystroke.
However, there is a special hack to clear the echo line without
doing anything else. If you type a \Text{Space} and there is a message in the
echo line, then the echo line is cleared, and the \Text{Space} does not
insert.

The echo line is also used for asking and answering questions.
After the
prompt, you can type any characters you like. The reply is always
terminated by a \Text{Return}. Before you commit to the reply, you can delete
characters by typing \Text{Backspace} or \Text{Rubout}, and you can delete the
entire line by typing \Text{Control-U}.
You can also abort the command in progress by typing \Text{Control-G}.
Command processors are written to ask all questions before doing nasty
things, so that you will never do damage by aborting a command.
\subsection{Command Arguments}
All commands can be given a numeric argument. Most commands use this
argument as a repeat count. Commands can tell if an argument has been
supplied, and will use a default argument if no user argument is present.
Usually, the default argument is \Text{1}.

A \Name{C-U} always introduces an argument.

If a numeric argument is just a string of \Name{C-U} characters, then the
value of the argument is
$4^{(number of \Name{C-U} characters)}$.
Therefore \Name{C-U} is 4, \Name{C-U C-U} is 16,
\Name{C-U C-U C-U} is 256, and so on.
Some commands
care if an argument is present, but don't look at its value; the \Name{C-U}
character is a good way to say ``here is an argument''.

If there is a signed decimal number after the string of
\Name{C-U} characters it specifies the value of the argument. In this case
the number of \Name{C-U} characters does not matter. An argument of 10 can be
represented by \Name{C-U 10},
\Name{C-U C-U 10}, or even \Name{C-U C-U C-U C-U 10}.
\subsection{Dot, Mark, and the Region}
Most editing commands act on the character or characters surrounding
the current location in the current buffer. The thing that marks the current
location is always called ``dot''. The ``dot'' always points between two
characters. This isn't the way that the cursor works on most display terminals,
so, by convention, dot is always immediately to the left of the character
on which the hardware cursor is placed.

There is also a second remembered position in the buffer, called the
mark. There are commands that set the value of mark.

Some commands act on a block of text called the region. The region is
all of the characters between dot and mark. The relative position of dot
and mark do not matter.
\subsection{Word Syntax}
Some MicroEMACS commands manipulate words. In full EMACS the characters
that make up words can be changed by the user. In MicroEMACS, they are
fixed, and include the upper and lower case letters, the dollar sign
\verb+$+ and the underline \verb+_+. This set of characters is intended more
for editing programs, but the word commands still work reasonably
when editing text.
\section{Quitting}
When you finish editing, you have to quit.
MicroEMACS never writes out a file because it thinks that this
is the right thing to do. However, it will attempt to protect you from
forgetting to do so.
\begin{cmdlist}
%% This was used to test the indentation
%% \item[Foobar]
%% \verb+\labelwidth    = + \\
%%	\framebox[\labelwidth]{*}	\\
%% \verb+\leftmargin    = + \\
%%	\framebox[\leftmargin]{*}	\\
%% \verb+\labelsep      = + \\
%%	\framebox[\labelsep]{*}	\\
%% \verb+\listparindent = + \\
%%	\framebox[\listparindent]{*}	\\
\Nameitem{C-X C-C}
This is the basic quit command. MicroEMACS exits, returning
control to the shell. If there are any changed buffers that have not been
written out, it will ask for permission to quit. Supplying an argument to
\Name{C-X C-C} makes the command quit unconditionally, without asking for
confirmation. The value of the argument is not important.
\Nameitem{C-C}
This command suspends the execution of MicroEMACS, and runs
a command interpreter in a subjob. When the subjob terminates, the screen
is cleared and repainted.

The \Name{C-C} command is the same as
\Name{C-X C-C} if the operating system cannot
create command interpreters in subjobs.

Subjobs are implemented in VAX/VMS and MS-DOS. Users of CP/M-86 are
out of luck.
\Nameitem{C-Z}
This is a slightly more elabourate quit command.
If the current buffer has been changed, \Name{C-Z} saves the contents of
the buffer in the associated file (it gets an error if there is no
associated file). If the current buffer has not been changed, \Name{C-Z}
acts like \Name{C-C}.
\end{cmdlist}
\section{Moving Around}
\begin{cmdlist}
 \item[Foobar]
 \verb+\labelwidth    = + \\
	\framebox[\labelwidth]{*}	\\
 \verb+\leftmargin    = + \\
	\framebox[\leftmargin]{*}	\\
 \verb+\labelsep      = + \\
	\framebox[\labelsep]{*}	\\
 \verb+\listparindent = + \\
	\framebox[\listparindent]{*}	\\

\Nameitem{C-A}
Move to the beginning of the current line. Any argument
is ignored. Always succeeds.
\Nameitem{C-B}
Move backwards character by character. The number of characters
to move is specified by the argument. If no argument is given it moves
backwards by 1 character. A newline counts as a single character. Fails
if executed at the beginning of the buffer.
\Nameitem{C-E}
Move to the end of the current line. Any argument is ignored.
Always succeeds.
\Nameitem{C-F}
Move forwards character by character. The number of characters
to move is specified by the argument. If no argument is given it moves
forwards by 1 character. A newline counts as a single character. Fails
if executed at the end of the buffer.
\Nameitem{C-N}
Move forward by lines. Attempt to preserve the current horizontal
position. The number of lines to move is specified by the argument. If no
argument is given it moves by 1 line. Fails if executed at the end of the
buffer.
\Nameitem{C-P}
Move backwards by lines. Attempt to preserve the current horizontal
position. The number of lines to move is specified by the argument. If no
argument is given it moves by 1 line. Fails if executed at the beginning of
the buffer.
\Nameitem{C-V}
Move forward by pages. If an argument is given, it specifies the
number of lines to move. If no argument is given, the size of the window
less 2 lines (for overlap) is used. If the window is too small to allow
2 lines of overlap, it moves by 1 line. Dot is set to the first character
of the first line on the new screen.

There is a compile time option that makes this command take an argument
in screenfulls instead of lines. Look in \Text{ed.h} for the gory details.
\Nameitem{M-V}
Move backwards by pages. If an argument is given, it specifies
the number of lines to move. If no argument is given, the size of the window
less 2 lines (for overlap) is used. If the window is too small to allow
2 lines of overlap, it moves by 1 line. Dot is set to the first character
of the first line on the new screen.

There is a compile time option that makes this command take an argument
in screenfulls instead of lines. Look in \Text{ed.h} for the gory details.
\Nameitem{M-\verb+<+}
Move to the beginning of the buffer. Any argument is ignored.
Dot is set to the first character of the first line in the buffer.
\Nameitem{M-\verb+>+}
Move to the end of the buffer. Any argument is ignored. Dot is
set to the first character in the fake line immediately after the buffer.
The window is set to display dot near the center of the screen.
\end{cmdlist}
\section{Inserting}
All characters between hexadecimal 20 and 7E (blank through tilde),
and all characters between hexadecimal A0 and FE (right hand graphics)
are self inserting. They are inserted into the buffer at the current
location of dot, and dot moves 1 character to the right.

Any self inserting character can be given an argument. This argument is
used as a repeat count, and the character is inserted that number of
times. This is useful for creating lines of \verb+*+ characters of a specific
length, and other pseudo-graphic things.
\begin{cmdlist}
 \item[Foobar]
 \verb+\labelwidth    = + \\
	\framebox[\labelwidth]{*}	\\
 \verb+\leftmargin    = + \\
	\framebox[\leftmargin]{*}	\\
 \verb+\labelsep      = + \\
	\framebox[\labelsep]{*}	\\
 \verb+\listparindent = + \\
	\framebox[\listparindent]{*}	\\

\Nameitem{C-M}
The return key works just like you would expect it to work;
it inserts a newline character. Lines can be split by moving into the middle
of the line, and inserting a newline.

If dot is positioned at end of line,
and the line after the current line is a blank line,
then \Name{C-M} does not insert a newline,
but simply moves to the first position
on the following line. This lets you create a block of blank space (perhaps
using \Name{C-O}) and then type text into it.
\Nameitem{C-O}
This command creates blank lines. To be precise, it inserts
a newline by doing a \Name{C-B},
and then backs up by doing a \Name{C-B}. If dot is at
the start of a line, this will leave dot sitting on the first character
of a new blank line.
\Nameitem{C-Q, M-Q}
Characters which are special to MicroEMACS can be inserted by
using this command.
The next character after the \Name{C-Q} or \Name{M-Q} is stripped of
any special meaning. It is simply inserted into the current buffer.
Any argument specified on the \Name{C-Q}
or \Name{M-Q} command is used as the insert
repeat count.

The\Name{C-Q} form of the command is the easiest to use. However, some terminals
demand that MicroEMACS perform XON/XOFF processing. If this is the case,
the \Name{C-Q} will be eaten by low level terminal support,
and will not be usable
as a command. The \Name{M-Q} form can always be used.
\end{cmdlist}
\section{Deleting and Killing}
There are two general classes of commands that remove text from the
buffer; delete commands and kill commands.
Delete commands remove text from the buffer, and throw it away.
Kill commands remove text from the buffer, but save the text in a special
place called the kill buffer. Kill commands clear the kill buffer only
if the previous command was not a kill command. Multiple kill commands
executed in a row append text to the kill buffer.
\begin{cmdlist}
 \item[Foobar]
 \verb+\labelwidth    = + \\
	\framebox[\labelwidth]{*}	\\
 \verb+\leftmargin    = + \\
	\framebox[\leftmargin]{*}	\\
 \verb+\labelsep      = + \\
	\framebox[\labelsep]{*}	\\
 \verb+\listparindent = + \\
	\framebox[\listparindent]{*}	\\

\Nameitem{C-D}
Delete characters to the right of dot. If an argument is
specified that number of characters is deleted. If no argument is specified
then 1 character is deleted. In addition, if an argument is specified,
then the command kills the text instead of deleting it. It fails if
there are not enough characters to delete between dot and the end of
the buffer.
\Nameitem{Rubout}
Delete characters to the left of dot. If an argument is
specified then that number of characters is deleted. If no argument is
specified then 1 character is deleted. In addition, if an argument is
specified, the rubout command kills the text instead of deleting it.
The command fails if there is not enough text between the start of the
buffer and dot.
\Nameitem{C-K}
This is the basic killing command. If there is no argument
it kills from dot to the end of the line, unless dot is at the
end of line, when it kills the end of line.
If a positive argument is specified, \Name{C-K} kills that many lines,
including the newlines.
If an argument of 0 is specified, \Name{C-K} kills from the start of the current
line to dot.
Finally, if a negative argument is specified, \Name{C-K} kills backwards over
abs(arg) newlines. This command fails if there arn't enough characters
left in the buffer to be killed.
\Nameitem{C-W}
Kill all of the text enclosed in the region. Any argument
is ignored.
\Nameitem{C-Y}
Insert the text in the kill buffer into the current buffer
at dot. If an argument is specified, it specifies the number of times
the text is yanked. If no argument is specified, yank the text back
once. Dot is advanced over the inserted text, as if the text had
been typed in normally. Always succeeds.
\Nameitem{C-X C-O}
... Yes? ...
\end{cmdlist}
\section{Words}
\begin{cmdlist}
\Nameitem{M-B}
The backward word command moves dot backward by words, stopping
on the first character of the word. If an argument is specified, it is the
number of words over which to move. The default argument is 1.
\Nameitem{M-C}
This command moves forward over a word, converting all characters
in the word to lower case except the first one, which is converted to
upper case. If an argument is supplied, it must be positive, and it specifies
the number of words over which to move.
\Nameitem{M-D}
The delete word command moves dot forward by words, and kills
any characters over which it moves. If an argument is specified, it is the
number of words over which to move. The default argument is 1.
\Nameitem{M-F}
The forward word command moves dot forward by words, stopping
on the first non-word character. If an argument is specified, it is the
number of words over which to move. The default argument is 1.
\Nameitem{M-L}
This command moves forward over a word, converting it to lower
case. If an argument is supplied, it must be positive, and it specifies the
number of words over which to move.
\Nameitem{C-M-H}
The backward delete word command moves backward by words, stopping
on the first character of the word, and killing as it moves. If an argument
is present, it specifies the number of words over which to move. The
default argument is 1.
\Nameitem{M-Rubout}
This command is the same as \Name{C-M-H}. It exists only to preserve the
symmetry between the word commands and the character commands (both backspace
and rubout are backward delete character).
\Nameitem{M-U}
This command mobes forward over a word, converting it to upper
case. If an argument is supplied, it must be positive, and it specifies the
number of words over which to move.
\end{cmdlist}
\section{Case Conversion}
In addition to the word mode case conversion commands, MicroEMACS has
commands to modify the case of large blocks of text. These commands should
be used with caution because they cause major damage to (potentially)
large areas of the buffer.
\begin{cmdlist}
\Nameitem{C-X C-L}
The lowercase region command converts all of the characters
between dot and mark into lower case.
\Nameitem{C-X C-U}
The uppercase region command converts all of the characters
between dot and mark into upper case.
\end{cmdlist}
\section{Searching}
Search commands look though the buffer, in either the forward or the reverse
direction, looking for text that matches a search pattern. Search commands
prompt for the search pattern in the echo line. The search pattern used
by the last search command is remembered, and displayed in the prompt. If
you want to use this pattern again, just hit carriage return at the
prompt.

In search strings, all characters stand for themselves, and all searches
are case insensitive. There are no regular expression searches, and there
probably should be. The newline characters at the ends of the lines are
considered to have hexadecimal value 0A, and can be matched by a linefeed
in a search string.
\begin{cmdlist}
\Nameitem{C-S, M-S}
Search forward, from the current location, toward the end of
the buffer. If found, dot is positioned after the matched text. If the
text is not found, dot does not move.

The \Name{C-S} form of the command is not usable if the terminal being used
required XON/XOFF support. In fact, if you use this format of the command
on such a terminal, it will hang until you type \Name{C-Q}.
\Nameitem{C-R, M-R}
Search reverse, from the current location, toward the front of
the buffer. If found, dot is positionedat the first character of the
matched text. If the text is not found, dot does not move.
\end{cmdlist}
\section{Files}
\begin{cmdlist}
\Nameitem{C-X C-F}
This command prompts in the echo line for a file name,
which becomes the new associated file name for the current buffer.
\Nameitem{C-X C-R}
This command prompts in the echo line for a file name,
then it deletes all of the text in the current buffer and reads in the
file. The associated file name is set to the name of the file just
read. The number of lines read is displayed in the echo line.
\Nameitem{C-X C-S}
This command writes the contents of the current buffer
to its associated file. The ``changed'' flag for the current buffer is
reset. It is an error to use this command in a buffer which lacks
an associated file name. This command is a no-operation if the
buffer has not been changed since the last write.
\Nameitem{C-X C-V}
This command selects a file for editing. It prompts for
a file name in the echo line. It then looks through all of the buffers
for a buffer whose associated file name is the same as the file being
selected. If a buffer is found, it just switches to that buffer.
Otherwise it crerates a new buffer, (fabricating a name from the last
part of the new file name), reads the file into it, and switches to the
buffer.

If the desired new buffer name is not unique (perhaps you tried to
visit a file in some other directory with the same name as a file already
read in) the command will prompt for a new buffer name. You can either
supply a buffer name, or just type newline to overwrite the old
buffer.
\Nameitem{C-X C-W}
This command prompts in the echo line for a file name,
then it writes the contents of the current buffer to that file. The
``changed'' flag for the current buffer is reset, and the supplied file
name becomes the associated file name for the current buffer.
\end{cmdlist}
\section{Keyboard Macros}
Keyboard macros simplify a large class of repitious editing tasks.
The basic idea is simple. A set of keystrokes can be collected into a
group, and then the group may be replayed any number of times.

There is only 1 keyboard macro. It would (perhaps) be nice if you could
have more than 1 of them.
\begin{cmdlist}
\Nameitem{C-X (}
This command starts the collection of a keyboard macro. All
keystrokes up to the next \Name{C-X )} will be gathered up,
and may be replayed by
the execute keyboard macro command.
\Nameitem{C-X )}
This command stops the collection of a keyboard macro.
\Nameitem{C-X E}
The execute keyboard macro command replays the current keyboard
macro. If an argument is present, it specifies the number of times the macro
should be executed. If no argument is present, it runs the macro once.
\end{cmdlist}
\section{Buffer Management}
Previous sections have made references to the text in ``the buffer'',
which implied that there is only one buffer. This is not true; MicroEMACS
allows any number of buffers, memory space permitting.

Each buffer has its own buffer name (a 16 character string),
and optional associated file name,
and a block of text. A value of dot and mark is also associated with
any buffer that is not currently being displayed. This remembered value
of dot and mark makes a buffer come back in approximately the same state
as it was when it was hidden.

Also associated with each buffer is a changed flag. This flag is set
when the text in the buffer is modified, and reset when the text in the
buffer is written out to its associated file. MicroEMACS will always ask
for confirmation before executing a command that would cause changed
text to be lost.
\begin{cmdlist}
\Nameitem{C-X C-B}
Create a pop-up window on the screen, and display it in
the name, size (in characters), associated file name, and changed flag
of all buffers. This command works by creating a special buffer which
contains the text of the display, and then selecting it in a window. You
can switch into this window if you like. You can even edit the text.
MicroEMACS makes no attempt to keep a buffer list which is on the screen
updated as other buffers are edited; however, another \Name{C-X C-B} command
will cause the display to be updated in place.
\Nameitem{C-X B}
This command prompts for a buffer name, and then switches
the buffer being displayed in the current window to that buffer. The
buffer will be created if it does not exist.
\Nameitem{C-X K}
This command prompts for a buffer name, and then destroys
the buffer with that name. It will ask for permission to destroy the
buffer if the text has been changed since it was written to the associated
file. You cannot delete a buffer that is being displayed.
\end{cmdlist}
\section{Window Management}
MicroEMACS lets you have multiple windows on the screen.
Each window has its own mode line, its own value of dot and mark,
and its own associated buffer. If you have a buffer displayed in more
that one window and you edit a line, it is updated in all windows.
\begin{cmdlist}
\Nameitem{C-L}
This command clears the screen, and completely redisplays all
of the text is all of the windows. It is useful if a line error has garbaged
your screen. If you give \Name{C-L} and argument
(any value will do) it will re-center
dot in the current window as well.
\Nameitem{C-X 2}
This is the basic window making command. The current window
is split in two. Each window is displaying the same buffer, and has
the same value of dot and mark. The window must be at least three lines
high, otherwise there isn't enough room for two text lines and the
new mode line. Any argument is ignored.

After the window is created, it is usually switched to a private
buffer using \Name{C-X B} or (perhaps more often) \Name{C-X C-V}.
Note that because
both windows are displaying the same buffer, reading a file in one
window with \Name{C-X C-R} probably does not do what you want.
\Nameitem{C-X 1}
This is the basic window destroying command. All windows but
the current window are removed from the screen. The current window grows
to fill the vacated screen space. Any argument is ignored.
\Nameitem{C-X N}
Move to the next window down the screen. If the current window
is the bottom window, move to the top window. Ignores any argument.
\Nameitem{C-X P}
Move to the previous window up the screen. If the current window
is the top window, move to the bottom window. Ignores any argument.
\Nameitem{C-X Z}
The current window is enlarged, if possible. Any argument is
used as a ``number of lines by which to grow''. The default argument is 1.
Screen space is stolen from the window immediately below the current
window. If the current window is the bottom window on the screen then
space is stolen from the window immediately above the current window.
You cannot steal all of the lines away from a window.
\Nameitem{C-X C-N}
Scroll the current window down. Any argument is used as a
``number of lines by which to scroll'' count. The default argument is 1.
\Nameitem{C-X C-P}
Scroll the current window up. Any argument is used as a
``number of lines by which to scroll'' count. The default argument is 1.
\Nameitem{C-X C-Z}
The current window is shrunk, if possible. Any argument is
used as a ``number of lines by which to shrink''. The default argument is 1.
Screen space is given to the window immediately below the current window.
If the current window is the bottom window on the screen, the space is
given to the window immediately above the current window. You cannot shrink
a window out of existance.
\Nameitem{M-!}
This command is used to control the line of a window upon
which dot is displayed. If its argument is positive, then that number is
taken to be the origin 1 line number of the current window upon which
the line containing dot should be placed. If its argument is 0, then dot
is moved to the center of the window. If the number is less then zero
then it is taken to be the negation of the origin 1 line number of the
current window starting at the bottom upon which the line containing dot
should be placed. If no argument is supplied, a default argument of 1 is
used; This lets \Name{M-!} function as a ``move dot to the top of the window''
command, which is very useful.
\end{cmdlist}
\section{Building a MicroEMACS}
All versions of MicroEMACS are built from the same set of
source files. Compile time options are selected by setting compilation
switches in \Text{ed.h}, and then letting conditional compilation do the
right thing.
\subsection{Operating System}
The three switches VMS, CPM, and MSDOS select the operating system.
Only one of these switches should be set to \Text{1} at any time. If you set
more than one of them to \Text{1}, you will get errors when you compile.

To date, only the \Text{termio.c} (low level terminal I/O) and \Text{spawn.c}
(subjob creation) contain code conditionalized on the operating system.
Adding a new operating system consists mostly of adding the appropriate
magic incantations to these two files.
\subsection{Terminal Support}
MicroEMACS currently supports two kinds of terminals. The \Text{ANSI}
compilation switch selects support for ANSI compatable terminals. The
\Text{VT52} compilation switch selects support for VT52 like terminals (this
includes the Zenith Z19). You should not set both of these.

The \Text{VT100} and \Text{LK201} compilation switches,
if \Text{1}, select keyboard
handling appropriate to VT100 and LK201 keyboards, respectfully. The
LK201 support has only been tested on the Rainbow.

The \Text{RAINBOW} compilation switch, when set, modifies the display
code to use fast video.
A special routine, written in assembler, copies data from the MicroEMACS
screen image directly into video RAM. This routine is guaranteed to
do bad things on anything but a Rainbow. On the Rainbow, you still need
to set \Text{ANSI}, and you will want to set \Text{LK201}.
\subsection{Randomness}
In ITS EMACS, arguments on the \Name{C-V}
and \Name{M-V} commands work in lines. In
Gosling EMACS, arguments on the \Name{C-V}
and \Name{M-V} commands work in screens.
The \Text{CVMVAS} compilation switch, if \Text{1}, makes the commands work
like in Gosling EMACS.
\end{document}
SHAR_EOF
if test 36749 -ne "`wc -c < 'uemacs.tex'`"
then
	echo shar: error transmitting "'uemacs.tex'" '(should have been 36749 characters)'
fi
fi
exit 0
#	End of shell archive