mirror of
https://gitlab.com/freepascal.org/fpc/source.git
synced 2025-05-19 17:42:36 +02:00
754 lines
23 KiB
TeX
754 lines
23 KiB
TeX
% \begin{meta-comment}
|
|
%
|
|
% $Id$
|
|
%
|
|
% Allow @-commands
|
|
%
|
|
% (c) 1995 Mark Wooding
|
|
%
|
|
%----- Revision history -----------------------------------------------------
|
|
%
|
|
% $Log$
|
|
% Revision 1.1 1998-09-21 10:18:06 michael
|
|
% Initial implementation
|
|
%
|
|
% Revision 1.3 1996/11/19 20:46:55 mdw
|
|
% Entered into RCS
|
|
%
|
|
%
|
|
% \end{meta-comment}
|
|
%
|
|
% \begin{meta-comment} <general public licence>
|
|
%%
|
|
%% at package -- support for `@' commands'
|
|
%% Copyright (c) 1996 Mark Wooding
|
|
%%
|
|
%% This program is free software; you can redistribute it and/or modify
|
|
%% it under the terms of the GNU General Public License as published by
|
|
%% the Free Software Foundation; either version 2 of the License, or
|
|
%% (at your option) any later version.
|
|
%%
|
|
%% This program is distributed in the hope that it will be useful,
|
|
%% but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
%% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
%% GNU General Public License for more details.
|
|
%%
|
|
%% You should have received a copy of the GNU General Public License
|
|
%% along with this program; if not, write to the Free Software
|
|
%% Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|
%%
|
|
% \end{meta-comment}
|
|
%
|
|
% \begin{meta-comment} <Package preamble>
|
|
%<+package>\NeedsTeXFormat{LaTeX2e}
|
|
%<+package>\ProvidesPackage{at}
|
|
%<+package> [1996/05/02 1.3 @-command support (MDW)]
|
|
% \end{meta-comment}
|
|
%
|
|
% \CheckSum{355}
|
|
%% \CharacterTable
|
|
%% {Upper-case \A\B\C\D\E\F\G\H\I\J\K\L\M\N\O\P\Q\R\S\T\U\V\W\X\Y\Z
|
|
%% Lower-case \a\b\c\d\e\f\g\h\i\j\k\l\m\n\o\p\q\r\s\t\u\v\w\x\y\z
|
|
%% Digits \0\1\2\3\4\5\6\7\8\9
|
|
%% Exclamation \! Double quote \" Hash (number) \#
|
|
%% Dollar \$ Percent \% Ampersand \&
|
|
%% Acute accent \' Left paren \( Right paren \)
|
|
%% Asterisk \* Plus \+ Comma \,
|
|
%% Minus \- Point \. Solidus \/
|
|
%% Colon \: Semicolon \; Less than \<
|
|
%% Equals \= Greater than \> Question mark \?
|
|
%% Commercial at \@ Left bracket \[ Backslash \\
|
|
%% Right bracket \] Circumflex \^ Underscore \_
|
|
%% Grave accent \` Left brace \{ Vertical bar \|
|
|
%% Right brace \} Tilde \~}
|
|
%%
|
|
%
|
|
% \begin{meta-comment} <driver>
|
|
%
|
|
%<*driver>
|
|
\input{mdwtools}
|
|
\describespackage{at}
|
|
\aton
|
|
\atlet p=\package
|
|
\atdef at{\package{at}}
|
|
\atdef={\mbox{-}}
|
|
\atdef-{@@@=}
|
|
\atlet.=\syntax
|
|
\mdwdoc
|
|
%</driver>
|
|
%
|
|
% \end{meta-comment}
|
|
%
|
|
% \section{User guide}
|
|
%
|
|
% The @at\ package is an attempt to remove a lot of tedious typing that
|
|
% ends up in \LaTeX\ documents, by expanding the number of short command
|
|
% names available. The new command names begin with the `|@|' character,
|
|
% rather than the conventional `|\|', so you can tell them apart.
|
|
%
|
|
% The package provides some general commands for defining @-commands, and
|
|
% then uses them to define some fairly simple ones which will be useful to
|
|
% most people.
|
|
%
|
|
% The rules for @-command names aren't terribly complex:
|
|
% \begin{itemize}
|
|
% \item If the first character of the name is a letter, then the command name
|
|
% consists of all characters up to, but not including, the first
|
|
% nonletter. Spaces following the command name are ignored.
|
|
% \item If the first character of the name is a backslash, then the @-command
|
|
% name consists of the control sequence introduced by the backslash.
|
|
% \item Otherwise, the command name consists only of that first character.
|
|
% Spaces following the name are not ignored, unless that character
|
|
% was itself a space character.
|
|
% \end{itemize}
|
|
%
|
|
% Usually, digits are not considered to be letters. However, the
|
|
% \package{at} package will consider digits to be letters if you give it the
|
|
% \textsf{digits} option in the |\usepackage| command. (Note that this
|
|
% only affects the \package{at} package; it won't change the characters
|
|
% allowed in normal command names.)
|
|
%
|
|
% \DescribeMacro{\atallowdigits}
|
|
% \DescribeMacro{\atdisallowdigits}
|
|
% You can enable and disable digits being considered as letters dynamically.
|
|
% The |\atallowdigits| command allows digits to be used as letters;
|
|
% |\atdisallowdigits| prevents this. Both declarations follow \LaTeX's
|
|
% usual scoping rules. Both of these commands have corresponding
|
|
% environments with the same names (without the leading `|\|', obviously).
|
|
%
|
|
% \subsection{Defining @-commands}
|
|
%
|
|
% \DescribeMacro{\newatcommand}
|
|
% \DescribeMacro{\renewatcommand}
|
|
% The |\newatcommand| command will define a new @-command using a syntax
|
|
% similar to |\newcommand|. For example, you could define
|
|
% \begin{listing}
|
|
%\newatcommand c[1]{\chapter{#1}}
|
|
% \end{listing}
|
|
% to make @.{"@c{"<name>"}"} equivalent to @.{"\\chapter{"<name>"}"}.
|
|
%
|
|
% A |\renewatcommand| is also provided to redefine existing commands, should
|
|
% the need arise.
|
|
%
|
|
% \DescribeMacro{\atdef}
|
|
% For \TeX\ hackers, the |\atdef| command defines @-commands using a syntax
|
|
% similar to \TeX's built-in |\def|.
|
|
%
|
|
% As an example, the following command makes @.{"@/"<text>"/"} write its
|
|
% argument \<text> in italics:
|
|
% \begin{listing}
|
|
%\atdef/#1/{\textit{#1}}
|
|
% \end{listing}
|
|
% The real implementation of the |@/|\dots|/| command is a bit more
|
|
% complex, and is given in the next section.
|
|
%
|
|
% You can use all of \TeX's features for defining the syntax of your
|
|
% command. (See chapter~20 of @/The \TeX book/ for more details.)
|
|
%
|
|
% \DescribeMacro{\atlet}
|
|
% Since |\atdef| is provided to behave similarly to |\def|, @at\ provides
|
|
% |\atlet| which works similarly to |\let|. For example you can say
|
|
% \begin{listing}
|
|
%\atlet!=\index
|
|
% \end{listing}
|
|
% to allow the short |@!| to behave exactly like |\index|.
|
|
%
|
|
% Note that all commands defined using these commands are robust even if you
|
|
% use fragile commands in their definitions. Unless you start doing very
|
|
% strange things, @-commands never need |\protect|ing.
|
|
%
|
|
% \subsection{Predefined @-commands}
|
|
%
|
|
% A small number of hopefully useful commands are provided by default.
|
|
% These are described in the table below:
|
|
%
|
|
% \bigskip \begin{center} \begin{tabular}{lp{3in}} \hline
|
|
% \bf Command & \bf Meaning \\ \hline
|
|
% @.{"@@"} & Typesets an `@@' character. \\
|
|
% @.{"@/"<text>"/"} & In text (LR or paragraph) mode, typesets its
|
|
% argument emphasised. In maths mode, it
|
|
% always chooses italics. \\
|
|
% @.{"@*"<text>"*"} & Typesets its argument \<text> in bold. \\
|
|
% @.{"@i{"<text>"}"} & Equivalent to `@.{"\\index{"<text>"}"}'. \\
|
|
% @.{"@I{"<text>"}"} & As for |@i|, but also writes its argument
|
|
% to the document. \\ \hline
|
|
% \end{tabular} \end{center} \bigskip
|
|
%
|
|
% Package writers should not rely on any predefined @-commands -- they're
|
|
% provided for users, and users should be able to redefine them without
|
|
% fear of messing anything up. (This includes the `standard' commands
|
|
% provided by the @at\ package, by the way. They're provided in the vague
|
|
% hope that they might be useful, and as examples.)
|
|
%
|
|
% \implementation
|
|
%
|
|
% \section{Implementation}
|
|
%
|
|
% \begin{macrocode}
|
|
%<*package>
|
|
% \end{macrocode}
|
|
%
|
|
% \subsection{Options handling}
|
|
%
|
|
% We need a switch to say whether digits should be allowed. Since this
|
|
% is a user thing, I'll avoid |\newif| and just define the thing by hand.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\atallowdigits{\let\ifat@digits\iftrue}
|
|
\def\atdisallowdigits{\let\ifat@digits\iffalse}
|
|
% \end{macrocode}
|
|
%
|
|
% Now define the options.
|
|
%
|
|
% \begin{macrocode}
|
|
\DeclareOption{digits}{\atallowdigits}
|
|
\DeclareOption{nodigits}{\atdisallowdigits}
|
|
\ExecuteOptions{nodigits}
|
|
\ProcessOptions
|
|
% \end{macrocode}
|
|
%
|
|
% \subsection{How the commands work}
|
|
%
|
|
% Obviously we make the `@@' character active. It inspects the next
|
|
% character (or argument, actually -- it can be enclosed in braces for
|
|
% longer commands, although this is a bit futile), and builds the command
|
|
% name from that.
|
|
%
|
|
% The |\at| command is equivalent to the active `@@' character always.
|
|
%
|
|
%
|
|
% \subsection{Converting command names}
|
|
%
|
|
% We need to be able to read an @-command name, and convert it to a normal
|
|
% \TeX\ control sequence. First, we declare some control sequences for
|
|
% braces, which we need later.
|
|
%
|
|
% \begin{macrocode}
|
|
\begingroup
|
|
\catcode`\<1
|
|
\catcode`\>2
|
|
\catcode`\{12
|
|
\catcode`\}12
|
|
\gdef\at@lb<{>
|
|
\gdef\at@rb<}>
|
|
\gdef\at@spc< >
|
|
\endgroup
|
|
% \end{macrocode}
|
|
%
|
|
% I'll set up some helper routines now, to help me read the command
|
|
% names. The way this works is that we |\futurelet| the token into
|
|
% |\@let@token|. These routines will then sort out what to do next.
|
|
%
|
|
% \begin{macro}{\at@test}
|
|
%
|
|
% Given an |\if|\dots\ test, does its first or second argument.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@test#1\then{%
|
|
#1\expandafter\@firstoftwo\else\expandafter\@secondoftwo\fi%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\at@ifcat}
|
|
%
|
|
% Checks the category code of the current character. If it matches the
|
|
% argument, it does its second argument, otherwise it does the third.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@ifcat#1{\at@test\ifcat#1\noexpand\@let@token\then}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\at@ifletter}
|
|
%
|
|
% This routine tests the token to see if it's a letter, and if so adds
|
|
% it to the token list and does the first argument; otherwise it does the
|
|
% second argument. It accepts digits as letters if the switch is turned
|
|
% on.
|
|
%
|
|
% There's some fun later, so I'll describe this slowly. First, we compare
|
|
% the category code to a letter, and if we have a match, we know we're done;
|
|
% we need to pick up the letter as an argument. If the catcode is `other',
|
|
% we must compare with numbers to see if it's in range.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@ifletter#1#2{%
|
|
\at@ifcat x%
|
|
{\at@ifletter@ii{#1}}%
|
|
{\at@ifcat 0%
|
|
{\at@ifletter@i{#1}{#2}}%
|
|
{#2}%
|
|
}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Right. It's `other' (so it's safe to handle as a macro argument) and we
|
|
% need to know if it's a digit. This is a little tricky: I use |\if| to
|
|
% compare two characters. The first character is~`1' or~`0' depending on the
|
|
% `digit' switch; the second is~`1' or~`x' depending on whether it's actually
|
|
% a digit. They'll only match if everything's worked out OK.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@ifletter@i#1#2#3{%
|
|
\at@test\if%
|
|
\ifat@digits1\else0\fi%
|
|
\ifnum`#3<`0x\else\ifnum`#3>`9x\else1\fi\fi%
|
|
\then%
|
|
{\at@ifletter@ii{#1}{#3}}%
|
|
{#2#3}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Right; we have the character, so add it to the list and carry on.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@ifletter@ii#1#2{\toks@\expandafter{\the\toks@#2}#1}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% Now we define the command name reading routines. We have @/almost/ the
|
|
% same behaviour as \TeX, although we can't support `|%|' characters for
|
|
% reasons to do with \TeX's tokenising algorithm.
|
|
%
|
|
% \begin{macro}{\at@read@name}
|
|
%
|
|
% The routine which actually reads the command name works as follows:
|
|
% \begin{enumerate}
|
|
% \item Have a peek at the next character. If it's a left or right brace,
|
|
% then use the appropriate character.
|
|
% \item If the character is not a letter, just use the character (or whole
|
|
% control sequence.
|
|
% \item Finally, if it's a letter, keep reading letters until we find one
|
|
% that wasn't.
|
|
% \end{enumerate}
|
|
%
|
|
% First, we do some setting up and read the first character
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@read@name#1{%
|
|
\let\at@next=#1%
|
|
\toks@{}%
|
|
\futurelet\@let@token\at@rn@i%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Next, sort out what to do, based on the category code.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@rn@i{%
|
|
\def\@tempa{\afterassignment\at@rn@iv\let\@let@token= }%
|
|
\at@ifletter%
|
|
{\futurelet\@let@token\at@rn@iii}%
|
|
{\at@ifcat\bgroup%
|
|
{\toks@\expandafter{\at@lb}\@tempa}%
|
|
{\at@ifcat\egroup%
|
|
{\toks@\expandafter{\at@rb}\@tempa}%
|
|
{\at@ifcat\at@spc%
|
|
{\toks@{ }\@tempa}%
|
|
{\at@rn@ii}%
|
|
}%
|
|
}%
|
|
}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Most types of tokens can be fiddled using |\string|.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@rn@ii#1{%
|
|
\toks@\expandafter{\string#1}%
|
|
\at@rn@iv%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% We've found a letter, so we should check for another one.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@rn@iii{%
|
|
\at@ifletter%
|
|
{\futurelet\@let@token\at@rn@iii}%
|
|
{\@ifnextchar.\at@rn@iv\at@rn@iv}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Finally, we need to pass the real string, as an argument, to the
|
|
% macro. We make |\@let@token| relax, since it might be something which will
|
|
% upset \TeX\ later, e.g., a |#| character.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@rn@iv{%
|
|
\let\@let@token\relax%
|
|
\expandafter\at@next\csname at.\the\toks@\endcsname%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\at@cmdname}
|
|
%
|
|
% Given a control sequence, work out which @-command it came from.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@cmdname#1{\expandafter\at@cmdname@i\string#1\@@foo}
|
|
% \end{macrocode}
|
|
%
|
|
% Now extract the trailing bits.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@cmdname@i#1.#2\@@foo{#2}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\at@decode}
|
|
%
|
|
% The |\at@decode| macro takes an extracted @-command name, and tries to
|
|
% execute the correct control sequence derived from it.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@decode#1{%
|
|
\at@test\ifx#1\relax\then{%
|
|
\PackageError{at}{Unknown @-command `@\at@cmdname#1'}{%
|
|
The @-command you typed wasn't recognised, so I've ignored it.
|
|
}%
|
|
}{%
|
|
#1%
|
|
}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\@at}
|
|
%
|
|
% We'd like a measure of compatibility with @p{amsmath}. The @-commands
|
|
% provided by @p{amsmath} work only in maths mode, so this gives us a way of
|
|
% distinguishing. If the control sequence |\Iat| is defined, and we're in
|
|
% maths mode, we'll call that instead of doing our own thing.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\@at{%
|
|
\def\@tempa{\at@read@name\at@decode}%
|
|
\ifmmode\ifx\Iat\not@@defined\else%
|
|
\let\@tempa\Iat%
|
|
\fi\fi%
|
|
\@tempa%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
%
|
|
% \subsection{Defining new commands}
|
|
%
|
|
% \begin{macro}{\at@buildcmd}
|
|
%
|
|
% First, we define a command to build these other commands:
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@buildcmd#1#2{%
|
|
\expandafter\def\csname\expandafter
|
|
\@gobble\string#1@decode\endcsname##1{#2##1}%
|
|
\edef#1{%
|
|
\noexpand\at@read@name%
|
|
\expandafter\noexpand%
|
|
\csname\expandafter\@gobble\string#1@decode\endcsname%
|
|
}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\newatcommand}
|
|
% \begin{macro}{\renewatcommand}
|
|
% \begin{macro}{\provideatcommand}
|
|
% \begin{macro}{\atdef}
|
|
% \begin{macro}{\atshow}
|
|
%
|
|
% Now we define the various operations on @-commands.
|
|
%
|
|
% \begin{macrocode}
|
|
\at@buildcmd\newatcommand\newcommand
|
|
\at@buildcmd\renewatcommand\renewcommand
|
|
\at@buildcmd\provideatcommand\providecommand
|
|
\at@buildcmd\atdef\def
|
|
\at@buildcmd\atshow\show
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
% \end{macro}
|
|
% \end{macro}
|
|
% \end{macro}
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\atlet}
|
|
%
|
|
% |\atlet| is rather harder than the others, because we want to allow people
|
|
% to say things like @.{"\\atlet"<name>"=@"<name>}. The following hacking
|
|
% does the trick. I'm trying very hard to duplicate |\let|'s behaviour with
|
|
% respect to space tokens here, to avoid any surprises, although there
|
|
% probably will be some differences. In particular, |\afterassignment|
|
|
% won't work in any sensible way.
|
|
%
|
|
% First, we read the name of the @-command we're defining. We also open
|
|
% a group, to stop messing other people up, and make `@@' into an `other'
|
|
% token, so that it doesn't irritatingly look like its meaning as a control
|
|
% sequence.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\atlet{%
|
|
\begingroup%
|
|
\@makeother\@%
|
|
\at@read@name\atlet@i%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Put the name into a scratch macro for later use. Now see if there's an
|
|
% equals sign up ahead. If not, this will gobble any spaces in between the
|
|
% @-command name and the argument.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\atlet@i#1{%
|
|
\def\at@temp{#1}%
|
|
\@ifnextchar=\atlet@ii{\atlet@ii=}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Now we gobble the equals sign (whatever catcode it is), and peek at the
|
|
% next token up ahead using |\let| with no following space.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\atlet@ii#1{\afterassignment\atlet@iii\global\let\at@gnext=}
|
|
% \end{macrocode}
|
|
%
|
|
% The control sequence |\at@gnext| is now |\let| to be whatever we want the
|
|
% @-command to be, unless it's picked up an `@@' sign. If it has, we've
|
|
% eaten the |@| token, so just read the name and pass it on. Otherwise,
|
|
% we can |\let| the @-command directly to |\at@gnext|. There's some
|
|
% nastiness here to make |\the\toks@| expand before we close the group and
|
|
% restore its previous definition.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\atlet@iii{%
|
|
\if @\noexpand\at@gnext%
|
|
\expandafter\at@read@name\expandafter\atlet@iv%
|
|
\else%
|
|
\expandafter\endgroup%
|
|
\expandafter\let\at@temp= \at@gnext%
|
|
\fi%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% We've read the source @-command name, so just copy the definitions over.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\atlet@iv#1{%
|
|
\expandafter\endgroup%
|
|
\expandafter\let\at@temp=#1%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
%
|
|
% \subsection{Robustness of @-commands}
|
|
%
|
|
% We want all @-commands to be robust. We could leave them all being
|
|
% fragile, although making robust @-commands would then be almost impossible.
|
|
% There are two problems which we must face:
|
|
%
|
|
% \begin{itemize}
|
|
%
|
|
% \item The `|\@at|' command which scans the @-command name is (very)
|
|
% fragile. I could have used |\DeclareRobustCommand| for it (and in
|
|
% fact I did in an earlier version), but that doesn't help the other
|
|
% problem at all.
|
|
%
|
|
% \item The `name' of the @-command may contain active characters or control
|
|
% sequences, which will be expanded at the wrong time unless we do
|
|
% something about it now.
|
|
%
|
|
% \end{itemize}
|
|
%
|
|
% We must also be careful not to introduce extra space characters into any
|
|
% files written, because spaces are significant in @-commands. Finally,
|
|
% we have a minor problem in that most auxiliary files are read in with
|
|
% the `@@' character set to be a letter.
|
|
%
|
|
% \begin{macro}{\at}
|
|
%
|
|
% Following the example of \LaTeX's `short' command handling, we'll define
|
|
% |\at| to decide what to do depending on what |\protect| looks like. If
|
|
% we're typesetting, we just call |\@at| (above) and expect it to cope.
|
|
% Otherwise we call |\at@protect|, which scoops up the |\fi| and the |\@at|,
|
|
% and inserts other magic.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at{\ifx\protect\@typeset@protect\else\at@protect\fi\@at}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\at@protect}
|
|
%
|
|
% Since we gobbled the |\fi| from the above, we must put that back. We then
|
|
% need to do things which are more complicated. If |\protect| is behaving
|
|
% like |\string|, then we do one sort of protection. Otherwise, we assume
|
|
% that |\protect| is being like |\noexpand|.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@protect\fi#1{%
|
|
\fi%
|
|
\ifx\protect\string%
|
|
\expandafter\at@protect@string%
|
|
\else%
|
|
\expandafter\at@protect@noexpand%
|
|
\fi%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\at@protect@string}
|
|
%
|
|
% When |\protect| is |\string|, we don't need to be able to recover the
|
|
% original text particularly accurately -- it's for the user to look at.
|
|
% Therefore, we just output a $|@|_{11}$ and use |\string| on the next
|
|
% token. This must be sufficient, since we only allow multi-token command
|
|
% names if the first token is a letter (code~11).
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@protect@string{@\string}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\at@protect@noexpand}
|
|
%
|
|
% This is a little more complex, since we're still expecting to be executed
|
|
% properly at some stage. However, there's a cheeky dodge we can employ
|
|
% since the |\at| command is thoroughly robustified (or at least it will be
|
|
% by the time we've finished this). All |\@unexpandable@protect| does
|
|
% is confer repeated robustness on a fragile command. Since our command
|
|
% is robust, we don't need this and we can get away with just using a
|
|
% single |\noexpand|, both for the |\@at@| command and the following token
|
|
% (which we must robustify, because no-one else can do it for us -- if
|
|
% anyone tries, they end up using the |@\protect| command which is rather
|
|
% embarassing).
|
|
%
|
|
% I'll give the definition, and then examine how this expands in various
|
|
% cases.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\at@protect@noexpand{\noexpand\@at@ @\noexpand}
|
|
\def\@at@#1{\at}
|
|
% \end{macrocode}
|
|
%
|
|
% A few points, before we go into the main examination of the protection.
|
|
% I've inserted a $|@|_{11}$ token, which is gobbled by |\@at@| when the
|
|
% thing is finally expanded fully. This prevents following space tokens
|
|
% in an |\input| file from being swallowed because they follow a control
|
|
% sequence. (I can't use the normal $|@|_{13}$ token, because when files
|
|
% like the |.aux| file are read in, |@| is given code~11 by
|
|
% |\makeatletter|.)
|
|
%
|
|
% \setbox0\hbox{|@at@|}
|
|
% Now for a description of why this works. When |\at| is expanded, it works
|
|
% out that |\protect| is either |\noexpand| or |\@unexpandable@protect|, and
|
|
% becomes |\at@protect@noexpand|. Because of the |\noexpand| tokens, this
|
|
% stops being expanded once it reaches $\fbox{\box0}\,|@|_{11}\,x$ (where
|
|
% $x$ is the token immediately following the $|@|_{13}$ character). If this
|
|
% is expanded again, for example in another |\edef|, or in a |\write| or a
|
|
% |\mark|, the |\@at@| wakes up, gobbles the following |@| (whatever catcode
|
|
% it is -- there may be intervening |\write| and |\input| commands) and
|
|
% becomes |\at|, and the whole thing can start over again.
|
|
%
|
|
% \end{macro}
|
|
%
|
|
%
|
|
% \subsection{Enabling and disabling @-commands}
|
|
%
|
|
% \begin{macro}{\aton}
|
|
%
|
|
% We define the |\aton| command to enable all of our magic. We store
|
|
% the old catcode in the |\atoff| command, make `@@' active, and make it
|
|
% do the stuff.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\aton{%
|
|
\ifnum\catcode`\@=\active\else%
|
|
\edef\atoff{\catcode`\noexpand\@\the\catcode`\@}%
|
|
\catcode`\@\active%
|
|
\lccode`\~`\@%
|
|
\lowercase{\let~\at}%
|
|
\fi%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\atoff}
|
|
%
|
|
% The |\atoff| command makes `@@' do the stuff it's meant to. We remember
|
|
% the old catcode and revert to it. This is largely unnecessary.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\atoff{\catcode`\@12}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\makeatother}
|
|
%
|
|
% Now we make our active `@@' the default outside of package files.
|
|
%
|
|
% \begin{macrocode}
|
|
\let\makeatother\aton
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% And we must make sure that the user can use all of our nice commands.
|
|
% Once the document starts, we allow @-commands.
|
|
%
|
|
% \begin{macrocode}
|
|
\AtBeginDocument{\aton}
|
|
% \end{macrocode}
|
|
%
|
|
% \begin{macro}{\dospecials}
|
|
% \begin{macro}{\@sanitize}
|
|
%
|
|
% We must add the `@@' character to the various specials lists.
|
|
%
|
|
% \begin{macrocode}
|
|
\expandafter\def\expandafter\dospecials\expandafter{\dospecials\do\@}
|
|
\expandafter\def\expandafter\@sanitize\expandafter{%
|
|
\@sanitize\@makeother\@}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
% \end{macro}
|
|
%
|
|
% \subsection{Default @-commands}
|
|
%
|
|
% We define some trivial examples to get the user going.
|
|
%
|
|
% \begin{macrocode}
|
|
\expandafter\chardef\csname at.@\endcsname=`\@
|
|
\atdef*#1*{\ifmmode\mathbf{#1}\else\textbf{#1}\fi}
|
|
\atdef/#1/{\ifmmode\mathit{#1}\else\emph{#1}\fi}
|
|
\atlet i=\index
|
|
\atdef I#1{#1\index{#1}}
|
|
%</package>
|
|
% \end{macrocode}
|
|
%
|
|
% \hfill Mark Wooding, \today
|
|
%
|
|
% \Finale
|
|
%
|
|
\endinput
|