mirror of
https://gitlab.com/freepascal.org/fpc/source.git
synced 2025-05-28 22:42:57 +02:00
1190 lines
34 KiB
TeX
1190 lines
34 KiB
TeX
% \begin{meta-comment}
|
|
%
|
|
% $Id$
|
|
%
|
|
% Verbatim typesetting done properly (ahem)
|
|
%
|
|
% (c) 1996 Mark Wooding
|
|
%
|
|
%----- Revision history -----------------------------------------------------
|
|
%
|
|
% $Log$
|
|
% Revision 1.1 1998-09-21 10:19:01 michael
|
|
% Initial implementation
|
|
%
|
|
% Revision 1.3 1996/11/19 21:01:18 mdw
|
|
% Entered into RCS
|
|
%
|
|
%
|
|
% \end{meta-comment}
|
|
%
|
|
% \begin{meta-comment} <general public licence>
|
|
%%
|
|
%% sverb package -- handling of verbatim text
|
|
%% 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{sverb}
|
|
%<+package> [1996/05/08 1.3 Verbatim typesetting]
|
|
% \end{meta-comment}
|
|
%
|
|
% \CheckSum{651}
|
|
%% \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>
|
|
\input{mdwtools}
|
|
\describespackage{sverb}
|
|
\mdwdoc
|
|
%</driver>
|
|
%
|
|
% \end{meta-comment}
|
|
%
|
|
% \section{User guide}
|
|
%
|
|
% The \package{sverb} package provides some useful commands and environments
|
|
% for doing things with verbatim text. I prefer this code to the standard
|
|
% \package{verbatim} package (by Rainer Sch\"opf et al.)\ although I'm
|
|
% biased.
|
|
%
|
|
% The package was written to fulfil a particular purpose: I wanted to be able
|
|
% to typeset ARM assembler code, 77~columns wide, on A5~paper, with the
|
|
% fields separated by \textit{tab} characters. It's grown up fairly
|
|
% organically from that, and I've tidied it when I've seen the code get too
|
|
% ugly.
|
|
%
|
|
% The current features are:
|
|
%
|
|
% \begin{itemize}
|
|
%
|
|
% \item A `listing' environment which typesets verbatim text nicely.
|
|
%
|
|
% \item A command to read verbatim text from an external file.
|
|
%
|
|
% \item Support for arbitrary-sized chunks of text without overflowing \TeX's
|
|
% memory.
|
|
%
|
|
% \item Support for \textit{tab} characters in the verbatim text.
|
|
%
|
|
% \item An environment for typesetting demonstrations of \LaTeX\ markup.
|
|
%
|
|
% \item It all works correctly with the \package{doc} system for documenting
|
|
% \LaTeX\ packages.
|
|
%
|
|
% \item A fairly hairy but quite powerful programmer interface to the yukky
|
|
% bits of the package.
|
|
%
|
|
% \end{itemize}
|
|
%
|
|
% The interface is described in its own section, so that more timid readers
|
|
% can avoid it. That said, some of the stuff in this section gets rather
|
|
% technical.
|
|
%
|
|
% Note that this package doesn't even try to do anything with short bits of
|
|
% verbatim text (as handled by the |\verb:...:| command). I have a separate
|
|
% package (\package{syntax}) which does all sorts of horrible things along
|
|
% those lines.
|
|
%
|
|
% \subsection{The \env{listing} environment}
|
|
%
|
|
% \DescribeEnv{listing}
|
|
% The main method for typesetting verbatim text is the \env{listing}
|
|
% environment. This works pretty much the same as the standard
|
|
% \env{verbatim} environment, with some exceptions, which are described
|
|
% below.
|
|
%
|
|
% So that you know exactly what you're getting, here are the rules by which
|
|
% \package{sverb} decides what the verbatim text actually is:
|
|
%
|
|
% \begin{itemize}
|
|
%
|
|
% \item If there's any text, other than spaces, on the same line as the
|
|
% `|\begin{listing}|', then the contents of the environment begins
|
|
% immediately after the closing brace (with all leading spaces
|
|
% preserved). Otherwise, the text begins on the following line.
|
|
%
|
|
% \item If there is any text, other than spaces, before the
|
|
% `|\end{listing}|', but on the same line, this is considered to be the
|
|
% last line of the text; otherwise the text is presumed to have ended
|
|
% at the end of the previous line.
|
|
%
|
|
% \item Any text following the |\end{listing}| on the same line is thrown
|
|
% away. There are good reasons for this, but they're technical.
|
|
% Essentially there's nothing I can do about it.
|
|
%
|
|
% \end{itemize}
|
|
%
|
|
% \begin{figure}
|
|
% \begin{demo}[w]{The \env{listing} environment}
|
|
%\dots in the following code:
|
|
%
|
|
%\begin{listing}
|
|
%init MOV R0,#200 ;Version 2.00 please
|
|
% LDR R1,=&4B534154 ;Magic number (`TASK')
|
|
% ADR R2,appName ;Find application name
|
|
% SWI Wimp_Initialise ;Register as a WIMP task
|
|
%\end{listing}
|
|
%
|
|
%The next step is to \dots
|
|
% \end{demo}
|
|
% \end{figure}
|
|
%
|
|
% Tab characters are supported within the environment: tab stops are set
|
|
% every eighth column, although this can be modified.
|
|
%
|
|
% \subsubsection{Configuring the \env{listing} environment}
|
|
%
|
|
% The text size used in the \env{listing} environment is set by the
|
|
% |\listingsize| command. By default, this is set to |\small|, although you
|
|
% can redefine it in the document preamble, or it can be set in the document
|
|
% class.
|
|
%
|
|
% The amount by which the listing text is indented is controlled by the
|
|
% |\listingindent| length parameter. This is a fixed length, whose default
|
|
% value is 1\,em.
|
|
%
|
|
% \subsubsection{Choosing a different end-text}
|
|
%
|
|
% \DescribeEnv{listing*}
|
|
% The \env{listing} environment is terminated by the exact character sequence
|
|
% `|\end{listing}|'. This isn't too much of a problem, unless you want to
|
|
% include this string in the text. This is achieved by the \env{listing$*$}
|
|
% environment, which allows you to specify the end-text to find as an
|
|
% argument.
|
|
%
|
|
% For example:
|
|
%
|
|
% \begin{demo}{The \env{listing$*$} environment}
|
|
%Type a listing as follows:
|
|
%
|
|
%\begin{listing*}{<end-listing*>}
|
|
%\begin{listing}
|
|
%This is a listing. Yes.
|
|
%\end{listing}
|
|
%<end-listing*>
|
|
%\end{demo}
|
|
%
|
|
% Don't include `special' characters in your chosen end-text unless you know
|
|
% what you're doing.
|
|
%
|
|
% \subsection{Writing text to a file}
|
|
%
|
|
% \DescribeEnv{verbwrite}
|
|
% You can write verbatim text to a file using the \env{verbwrite}
|
|
% environment. The syntax is fairly straightforward:
|
|
%
|
|
% \begin{quote}
|
|
% \syntax{"\\begin{verbwrite}{"<file-name>"}" \dots "\\end{verbwrite}"}
|
|
% \end{quote}
|
|
%
|
|
% The text of the environment is written to the named file. The rules about
|
|
% where the text actually starts and ends are the same as for the
|
|
% \env{listing} environment.
|
|
%
|
|
% There is also a $*$-variant, like \env{listing$*$}, which allows you to
|
|
% choose the end-text. The end-text is the first argument, the filename
|
|
% comes second.
|
|
%
|
|
% There is a restriction on the characters you can write to the file: they
|
|
% must all be considered `printable' by \TeX; otherwise they will be read
|
|
% back in as `\syntax{"^^"<chars>}' which isn't too good. Unfortunately,
|
|
% this includes tab characters, so you can't write them.\footnote{^^A
|
|
% Well, not without doing serious surgery on \TeX\ itself, anyway. }
|
|
%
|
|
% \iffalse [Example time... Ho hum. There is evilness here.] \fi
|
|
%\begin{verbwrite*}{<end-write>}{wrdemo1.tmp}
|
|
%\begin{verbwrite}{wrdemo.tmp}
|
|
%This is some text written to
|
|
%a file near the beginning of
|
|
%the file.
|
|
%\end{verbwrite}
|
|
%<end-write>
|
|
%
|
|
% For example: \verbinput{wrdemo1.tmp}
|
|
%
|
|
% \input{wrdemo1.tmp} \iffalse [Now build the file ;-) ] \fi
|
|
%
|
|
% \subsection{The \cmd\verbinput\ command}
|
|
%
|
|
% \DescribeMacro{\verbinput}
|
|
% You can input a pre-prepared text file exactly as it is in the input using
|
|
% the |\verbinput| command. The filename is given as an argument. For
|
|
% example:
|
|
%
|
|
% \begin{demo}{The \cmd\verbinput\ command}
|
|
%\verbinput{wrdemo.tmp}
|
|
% \end{demo}
|
|
%
|
|
% \subsection{The \env{demo} environment}
|
|
%
|
|
% Package authors need to document their packages, and it's common to want
|
|
% to display examples showing the original text and the output side-by-side
|
|
% (or, when space doesn't permit this, one above the other). Both the
|
|
% \LaTeX\ book and \textit{The \LaTeX\ Companion} contain such examples.
|
|
%
|
|
% The \env{demo} environment allows such displays to be created easily. The
|
|
% syntax of the environment is as follows:
|
|
%
|
|
% \begin{quote}
|
|
% \syntax{"\\begin{demo}["<shape>"]{"<title>"}" \dots "\\end{demo}"}
|
|
% \end{quote}
|
|
%
|
|
% The optional \synt{shape} argument can be either `|w|' (wide), or `|n|'
|
|
% (narrow). A `wide' shape places the input and output one above the other,
|
|
% while the `narrow' shape puts them side-by-side. The default shape is
|
|
% `narrow'. An attractive border is drawn around the display to finish it
|
|
% off nicely.
|
|
%
|
|
% An example:
|
|
%
|
|
%\begin{demo*}{<end-demo>}[w]{The \env{demo} environment}
|
|
%\begin{demo}{From the \textit{\TeX book}}
|
|
%\[ \sum_{p\;\rm prime}
|
|
% f(p) = \int_{t>1}
|
|
% f(t)\,{\rm d}\pi(t) \]
|
|
%\end{demo}
|
|
%<end-demo>
|
|
%
|
|
% \DescribeEnv{demo*}
|
|
% As with the other environments created by this package, there's a
|
|
% $*$-variant which takes the end-text as an argument.
|
|
%
|
|
%
|
|
% \section{Programmer interface}
|
|
%
|
|
% This section describes the publicly available routines provided by the
|
|
% \package{sverb} package. Routines not described here are libable to be
|
|
% changed or even removed without warning, so don't use them.
|
|
%
|
|
% \subsection{Environment hooks}
|
|
%
|
|
% Each of the environments created here works in the same way. For each
|
|
% environment \env{foo}, there's a main command responsible for doing the
|
|
% work, called |\sv@foo|. This is given all the arguments of the normal
|
|
% environment, and two more:
|
|
%
|
|
% \begin{itemize}
|
|
%
|
|
% \item The `end-text' to search for, which marks the end of the environment.
|
|
%
|
|
% \item Some actions to perform after the text has been read and processed.
|
|
% This allows the calling macro to do some extra actions, like closing
|
|
% boxes, etc.
|
|
%
|
|
% \end{itemize}
|
|
%
|
|
% All the environments do is call the main command with appropriate
|
|
% arguments.
|
|
%
|
|
% \subsection{Reading the verbatim text}
|
|
%
|
|
% \DescribeMacro{\sv@read}
|
|
% The main scanning routine is |\sv@read|. It is called with three
|
|
% arguments:
|
|
%
|
|
% \begin{itemize}
|
|
%
|
|
% \item The end-text marking the end of the environment.
|
|
%
|
|
% \item The name of a macro (which must be a single token) which is called
|
|
% with a line of text as its single argument. This is given each
|
|
% line of text which is read from the environment in turn.
|
|
%
|
|
% \item A macro, or other sort of action, which is to be done when the text
|
|
% has been read and processed.
|
|
%
|
|
% \end{itemize}
|
|
%
|
|
% The macro |\sv@read| assumes that the caller has already made some
|
|
% provision for removing the category codes of the following text, by either
|
|
% calling |\@verbatim| or using the construction
|
|
% \begin{listing}
|
|
%\let\do=\@makeother
|
|
%\dospecials
|
|
% \end{listing}
|
|
%
|
|
% \DescribeMacro{\sv@safespc}
|
|
% Note that any space characters you read using |\sv@read| will be catcoded
|
|
% as |\active|. Normally this is OK because |\obeyspaces| (or
|
|
% |\@vobeyspaces|) will be in effect. If you're doing something more exotic,
|
|
% like writing text to a file or building a command string, you can call
|
|
% |\sv@safespc| which defines the active-space character to be a normal
|
|
% whitespace-space when expanded.
|
|
%
|
|
% \implementation
|
|
%
|
|
% \section{Implementation}
|
|
%
|
|
% This section defines several macros and environments which allow verbatim
|
|
% typing, with a high degree of configurability. OK, so this sort of
|
|
% thing's been done so often before that it isn't true, but I don't really
|
|
% care.
|
|
%
|
|
% \begin{macrocode}
|
|
%<*package>
|
|
% \end{macrocode}
|
|
%
|
|
% \subsection{Simple things}
|
|
%
|
|
% To help us build funny macros which involve strange and different category
|
|
% codes, I'll write some simple macros which I can use while building my
|
|
% complicated and clever ones.
|
|
%
|
|
% \begin{macro}{\@cspecials}
|
|
%
|
|
% This macro is used to assist the definition of some of the environments.
|
|
% It makes `|\|', `|{|' and `|}|' into `other' characters, and replaces them
|
|
% with `\verb"|"', `|<|' and `|>|' respectively. Note that `|[|' and `|]|'
|
|
% aren't used, because they make defining commands which take optional
|
|
% arguments awkward. Note that we open a group here. This should be closed
|
|
% using \verb"|endgroup" at the end of the special section.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\@cspecials{%
|
|
\begingroup%
|
|
\catcode`|0%
|
|
\catcode`<1%
|
|
\catcode`>2%
|
|
\catcode`\{12%
|
|
\catcode`\}12%
|
|
\catcode`\\12%
|
|
}
|
|
% \end{macrocode}
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\sv@startlisting}
|
|
%
|
|
% This macro sets everything up nicely for a \env{listing}-type verbatim
|
|
% environment.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@startlisting{%
|
|
\def\par{\@@par\penalty\interlinepenalty}%
|
|
\@@par%
|
|
\leftskip\@totalleftmargin%
|
|
\obeylines%
|
|
\@noligs%
|
|
\let\do\@makeother\dospecials%
|
|
\verbatim@font%
|
|
\frenchspacing%
|
|
\@vobeyspaces%
|
|
\settabwidth%
|
|
\catcode9\active%
|
|
\lccode`\~9\lowercase{\let~\sv@vtab}%
|
|
\lccode`\~13\lowercase{\let~\vinput@cr}%
|
|
\interlinepenalty500%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \subsection{Tab character handling}
|
|
%
|
|
% One of the things we want to do here is handle tab characters properly.
|
|
% (Here, `properly' means `moving to the next column which is a multiple of
|
|
% eight', the way these things were always meant to.)
|
|
%
|
|
% \begin{macro}{\settabwidth}
|
|
%
|
|
% The tabs used by our tabbed verbatim environments are set up by this
|
|
% routine. It sets the tab width parameter |\svtab| to 8 times the width
|
|
% of a |\tt| space. If you really want, you can redefine this macro.
|
|
%
|
|
% \begin{macrocode}
|
|
\newdimen\svtab
|
|
\def\settabwidth{\setbox\z@\hbox{\texttt{\space}}\svtab8\wd\z@}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\sv@vtab}
|
|
%
|
|
% Here we handle tabs inside verbatim environments. We expect each line to
|
|
% be typeset as a box, using something like
|
|
%
|
|
% \begin{listing}
|
|
%\setbox0\hbox{#1}
|
|
%\leavevmode
|
|
%\box0
|
|
%\par
|
|
% \end{listing}
|
|
%
|
|
% The idea is that you make tab active, and set it to this macro. We stop
|
|
% the current box, stretch it to the right width, and start another one
|
|
% straight after, so nobody know the difference. The code here is straight
|
|
% from Appendix~D of \textit{The \TeX book}.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@vtab{%
|
|
\hfill\egroup%
|
|
\@tempdima\wd\z@%
|
|
\divide\@tempdima\svtab%
|
|
\multiply\@tempdima\svtab%
|
|
\advance\@tempdima\svtab%
|
|
\wd\z@\@tempdima%
|
|
\leavevmode\box\z@%
|
|
\setbox\z@\hbox\bgroup%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\verbinput}
|
|
%
|
|
% We allow input from a file, by the |\verbinput| command. We display the
|
|
% text pretty much the same as the \env{listing} environment below.
|
|
%
|
|
% We set tab and return active, and get them to do appropriate things. This
|
|
% isn't actually all that hard.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\verbinput#1{%
|
|
\begin{listinglist}%
|
|
\listingsize%
|
|
\sv@startlisting%
|
|
\setbox\z@\hbox\bgroup%
|
|
\input{#1}%
|
|
\sv@stripspc%
|
|
\egroup%
|
|
\ifdim\wd\z@=\z@%
|
|
\ifhmode\par\fi%
|
|
\else%
|
|
\leavevmode\box\z@\par%
|
|
\fi%
|
|
\end{listinglist}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\vinput@cr}
|
|
%
|
|
% This macro handles return characters while inputting text in |\verbinput|.
|
|
% We just output our current box, and start another.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\vinput@cr{%
|
|
\egroup%
|
|
\leavevmode\box\z@%
|
|
\par%
|
|
\setbox\z@\hbox\bgroup%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \subsection{Reading verbatim text}
|
|
%
|
|
% The traditional way of reading verbatim text is to use a delimited
|
|
% argument, as described in the \textit{\TeX book}. This works well-ish if
|
|
% the text isn't very long. A better solution would be to pick out the text
|
|
% line-by-line and process it like that. So this is what we do.
|
|
%
|
|
% \begin{macro}{\matcher}
|
|
%
|
|
% For long verbatim environments, we need to be able to find the end text.
|
|
% This is rather tricky. The solution here is rather horrible. The
|
|
% environment picks out each line of the text at a time, as an argument, and
|
|
% tests to see if it contains the text we're after. We do the test in a
|
|
% particularly yukky way: we add the actual target text to the end of the
|
|
% line, and inspect the text following the match to see if the match is at
|
|
% the end.
|
|
%
|
|
% The |\matcher| macro creates a `matcher' which will test strings to see if
|
|
% they contain something interesting.
|
|
%
|
|
% To create a matcher, say
|
|
% \syntax{"\\matcher{"<cmd-name>"}{"<target>"}{"<process-cmd>"}"}. The
|
|
% command \synt{cmd-name} accepts a line of text as an argument and calls
|
|
% the \synt{process-cmd} with the text of the line before the match, or the
|
|
% whole lot. It also sets |\@ifmatched| appropriately.
|
|
%
|
|
% (Having spent ages coming up with this cruft myself, I found some very
|
|
% similar, but slightly better, code in Appendix~D. So I've changed mine to
|
|
% match Donald's. Anyway, credit where it's due: cheers Don.)
|
|
%
|
|
% \begin{macrocode}
|
|
\newif\if@matched
|
|
\def\matcher#1#2#3{%
|
|
\expandafter\def\csname\string#1$match\endcsname##1#2##2##3\end{%
|
|
\ifx##2\relax%
|
|
\@matchedfalse%
|
|
\else%
|
|
\@matchedtrue%
|
|
\fi%
|
|
#3{##1}%
|
|
}%
|
|
\expandafter\def\expandafter#1\expandafter##\expandafter1\expandafter{%
|
|
\csname\string#1$match\endcsname##1#2\relax\end%
|
|
}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\sv@stripspc}
|
|
%
|
|
% This macro strips any trailing glue in the current horizontal list. This
|
|
% is fairly simple, actually: we just loop while glue is the last item. It's
|
|
% slightly complicated by penalties which \TeX\ puts into the list between
|
|
% the glue items, but we just remove them too.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@stripspc{%
|
|
\unpenalty%
|
|
\ifdim\lastskip=\z@\else%
|
|
\unskip\expandafter\sv@stripspc%
|
|
\fi%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\sv@percent}
|
|
%
|
|
% This macro strips a single leading percent character if there is one, and
|
|
% if the \env{doc} package is loaded. We store the possibly stripped text in
|
|
% |\@tempa|.
|
|
%
|
|
% \begin{macrocode}
|
|
\begingroup
|
|
\catcode`\%=12
|
|
\gdef\sv@percent#1#2\relax
|
|
{\ifx\check@percent\@@undefined
|
|
\ifx#1\relax\def\@tempa{}\else
|
|
\def\@tempa{#1#2}\fi\else
|
|
\ifx#1\relax\def\@tempa{}\else
|
|
\ifx#1%\def\@tempa{#2}\else
|
|
\def\@tempa{#1#2}\fi\fi\fi}
|
|
\endgroup
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\@isspaces}
|
|
%
|
|
% We want to avoid writing the first and last lines of the environment to the
|
|
% file if there's nothing in them. To do this, we need to know whether a
|
|
% piece of text contains only space characters. This macro does this, in a
|
|
% rather nasty way. See the other macros below for details of how this
|
|
% works.
|
|
%
|
|
% We define |\sv@safespc| at the same time: this makes space active and
|
|
% expand to a space character which is not active. Neat, huh?
|
|
%
|
|
% \begin{macrocode}
|
|
\lccode`\~32
|
|
\lccode`\!32
|
|
\lowercase{%
|
|
\def\@isspaces#1{%
|
|
\ifx#1\relax%
|
|
\def\@tempb{\@tempswafalse}%
|
|
\else\ifx#1~%
|
|
\let\@tempb\@isspaces%
|
|
\else%
|
|
\def\@tempb##1\relax{}%
|
|
\fi\fi%
|
|
\@tempb%
|
|
}
|
|
\def\sv@safespc{%
|
|
\catcode32\active%
|
|
\def~{ }%
|
|
}
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\sv@read}
|
|
%
|
|
% This macro does the main job of reading a chunk of verbatim text. You call
|
|
% it like this:
|
|
%
|
|
% \begin{quote}
|
|
% \syntax{"\\sv@read{"<end-text>"}{"<process-line-proc>"}{"<end-proc>"}"}
|
|
% \end{quote}
|
|
%
|
|
% The \synt{end-text} is the text to find at the end of the `environment': we
|
|
% stop when we find it.
|
|
%
|
|
% The \synt{process-line-proc} is a macro which is passed as an argument each
|
|
% line which we read from the text.
|
|
%
|
|
% The \synt{end-proc} is a macro to call once we've finished reading all of
|
|
% the text. This can tidy up an environment or close a file or whatever.
|
|
%
|
|
% We read the text by picking out newlines using a delimited macro. We have
|
|
% to be a little clever, because newlines are active in verbatim text.
|
|
%
|
|
% We will also strip `|%|' signs off the beginning if the \package{doc}
|
|
% package is here (\package{doc} tries to play with \LaTeX's verbatim stuff,
|
|
% and doesn't understand the way we do things).
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@read#1#2#3{%
|
|
% \end{macrocode}
|
|
%
|
|
% This code does all sorts of evil things, so I'll start by opening a group.
|
|
%
|
|
% \begin{macrocode}
|
|
\begingroup%
|
|
% \end{macrocode}
|
|
%
|
|
% So that I can spot the end-text, I'll create a matcher macro.
|
|
%
|
|
% \begin{macrocode}
|
|
\matcher\@match{#1}\sv@read@ii%
|
|
% \end{macrocode}
|
|
%
|
|
% So that I can identify line ends, I'll make them active. I'll also make
|
|
% spaces active so that they can expand to whatever they ought to expand
|
|
% to (spaces in files, or funny \verb*" " characters or whatever.
|
|
%
|
|
% \begin{macrocode}
|
|
\catcode13\active%
|
|
\catcode32\active%
|
|
% \end{macrocode}
|
|
%
|
|
% I'll use the |\if@tempswa| flag to tell me whether I ought to output the
|
|
% current line. This is a little messy, so I'll describe it later. I'll
|
|
% initialise it to false because this is the correct thing to do.
|
|
%
|
|
% \begin{macrocode}
|
|
\@tempswafalse%
|
|
% \end{macrocode}
|
|
%
|
|
% Most of the job is done by two submacros. I'll define them in terms of
|
|
% my current arguments (to save lots of token munging). The first just
|
|
% extracts the next line (which ends at the next newline character) and
|
|
% tries to match it.
|
|
%
|
|
% \begin{macrocode}
|
|
\lccode`\~13\lowercase{%
|
|
\def\sv@read@i##1~{\@match{##1}}%
|
|
}%
|
|
% \end{macrocode}
|
|
%
|
|
% The results of the match get passed here, along with the text of the
|
|
% line up to the matched text.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@read@ii##1{%
|
|
% \end{macrocode}
|
|
%
|
|
% The first job to do is to maybe strip off percent signs from the beginning,
|
|
% to keep \package{doc} happy.
|
|
%
|
|
% \begin{macrocode}
|
|
\sv@percent##1\relax\relax%
|
|
% \end{macrocode}
|
|
%
|
|
% Now I need to decide whether I ought to output this line. The method goes
|
|
% like this: if this is the first line (|\if@tempswa| is false) or the last
|
|
% (|\if@matched| is true), \emph{and} the text consists only of spaces, then
|
|
% I'll ignore it.
|
|
%
|
|
% The first thing to do is to notice the last line -- if |\if@matched| is
|
|
% true, then I'll make |\if@tempswa| false to make the first-line and
|
|
% last-line cases work the same way.
|
|
%
|
|
% \begin{macrocode}
|
|
\if@matched\@tempswafalse\fi%
|
|
% \end{macrocode}
|
|
%
|
|
% Now if this is the first or last line, I'll examine it for spaces. This
|
|
% is done in a separate macro. It will set |\if@tempswa| false if the
|
|
% text contains only spaces.
|
|
%
|
|
% \begin{macrocode}
|
|
\if@tempswa\else\@tempswatrue\expandafter\@isspaces\@tempa\relax\fi%
|
|
% \end{macrocode}
|
|
%
|
|
% Now, if |\if@tempswa| is still true, perform the \<process-line-proc> on
|
|
% the line of text. I'll provide a group, so that it doesn't upset me
|
|
% too much.
|
|
%
|
|
% \begin{macrocode}
|
|
\if@tempswa%
|
|
\begingroup%
|
|
\expandafter#2\expandafter{\@tempa}%
|
|
\endgroup%
|
|
\fi%
|
|
% \end{macrocode}
|
|
%
|
|
% The next line won't be the first one, so I'll set the flag true in
|
|
% readiness.
|
|
%
|
|
% \begin{macrocode}
|
|
\@tempswatrue%
|
|
% \end{macrocode}
|
|
%
|
|
% Now, if that wasn't the last line, go round again; otherwise end the group
|
|
% I started ages ago, and do the user's \<end-proc>.
|
|
%
|
|
% \begin{macrocode}
|
|
\if@matched\def\@tempa{\endgroup#3}\else\let\@tempa\sv@read@i\fi%
|
|
\@tempa%
|
|
}%
|
|
% \end{macrocode}
|
|
%
|
|
% Now to start the thing up. I'll read the first line.
|
|
%
|
|
% \begin{macrocode}
|
|
\sv@read@i%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\sv@readenv}
|
|
%
|
|
% This macro works out an appropriate end-text for the current environment.
|
|
% If you say \syntax{"\\sv@readenv{"<macro-name>"}"}, it will expand do
|
|
% \begin{listinglist} \listingsize \synshorts
|
|
% <macro-name>"{\\"$_{12}$"end{"$_{12}$<current-env-name>"}"$_{12}$"}"^^A
|
|
% "{\\end{"<current-env-name>"}}"
|
|
% \end{listinglist}
|
|
% Easy, no?
|
|
%
|
|
% This is all done with mirrors. No, err\dots\ it's done with
|
|
% |\expandafter|.
|
|
%
|
|
% \begin{macrocode}
|
|
\begingroup
|
|
\lccode`\<=`\{
|
|
\lccode`\>=`\}
|
|
\lccode`\|=`\\
|
|
\lowercase{\endgroup
|
|
\def\sv@readenv#1{%
|
|
\expandafter\expandafter\expandafter%
|
|
#1\expandafter\sv@readenv@i\@currenvir\@@%
|
|
}
|
|
\def\sv@readenv@i#1\@@{{|end<#1>}{\end{#1}}}
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\sv@verbline}
|
|
%
|
|
% This macro typesets a line in a verbatim way, so you can construct a real
|
|
% verbatim environment from it. It's a bit tricky in the way that it catches
|
|
% the last line. Don't worry about this: it's easy really. Note the
|
|
% |\relax| after the |\par| -- this is because \package{doc} tries to do
|
|
% clever things with |\par| to strip `|%|' signs out.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@verbline#1{%
|
|
\setbox\z@\hbox{#1\sv@stripspc}%
|
|
\ifdim\wd\z@=\z@%
|
|
\if@matched\ifhmode\par\relax\fi\else\leavevmode\par\relax\fi%
|
|
\else%
|
|
\leavevmode\box\z@\par\relax%
|
|
\fi%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \subsection{Listing environments}
|
|
%
|
|
% The \env{listing} environment is our equivalent of the standard
|
|
% \env{verbatim} environment. We do some slightly cleverer things, though,
|
|
% to make sure (for example) that even text which contains |\end{listing}|
|
|
% can be typeset.
|
|
%
|
|
% \begin{macro}{\listinglist}
|
|
% \begin{environment}{listinglist}
|
|
%
|
|
% This defines the layout for the \env{listing} environment. It starts a
|
|
% list with the appropriate shape. It's also made into an environment, so
|
|
% that the end-paragraph-environment bits work correctly.
|
|
%
|
|
% The |\listingindent| length parameter sets up the indentation of the
|
|
% listings. If there's a |\parindent| setting, I'll line listings up with
|
|
% that; otherwise I'll just choose something which looks right.
|
|
%
|
|
% \begin{macrocode}
|
|
\newdimen\listingindent
|
|
\AtBeginDocument{%
|
|
\ifdim\parindent=\z@\listingindent1em\else\listingindent\parindent\fi%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Now to define a size hook for the environment. This is fairly simple
|
|
% stuff.
|
|
%
|
|
% \begin{macrocode}
|
|
\ifx\listingsize\@@undefined
|
|
\let\listingsize\small
|
|
\fi
|
|
% \end{macrocode}
|
|
%
|
|
% Now to define the environment itself. Suppress the indentation if we're
|
|
% first thing on a new list item, so that the listing lines up with
|
|
% everything else.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\listinglist{%
|
|
\list{}{%
|
|
\if@inlabel%
|
|
\leftmargin\z@%
|
|
\else%
|
|
\leftmargin\listingindent%
|
|
\fi%
|
|
\rightmargin\z@%
|
|
\labelwidth\z@%
|
|
\labelsep\z@%
|
|
\itemindent\z@%
|
|
\listparindent\z@%
|
|
\let\makelabel\relax%
|
|
\parsep\z@skip%
|
|
}%
|
|
\parfillskip\@flushglue%
|
|
\item\relax%
|
|
}
|
|
\let\endlistinglist\endlist
|
|
% \end{macrocode}
|
|
%
|
|
% \end{environment}
|
|
% \end{macro}
|
|
%
|
|
% \begin{environment}{listing}
|
|
%
|
|
% The \env{listing} environment is the only real verbatim-like environment we
|
|
% create will all this kit, although it does the job very nicely.
|
|
%
|
|
% The environment indents its contents slightly, unlike \env{verbatim}, and
|
|
% uses a smaller typeface in an attempt to fit 77-column text on an A5~page.
|
|
% There is also a $*$-variant, which allows you to specify the terminating
|
|
% text. This enables you to include absolutely any text in the environment,
|
|
% including |\end{listing}|.
|
|
%
|
|
% First, we must define the |\listing| command.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\listing{%
|
|
\listinglist%
|
|
\listingsize%
|
|
\sv@readenv\sv@listing%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Now we define the |\@listing| command, which does most of the work. We
|
|
% base the \env{listing} environment on a \env{list}.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@listing#1#2{%
|
|
\sv@startlisting%
|
|
\sv@read{#1}\sv@verbline{\endlistinglist#2}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% Now we define the starred version. The command name needs to include the
|
|
% `|*|' character, so we must use |\csname|. There's some hacking here to
|
|
% allow us to read the name using the appropriate catcodes for otherwise
|
|
% normal characters: \LaTeX\ activates some characters and makes them typeset
|
|
% themselves to suppress some ligaturing.
|
|
%
|
|
% \begin{macrocode}
|
|
\expandafter\def\csname listing*\endcsname{%
|
|
\listinglist%
|
|
\listingsize%
|
|
\begingroup%
|
|
\@noligs%
|
|
\def\@tempa##1{\endgroup\sv@listing{##1}{\end{listing*}}}%
|
|
\@tempa%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{environment}
|
|
%
|
|
% \begin{environment}{ignore}
|
|
%
|
|
% The \env{ignore} environment entirely ignores its contents. Anything at
|
|
% all may be put into the environment: it is discarded utterly.
|
|
%
|
|
% We define some macros for defining ignoring environments, because this can
|
|
% be useful for version control, possibly.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@ignore#1#2{%
|
|
\@bsphack%
|
|
\let\do\@makeother\dospecials%
|
|
\sv@read{#1}\@gobble{\@esphack#2}%
|
|
}
|
|
\def\ignore{\sv@readenv\sv@ignore}
|
|
\def\ignoreenv#1{%
|
|
\expandafter\let\csname #1\endcsname\ignore%
|
|
}
|
|
\def\unignoreenv#1{%
|
|
\expandafter\def\csname #1\endcsname{\endgroup}%
|
|
\expandafter\def\csname end#1\endcsname%
|
|
{\begingroup\def\@currenvir{#1}}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{environment}
|
|
%
|
|
% \subsection{The \env{verbwrite} environment}
|
|
%
|
|
% The \env{verbwrite} environment allows text to be written to a file in a
|
|
% verbatim way. Note that tab characters don't work, because \TeX\ refuses
|
|
% to be nice.
|
|
%
|
|
% \begin{macro}{\sv@write}
|
|
%
|
|
% As seems to be traditional now, we first define a general hookable macro
|
|
% which allows a caller to specify the end-text and what to do afterwards.
|
|
%
|
|
% \begin{macrocode}
|
|
\newwrite\sv@writefile
|
|
\def\sv@write#1#2{%
|
|
\begingroup%
|
|
\@bsphack%
|
|
\let\do\@makeother\dospecials%
|
|
\sv@safespc%
|
|
\sv@read{#1}\sv@writeline{\sv@endwrite#2}%
|
|
}
|
|
\def\sv@writeline#1{%
|
|
\immediate\write\sv@writefile{#1}%
|
|
}
|
|
\def\sv@endwrite{%
|
|
\@esphack%
|
|
\endgroup%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{environment}{verbwrite}
|
|
%
|
|
% Now we can define the actual environment. We define a $*$-variant which
|
|
% allows the user to specify the end-text, just to make sure.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\verbwrite#1{%
|
|
\immediate\openout\sv@writefile#1\relax%
|
|
\sv@readenv\sv@write%
|
|
}
|
|
\def\endverbwrite{\immediate\closeout\sv@writefile}
|
|
\expandafter\def\csname verbwrite*\endcsname#1#2{%
|
|
\immediate\openout\sv@writefile#2\relax%
|
|
\sv@write{#1}{\immediate\closeout\sv@writefile\end{verbwrite*}}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{environment}
|
|
%
|
|
% \subsection{The \env{demo} environment}
|
|
%
|
|
% By way of tying all of this together, I present an environment for
|
|
% displaying demonstrations of \LaTeX\ markup. We read the contents of the
|
|
% environment, write it to a temporary file, and read it back twice,
|
|
% typesetting it the first time and displaying it verbatim the second time.
|
|
%
|
|
% \begin{macro}{\sv@demoname}
|
|
%
|
|
% This macro expands to the filename to use for the temporary data. To
|
|
% allow the package documentation to demonstrate the \env{demo} environment
|
|
% itself, we need to keep a nesting count. This avoids too much hackery,
|
|
% which unfortunately appears to plague all of my \TeX\ code.
|
|
%
|
|
% \begin{macrocode}
|
|
\newcount\sv@nestcount
|
|
\def\sv@demoname{demo\number\sv@nestcount.tmp}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{macro}{\sv@demo}
|
|
%
|
|
% As for listing, we do all the business through a private macro. This is
|
|
% good because it means we can leave the main macro readable. The argument
|
|
% is the end-text to spot.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@demo#1#2{%
|
|
\@ifnextchar[{\sv@demo@i{#1}{#2}}{\sv@demo@i{#1}{#2}[n]}%
|
|
}
|
|
\def\sv@demo@i#1#2[#3]#4{%
|
|
\advance\sv@nestcount by\@ne%
|
|
\immediate\openout\sv@writefile\sv@demoname\relax%
|
|
\sv@write{#1}{%
|
|
\immediate\closeout\sv@writefile%
|
|
\sv@dodemo{#2}{#3}{#4}%
|
|
}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% \begin{environment}{demo}
|
|
%
|
|
% This is the real environment. We provide \env{demo$*$} too, to allow the
|
|
% user to choose the end-text.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\demo{\sv@readenv\sv@demo}
|
|
\expandafter\def\csname demo*\endcsname#1{\sv@demo{#1}{\end{demo*}}}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{environment}
|
|
%
|
|
% \begin{macro}{\sv@dodemo}
|
|
%
|
|
% First, let's define some common bits of code in the stuff below. The
|
|
% minipages used to typeset the material has some clever stuff to avoid
|
|
% strange spacing in the output.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@demosmp{%
|
|
\begin{minipage}[t]{\@tempdima}%
|
|
\vskip8\p@%
|
|
\hrule\@height\z@%
|
|
\raggedright%
|
|
\vbox\bgroup%
|
|
}
|
|
\def\sv@demoemp{%
|
|
\par\unpenalty\unskip%
|
|
\egroup%
|
|
\vskip8\p@%
|
|
\hrule\@height\z@%
|
|
\end{minipage}%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% This is the macro which actually typesets the demonstration.
|
|
%
|
|
% \begin{macrocode}
|
|
\def\sv@dodemo#1#2#3{%
|
|
% \end{macrocode}
|
|
%
|
|
% Now work out some values. We set |\hsize| to the line width leaving 2\,em
|
|
% of space on either side. The size of the minipages is calculated depending
|
|
% on the shape of the demonstration. This is all fairly simple.
|
|
%
|
|
% \begin{macrocode}
|
|
\begingroup%
|
|
\@tempdima\linewidth%
|
|
\advance\@tempdima-2em%
|
|
\hsize\@tempdima%
|
|
\if#2w%
|
|
\advance\@tempdima-2em%
|
|
\else%
|
|
\advance\@tempdima-3em%
|
|
\divide\@tempdima2%
|
|
\fi%
|
|
% \end{macrocode}
|
|
%
|
|
% Now we open a big vertical box, and put in a header to mark off the
|
|
% demonstration.
|
|
%
|
|
% \begin{macrocode}
|
|
\par%
|
|
\setbox\z@\hbox{\strut\enspace#3\enspace\strut}%
|
|
\@tempdimb.5\dp\z@%
|
|
\advance\@tempdimb-.5\ht\z@%
|
|
\ht\z@\@tempdimb\dp\z@\@tempdimb%
|
|
\noindent\hskip1em\vtop{%
|
|
\hb@xt@\hsize{%
|
|
\hrulefill%
|
|
\raise\@tempdimb\box\z@%
|
|
\hrulefill%
|
|
}%
|
|
\nointerlineskip%
|
|
\hb@xt@\hsize{\vrule\@height5\p@\hfil\vrule\@height5\p@}%
|
|
\nointerlineskip%
|
|
% \end{macrocode}
|
|
%
|
|
% Now we insert the output text in the first minipage. I'll force `|%|'
|
|
% to be a comment character, in case something like \package{doc} has had its
|
|
% wicked way.
|
|
%
|
|
% \begin{macrocode}
|
|
\vskip-\parskip%
|
|
\noindent\hbox{}\hskip1em%
|
|
\sv@demosmp%
|
|
\catcode`\%14\relax%
|
|
\input{\sv@demoname}%
|
|
\sv@demoemp%
|
|
% \end{macrocode}
|
|
%
|
|
% Insert some kind of separation between the two. In `wide' format, we start
|
|
% a new line, and put a ruleoff between the two. In `narrow' format, we just
|
|
% leave some space.
|
|
%
|
|
% \begin{macrocode}
|
|
\if#2w%
|
|
\vskip8\p@\hrule\vskip8\p@%
|
|
\noindent\hbox{}%
|
|
\fi%
|
|
\hskip1em%
|
|
% \end{macrocode}
|
|
%
|
|
% Now we put the verbatim copy of the text in the other minipage.
|
|
%
|
|
% \begin{macrocode}
|
|
\sv@demosmp%
|
|
\listingindent\z@%
|
|
\verbinput\sv@demoname%
|
|
\sv@demoemp%
|
|
\par%
|
|
\nointerlineskip%
|
|
\hb@xt@\hsize{\vrule\@height5\p@\hfil\vrule\@height5\p@}%
|
|
\hrule%
|
|
}%
|
|
\endgroup%
|
|
\par%
|
|
\vskip\baselineskip%
|
|
#1%
|
|
}
|
|
% \end{macrocode}
|
|
%
|
|
% \end{macro}
|
|
%
|
|
% That's all there is. Have fun.
|
|
%
|
|
% \begin{macrocode}
|
|
%</package>
|
|
% \end{macrocode}
|
|
%
|
|
% \hfill Mark Wooding, \today
|
|
%
|
|
% \Finale
|
|
%
|
|
\endinput
|