| 1 |
% \iffalse |
| 2 |
%<*never> |
| 3 |
\documentclass{ltxdoc} |
| 4 |
|
| 5 |
\AtBeginDocument{\CodelineIndex\EnableCrossrefs} |
| 6 |
\AtEndDocument{\PrintIndex} |
| 7 |
\GetFileInfo{stack.cls} |
| 8 |
\def\pack#1{\texttt{#1}} |
| 9 |
\begin{document} |
| 10 |
\def\docdate{2002/01/20} |
| 11 |
\CheckSum{12} |
| 12 |
\DocInput{stack.dtx} |
| 13 |
\end{document} |
| 14 |
%</never> |
| 15 |
% \fi |
| 16 |
% \iffalse |
| 17 |
% |
| 18 |
% (c) 2002 Benjamin BAYART. Freely redistributable under lppl. |
| 19 |
% |
| 20 |
% \fi |
| 21 |
% \CheckSum{110} |
| 22 |
% \DoNotIndex{\edef,\advance,\else,\csname,\endcsname,\expandafter} |
| 23 |
% \DoNotIndex{\ifx,\fi,\input,\ifnum,\relax,\newcount,\romannumeral} |
| 24 |
% \DoNotIndex{\the,\typeout,\xdef,\global,\newcommand,\RequirePackage} |
| 25 |
% \DoNotIndex{\PackageError,\ProvidesPackage,\gdef} |
| 26 |
%% \CharacterTable |
| 27 |
%% {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 |
| 28 |
%% 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 |
| 29 |
%% Digits \0\1\2\3\4\5\6\7\8\9 |
| 30 |
%% Exclamation \! Double quote \" Hash (number) \# |
| 31 |
%% Dollar \$ Percent \% Ampersand \& |
| 32 |
%% Acute accent \' Left paren \( Right paren \) |
| 33 |
%% Asterisk \* Plus \+ Comma \, |
| 34 |
%% Minus \- Point \. Solidus \/ |
| 35 |
%% Colon \: Semicolon \; Less than \< |
| 36 |
%% Equals \= Greater than \> Question mark \? |
| 37 |
%% Commercial at \@ Left bracket \[ Backslash \\ |
| 38 |
%% Right bracket \] Circumflex \^ Underscore \_ |
| 39 |
%% Grave accent \` Left brace \{ Vertical bar \| |
| 40 |
%% Right brace \} Tilde \~} |
| 41 |
% |
| 42 |
% \title{Using stacks independently from \TeX's one} |
| 43 |
% \author{Benjamin \textsc{Bayart}} |
| 44 |
% \date{\filedate} |
| 45 |
% \maketitle |
| 46 |
% \begin{abstract} |
| 47 |
% This package is aimed to provide a new stack mechanisme independent |
| 48 |
% from the \TeX\ stack itself. The first use was to have a system where |
| 49 |
% |\input| become directory recursive. I.e.~to provide a kind of |
| 50 |
% |\chdir| command: when you include a file ``\emph{a}'', it's directory become |
| 51 |
% the current one, until the end of ``\emph{a}''. See the ``sample package'' |
| 52 |
% section for full explanation. |
| 53 |
% |
| 54 |
% This package by itself is of no use, except for some macro-developpers. |
| 55 |
% \end{abstract} |
| 56 |
% \tableofcontents |
| 57 |
% |
| 58 |
% \section{User's guide} |
| 59 |
% |
| 60 |
% \subsection{The \pack{stack} package} |
| 61 |
% |
| 62 |
% All of the macros provided by this package are supposed for |
| 63 |
% package developpers (the is no meaning for a stack in a final |
| 64 |
% document, only to be used in a package), so, all of those macros |
| 65 |
% have capital letters in their names, following old guidelines |
| 66 |
% for \LaTeX3 project. |
| 67 |
% |
| 68 |
% The first macro is the one to create a new stack:\newline |
| 69 |
% \cs{NewStack}\marg{stack}\marg{default}\newline |
| 70 |
% Well, it just initializes all the required data for the stack. |
| 71 |
% \emph{default} is the value of the top of the stack if the stack is |
| 72 |
% empty. |
| 73 |
% |
| 74 |
% The second one is the macro to push a new value on top of the |
| 75 |
% stack:\newline |
| 76 |
% \cs{Push}\marg{stack}\marg{value}\newline |
| 77 |
% |
| 78 |
% The third one is to remove the value on the top (i.e. to pop):\newline |
| 79 |
% \cs{Pop}\marg{stack} |
| 80 |
% |
| 81 |
% Finally, the most usefull one is the command to obtain the value |
| 82 |
% on the top of the stack:\newline |
| 83 |
% \cs{Stack}\marg{stack} |
| 84 |
% |
| 85 |
% Another one is provided, but only for debugging purpose, it will |
| 86 |
% |\typeout| the current value on the top of the stack:\newline |
| 87 |
% \cs{ShowTop}\marg{stack} |
| 88 |
% |
| 89 |
% \subsection{The \pack{relinput} package} |
| 90 |
% |
| 91 |
% |relinput| provides only one macro which syntax is:\newline |
| 92 |
% \cs{relinput}\marg{directory}\marg{file}. |
| 93 |
% |
| 94 |
% This will move the ``current directory'' into \emph{directory}, |
| 95 |
% from there include \emph{file}, and then go back to what was |
| 96 |
% the current directory before you call the macro. |
| 97 |
% |
| 98 |
% A good way to understand it is to try to split a large document |
| 99 |
% in a large number of files. Say a first level of directories |
| 100 |
% for the parts, a second one for the chapters, a third one for |
| 101 |
% the sections, and in each one a file per subsection. |
| 102 |
% |
| 103 |
% So, we will have this kind of directory tree: |
| 104 |
% \begin{macrocode} |
| 105 |
%<*never> |
| 106 |
part-one/ |
| 107 |
part-one/thepart.tex |
| 108 |
part-one/chapter-one/ |
| 109 |
part-one/chapter-one/thechapter.tex |
| 110 |
part-one/chapter-one/section-one/ |
| 111 |
part-one/chapter-one/section-one/thesection.tex |
| 112 |
part-one/chapter-one/section-one/subsec1.tex |
| 113 |
part-one/chapter-one/section-one/subsec2.tex |
| 114 |
part-one/chapter-one/section-two/ |
| 115 |
part-one/chapter-one/section-three/ |
| 116 |
part-one/chapter-two/ |
| 117 |
part-one/chapter-two/thechapter.tex |
| 118 |
part-two/ |
| 119 |
part-two/thepart.tex |
| 120 |
part-two/chapter-one/ |
| 121 |
part-two/chapter-one/thechapter.tex |
| 122 |
% \end{macrocode} |
| 123 |
% |
| 124 |
% And so on. |
| 125 |
% |
| 126 |
% The main file only says: |
| 127 |
% \begin{macrocode} |
| 128 |
\relinput{part-one}{thepart.tex} |
| 129 |
\relinput{part-two}{thepart.tex} |
| 130 |
% \end{macrocode} |
| 131 |
% |
| 132 |
% The file |part-one/thepart.tex| states: |
| 133 |
% \begin{macrocode} |
| 134 |
\relinput{chapter-one}{thechapter.tex} |
| 135 |
\relinput{chapter-two}{thechapter.tex} |
| 136 |
% \end{macrocode} |
| 137 |
% |
| 138 |
% And so on for the other files, untill |
| 139 |
% |part-one/chapter-one/section-one/thesection.tex| which says: |
| 140 |
% \begin{macrocode} |
| 141 |
\relinput{.}{subsec1.tex} |
| 142 |
\relinput{.}{subsec2.tex} |
| 143 |
%</never> |
| 144 |
% \end{macrocode} |
| 145 |
% |
| 146 |
% With such a scheme, when you work on a chapter, you don't have to |
| 147 |
% remember in which directory it is, and, even better, if you |
| 148 |
% decide to rename the |chapter-one| directory to |introduction| |
| 149 |
% you have only one file to modify. |
| 150 |
% |
| 151 |
% By using the usual \LaTeX\ mechanism you would have to edit the |
| 152 |
% same file, plus all the files in this directory which include |
| 153 |
% others (that's a rather large number of files). |
| 154 |
% |
| 155 |
% \section{Test file} |
| 156 |
% |
| 157 |
% The test application for |stack| is a ``relative dir input''. The basic |
| 158 |
% macro is |\relinput| which takes two arguments: first a directory |
| 159 |
% where to go (from the previous one) and then a file to include. |
| 160 |
% |
| 161 |
% In your main file (|./file.tex|) if you do |\relinput{dir}{file2.tex}| |
| 162 |
% then \TeX\ will include |./dir/file2.tex|. If, in |file2|, you do |
| 163 |
% |\relinput{dir2}{file3.tex}|, then, \TeX\ will include |./dir/dir2/file3.tex|. |
| 164 |
% |
| 165 |
% Really interesting if you have a large document with a lot of files |
| 166 |
% in which case you can move a directory and have only minors corrections |
| 167 |
% to do in a few files, and, anyway, no corrections in the files \emph{in} |
| 168 |
% this directory, but only to files \emph{out} of the directory. |
| 169 |
% |
| 170 |
% This test file is provided as a small package |relinput|. |
| 171 |
% \begin{macrocode} |
| 172 |
%<*test> |
| 173 |
\ProvidesPackage{relinput}[2002/01/20 v1.00 Relative input] |
| 174 |
\RequirePackage{stack} |
| 175 |
\NewStack{reldir}{.} |
| 176 |
% \end{macrocode} |
| 177 |
% \begin{macro}{\relinput} |
| 178 |
% \begin{macrocode} |
| 179 |
\newcommand\relinput[2]{% |
| 180 |
\Push{reldir}{\Stack{reldir}/#1}% |
| 181 |
\input{\Stack{reldir}/#2}% |
| 182 |
\Pop{reldir}% |
| 183 |
} |
| 184 |
%</test> |
| 185 |
% \end{macrocode} |
| 186 |
% \end{macro} |
| 187 |
% |
| 188 |
% \StopEventually{} |
| 189 |
% |
| 190 |
% \section{Code for the \pack{stack} package} |
| 191 |
% |
| 192 |
% \subsection{Identification} |
| 193 |
% |
| 194 |
% \begin{macrocode} |
| 195 |
%<*package> |
| 196 |
\ProvidesPackage{stack}[2002/01/20 v1.0 Multi-stack system] |
| 197 |
% \end{macrocode} |
| 198 |
% |
| 199 |
% \subsection{Macros by itself} |
| 200 |
% |
| 201 |
% In the following code, to use only one counter (in the \TeX\ meaning) |
| 202 |
% we store the various values in macros, and use only the counter when |
| 203 |
% we need to do computations. |
| 204 |
% |
| 205 |
% \begin{macro}{\tmp@stack} |
| 206 |
% \begin{macrocode} |
| 207 |
\newcount\tmp@stack |
| 208 |
% \end{macrocode} |
| 209 |
% \end{macro} |
| 210 |
% |
| 211 |
% Creating a new stack is, essentially, creating a new counter-like |
| 212 |
% macro, initializing it to 0, and creating the macro for stack-top |
| 213 |
% evaluation. |
| 214 |
% |
| 215 |
% \begin{macro}{\NewStack} |
| 216 |
% This stack-top evaluation macro is required due to an ugly trick. |
| 217 |
% During the push, we ``evaluate'' the value to push with an |
| 218 |
% |\edef|, because one might want to push something which contain the |
| 219 |
% previous top of the stack (see the example package provided). |
| 220 |
% If the |\Stack| macro computes the real value on the top of the |
| 221 |
% stack, then, the expanded definition in |\Push| will contain more |
| 222 |
% or less the \emph{content} of |\Stack| and not the value. |
| 223 |
% |
| 224 |
% Another way would be to have a |\csname...\endcsname| pair within |
| 225 |
% another one in an |\ifx| condition, which doesn't work. |
| 226 |
% |
| 227 |
% \begin{macrocode} |
| 228 |
\newcommand\NewStack[2]{% |
| 229 |
\expandafter\gdef\csname #1@count\endcsname{0}% |
| 230 |
\expandafter\do@stack@macro\csname #1@count\endcsname{#1}{#2}% |
| 231 |
} |
| 232 |
\newcommand\do@stack@macro[3]{% |
| 233 |
\expandafter\newcommand\expandafter*% |
| 234 |
\csname stack@#2\endcsname{% |
| 235 |
\expandafter\ifx\csname #2@\romannumeral#1\endcsname\relax |
| 236 |
#3% |
| 237 |
\else |
| 238 |
\csname #2@\romannumeral#1\endcsname |
| 239 |
\fi |
| 240 |
}% |
| 241 |
} |
| 242 |
% \end{macrocode} |
| 243 |
% \end{macro} |
| 244 |
% |
| 245 |
% \begin{macro}{\Stack} |
| 246 |
% I'd rather remove the ``check'' code, but, well, it would suppose |
| 247 |
% the end-user of the macro is a fair developper, which is not |
| 248 |
% realistic. |
| 249 |
% |
| 250 |
% If the stack has been properly created, we call the top-stack evaluation |
| 251 |
% macro. |
| 252 |
% |
| 253 |
% \begin{macrocode} |
| 254 |
\newcommand\Stack[1]{% |
| 255 |
\ifx\csname #1@count\endcsname\relax |
| 256 |
\PackageError{stack}{Undefined stack #1}{You should first create the |
| 257 |
stack with \NewStack}% |
| 258 |
\else |
| 259 |
\csname stack@#1\endcsname |
| 260 |
\fi |
| 261 |
} |
| 262 |
% \end{macrocode} |
| 263 |
% \end{macro} |
| 264 |
% |
| 265 |
% \begin{macro}{\Push} |
| 266 |
% \begin{macro}{\Pop} |
| 267 |
% The |\Push| macro is one of the most complex ones. |
| 268 |
% |
| 269 |
% First, we check if the stack is properly defined. |
| 270 |
% \begin{macrocode} |
| 271 |
\newcommand\Push[2]{% |
| 272 |
\ifx\csname #1@count\endcsname\relax |
| 273 |
\PackageError{stack}{Undefined stack #1}{You should first create the |
| 274 |
stack with \NewStack}% |
| 275 |
\else |
| 276 |
% \end{macrocode} |
| 277 |
% |
| 278 |
% We store the new top-value in a macro, just because one might do |
| 279 |
% |\Push{stack}{\Stack{stack}.ext}| in witch case the expansion of |
| 280 |
% the value to push is something rather tricky. |
| 281 |
% |
| 282 |
% Then, we step the depth of the stack (put the depth in the scractch |
| 283 |
% counter, advance this counter, have the value back in the macro). |
| 284 |
% |
| 285 |
% Only after that the value to be pushed is stored in the corresponding |
| 286 |
% slot. |
| 287 |
% \begin{macrocode} |
| 288 |
\edef\stack@newtop{#2}% |
| 289 |
\tmp@stack\csname #1@count\endcsname\relax |
| 290 |
\global\advance\tmp@stack by 1\relax |
| 291 |
\expandafter\xdef\csname #1@count\endcsname{\the\tmp@stack}% |
| 292 |
\expandafter\xdef\csname #1@\romannumeral\tmp@stack\endcsname |
| 293 |
{\stack@newtop}% |
| 294 |
\fi |
| 295 |
} |
| 296 |
% \end{macrocode} |
| 297 |
% |
| 298 |
% The |\Pop| macro is more simple: if the stack is properly defined and |
| 299 |
% the depth is more than zero, then down-step the depth. |
| 300 |
% \begin{macrocode} |
| 301 |
\newcommand\Pop[1]{% |
| 302 |
\ifx\csname #1@count\endcsname\relax |
| 303 |
\PackageError{stack}{Undefined stack #1}{You should first create the |
| 304 |
stack with \NewStack}% |
| 305 |
\else |
| 306 |
\tmp@stack\csname #1@count\endcsname\relax |
| 307 |
\ifnum\tmp@stack>0\relax |
| 308 |
\advance\tmp@stack by -1\relax |
| 309 |
\expandafter\xdef\csname #1@count\endcsname{\the\tmp@stack}% |
| 310 |
\fi |
| 311 |
\fi |
| 312 |
} |
| 313 |
% \end{macrocode} |
| 314 |
% \end{macro} |
| 315 |
% \end{macro} |
| 316 |
% |
| 317 |
% Now, a debug only macro. |
| 318 |
% \begin{macro}{ShowTop} |
| 319 |
% \begin{macrocode} |
| 320 |
\newcommand\ShowTop[1]{% |
| 321 |
{\edef\top{\Stack{#1}}% |
| 322 |
\typeout{The top of #1: \top}}} |
| 323 |
% \end{macrocode} |
| 324 |
% \end{macro} |
| 325 |
% |
| 326 |
% That's all folks. |
| 327 |
% \begin{macrocode} |
| 328 |
%</package> |
| 329 |
% \end{macrocode} |
| 330 |
% |
| 331 |
% \Finale |
| 332 |
% \end{document} |