| 1 | % \begin{meta-comment} |
| 2 | % |
| 3 | % $Id: mdwtab.dtx,v 1.3 2003/11/10 14:43:48 mdw Exp $ |
| 4 | % |
| 5 | % Another rewrite of the tabular environment, and maths alignments |
| 6 | % |
| 7 | % (c) 1996 Mark Wooding |
| 8 | % |
| 9 | % \end{meta-comment} |
| 10 | % |
| 11 | % \begin{meta-comment} <general public licence> |
| 12 | %% |
| 13 | %% mdwtab package -- another rewrite of the tabular environment, etc. |
| 14 | %% Copyright (c) 1996 Mark Wooding |
| 15 | %% |
| 16 | %% This program is free software; you can redistribute it and/or modify |
| 17 | %% it under the terms of the GNU General Public License as published by |
| 18 | %% the Free Software Foundation; either version 2 of the License, or |
| 19 | %% (at your option) any later version. |
| 20 | %% |
| 21 | %% This program is distributed in the hope that it will be useful, |
| 22 | %% but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 23 | %% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 24 | %% GNU General Public License for more details. |
| 25 | %% |
| 26 | %% You should have received a copy of the GNU General Public License |
| 27 | %% along with this program; if not, write to the Free Software |
| 28 | %% Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
| 29 | %% |
| 30 | % \end{meta-comment} |
| 31 | % |
| 32 | % \begin{meta-comment} <Package preambles> |
| 33 | %<+mdwtab>\NeedsTeXFormat{LaTeX2e} |
| 34 | %<+mdwtab>\ProvidesPackage{mdwtab} |
| 35 | %<+mdwtab> [2003/08/24 1.10 Table typesetting with style] |
| 36 | %<+mathenv>\NeedsTeXFormat{LaTeX2e} |
| 37 | %<+mathenv>\ProvidesPackage{mathenv} |
| 38 | %<+mathenv> [2003/08/24 1.10 Various maths environments] |
| 39 | %<+colour>\NeedsTeXFormat{LaTeX2e} |
| 40 | %<+colour>\ProvidesPackage{mtcolour} |
| 41 | %<+colour> [2003/08/24 1.10 Colour support for mdwtab] |
| 42 | %<+color>\NeedsTeXFormat{LaTeX2e} |
| 43 | %<+color>\ProvidesPackage{mtcolor} |
| 44 | %<+color> [2003/08/24 1.10 Fix for people who can't spell] |
| 45 | % \end{meta-comment} |
| 46 | % |
| 47 | % \CheckSum{3402} |
| 48 | %% \CharacterTable |
| 49 | %% {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 |
| 50 | %% 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 |
| 51 | %% Digits \0\1\2\3\4\5\6\7\8\9 |
| 52 | %% Exclamation \! Double quote \" Hash (number) \# |
| 53 | %% Dollar \$ Percent \% Ampersand \& |
| 54 | %% Acute accent \' Left paren \( Right paren \) |
| 55 | %% Asterisk \* Plus \+ Comma \, |
| 56 | %% Minus \- Point \. Solidus \/ |
| 57 | %% Colon \: Semicolon \; Less than \< |
| 58 | %% Equals \= Greater than \> Question mark \? |
| 59 | %% Commercial at \@ Left bracket \[ Backslash \\ |
| 60 | %% Right bracket \] Circumflex \^ Underscore \_ |
| 61 | %% Grave accent \` Left brace \{ Vertical bar \| |
| 62 | %% Right brace \} Tilde \~} |
| 63 | %% |
| 64 | % |
| 65 | % \begin{meta-comment} |
| 66 | % |
| 67 | %<*driver> |
| 68 | \input{mdwtools} |
| 69 | \describespackage{mdwtab} |
| 70 | \describespackage{mtcolour} |
| 71 | \describespackage{mathenv} |
| 72 | \addcontents{lot}{\listoftables} |
| 73 | \mdwdoc |
| 74 | %</driver> |
| 75 | % |
| 76 | % \end{meta-comment} |
| 77 | % |
| 78 | %^^A------------------------------------------------------------------------- |
| 79 | % \renewcommand{\tabstyle}{\small} |
| 80 | % |
| 81 | % \section{User guide} |
| 82 | % |
| 83 | % |
| 84 | % The \package{mdwtab} package contains a reimplementation of the standard |
| 85 | % \LaTeX\ \env{tabular} and \env{array} environments. This is not just an |
| 86 | % upgraded version: it's a complete rewrite. It has several advantages over |
| 87 | % the official \package{array} package (not raw \LaTeX's, which is even less |
| 88 | % nice), and it's more-or-less compatible. Most of these are rather |
| 89 | % technical, I'll admit. |
| 90 | % |
| 91 | % \begin{itemize} |
| 92 | % |
| 93 | % \item The newcolumn system is properly and perfectly integrated into the |
| 94 | % system. There are now \emph{no} `primitive' column types -- all the |
| 95 | % standard types are created as user-defined columns. |
| 96 | % |
| 97 | % \item You can define entirely different table-like environments using the |
| 98 | % equipment here. It's still hard work, although less so than before. |
| 99 | % I'll do an example of this some time. |
| 100 | % |
| 101 | % \item Construction of the preamble is generally much tidier. I've used |
| 102 | % token registers rather than |\edef|, and it's all done very nicely. |
| 103 | % |
| 104 | % \item Fine spacing before and after rules (described by DEK as `a mark of |
| 105 | % quality') is now utterly trivial, since the preamble-generator will |
| 106 | % store the appropriate information. |
| 107 | % |
| 108 | % \item You can use \env{array} in LR and paragraph modes without having |
| 109 | % to surround it with `|$|' signs. |
| 110 | % |
| 111 | % \item Usually you don't want tables in the middle of paragraphs. For these |
| 112 | % cases, I've provided a simpler way to position the table |
| 113 | % horizontally. |
| 114 | % |
| 115 | % \item Footnotes work properly inside \env{tabular} environments (hoorah!). |
| 116 | % You can `catch' footnotes using the \env{minipage} environment if |
| 117 | % you like. (It uses an internal version of the \package{footnote} |
| 118 | % package to handle footnotes, which doesn't provide extra goodies like |
| 119 | % the \env{footnote} environment; you'll need to load the full package |
| 120 | % explicitly to get them.) |
| 121 | % |
| 122 | % \item Standard \LaTeX\ tabular environments have a problem with lining up |
| 123 | % ruled tables. The |\firsthline| command given in the \textit{\LaTeX\ |
| 124 | % Companion} helps a bit, but it's not really good enough, and besides, |
| 125 | % it doesn't \emph{actually} line the text up right after all. The |
| 126 | % \package{mdwtab} package does the job properly to begin with, so you |
| 127 | % don't need to worry. |
| 128 | % |
| 129 | % \end{itemize} |
| 130 | % |
| 131 | % I've tested the following packages with \package{mdwtab}, and they all |
| 132 | % work. Some of the contortions required to make them work weren't pleasant, |
| 133 | % but you don't need to know about them. By a strange coincidence, all the |
| 134 | % packages were written by David Carlisle. Anyway, here's the list: |
| 135 | % \begin{itemize} |
| 136 | % \item The quite nice \package{dcolumn} package. |
| 137 | % \item The more useful \package{delarray} package. |
| 138 | % \item The rather spiffy \package{hhline} package. |
| 139 | % \item The truly wonderful \package{tabularx} package. |
| 140 | % \item The utterly magnificent \package{longtable} package. |
| 141 | % \end{itemize} |
| 142 | % |
| 143 | % Note that I've looked at \package{supertabular} as well: it won't work, so |
| 144 | % use \package{longtable} instead, 'cos it's much better. |
| 145 | % |
| 146 | % |
| 147 | % \subsection{The downside} |
| 148 | % |
| 149 | % There's no such thing as a free lunch. The \package{mdwtab} environment |
| 150 | % is not 100\% compatible with the \env{tabular} environment found in |
| 151 | % \LaTeXe\ or the \package{array} package. |
| 152 | % |
| 153 | % The differences between \package{mdwtab} and \LaTeXe's \env{tabular} |
| 154 | % environment are as follows: |
| 155 | % |
| 156 | % \begin{itemize} \synshorts \let\`=\lq |
| 157 | % |
| 158 | % \item The vertical spacing in \env{array} environments is different to |
| 159 | % that in \env{tabular} environments. This produces more attractive |
| 160 | % results in most mathematical uses of \env{array}s, in the author's |
| 161 | % opinion. The spacing can be modified by playing with length |
| 162 | % parameters. |
| 163 | % |
| 164 | % \item The presence of horizontal and vertical rules will alter the spacing |
| 165 | % of the table (so a pair of columns separated by a `|' is wider than |
| 166 | % a pair with no separation by "\\arrayrulewidth". This does mean that |
| 167 | % horizontal and vertical rules match up properly -- the usual \LaTeX\ |
| 168 | % environment makes the horizontal rules stop just short of the edge |
| 169 | % of the table, making an ugly mess (check out the \textit{\LaTeX\ |
| 170 | % book} if you don't believe me -- page~62 provides a good example). |
| 171 | % The \package{array} package handles rules in the same way as |
| 172 | % \package{mdwtab}. |
| 173 | % |
| 174 | % \setbox0=\hbox{\footnotesize`\\def\\xcs{\\tabskip=\\fill}'} |
| 175 | % \setbox2=\hbox{\footnotesize`...@{\\span\\xcs}...'} |
| 176 | % \item In common with the \package{array} package, there are some |
| 177 | % restrictions on the use of the "\\extracolsep" command in preambles: |
| 178 | % you may use at most one "\\extracolsep" command in each `@' or `!' |
| 179 | % expression. Also, you can't say |
| 180 | % \begin{listing} |
| 181 | %\newcommand{\xcs}{\extracolsep{\fill}} |
| 182 | % \end{listing} |
| 183 | % and then expect something like `...@{\\xcs}...' to actually work -- |
| 184 | % the "\\extracolsep" mustn't be hidden inside any other |
| 185 | % commands. Because things like `@' expressions aren't expanded at |
| 186 | % the time, "\\extracolsep" has to be searched and processed |
| 187 | % \`by hand'.\footnote{^^A |
| 188 | % All \cs{extracolsep} does is modify the \cs{tabskip} glue, so |
| 189 | % if you were an evil \TeX\ hacker like me, you could just say |
| 190 | % \unhbox0\ and put \unhbox2\ in your preamble. That'd work nicely. |
| 191 | % It also works with the \package{array} package.} |
| 192 | % |
| 193 | % \item Control sequences (commands) in a table's preamble aren't expanded |
| 194 | % before the preamble is read. In fact, commands in the preamble are |
| 195 | % considered to be column types, and their names are entirely |
| 196 | % independent of normal \LaTeX\ commands. No column types of this |
| 197 | % nature have yet been defined\footnote{^^A |
| 198 | % There used to be an internal \cs{@magic} type used by |
| 199 | % \env{eqnarray}, but you're not supposed to know about that. |
| 200 | % Besides, it's not there any more.} |
| 201 | % but the possibility's always there. Use the "\\newcolumntype" or |
| 202 | % "\\coldef" commands to define new column types. |
| 203 | % |
| 204 | % \item The preamble parsing works in a completely different way. There is |
| 205 | % a certain amount of compatibility provided, although it's heavily |
| 206 | % geared towards keeping \package{longtable} happy and probably won't |
| 207 | % work with other packages. |
| 208 | % |
| 209 | % \item Obscure constructs which were allowed by the old preamble parser but |
| 210 | % violate the syntax shown in the next section (e.g., `|@{}|' to |
| 211 | % suppress the "\\doublerulesep" space between two vertical rules, |
| 212 | % described in \textit{The \LaTeX\ Companion} as \`a misuse of the |
| 213 | % `@{...}' qualifier') are now properly outlawed. You will be given |
| 214 | % an error message if you attempt to use such a construction. |
| 215 | % |
| 216 | % \item The `*' forms (which repeat column types) are now expanded at a |
| 217 | % different time. Previously, preambles like `c@*{4}{{:}@}{--}c' |
| 218 | % were considered valid (the example would expand to |
| 219 | % `c@{:}@{:}@{:}@{:}@{--}c'), because `*'s were expanded before the |
| 220 | % preamble was actually parsed. In the new system, `*' is treated |
| 221 | % just like any other preamble character (it just has a rather odd |
| 222 | % action), and preambles like this will result in an error (and |
| 223 | % probably a rather confusing one). |
| 224 | % |
| 225 | % \item David Carlisle's \package{colortbl} package entirely fails to work |
| 226 | % with \package{mdwtab}. However, we now have colour support of our |
| 227 | % own which is at times similar in style. |
| 228 | % |
| 229 | % \end{itemize} |
| 230 | % |
| 231 | % There are also several incompatibilities between \package{mdwtab} and |
| 232 | % \package{array}: |
| 233 | % |
| 234 | % \begin{itemize} \synshorts \let\`=\lq |
| 235 | % |
| 236 | % \item Because of the way "\\newcolumntype" works in the \package{array} |
| 237 | % package, a horrid construction like |
| 238 | % \begin{listing} |
| 239 | %\newcolumntype{x}{{:}} |
| 240 | %\begin{tabular}{|c!xc|} |
| 241 | % \end{listing} |
| 242 | % is considered to be valid, and is interpreted as `|c!{:}c|'. My |
| 243 | % reading of pages~54 and~55 of the \textit{\LaTeX\ book} tells me |
| 244 | % that this sort of thing is forbidden in normal \LaTeX\ commands. |
| 245 | % The \package{mdwtab} preamble parser now treats column type letters |
| 246 | % much more like commands with the result that the hacking above won't |
| 247 | % work any more. The construction above would actually be interpreted |
| 248 | % as `|c!{x}c|' (i.e., the `x' column type wouldn't be expanded to |
| 249 | % `{:}' because the parser noticed that it was the argument to the |
| 250 | % `!' modifier\footnote{^^A |
| 251 | % This is a direct result of the way \TeX\ treats undelimited |
| 252 | % arguments. See chapters~5 and~20 of \textit{The \TeX book} for |
| 253 | % more information about how grouping affects argument reading.}). |
| 254 | % |
| 255 | % \item Most of the points above, particularly those relating to the |
| 256 | % handling of the preamble, also apply to the \package{array} package. |
| 257 | % it's not such an advance over the \LaTeXe\ version as everyone said |
| 258 | % it was. |
| 259 | % |
| 260 | % \end{itemize} |
| 261 | % |
| 262 | % |
| 263 | % \subsection{Syntax} |
| 264 | % |
| 265 | % \DescribeEnv{tabular} |
| 266 | % \DescribeEnv{tabular*} |
| 267 | % \DescribeEnv{array} |
| 268 | % So that everyone knows where I stand, here's a complete syntax for my |
| 269 | % version of the \env{tabular} environment, and friends |
| 270 | % |
| 271 | % \begin{grammar} |
| 272 | % |
| 273 | % <tabular-env> ::= \[[ |
| 274 | % "\\begin" |
| 275 | % \begin{stack} |
| 276 | % "{tabular}" \\ "{tabular*}" "{" <length> "}" \\ |
| 277 | % "{array}" \\ "{smarray}" |
| 278 | % \end{stack} |
| 279 | % \[ "[" <position-arg> "]" \] |
| 280 | % "{" <preamble> "}" <text> |
| 281 | % "\\end" |
| 282 | % \( "{tabular}" \\ "{tabular*}" \\ "{array}" \\ "{smarray}" \) |
| 283 | % \]] |
| 284 | % |
| 285 | % <position-arg> ::= (see below) |
| 286 | % |
| 287 | % <preamble> ::= \[[ |
| 288 | % <first-column> |
| 289 | % \[ \< <column> \> \] |
| 290 | % \]] |
| 291 | % |
| 292 | % <first-column> ::= \[[ \[ <rule> \] <column> \]] |
| 293 | % |
| 294 | % <column> ::= \[[ |
| 295 | % \[ <spacing> \] \[ \< <user-pre-text> \> \] <column-type> |
| 296 | % \[ \< <user-post-text> \> \] \[ <spacing> \] \[ <rule> \] |
| 297 | % \]] |
| 298 | % |
| 299 | % <spacing> ::= \[[ "@" "{" <text> "}" \]] |
| 300 | % |
| 301 | % <user-pre-text> ::= \[[ \[ "?" \] ">" "{" <text> "}" \]] |
| 302 | % |
| 303 | % <column-type> ::= \[[ |
| 304 | % \begin{stack} |
| 305 | % \[ "T" \\ "M" \] \( "l" \\ "c" \\ "r" \) \\ |
| 306 | % \( "p" \\ "m" \\ "b" \) "{" <length> "}" \\ |
| 307 | % "#" "{" <raw-pre-text> "}" "{" <raw-post-text> "}" |
| 308 | % \end{stack} |
| 309 | % \]] |
| 310 | % |
| 311 | % <user-post-text> ::= \[[ \[ "?" \] "<" "{" <text> "}" \]] |
| 312 | % |
| 313 | % <rule> ::= \[[ \( "|" \\ "!" "{" <text> "}" \) \]] |
| 314 | % |
| 315 | % \end{grammar} |
| 316 | % |
| 317 | % If you examine the above very carefully, you'll notice a slight deviation |
| 318 | % from the original -- an |@|-expression \emph{following} a rule is |
| 319 | % considered to be part of the \emph{next} column, not the current one. This |
| 320 | % is, I think, an almost insignificant change, and essential for some of the |
| 321 | % new features. You'll also notice the new |#| column type form, which |
| 322 | % allows you to define new real column types instead of just modifying |
| 323 | % existing ones. It's not intended for direct use in preambles -- it's |
| 324 | % there mainly for the benefit of people who know what they're doing and |
| 325 | % insist on using |\newcolumntype| anyway. |
| 326 | %% |
| 327 | % The actual column types are shown in table~\ref{tbl:columns}. |
| 328 | % |
| 329 | % \begin{table} |
| 330 | % \begin{tabular}[C]{| >{\synshorts} c | m{3in} |} \hlx{hv[1]} |
| 331 | % |
| 332 | % \multicolumn{2}{|c|}{\bf Column types} \\ \hlx{v[1]hv} |
| 333 | % \bf Name & \bf Meaning \\ \hlx{vhv.} |
| 334 | % "l" & Left aligned text (\env{tabular}) or |
| 335 | % equation (\env{array}). \\ \hlx{.} |
| 336 | % "c" & Centred text (\env{tabular}) or |
| 337 | % equation (\env{array}). \\ \hlx{.} |
| 338 | % "r" & Right aligned text (\env{tabular}) or |
| 339 | % equation (\env{array}). \\ \hlx{vhv.} |
| 340 | % "Ml", "Mc" and "Mr" & Left, centre and right aligned |
| 341 | % equations.* \\ \hlx{.} |
| 342 | % "Tl", "Tc" and "Tr" & Left, centre and right aligned |
| 343 | % text.* \\ \hlx{vhv.} |
| 344 | % "p{"<width>"}" & Top aligned paragraph with the given |
| 345 | % width. \\ \hlx{.} |
| 346 | % "m{"<width>"}" & Vertically centred paragraph with |
| 347 | % the given width. \\ \hlx{.} |
| 348 | % "b{"<width>"}" & Bottom aligned paragraph with the |
| 349 | % given width. \\ \hlx{vhv.} |
| 350 | % "#{"<pre>"}{"<post>"}" & User defined column type: |
| 351 | % \<pre> is inserted before the |
| 352 | % cell entry, \<post> is inserted |
| 353 | % afterwards.* \\ \hlx{vhhv[1]} |
| 354 | % |
| 355 | % \multicolumn{2}{|c|}{\bf Other modifier characters} \\ \hlx{v[1]hv} |
| 356 | % \bf Name & \bf Meaning \\ \hlx{vhv.} |
| 357 | % "|" & Inserts a vertical rule between |
| 358 | % columns. \\ \hlx{.} |
| 359 | % "$*["<params>"]" & Inserts a vertical rule of given |
| 360 | % width between columns; "*" selects |
| 361 | % "\arraythickrulewidth".* \\ \hlx{.} |
| 362 | % "!{"<text>"}" & Inserts \<text> between columns, |
| 363 | % treating it as a vertical rule. \\ \hlx{vhv.} |
| 364 | % "@{"<text>"}" & Inserts \<text> instead of the |
| 365 | % usual intercolumn space. \\ \hlx{vhv.} |
| 366 | % ">{"<text>"}" & Inserts \<text> just before the |
| 367 | % actual column entry. \\ \hlx{.} |
| 368 | % "<{"<text>"}" & Inserts \<text> just after the |
| 369 | % actual column entry. \\ \hlx{.} |
| 370 | % "?<{"<text>"}" & Inserts \<text> before the column |
| 371 | % entry \emph{and} the rules list.* \\ \hlx{.} |
| 372 | % "?>{"<text>"}" & Inserts \<text> after the column |
| 373 | % entry \emph{and} the rules list.* \\ \hlx{vhv.} |
| 374 | % "*{"<count>"}{"<chars>"}" & Inserts \<count> |
| 375 | % copies of the \<chars> into the |
| 376 | % preamble. \\ \hlx{vhs} |
| 377 | % |
| 378 | % \multicolumn{2}{@{}l}{* This column type is a new feature} |
| 379 | % \end{tabular} |
| 380 | % |
| 381 | % \caption{\package{array} and \package{tabular} column types and modifiers} |
| 382 | % \label{tbl:columns} |
| 383 | % \end{table} |
| 384 | % |
| 385 | % Now that's sorted everything out, there shouldn't be any arguments at all |
| 386 | % about what a column means. |
| 387 | % |
| 388 | % The lowercase \<position-arg>s \lit{t}, \lit{c} and \lit{b} do exactly |
| 389 | % what they did before: control the vertical positioning of the table. The |
| 390 | % uppercase ones control the \emph{horizontal} positioning -- this is how you |
| 391 | % create \emph{unboxed} tables. You can only create unboxed tables in |
| 392 | % paragraph mode. |
| 393 | % |
| 394 | % Note that unboxed tables still can't be broken across pages. Use |
| 395 | % the \package{longtable} package for this, because it already does an |
| 396 | % excellent job. |
| 397 | % |
| 398 | % \DescribeMacro{\tabpause} |
| 399 | % One thing you can to with unboxed tables, however, is to `interrupt' them, |
| 400 | % do some normal typesetting, and then continue. This is achieved by the |
| 401 | % |\tabpause| command: its argument is written out in paragraph mode, and |
| 402 | % the table is continued after the argument finishes. |
| 403 | % Note that it isn't a real argument as far as commands like |\verb| are |
| 404 | % concerned -- they'll work inside |\tabpause| without any problems. |
| 405 | % |
| 406 | % \DescribeMacro{\vline} |
| 407 | % The |\vline| command draws a vertical rule the height of the current table |
| 408 | % cell (unless the current cell is being typeset in paragraph mode -- it |
| 409 | % only works in the simple LR-mode table cells, or in \lit{@} or \lit{!} |
| 410 | % modifiers). It's now been given an optional argument which describes |
| 411 | % parameters for the line. See section~\ref{sec:ruleparams}. |
| 412 | % |
| 413 | % { \let\tabstyle=\relax |
| 414 | % \begin{demo}{An example of \cmd\vline} |
| 415 | %\large |
| 416 | %\begin{tabular} |
| 417 | % {| c !{\vline[5pt]} c | c |} |
| 418 | % \hlx{hv} |
| 419 | % \bf A & \it B & \sf C \\ |
| 420 | % \hlx[2pt]{vhv} |
| 421 | % \bf D & \it E & \sf F \\ |
| 422 | % \hlx{vh} |
| 423 | %\end{tabular} |
| 424 | % \end{demo} |
| 425 | % } |
| 426 | % |
| 427 | % \DescribeMacro{smarray} |
| 428 | % You've probably noticed that there's an unfamiliar environment mentioned |
| 429 | % in the syntax shown above. The \env{smarray} environment produces a |
| 430 | % `small' array, with script size cells rather than the normal full text |
| 431 | % size cells. I've seen examples of this sort of construction\footnote{^^A |
| 432 | % There's a nasty use of \env{smallmatrix} in the |testmath.tex| file which |
| 433 | % comes with the \package{amslatex} distribution. It's actually there to |
| 434 | % simulate a `smallcases' environment, which the \package{mathenv} package |
| 435 | % includes, based around \env{smarray}.} ^^A |
| 436 | % being implemented by totally unsuitable commands. Someone may find it |
| 437 | % handy. |
| 438 | % |
| 439 | % |
| 440 | % \subsection{An updated \cs{cline} command} |
| 441 | % |
| 442 | % \DescribeMacro{\cline} |
| 443 | % \DescribeMacro{\hline} |
| 444 | % The standard \LaTeX\ |\cline| command has been updated. As well as just |
| 445 | % passing a range of columns to draw lines through, you can now pass a comma |
| 446 | % separated list of column numbers and ranges: |
| 447 | % |
| 448 | % \begin{grammar} |
| 449 | % <cline-cmd> ::= \[[ |
| 450 | % "\\cline" |
| 451 | % \[ "*" \] \\ \[ "[" <rule-params> "]" \] |
| 452 | % "{" \< <number> \[ "-" <number> \] \\ "," \> "}" |
| 453 | % \]] |
| 454 | % \end{grammar} |
| 455 | % |
| 456 | % The rules printed by |\cline| and |\hline| can be modified by rule |
| 457 | % parameters: see section~\ref{sec:ruleparams}. |
| 458 | % |
| 459 | % Note that |\cline| rules are rather bodgy. Other rules now have |
| 460 | % \emph{thickness}, but |\cline|s don't. Instead, they backspace over |
| 461 | % previous stuff. If you don't like that, insert an appropriate gap, using |
| 462 | % |\vgap|. The \lit{z} rune in |\hlx| is perfect for this kind of thing -- |
| 463 | % precede your \lit{c} lines by \lit{z} lines for best results. |
| 464 | % |
| 465 | % {\let\tabstyle\relax |
| 466 | % \begin{demo}[w]{A \cs{cline} example} |
| 467 | %\newcommand{\mc}{\multicolumn{1}} |
| 468 | %\begin{tabular}[C]{|c|c|c|c|} \cline{2,4} |
| 469 | % \mc{c|}{one} & two & three & four \\ \hline |
| 470 | % five & six & seven & \mc{c}{eight} \\ \cline{1,3} |
| 471 | %\end{tabular} |
| 472 | % \end{demo} |
| 473 | % } |
| 474 | % |
| 475 | % \subsection{Other stuff} |
| 476 | % |
| 477 | % \DescribeMacro\nextrow |
| 478 | % The \env{tabular} and \env{array} environments maintain a counter |
| 479 | % \textsf{tabrow}. The counter is reset to zero at the start of each table. |
| 480 | % It is stepped by one by default; you can provide an optional argument which |
| 481 | % is the amount to add. |
| 482 | % |
| 483 | % |
| 484 | % \subsection{Spacing control} |
| 485 | % |
| 486 | % One of the most irritating things about \LaTeX's tables is that there isn't |
| 487 | % enough space around horizontal rules. Donald Knuth, in \textit{The |
| 488 | % \TeX book}, describes addition of some extra vertical space here as `a mark |
| 489 | % of quality', and since \TeX\ was designed to produce `beautiful documents' |
| 490 | % it seems a shame that \LaTeX\ doesn't allow this to be done nicely. Well, |
| 491 | % it does now. |
| 492 | % |
| 493 | % \DescribeMacro{\vgap} |
| 494 | % The extra vertical space is added using a command |\vgap|, with the |
| 495 | % following syntax: |
| 496 | % |
| 497 | % \begin{grammar} |
| 498 | % |
| 499 | % <vgap-cmd> ::= \[[ |
| 500 | % "\\vgap" \[ "[" <which-cols> "]" \] "{" <length> "}" |
| 501 | % \]] |
| 502 | % |
| 503 | % <which-cols> ::= \[[ \< <number> \[ "-" <number> \] \\ "," \> \]] |
| 504 | % |
| 505 | % \end{grammar} |
| 506 | % |
| 507 | % This command must appear either immediately after the beginning of the |
| 508 | % table or immediately after the |\\| which ends a row. (Actually, there are |
| 509 | % other commands which also have this requirement -- you can specify a |
| 510 | % collection of them wherever you're allowed to give any one.) It adds some |
| 511 | % vertical space (the amount is given by the \<length>) to the table, |
| 512 | % making sure that the vertical rules of the table are extended correctly. |
| 513 | % |
| 514 | % The |\vgap| command relies on information stored while your table preamble |
| 515 | % is being examined. However, it's possible that you might not want some |
| 516 | % of the rules drawn (e.g., if you've used |\multicolumn|). The optional |
| 517 | % \<which-cols> argument allows you to specify which rules are \emph{not} |
| 518 | % to be drawn. You can specify either single column numbers or ranges. The |
| 519 | % rule at the very left hand side is given the number~0; the rules at the |
| 520 | % end of column~$n$ are numbered~$n$. It's easy really. |
| 521 | % |
| 522 | % \DescribeMacro{\hlx} |
| 523 | % Using |\vgap| is all very well, but it's a bit cumbersome, and takes up a |
| 524 | % lot of typing, especially when combined with |\hline| commands. The |\hlx| |
| 525 | % command tries to tidy things. |
| 526 | % |
| 527 | % The syntax is simple: |
| 528 | % \begin{grammar} |
| 529 | % |
| 530 | % <hlx-cmd> ::= \[[ |
| 531 | % "\\hlx" |
| 532 | % \[ "*" \] \[ "[" <rule-params> "]" \] |
| 533 | % "{" |
| 534 | % \begin{rep} |
| 535 | % \begin{stack} |
| 536 | % "h" \\ |
| 537 | % \tok{"v["<which-cols>"]["<length>"]"} \\ |
| 538 | % \tok{"z["<which-cols>"]["<length>"]"} \\ |
| 539 | % \tok{"s["<length>"]"} \\ |
| 540 | % \tok{"c{"<which-cols>"}"} \\ |
| 541 | % "b" \\ |
| 542 | % \tok{"/["<number>"]"} \\ |
| 543 | % \tok{"!{"<rule-params>"}"} \\ |
| 544 | % \tok{"?{"<stuff>"}"} \\ |
| 545 | % \tok{"+["<step>"]"} \\ |
| 546 | % "." |
| 547 | % \end{stack} |
| 548 | % \end{rep} |
| 549 | % "}" |
| 550 | % \]] |
| 551 | % |
| 552 | % \end{grammar} |
| 553 | % The |*| or optional \<rule-params> give rule-drawing parameters for the |h| |
| 554 | % and |c| subcommands. (Note that you can't pass a |*| or an optional |
| 555 | % parameters argument to the |h| or |c| subcommands directly.) See |
| 556 | % section~\ref{sec:ruleparams}. |
| 557 | % |
| 558 | % The argument works a bit like a table preamble, really. Each letter is a |
| 559 | % command. The following are supported: |
| 560 | % |
| 561 | % \begin{description} |
| 562 | % |
| 563 | % \item [\lit*{h}] Works just like |\hline|. If you put two adjacent to each |
| 564 | % other, a gap will be put between them. |
| 565 | % |
| 566 | % \item [\lit*{v[}\<which-cols>\lit*{][}\<length>\lit*{]}] Works |
| 567 | % like \syntax{"\\vgap["<which-cols>"]{"<length>"}"}. If the |
| 568 | % \<length> is omitted, the value of |\doublerulesep| is used. |
| 569 | % This usually looks right. |
| 570 | % |
| 571 | % \item [\lit*{z[}\<which-cols>\lit*{][}\<length>\lit*{]}] Like \lit{v}, |
| 572 | % except that the default gap is the current rule width (set by the |
| 573 | % \<rule-params>) rather than |\doublerulesep|. This is a good thing |
| 574 | % to insert before a |\cline| row. |
| 575 | % |
| 576 | % \item [\lit*{s[}\<length>\lit*{]}] Leaves a vertical gap with the |
| 577 | % given size. If you omit the \<length> then |\doublerulesep| is |
| 578 | % used. This is usually right. |
| 579 | % |
| 580 | % \item [\lit*{c\char`\{}\<which-cols>\lit*{\char`\}}] Works just like |
| 581 | % |\cline|. |
| 582 | % |
| 583 | % \item [\lit*{b}] Inserts a backspace the width of a rule. This is useful |
| 584 | % when doing \package{longtable}s. |
| 585 | % |
| 586 | % \item [\lit*{/[}\<number>\lit*{]}] Allows a page break in a table. Don't |
| 587 | % use this except in a \env{longtable} environment. The \<number> |
| 588 | % works exactly the same as it does in the |\pagebreak| command, |
| 589 | % except that the default is 0, which just permits a break without |
| 590 | % forcing it. |
| 591 | % |
| 592 | % \item [\lit*{!\char`\{}\<rule-params>\lit*{\char`\}}] Change the rule |
| 593 | % parameters to be used for subsequent subcommands. |
| 594 | % |
| 595 | % \item [\lit*{?\char`\{}\<stuff>\lit*{\char`\}}] Do \<stuff>, which can be |
| 596 | % any commands which \emph{don't} typeset anything. |
| 597 | % |
| 598 | % \item [\lit*{+[\<step>]}] Add \<step> (default is 1) to the value of the |
| 599 | % \textsf{tabrow} counter. |
| 600 | % |
| 601 | % \item [\lit*{.}] (That's a dot) Starts the next row of the table. No |
| 602 | % more characters may follow the dot, and no |\hline|, |\hlx|, |\vgap| |
| 603 | % or |\multicolumn| commands may be used after it. You don't have to |
| 604 | % include it, and most of the time it's totally useless. It can be |
| 605 | % handy for some macros, though. I used it in (and in fact added it |
| 606 | % especially for) the table of column types. |
| 607 | % |
| 608 | % \end{description} |
| 609 | % |
| 610 | % An example of the use of |\hlx| is given, so you can see what's going on. |
| 611 | % |
| 612 | % \begin{figure} |
| 613 | % \let\tabstyle\relax |
| 614 | % \begin{demo}[w]{Beautiful table example} |
| 615 | %\newcommand{\zerowidth}[1]{\hbox to 0pt{\hss#1\hss}} |
| 616 | %\setlength{\tabcolsep}{1.5em} |
| 617 | %\begin{tabular}[C]{| r | c | r |} \hlx{hv[1,2]} |
| 618 | % \multicolumn{3}{|c|}{\bf AT\&T Common Stock} \\ \hlx{v[1,2]hv} |
| 619 | % \multicolumn{1}{|c|}{\zerowidth{\bf Year}} & |
| 620 | % \multicolumn{1}{c|}{\zerowidth{\bf Price}} & |
| 621 | % \multicolumn{1}{c|}{\zerowidth{\bf Dividend}} \\ \hlx{vhv} |
| 622 | % 1971 & 41--54 & \$2.60 \\ |
| 623 | % 2 & 41--54 & 2.70 \\ |
| 624 | % 3 & 46--55 & 2.87 \\ |
| 625 | % 4 & 40--53 & 3.24 \\ |
| 626 | % 5 & 45--52 & 3.40 \\ |
| 627 | % 6 & 51--59 & .95\rlap{*} \\ \hlx{vhs} |
| 628 | % \multicolumn{3}{@{}l}{* (first quarter only)} |
| 629 | %\end{tabular} |
| 630 | % \end{demo} |
| 631 | % \end{figure} |
| 632 | % |
| 633 | % |
| 634 | % \subsection{Creating beautiful long tables} |
| 635 | % |
| 636 | % You can use the |\vgap| and |\hlx| commands with David Carlisle's |
| 637 | % stunning \package{longtable} package. However, there are some things you |
| 638 | % should be away of to ensure that your tables always come out looking |
| 639 | % lovely. |
| 640 | % |
| 641 | % The \package{longtable} package will break a table at an |\hline| command, |
| 642 | % leaving a rule at the bottom of the page and another at the top of the |
| 643 | % next page. This means that a constructions like |\hlx{vhv}| will be |
| 644 | % broken into something like |\hlx{vh}| at the bottom of the page and |
| 645 | % |\hlx{hv}| at the top of the next. You need to design the table headers |
| 646 | % and footers with this in mind. |
| 647 | % |
| 648 | % However, there appears to be a slight problem:\footnote |
| 649 | % {You might very well call it a bug. I couldn't possibly comment.} |
| 650 | % if the footer starts with an |\hline|, and a page is broken at an |\hline|, |
| 651 | % then you get an extra thick rule at the bottom of the page. This is a bit |
| 652 | % of a problem, because if the rule isn't there in the footer and you get |
| 653 | % a break between two rows \emph{without} a rule between them, then the page |
| 654 | % looks very odd. |
| 655 | % |
| 656 | % If you want to do ruled longtables, I'd recommend that you proceed as |
| 657 | % follows: |
| 658 | % \begin{itemize} |
| 659 | % \item End header sections with an |\hlx{vh}|. |
| 660 | % \item Begin footer sections with an |\hlx{bh}|. |
| 661 | % \item Begin the main table with |\hlx{v}|. |
| 662 | % \item Insert |\hlx{vhv}| commands in the main table body as usual. |
| 663 | % \end{itemize} |
| 664 | % If \package{longtable} gets modified appropriately, the use of the \lit{b} |
| 665 | % command won't be necessary. |
| 666 | % |
| 667 | % Here's an example of the sort of thing you'd type. |
| 668 | % |
| 669 | % \begin{listinglist} \listingsize |
| 670 | % \verb"\begin{longtable}[c]{|c|l|} \hlx{hv}" \\ |
| 671 | % \verb"\bf Heading & \bf Also heading \\ \hlx{vh}" \\ |
| 672 | % \verb"\endhead" \\ |
| 673 | % \verb"\hlx{bh}" \\ |
| 674 | % \verb"\endfoot" \\ |
| 675 | % \verb"\hlx{v}" \\ |
| 676 | % \verb"First main & table line \\ \hlx{vhv}" \\ |
| 677 | % \verb"Lots of text & like this \\ \hlx{vhv}" \\ |
| 678 | % \null\quad\vdots \\ |
| 679 | % \verb"Lots of text & like this \\ \hlx{vhv}" \\ |
| 680 | % \verb"Last main & table line \\ \hlx{vh}" \\ |
| 681 | % \verb"\end{longtable}" |
| 682 | % \end{listinglist} |
| 683 | % |
| 684 | % |
| 685 | % \subsection{Rules and vertical positioning} |
| 686 | % |
| 687 | % In the \LaTeXe\ and \package{array.sty} versions of \env{tabular}, you run |
| 688 | % into problems if you try to use ruled tables together with the \lit{[t]} or |
| 689 | % \lit{[b]} position specifiers -- the top or bottom rule ends up being |
| 690 | % nicely lined up with the text baseline, giving you an effect which is |
| 691 | % nothing like the one you expected. The \textit{\LaTeX\ Companion} gives |
| 692 | % two commands |\firsthline| and |\lasthline| which are supposed to help with |
| 693 | % this problem. (These commands have since migrated into the \package{array} |
| 694 | % package.) Unfortunately, |\firsthline| doesn't do its job properly -- |
| 695 | % it gets the text position wrong by exactly the width of the table rules. |
| 696 | % |
| 697 | % The \package{mdwtab} package makes all of this automatic. It gets the |
| 698 | % baseline positions exactly right, whether or not you use rules. Earlier |
| 699 | % versions of this package required that you play with a length parameter |
| 700 | % called |\rulefudge|; this is no longer necessary (or even possible -- the |
| 701 | % length parameter no longer exists). The package now correctly compensates |
| 702 | % for all sorts of rules and |\vgap|s at the top and bottom of a table and |
| 703 | % it gets the positioning right all by itself. You've never had it so good. |
| 704 | % |
| 705 | % |
| 706 | % \subsection{Rule parameters} |
| 707 | % \label{sec:ruleparams} |
| 708 | % |
| 709 | % The rule-drawing commands |\hline|, |\vline|, |\cline| and |\hlx|, and the |
| 710 | % |$| column type (which is otherwise a synonym for "|") accept \emph{rule |
| 711 | % parameters}. If the command is followed by a |*|, then the rules are a bit |
| 712 | % thicker than usual -- they use |\arraythickrulewidth| rather than |
| 713 | % |\arrayrulewidth|. However, there's an optional argument which can contain |
| 714 | % one of: |
| 715 | % |
| 716 | % \begin{description} |
| 717 | % \renewcommand\makelabel[1]{\kern\labelsep\ttfamily#1} |
| 718 | % \item[thin] Use |\arrayrulewidth| as the line width. This is the default. |
| 719 | % \item[thick] Use |\arraythickrulewidth| as the line width. This is the |
| 720 | % same as giving a |*| after the command. |
| 721 | % \item[width=\<length>] Make the rules exactly \<length> wide. |
| 722 | % \item[\<length>] The same as \texttt{width=\<length>}, for compatibility. |
| 723 | % \end{description} |
| 724 | % |
| 725 | % More of these keywords will be added later if past experience is anything |
| 726 | % to go by. Note that the individual |\hlx| subcommands \emph{don't} take |
| 727 | % rule parameters, but see the |!| subcommand for updating the current |
| 728 | % parameters. |
| 729 | % |
| 730 | % \DescribeMacro\tabsetruleparams |
| 731 | % If you say \syntax{"\\tabsetruleparams{"<rule-params>"}"} then the |
| 732 | % \<rule-params> will be prepended to any parameters provided to specific |
| 733 | % rule-drawing commands (including the \lit{\char`\|} preamble command). For |
| 734 | % example, |\tabsetruleparams{thick}| makes all rules thick. This is a local |
| 735 | % declaration. |
| 736 | % |
| 737 | % |
| 738 | % \subsection{User serviceable parts} |
| 739 | % |
| 740 | % There are a lot of parameters which you can modify in order to make arrays |
| 741 | % and tables look nicer. They are all listed in table~\ref{tbl:config}. |
| 742 | % |
| 743 | % \begin{table} |
| 744 | % \begin{tabular}[C]{| l | m{3in} |} \hlx{hv} |
| 745 | % \bf Parameter & \bf Meaning \\ \hlx{vhv} |
| 746 | % |\tabstyle| & A command executed at the beginning of |
| 747 | % a \env{tabular} or \env{tabular$*$} |
| 748 | % environment. By default does nothing. |
| 749 | % Change using |\renewcommand|. \\ \hlx{vhv} |
| 750 | % |\extrarowheight| & A length added to the height of every |
| 751 | % row, used to stop table rules |
| 752 | % overprinting ascenders. Default 0\,pt. |
| 753 | % Usage is deprecated now: use |\hlx| |
| 754 | % instead. \\ \hlx{vhv} |
| 755 | % |\tabextrasep| & Extra space added between rows in a |
| 756 | % \env{tabular} or \env{tabular$*$} |
| 757 | % environment (added \emph{before} any |
| 758 | % following |\hline|). Default 0\,pt. \\ |
| 759 | % |\arrayextrasep| & Analogous to |\tabextrasep|, but for |
| 760 | % \env{array} environments. Default |
| 761 | % 1\,jot (3\,pt). \\ |
| 762 | % |\smarrayextrasep| & Analogous to |\tabextrasep|, but for |
| 763 | % \env{smarray} environments. Default |
| 764 | % 1\,pt. \\ \hlx{vhv} |
| 765 | % |\tabcolsep| & Space added by default on each side of |
| 766 | % a table cell (unless suppressed by an |
| 767 | % \lit{@}-expression) in \env{tabular} |
| 768 | % environments. Default is defined by |
| 769 | % your document class. \\ |
| 770 | % |\arraycolsep| & Analogous to |\tabcolsep|, but for |
| 771 | % \env{array} environments. Default is |
| 772 | % defined by your document class. \\ |
| 773 | % |\smarraycolsep| & Analogous to |\tabcolsep|, but for |
| 774 | % \env{smarray} environments. Default |
| 775 | % is 3\,pt. \\ \hlx{vhv} |
| 776 | % |\arrayrulewidth| & The width of horizontal and vertical |
| 777 | % rules in tables. \\ |
| 778 | % |\arraythickrulewidth|& The width of starred rules in tables. \\ |
| 779 | % |\doublerulesep| & Space added between two adjacent |
| 780 | % vertical or horizontal rules. Also |
| 781 | % used by |\hlx{v}|. \\ \hlx{vhv} |
| 782 | % |\arraystretch| & Command containing a factor to |
| 783 | % multiply the default row height. |
| 784 | % Default is defined by your document |
| 785 | % class (usually 1). \\ \hlx{vh} |
| 786 | % \end{tabular} |
| 787 | % |
| 788 | % \caption{Parameters for configuring table environments} |
| 789 | % \label{tbl:config} |
| 790 | % |
| 791 | % \end{table} |
| 792 | % |
| 793 | % |
| 794 | % \subsection{Defining column types} |
| 795 | % |
| 796 | % \DescribeMacro{\newcolumntype} |
| 797 | % The easy way to define new column types is using |\newcolumntype|. It |
| 798 | % works in more or less the same way as |\newcommand|: |
| 799 | % |
| 800 | % \begin{grammar} |
| 801 | % |
| 802 | % <new-col-type-cmd> ::= \[[ |
| 803 | % "\\newcolumntype" |
| 804 | % "{" <column-name> "}" |
| 805 | % \[ "[" <num-args> "]" \] |
| 806 | % \[ "[" <default-arg> "]" \] |
| 807 | % "{" <first-column> \[ \< <column> \> \] "}" |
| 808 | % \]] |
| 809 | % |
| 810 | % \end{grammar} |
| 811 | % |
| 812 | % (The \env{array.sty} implementation doesn't accept the \<default-arg> |
| 813 | % argument. I've no idea why not, 'cos it was very easy to implement.) |
| 814 | % |
| 815 | % \DescribeMacro{\colset} |
| 816 | % This implementation allows you to define lots of different sets of columns. |
| 817 | % You can change the current set using the |\colset| declaration: |
| 818 | % \begin{grammar} |
| 819 | % <colset-cmd> ::= \[[ "\\colset" "{" <set-name> "}" \]] |
| 820 | % \end{grammar} |
| 821 | % This leaves a problem, though: at any particular moment, the current |
| 822 | % column set could be anything, since other macros and packages can change |
| 823 | % it. |
| 824 | % |
| 825 | % \DescribeMacro{\colpush} |
| 826 | % \DescribeMacro{\colpop} |
| 827 | % What actually happens is that a stack of column sets is maintained. The |
| 828 | % |\colset| command just replaces the item at the top of the stack. The |
| 829 | % command |\colpush| pushes its argument onto the top of the stack, making |
| 830 | % it the new current set. The corresponding |\colpop| macro (which doesn't |
| 831 | % take any arguments) removes the top item from the stack, reinstating the |
| 832 | % previous current column set. |
| 833 | % |
| 834 | % \begin{grammar} |
| 835 | % <colpush-cmd> ::= \[[ "\\colpush" "{" <set-name> "}" \]] |
| 836 | % <colpop-cmd> ::= \[[ "\\colpop" \]] |
| 837 | % \end{grammar} |
| 838 | % |
| 839 | % The macros which manipulate the column set stack work \emph{locally}. |
| 840 | % The contents of the stack are saved when you open a new group. |
| 841 | % |
| 842 | % To make sure everyone behaves themselves properly, these are the rules for |
| 843 | % using the column set stack: |
| 844 | % |
| 845 | % \begin{itemize} |
| 846 | % |
| 847 | % \item Packages defining column types must ensure that they preserve the |
| 848 | % current column set. Either they must push their own column type |
| 849 | % and pop it off when they're finished defining columns, or they must |
| 850 | % avoid changing the stack at all, and use the optional arguments to |
| 851 | % |\coldef| and |\collet|. |
| 852 | % |
| 853 | % \item Packages must not assume that any particular column set is current |
| 854 | % unless they have made sure of it themselves. |
| 855 | % |
| 856 | % \item Packages must ensure that they pop exactly as much as they push. |
| 857 | % There isn't much policing of this (perhaps there should be more), |
| 858 | % so authors are encouraged to behave responsibly. |
| 859 | % |
| 860 | % \item Packages must change the current column set (using |\colset|) when |
| 861 | % they start up their table environment. This will be restored when |
| 862 | % the environment closes. |
| 863 | % |
| 864 | % \end{itemize} |
| 865 | % |
| 866 | % \DescribeMacro{\coldef} |
| 867 | % |\newcolumntype| is probably enough for most purposes. However, Real |
| 868 | % \TeX nicians, and people writing new table-generating environments, require |
| 869 | % something lower-level. |
| 870 | % |
| 871 | % \begin{grammar} |
| 872 | % <coldef-cmd> ::= \[[ |
| 873 | % "\\coldef" |
| 874 | % \[ "[" <set-name> "]" \] |
| 875 | % <col-name> <arg-template> "{" <replacement-text> "}" |
| 876 | % \]] |
| 877 | % \end{grammar} |
| 878 | % |
| 879 | % Note that this defines a column type in the current colset. It works |
| 880 | % almost exactly the same way as \TeX's primitive |\def|. There is a |
| 881 | % potential gotcha here: a |\tab@mkpream| token is inserted at the end of |
| 882 | % your replacement text. If you need to read an optional argument or |
| 883 | % something, you'll need to gobble this token before you carry on. The |
| 884 | % |\@firstoftwo| macro could be handy here: |
| 885 | % \begin{listing} |
| 886 | %\coldef x{\@firstoftwo{\@ifnextchar[\@xcolumn@i\@xcolumn@ii}}} |
| 887 | % \end{listing} |
| 888 | % This isn't a terribly pretty state of affairs, and I ought to do something |
| 889 | % about it. I've not seen any use for an optional argument yet, though. |
| 890 | % Note that if you do gobble the |\tab@mkpream|, it's your responsibility to |
| 891 | % insert another one at the very end of your macro's expansion (so that |
| 892 | % further preamble characters can be read). |
| 893 | % |
| 894 | % The replacement text is inserted directly. It's normal to insert preamble |
| 895 | % elements here. There are several to choose from: |
| 896 | % |
| 897 | % \begin{description} |
| 898 | % |
| 899 | % \item [Column items] provide the main `meat' of a column. You insert a |
| 900 | % column element by saying |
| 901 | % \syntax{"\\tabcoltype{"<pre-text>"}{"<post-text>"}"}. |
| 902 | % The user's text gets inserted between these two. (So do user pre- |
| 903 | % and post-texts. Bear this in mind.) |
| 904 | % |
| 905 | % \item [User pre-text items] work like the \lit{>} preamble command. You |
| 906 | % use the \syntax{"\\tabuserpretype{"<text>"}"} command to insert it. |
| 907 | % User pre-texts are written in \emph{reverse} order between the |
| 908 | % pre-text of the column item and the text from the table cell. |
| 909 | % |
| 910 | % \item [User post-text items] work like the \lit{<} preamble command. You |
| 911 | % use the \syntax{"\\tabuserposttype{"<text>"}"} command to insert it. |
| 912 | % Like user pre-texts, user post-texts are written in reverse order, |
| 913 | % between the table cell text and the column item post-text. |
| 914 | % |
| 915 | % \item [Space items] work like the \lit{@} preamble command. They're |
| 916 | % inserted with the \syntax{"\\tabspctype{"<text>"}"} command. |
| 917 | % |
| 918 | % \item [Rule items] work like the `\verb"|"' and \lit{!} commands. You |
| 919 | % insert them with the \syntax{"\\tabruletype{"<text>"}"} command. |
| 920 | % Note that the text is inserted by |\vgap| too, so it should contain |
| 921 | % things which adjust their vertical size nicely. If you really need |
| 922 | % to, you can test |\iftab@vgap| to see if you're in a |\vgap|. |
| 923 | % |
| 924 | % \end{description} |
| 925 | % |
| 926 | % \DescribeMacro{\collet} |
| 927 | % As well as defining columns, you can copy definitions (rather like |\let| |
| 928 | % allows you to copy macros). The syntax is like this: |
| 929 | % |
| 930 | % \begin{grammar} |
| 931 | % |
| 932 | % <collet-cmd> ::= \[[ |
| 933 | % \[ "[" <set-name> "]" \] <col-name> \[ "=" \] \[ "[" <set-name> "]" \] |
| 934 | % <col-name> |
| 935 | % \]] |
| 936 | % |
| 937 | % \end{grammar} |
| 938 | % |
| 939 | % (In other words, you can copy definitions from other column sets.) |
| 940 | % |
| 941 | % |
| 942 | % \subsection{Defining new table-generating environments} |
| 943 | % |
| 944 | % Quite a few routines are provided specifically to help you to define new |
| 945 | % environments which do alignment in a nice way. |
| 946 | % |
| 947 | % \subsubsection{Reading preambles} |
| 948 | % |
| 949 | % The main tricky bit in doing table-like environments is parsing preambles. |
| 950 | % No longer. |
| 951 | % |
| 952 | % \DescribeMacro{\tab@readpreamble} |
| 953 | % \DescribeMacro{\tab@doreadpream} |
| 954 | % The main parser routine is called |\tab@doreadpream|. Given a user |
| 955 | % preamble string as an argument, it will build an |\halign| preamble to |
| 956 | % return to you. However, the preamble produced won't be complete. This is |
| 957 | % because you can actually make multiple calls to |\tab@doreadpream| with |
| 958 | % bits of user preambles. The |\newcolumntype| system uses this mechanism, |
| 959 | % as does the \lit{*} (repeating) modifier. When there really is no more |
| 960 | % preamble to read, you need to \emph{commit} the heldover tokens to the |
| 961 | % output. The |\tab@readpreamble| routine will do this for you -- given a |
| 962 | % user preamble, it builds a complete output from it. |
| 963 | % |
| 964 | % A token register |\tab@preamble| is used to store the generated preamble. |
| 965 | % Before starting, you must initialise this token list to whatever you want. |
| 966 | % There's another token register, |\tab@shortline|, which is used to store |
| 967 | % tokens used by |\vgap|. For each column in the table, the list contains |
| 968 | % an |\omit| (to override the standard preamble) and an |\hfil| space taking |
| 969 | % up most of the column. Finally, for each rule item in the user preamble, |
| 970 | % the shortline list contains an entry of the form: |
| 971 | % \begin{quote} \synshorts |
| 972 | % "\\tab@ckr{"<column-number>"}{"<rule-text>"}" |
| 973 | % \end{quote} |
| 974 | % This is used to decide whether to print the rule or an empty thing of the |
| 975 | % same width. You probably ought to know that the very first column does |
| 976 | % \emph{not} have a leading |\omit| -- this is supplied by |\vgap| so that |
| 977 | % it can then look for optional arguments. |
| 978 | % |
| 979 | % \DescribeMacro{\tab@initread} |
| 980 | % As well as initialising |\tab@preamble| and emptying |\tab@shortline|, |
| 981 | % there are several other operations required to initialise a preamble read. |
| 982 | % These are all performed by the |\tab@initread| macro, although you may want |
| 983 | % to change some of the values for your specific application. For reference, |
| 984 | % the actions performed are: |
| 985 | % \begin{itemize} |
| 986 | % \item initialising the parser state by setting $|\tab@state| = |
| 987 | % |\tab@startstate|$; |
| 988 | % \item clearing the token lists |\tab@preamble| and |\tab@shortlist|; |
| 989 | % \item initialising the macros |\tab@tabtext|, |\tab@midtext|, and |
| 990 | % |\tab@multicol| to their default values of `|&|', |
| 991 | % `|\ignorespaces#\unskip|' and the empty token list respectively.^^A |
| 992 | % \footnote{^^A |
| 993 | % These are macros rather than token lists to avoid hogging all |
| 994 | % the token list registers. Actually, the package only allocates |
| 995 | % two, although it does use almost all of the temporary registers as |
| 996 | % well. Also, there's a lie: \cs{unskip} is too hamfisted to remove |
| 997 | % trailing spaces properly; I really use a macro called |
| 998 | % \cs{@maybe@unskip}} |
| 999 | % \item clearing the internal token list registers |\tab@pretext|, |
| 1000 | % |tab@userpretext| and |\tab@posttext|; |
| 1001 | % \item clearing the column counter |\tab@columns| to zero; |
| 1002 | % \item clearing the action performed when a new column is started (by making |
| 1003 | % the |\tab@looped| macro equal to |\relax|; this is used to make |
| 1004 | % |\multicolumn| macro raise an error if you try to do more than one |
| 1005 | % column); and |
| 1006 | % \item setting up some other switches used by the parser (|\iftab@rule|, |
| 1007 | % |\iftab@initrule| and |\iftab@firstcol|, all of which are set to be |
| 1008 | % |true|). |
| 1009 | % \end{itemize} |
| 1010 | % |
| 1011 | % The macro |\tab@multicol| is used by the |\multicolumn| command to insert |
| 1012 | % any necessary items (e.g., struts) before the actual column text. If you |
| 1013 | % set this to something non-empty, you should probably consider adding a |
| 1014 | % call to the macro to the beginning of |\tab@preamble|. |
| 1015 | % |
| 1016 | % When parsing is finally done, the count register |\tab@columns| contains |
| 1017 | % the number of columns in the alignment. Don't corrupt this value, because |
| 1018 | % it's used for handling |\hline| commands. |
| 1019 | % |
| 1020 | % \subsubsection{Starting new lines} |
| 1021 | % |
| 1022 | % The other messy bit required by table environments is the newline command |
| 1023 | % |\\|. There are nasty complications involved with starting new lines, some |
| 1024 | % of which can be handled by this package, and some on which I can only give |
| 1025 | % advice. |
| 1026 | % |
| 1027 | % \DescribeMacro{\tab@cr} |
| 1028 | % The optional arguments and star-forms etc. can be read fairly painlessly |
| 1029 | % using the |\tab@cr| command: |
| 1030 | % |
| 1031 | % \begin{grammar} |
| 1032 | % <tabcr-cmd> ::= \[[ |
| 1033 | % "\\tab@cr" <command> "{" <non-star-text> "}" "{" <star-text> "}" |
| 1034 | % \]] |
| 1035 | % \end{grammar} |
| 1036 | % |
| 1037 | % This will call your \<command> with two arguments. The first is the |
| 1038 | % contents of the optional argument, or `|\z@|' if there wasn't one. The |
| 1039 | % second is either \<star-text> or \<non-star-text> depending on |
| 1040 | % whether the user wrote the $*$-form or not. |
| 1041 | % |
| 1042 | % Somewhere in your \<command>, you'll have to use the |\cr| primitive to |
| 1043 | % end the table row. After you've done this, you \emph{must} ensure that you |
| 1044 | % don't do anything that gets past \TeX's mouth without protecting it -- |
| 1045 | % otherwise |\hline| and co.\ won't work. I usually wrap things up in a |
| 1046 | % |\noalign| to protect them, although there are other methods. Maybe. |
| 1047 | % |
| 1048 | % You might like to have a look at the \env{eqnarray} implementation provided |
| 1049 | % to see how all this gets put into practice. |
| 1050 | % |
| 1051 | % |
| 1052 | % \subsection{Colour support} |
| 1053 | % |
| 1054 | % I've now added colour support to \package{mdwtab}. That is, you can play |
| 1055 | % with the colours of table cell backgrounds, rules and text. The support |
| 1056 | % isn't there by default: you have to either give the \textsf{colour} option |
| 1057 | % when you load \package{mdwtab}, or include the \package{mtcolour} package |
| 1058 | % yourself. It's very new, and might break. It's probably not as good as |
| 1059 | % \package{colortbl}. I prefer English spellings for the commands and |
| 1060 | % declarations: to reduce confusion, I've provided synonyms with fewer `u's. |
| 1061 | % If only American package authors were so thoughtful. The examples in this |
| 1062 | % part of the documentation may not display correctly in some DVI viewers: |
| 1063 | % for best results, run |dvips| and view the PostScript using (say) |
| 1064 | % GhostScript. |
| 1065 | % |
| 1066 | % \subsubsection{New commands and features} |
| 1067 | % |
| 1068 | % \DescribeMacro\cellcolour |
| 1069 | % The |\cellcolour| command changes the background colour for the current |
| 1070 | % cell. You can use it directly in a table cell, or in the table preamble. |
| 1071 | % It doesn't matter whereabouts in the table cell it goes. Note that |
| 1072 | % unlike the \package{colortbl}, the |\cellcolour| command works on the |
| 1073 | % \emph{entire} contents of the cell, including the |\tabcolsep| space and |
| 1074 | % the rules, if any. That means that it's robust even if there are |@{...}| |
| 1075 | % preamble commands present. |
| 1076 | % |
| 1077 | % The actual syntax is like this: |
| 1078 | % |
| 1079 | % \begin{grammar} |
| 1080 | % <cell-colour-cmd> ::= \[[ |
| 1081 | % \( "\\cellcolour" \\ "\\cellcolor" \) |
| 1082 | % \[ "*" \] |
| 1083 | % \[ "[" <colour-model> "]" \] |
| 1084 | % "{" <colour> "}" |
| 1085 | % \[ "[" <left-overhang> "]" |
| 1086 | % \[ "[" <right-overhang> "]" \] \] |
| 1087 | % \]] |
| 1088 | % \end{grammar} |
| 1089 | % |
| 1090 | % The \lit{*} makes |\cellcolour| override an extant |\rowcolour| command |
| 1091 | % (see below). The \<colour-model> and \<colour> are as for the |\color| |
| 1092 | % command. The \<left-overhang> is how much the colour band should stick out |
| 1093 | % to the left of the cell; and similarly for the \<right-overhang>. If you |
| 1094 | % don't give a \<right-overhang> then the same value is used for both; if you |
| 1095 | % give neither then there's no overhang. The reason you might want overhang |
| 1096 | % is to deal with |\extracolsep| glue. I shouldn't worry about it if I were |
| 1097 | % you. |
| 1098 | % |
| 1099 | % It's very useful to use |\cellcolour| in a preamble, in particular, in the |
| 1100 | % |?>| preamble command (which was added specifically). (If you use only |>| |
| 1101 | % then |\vgap| leaves very odd-looking gaps in the table.) |
| 1102 | % |
| 1103 | % { \let\tabstyle=\relax |
| 1104 | % \begin{demo}{A coloured table} |
| 1105 | %\newcolumntype{\c}[2]{% |
| 1106 | % >{\color{#1}}% |
| 1107 | % ?>{\cellcolour{#2}}% |
| 1108 | %} |
| 1109 | %\begin{tabular} |
| 1110 | % {|\c{cyan}{red}c| |
| 1111 | % \c{magenta}{green}c| |
| 1112 | % \c{yellow}{blue}c|} |
| 1113 | % \hlx{hv} |
| 1114 | % One &Two &Three \\ \hlx{vhv} |
| 1115 | % Four &Five &Six \\ \hlx{vhv} |
| 1116 | % Seven&Eight&Nine \\ \hlx{vh} |
| 1117 | %\end{tabular} |
| 1118 | % \end{demo} |
| 1119 | % } |
| 1120 | % |
| 1121 | % Obviously, judicious use of |\newcolumntype| would abbreviate the above |
| 1122 | % considerably. |
| 1123 | % |
| 1124 | % \DescribeMacro\rowcolour |
| 1125 | % \DescribeMacro\rowcolouroff |
| 1126 | % The |\rowcolour| command changes the background colour in the same way as |
| 1127 | % |\cellcolour|; however, its effect takes precedence over |\cellcolour| (but |
| 1128 | % not |\cellcolour*|) if both are active, and isn't automatically turned off |
| 1129 | % at the start of the next cell. To actually turn it off again, say |
| 1130 | % |\rowcolouroff|. |
| 1131 | % |
| 1132 | % \begin{grammar} |
| 1133 | % <row-colour-cmd> ::= \[[ |
| 1134 | % \( "\\rowcolour" \\ "\\rowcolor" \) |
| 1135 | % \[ "[" <colour-model> "]" \] |
| 1136 | % "{" <colour> "}" |
| 1137 | % \]] |
| 1138 | % \end{grammar} |
| 1139 | % |
| 1140 | % Note that you don't get to specify overhang parameters here. The ones from |
| 1141 | % the |\cellcolour| declaration are used, unless there isn't one in which |
| 1142 | % case there aren't any. |
| 1143 | % |
| 1144 | % \DescribeMacro\ifmod |
| 1145 | % A common thing to do is colour alternate rows of the table differently. |
| 1146 | % This is a bit tricker for \package{mdwtab} than it would be for, say, |
| 1147 | % \package{array}, since it's hard to spot where the `rows' actually change. |
| 1148 | % The solution is to use the \textsf{tabrow} counter, and |\ifmod|. Saying |
| 1149 | % say \syntax{"\\ifmod{"$x$"}{"$m$"}{"$y$"}{"<yes>"}{"<no>"}"} is the same as |
| 1150 | % saying \<yes> if $x \bmod m = y$, and \<no> otherwise. This is typically |
| 1151 | % used as follows. |
| 1152 | % |
| 1153 | % % { \let\tabstyle=\relax |
| 1154 | % \begin{demo}{Alternating row colours} |
| 1155 | %\begin{tabular} |
| 1156 | % {|?>{\ifmod |
| 1157 | % {\value{tabrow}}{2}{1} |
| 1158 | % {\rowcolour{white}} |
| 1159 | % {\rowcolour[gray]{0.9}}} |
| 1160 | % c|c|} |
| 1161 | % \hlx{h+v} |
| 1162 | % One & Two \\ \hlx{vh+v} |
| 1163 | % Three & Four \\ \hlx{vh+v} |
| 1164 | % Five & Six \\ \hlx{vh+v} |
| 1165 | % Seven & Eight \\ \hlx{vh+v} |
| 1166 | % Nine & Ten \\ \hlx{vh+} |
| 1167 | %\end{tabular} |
| 1168 | % \end{demo} |
| 1169 | % } |
| 1170 | % |
| 1171 | % There are new rule parameters for colours. You get a colourful rule if you |
| 1172 | % say \syntax{"colour" "=" <colour>}. You can also say \syntax{"colourmodel" |
| 1173 | % "=" <colour-model>} to choose unnamed colours. |
| 1174 | % |
| 1175 | % When I've thought of what other things need doing, I'll do some of them. |
| 1176 | % The kit I've provided \emph{can} do most interesting things, but it might |
| 1177 | % require a certain level of hacking. Ask me if you want something and it's |
| 1178 | % not obvious how to do it. |
| 1179 | % |
| 1180 | % \subsubsection{Dirty tricks} |
| 1181 | % |
| 1182 | % The colour support interacts with |\vgap| very badly. The preamble rune |
| 1183 | % |?>{\cellcolour{...}}| works well if you want to colour a column, and |
| 1184 | % |\rowcolour| works either in the preamble or as |
| 1185 | % |\hlx{?{\rowcolour{...}}}|. But what if you want to just colour one table |
| 1186 | % cell? You can, as suggested above, just say |\cellcolour{...}| in the |
| 1187 | % table text, but that leaves really nasty-looking gaps above and below if |
| 1188 | % there are adjacent |\vgap| rows. |
| 1189 | % |
| 1190 | % This is what |\hlx{?{...}}| was invented for. Here's a demo. |
| 1191 | % |
| 1192 | % \begin{demo}[w]{Colouring just one cell} |
| 1193 | %\let\hack=\relax |
| 1194 | %\begin{tabular}[C]{|c|?>{\hack}c|} \hlx{hv} |
| 1195 | %Uncoloured & cells here \\ \hlx{vhv} |
| 1196 | %And some & more \\ |
| 1197 | % \hlx{vh?{\gdef\hack{\cellcolour{red}}}v} |
| 1198 | %Yet more & This one's red! \\ |
| 1199 | % \hlx{vh?{\global\let\hack=\relax}v} |
| 1200 | %And more & uncoloured cells \\ \hlx{vh} |
| 1201 | %\end{tabular} |
| 1202 | % \end{demo} |
| 1203 | % |
| 1204 | % |
| 1205 | % \subsection{The \env{mathenv} package alignment environments} |
| 1206 | % |
| 1207 | % The \env{mathenv} package provides several environments for aligning |
| 1208 | % equations in various ways. They're mainly provided as a demonstration of |
| 1209 | % the table handling macros in \package{mdwtab}, so don't expect great |
| 1210 | % things. If you want truly beautiful mathematics, use |
| 1211 | % \package{amsmath}.\footnote{^^A |
| 1212 | % Particularly since nice commands like \cmd\over\ are being reactivated |
| 1213 | % in a later release of \package{amsmath}.} |
| 1214 | % However, the various environments do nest in an approximately useful way. |
| 1215 | % I also think that the \env{matrix} and \env{script} environments provided |
| 1216 | % here give better results than their \package{amsmath} equivalents, and |
| 1217 | % they are certainly more versatile. |
| 1218 | % |
| 1219 | % \subsubsection{The new \env{eqnarray} environment} |
| 1220 | % |
| 1221 | % \DescribeEnv{eqnarray} |
| 1222 | % \DescribeEnv{eqnarray*} |
| 1223 | % As an example of the new column defining features, and because the original |
| 1224 | % isn't terribly good, I've included a rewritten version of the |
| 1225 | % \env{eqnarray} environment. The new implementation closes the gap between |
| 1226 | % \env{eqnarray} and \AmSTeX\ alignment features. It's in a separate, |
| 1227 | % package called \package{mathenv}, to avoid wasting your memory. |
| 1228 | % |
| 1229 | % \begin{grammar} |
| 1230 | % |
| 1231 | % <eqnarray-env> ::= \[[ |
| 1232 | % <begin-eqnarray> \< <row> \\ "\\\\" \> <end-eqnarray> |
| 1233 | % \]] |
| 1234 | % |
| 1235 | % <begin-eqnarray> ::= \[[ |
| 1236 | % "\\begin" \( "{eqnarray}" \\ "{eqnarray*}" \) |
| 1237 | % \[ "[" \< <eqa-column> \> "]" \] |
| 1238 | % \]] |
| 1239 | % |
| 1240 | % <eqa-column> ::= \[[ |
| 1241 | % \[ "q" \\ ":" \] |
| 1242 | % \[ \< ">" "{" <pre-text> "}" \> \] |
| 1243 | % \begin{stack} |
| 1244 | % \[ "T" \] \( "r" \\ "c" \\ "l" \) \\ |
| 1245 | % "L" \\ |
| 1246 | % "x" |
| 1247 | % \end{stack} |
| 1248 | % \[ \< "<" "{" <post-text> "}" \> \] |
| 1249 | % \]] |
| 1250 | % |
| 1251 | % <end-eqnarray> ::= \[[ |
| 1252 | % "\\end" \begin{stack} "{eqnarray}" \\ "{eqnarray*}" \end{stack} |
| 1253 | % \]] |
| 1254 | % |
| 1255 | % \end{grammar} |
| 1256 | % |
| 1257 | % Descriptions of the various column types are given in |
| 1258 | % table~\ref{tbl:eqnarray}. |
| 1259 | % |
| 1260 | % \begin{table} |
| 1261 | % \begin{tabular}[C]{| >{\synshorts} c | m{3in} |} \hlx{hv[1]} |
| 1262 | % |
| 1263 | % \multicolumn{2}{|c|}{\bf Column types} \\ \hlx{v[1]hv} |
| 1264 | % \bf Name & \bf Meaning \\ \hlx{vhv.} |
| 1265 | % "l" & Left aligned piece of equation. \\ \hlx{.} |
| 1266 | % "c" & Centred piece of equation. \\ \hlx{.} |
| 1267 | % "x" & Centred or flush-left whole equation |
| 1268 | % (depending on \textsf{fleqn} option). \\ \hlx{.} |
| 1269 | % "r" & Right aligned piece of equation. \\ \hlx{vhv.} |
| 1270 | % "L" & Left aligned piece of equation whose |
| 1271 | % width is considered to be 2\,em. \\ \hlx{vhv.} |
| 1272 | % "Tl", "Tc" and "Tr" & Left, centre and right aligned |
| 1273 | % text. \\ \hlx{vhhv[1]} |
| 1274 | % |
| 1275 | % \multicolumn{2}{|c|}{\bf Other modifier characters} \\ \hlx{v[1]hv} |
| 1276 | % \bf Name & \bf Meaning \\ \hlx{vhv.} |
| 1277 | % ":" & Leaves a big gap between equations. |
| 1278 | % By default, the `chunks' separated by |
| 1279 | % \lit{:}s are equally spaced on the |
| 1280 | % line. \\ \hlx{.} |
| 1281 | % "q" & Inserts 1\,em of space \\ \hlx{vhv.} |
| 1282 | % ">{"<text>"}" & Inserts \<text> just before the |
| 1283 | % actual column entry. \\ \hlx{.} |
| 1284 | % "<{"<text>"}" & Inserts \<text> just after the |
| 1285 | % actual column entry. \\ \hlx{vhv.} |
| 1286 | % "*{"<count>"}{"<chars>"}" & Inserts \<count> |
| 1287 | % copies of the \<chars> into the |
| 1288 | % preamble. \\ \hlx{vh} |
| 1289 | % \end{tabular} |
| 1290 | % |
| 1291 | % \caption{\package{eqnarray} column types and modifiers} |
| 1292 | % \label{tbl:eqnarray} |
| 1293 | % \end{table} |
| 1294 | % |
| 1295 | % The default preamble, if you don't supply one of your own, is \lit{rcl}. |
| 1296 | % Most of the time, \lit{rl} is sufficient, although compatibility is more |
| 1297 | % important to me. |
| 1298 | % |
| 1299 | % By default, there is no space between columns, which makes formul\ae\ in an |
| 1300 | % \env{eqnarray} environment look just like formul\ae\ typeset on their own, |
| 1301 | % except that things get aligned in columns. This is where the default |
| 1302 | % \env{eqnarray} falls down: it leaves |\arraycolsep| space between each |
| 1303 | % column making the thing look horrible. |
| 1304 | % |
| 1305 | % An example would be good here, I think. This one's from exercise 22.9 of |
| 1306 | % the \textit{\TeX book}. |
| 1307 | % |
| 1308 | % \begin{demo}[w]{Simultaneous equations} |
| 1309 | %\begin{eqnarray}[*3{rc}rl] |
| 1310 | % 10w & + & 3x & + & 3y & + & 18z & = 1 \\ |
| 1311 | % 6w & - & 17x & & & - & 5z & = 2 |
| 1312 | %\end{eqnarray} |
| 1313 | % \end{demo} |
| 1314 | % |
| 1315 | % Choosing a more up-to-date example, here's some examples from the |
| 1316 | % \textit{\LaTeX\ Companion}. |
| 1317 | % |
| 1318 | % \begin{demo}[w]{Lots of equations} |
| 1319 | %\begin{eqnarray}[rl:rl:lq] |
| 1320 | % V_i &= v_i - q_i v_j, & X_i &= x_i - q_i x_j, & |
| 1321 | % U_i = u_i, \qquad \mbox{for $i \ne j$} \\ |
| 1322 | % V_j &= v_j, & X_j &= x_j & |
| 1323 | % U_j u_j + \sum_{i \ne j} q_i u_i. \label{eq:A} |
| 1324 | %\end{eqnarray} |
| 1325 | % \end{demo} |
| 1326 | % |
| 1327 | % \begin{figure} |
| 1328 | % \begin{demo}[w]{Plain text column and \cs{tabpause}} |
| 1329 | %\begin{eqnarray}[rlqqTl] |
| 1330 | % x &= y & by (\ref{eq:A}) \\ |
| 1331 | % x' &= y' & by definition \\ |
| 1332 | %\tabpause{and} |
| 1333 | % x + x' &= y + y' & by Axiom~1 |
| 1334 | %\end{eqnarray} |
| 1335 | % \end{demo} |
| 1336 | % \end{figure} |
| 1337 | % |
| 1338 | % The new features also mean that you don't need to mess about with |
| 1339 | % |\lefteqn| any more. This is handled by the \lit{L} column type: |
| 1340 | % |
| 1341 | % \begin{demo}{Splitting example} |
| 1342 | %\begin{eqnarray*}[Ll] |
| 1343 | % w+x+y+z = \\ |
| 1344 | % & a+b+c+d+e+{} \\ |
| 1345 | % & f+g+h+i+j |
| 1346 | %\end{eqnarray*} |
| 1347 | % \end{demo} |
| 1348 | % |
| 1349 | % Finally, just to prove that the spacing's right at last, here's another one |
| 1350 | % from the \textit{Companion}. |
| 1351 | % |
| 1352 | % \begin{demo}{Spacing demonstration} |
| 1353 | %\begin{equation} |
| 1354 | % x^2 + y^2 = z^2 |
| 1355 | %\end{equation} |
| 1356 | %\begin{eqnarray}[rl] |
| 1357 | % x^2 + y^2 &= z^2 \\ |
| 1358 | % y^2 &< z^2 |
| 1359 | %\end{eqnarray} |
| 1360 | % \end{demo} |
| 1361 | % |
| 1362 | % Well, that was easy enough. Now on to numbering. As you've noticed, the |
| 1363 | % equations above are numbered. You can use the \env{eqnarray$*$} |
| 1364 | % environment to turn off the numbering in the whole environment, or say |
| 1365 | % |\nonumber| on a line to suppress numbering of that one in particular. |
| 1366 | % |
| 1367 | % \DescribeMacro{\eqnumber} |
| 1368 | % More excitingly, you can say |\eqnumber| to enable numbering for a |
| 1369 | % particular equation, or \syntax{"\\eqnumber["<text>"]"} to choose what to |
| 1370 | % show instead of the line number. This works for both starred and unstarred |
| 1371 | % versions of the environment. Now |\nonumber| becomes merely a synonym for |
| 1372 | % `|\eqnumber[]|'. |
| 1373 | % |
| 1374 | % A note for cheats: you can use the sparkly new \env{eqnarray} for simple |
| 1375 | % equations by specifying \lit{x} as the column description. Who needs |
| 1376 | % \AmSTeX?\ |;-)| |
| 1377 | % |
| 1378 | % \DescribeEnv{eqlines} |
| 1379 | % \DescribeEnv{eqlines*} |
| 1380 | % In fact, there's a separate environment \env{eqlines}, which is equivalent |
| 1381 | % to \env{eqnarray} with a single \lit{x} column; the result is that you can |
| 1382 | % insert a collection of displayed equations separated by |\\| commands. If |
| 1383 | % you don't like numbering, use \env{eqlines$*$} instead. |
| 1384 | % |
| 1385 | % \subsubsection{The \env{eqnalign} environment} |
| 1386 | % |
| 1387 | % \DescribeEnv{eqnalign} |
| 1388 | % There's a new environment, \env{eqnalign}, which does almost the same |
| 1389 | % thing as \env{eqnarray} but not quite. It doesn't do equation numbers, |
| 1390 | % and it wraps its contents up in a box. The result of this is that: |
| 1391 | % |
| 1392 | % \begin{itemize} |
| 1393 | % |
| 1394 | % \item You can use \env{eqnalign} for just a part of a formula. |
| 1395 | % The \env{eqnarray} environment must take up the whole display. |
| 1396 | % |
| 1397 | % \item You can use \env{eqnalign} within \env{eqnarray} for extra fine |
| 1398 | % alignment of subsidiary bits. |
| 1399 | % |
| 1400 | % \item You can break off from doing an \env{eqnarray} using the |\tabpause| |
| 1401 | % command. You can't use |\tabpause| inside |
| 1402 | % \env{eqnalign}.\footnote{^^A |
| 1403 | % Well, technically speaking there's nothing to stop you. However, |
| 1404 | % the results won't be pretty.} |
| 1405 | % |
| 1406 | % \end{itemize} |
| 1407 | % |
| 1408 | % The \env{eqnalign} environment works like this: |
| 1409 | % |
| 1410 | % \begin{grammar} |
| 1411 | % |
| 1412 | % <eqnalign-env> ::= \[[ |
| 1413 | % <begin-eqnalign> <contents> <end-eqnalign> |
| 1414 | % \]] |
| 1415 | % |
| 1416 | % <begin-eqnalign> ::= \[[ |
| 1417 | % "\\begin" "{eqnalign}" |
| 1418 | % \[ "[" \< <eqa-column> \> "]" \] |
| 1419 | % \[ "[" \( "t" \\ "c" \\ "b" \) "]" \] |
| 1420 | % \]] |
| 1421 | % |
| 1422 | % <end-eqnalign> ::= \[[ "\\end" "{eqnalign}" \]] |
| 1423 | % |
| 1424 | % \end{grammar} |
| 1425 | % |
| 1426 | % As the syntax suggests, the preamble for the \env{eqnalign} environment |
| 1427 | % works exactly the same way as for \env{eqnarray}. Example time: another |
| 1428 | % one from the \textit{\TeX book}. |
| 1429 | % |
| 1430 | % \begin{figure} |
| 1431 | % \begin{demo}[w]{Example of \env{eqnalign}} |
| 1432 | %\[ |
| 1433 | % \left\{ \begin{eqnalign}[rl] |
| 1434 | % \alpha &= f(z) \\ \beta &= f(z^2) \\ |
| 1435 | % \gamma &= f(z^3) |
| 1436 | % \end{eqnalign} \right\} |
| 1437 | % \qquad |
| 1438 | % \left\{ \begin{eqnalign}[rl] |
| 1439 | % x &= \alpha^2 - \beta \\ y &= 2\gamma |
| 1440 | % \end{eqnalign} \right\}. |
| 1441 | %\] |
| 1442 | % \end{demo} |
| 1443 | % \end{figure} |
| 1444 | % |
| 1445 | % \DescribeMacro{\multicolumn} |
| 1446 | % The |\multicolumn| command works correctly in both the \env{eqnarray} and |
| 1447 | % \env{eqnalign} environments, although you should bear in mind that you |
| 1448 | % should give \env{eqnarray} column types, not \env{array} ones. |
| 1449 | % |
| 1450 | % \subsubsection{A note on spacing in alignment environments} |
| 1451 | % |
| 1452 | % Most of the time, equations in \env{eqnarray} and \env{eqnalign} |
| 1453 | % environments will be beautiful. However, there are some things you should |
| 1454 | % bear in mind when you produce beautiful equations. |
| 1455 | % |
| 1456 | % The main problem with spacing is making sure that binary relations and |
| 1457 | % binary operators have the correct amount of space on each side of them. |
| 1458 | % The alignment environments insert `hidden' objects at the ends of table |
| 1459 | % cells to assist with the spacing: \lit{l} column types have a hidden object |
| 1460 | % on the left, \lit{r} types have a hidden object on the right, and \lit{c} |
| 1461 | % types have a hidden object on \emph{both} ends. These hidden objects add |
| 1462 | % the correct space when there's a binary operator or relation next to them. |
| 1463 | % If some other sort of object is lurking there, no space is added. So far, |
| 1464 | % so good. |
| 1465 | % |
| 1466 | % The only problem comes when you have something like this: |
| 1467 | % |
| 1468 | % \begin{demo}{How not to do an \env{eqnarray}} |
| 1469 | %\begin{eqnarray*}[rcl] |
| 1470 | % x + y & = & 12 \\ |
| 1471 | % 2x - 5y & = & -6 |
| 1472 | %\end{eqnarray*} |
| 1473 | % \end{demo} |
| 1474 | % |
| 1475 | % The `$-$' sign in the second equation has been treated as a binary operator |
| 1476 | % when really it should be a unary prefix operator, but \TeX\ isn't clever |
| 1477 | % enough to know the difference. (Can you see the difference in the spacing |
| 1478 | % between $-6$~and~${}-6$?) There are two possible solutions to the |
| 1479 | % problem. You could wrap the `|-6|' up in a group (`|{-6}|'), or just the |
| 1480 | % $-$ sign (`|{-}6|'). A better plan, though, is to get rid of the middle |
| 1481 | % column altogether: |
| 1482 | % |
| 1483 | % \begin{demo}{How to do an \env{eqnarray}} |
| 1484 | %\begin{eqnarray*}[rl] |
| 1485 | % x + y & = 12 \\ |
| 1486 | % 2x - 5y & = -6 |
| 1487 | %\end{eqnarray*} |
| 1488 | % \end{demo} |
| 1489 | % |
| 1490 | % Since the things in the middle column were the same width, it's not |
| 1491 | % actually doing any good. Also, now that \TeX\ can see that the thing on |
| 1492 | % the left of the `$-$' sign is a relation (the `$=$' sign), it will space |
| 1493 | % the formula correctly. |
| 1494 | % |
| 1495 | % In this case, it might be even better to add some extra columns, and line |
| 1496 | % up the $x$ and $y$ terms in the left hand side: |
| 1497 | % |
| 1498 | % \begin{demo}{Extra beautiful \env{eqnarray}} |
| 1499 | %\begin{eqnarray*}[rrl] |
| 1500 | % x + & y & = 12 \\ |
| 1501 | % 2x - & 5y & = -6 |
| 1502 | %\end{eqnarray*} |
| 1503 | % \end{demo} |
| 1504 | % |
| 1505 | % ^^A Some hacking now to display box sizes. |
| 1506 | % |
| 1507 | % { |
| 1508 | % \catcode`p=12 \catcode`t=12 |
| 1509 | % \gdef\magni#1pt{#1} |
| 1510 | % } |
| 1511 | % |
| 1512 | % \newcommand{\widthof}[1]{^^A |
| 1513 | % \settowidth{\dimen0 }{#1}^^A |
| 1514 | % \expandafter\magni\the\dimen0\,pt^^A |
| 1515 | % } |
| 1516 | % |
| 1517 | % ^^A The text below makes an assumption which looks correct to me (I asked |
| 1518 | % ^^A TeX, and it agreed with me), although in case anything changes, I want |
| 1519 | % ^^A to be informed. |
| 1520 | % |
| 1521 | % \sbox0{$+$} \sbox2{$-$} \ifdim\wd0=\wd2\else% |
| 1522 | % \errmessage{Assertion failed: `+' and `-' are different widths!} |
| 1523 | % \fi |
| 1524 | % |
| 1525 | % There's no need to put the `$+$' and `$-$' operators in their own column |
| 1526 | % here, because they're both \widthof{$+$} wide, even though they don't |
| 1527 | % look it. |
| 1528 | % |
| 1529 | % \subsubsection{Configuring the alignment environments} |
| 1530 | % |
| 1531 | % There are a collection of parameters you can use to make the equation |
| 1532 | % alignment environments (\env{eqnarray} and \env{eqnalign}) look the way |
| 1533 | % you like them. These are all shown in table~\ref{tbl:eqnparms}. |
| 1534 | % |
| 1535 | % \begin{table} |
| 1536 | % \begin{tabular}[C]{| l | p{3in} |} \hlx{hv} |
| 1537 | % \bf Parameter & \bf Use \\ \hlx{vhv} |
| 1538 | % |\eqaopenskip| & Length put on the left of an |
| 1539 | % \env{eqnarray} environment. By |
| 1540 | % default, this is |\@centering| (to |
| 1541 | % centre the alignment) or |\mathindent| |
| 1542 | % (to left align) depending on whether |
| 1543 | % you're using the \textsf{fleqn} |
| 1544 | % document class option. \\ |
| 1545 | % |\eqacloseskip| & Length put on the right of an |
| 1546 | % \env{eqnarray} environment. By |
| 1547 | % default, this is |\@centering|, to |
| 1548 | % align the environment correctly. \\ \hlx{vhv} |
| 1549 | % |\eqacolskip| & Space added by the \lit{:} column |
| 1550 | % modifier. This should be a rubber |
| 1551 | % length, although it only stretches in |
| 1552 | % \env{eqnarray}, not in \env{eqnalign}. |
| 1553 | % The default value is 1\smallf1/2\,em |
| 1554 | % with 1000\,pt of stretch. \\ |
| 1555 | % |\eqainskip| & Space added at each side of a normal |
| 1556 | % column. By default this is 0\,pt. \\ \hlx{vhv} |
| 1557 | % |\eqastyle| & The maths style used in the alignment. |
| 1558 | % By default, this is |\textstyle|, |
| 1559 | % and you probably won't want to change |
| 1560 | % it. \\ \hlx{vh} |
| 1561 | % \end{tabular} |
| 1562 | % |
| 1563 | % \caption{Parameters for the \env{eqnarray} and \env{eqnalign} environments} |
| 1564 | % \label{tbl:eqnparms} |
| 1565 | % \end{table} |
| 1566 | % |
| 1567 | % |
| 1568 | % \subsection{Other multiline equations} |
| 1569 | % |
| 1570 | % Sometimes there's no sensible alignment point for splitting equations. The |
| 1571 | % normal thing to do under these circumstances is to put the first line way |
| 1572 | % over to the left of the page, and the last line over to the right. (If |
| 1573 | % there are more lines, I imagine we put them in the middle.) |
| 1574 | % |
| 1575 | % \DescribeEnv{spliteqn} |
| 1576 | % \DescribeEnv{spliteqn*} |
| 1577 | % The \env{spliteqn} environment allows you to do such splitting of |
| 1578 | % equations. Rather than tediously describe it, I'll just give an example, |
| 1579 | % because it's really easy. The $*$-version works the same, except it |
| 1580 | % doesn't put an equation number in. |
| 1581 | % |
| 1582 | % \begin{figure} |
| 1583 | % \begin{demo}[w]{A split equation} |
| 1584 | %\begin{spliteqn} |
| 1585 | % \sum_{1\le j\le n} |
| 1586 | % \frac {1} { (x_j - x_1) \ldots (x_j - x_{j-1}) |
| 1587 | % (x - x_j) (x_j - x_{j+1}) \ldots (x_j - x_n) } |
| 1588 | % \\ |
| 1589 | % = \frac {1} { (x - x_1) \ldots (x - x_n) }. |
| 1590 | %\end{spliteqn} |
| 1591 | % \end{demo} |
| 1592 | % \end{figure} |
| 1593 | % |
| 1594 | % \DescribeEnv{subsplit} |
| 1595 | % If you have a very badly behaved equation, you might want to split a part |
| 1596 | % of it (say, a bit of a fraction), particularly if you're doing things in |
| 1597 | % narrow columns. |
| 1598 | % |
| 1599 | % \begin{figure} |
| 1600 | % \begin{demo}[w]{A \env{subsplit} environment} |
| 1601 | %\begin{equation} |
| 1602 | % \frac{ |
| 1603 | % \begin{subsplit} |
| 1604 | % q^{\frac{1}{2} n(n+1)}(ea; q^2)_\infty (eq/a; q^2)_\infty \\ |
| 1605 | % (caq/e; q^2)_\infty (cq^2/ae; q^2)_\infty |
| 1606 | % \end{subsplit} |
| 1607 | % }{ |
| 1608 | % (e; q)_\infty (cq/e; q)_\infty |
| 1609 | % } |
| 1610 | %\end{equation} |
| 1611 | % \end{demo} |
| 1612 | % \end{figure} |
| 1613 | % |
| 1614 | % \subsection{Matrices} |
| 1615 | % |
| 1616 | % Also included in the \package{mathenv} package is a collection of things |
| 1617 | % for typesetting matrices. The standard \env{array} doesn't (in my opinion) |
| 1618 | % provide the right sort of spacing for matrices. \PlainTeX\ provides some |
| 1619 | % quite nice matrix handling macros, but they don't work in the appropriate |
| 1620 | % \LaTeX\ way. |
| 1621 | % |
| 1622 | % \textbf{Warning:} These definitions will make old versions of |
| 1623 | % \package{plain.sty} unhappy; newer versions correctly restore the |
| 1624 | % Plain~\TeX\ macros |\matrix| and |\pmatrix|. |
| 1625 | % |
| 1626 | % \DescribeEnv{matrix} |
| 1627 | % The simple way to do matrices is with the \env{matrix} environment. |
| 1628 | % |
| 1629 | % \begin{grammar} |
| 1630 | % |
| 1631 | % <matrix-env> ::= \[[ <begin-matrix> <contents> <end-matrix> \]] |
| 1632 | % |
| 1633 | % <begin-matrix> ::= \[[ "\\begin{matrix}" \[ "[" <matrix-cols> "]" \] \]] |
| 1634 | % |
| 1635 | % <matrix-cols> ::= \[[ |
| 1636 | % \< \[ "[" \] \[ "T" \] \( "l" \\ "c" \\ "r" \) \> |
| 1637 | % \]] |
| 1638 | % |
| 1639 | % <end-matrix> ::= \[[ "\\end{stack}" \]] |
| 1640 | % |
| 1641 | % \end{grammar} |
| 1642 | % |
| 1643 | % The \lit{l}, \lit{c} and \lit{r} columns are fairly obvious -- they align |
| 1644 | % their contents in the appropriate way. The \lit{[} character is more |
| 1645 | % complicated. It means `repeat the remaining column types forever', so a |
| 1646 | % preamble of \lit{cc[lr} means `two centred columns, then alternating left- |
| 1647 | % and right-aligned columns for as often as needed'. The default preamble, |
| 1648 | % if you don't specify one, is \lit{[c} -- `any number of centred columns'. |
| 1649 | % |
| 1650 | % \DescribeMacro{\multicolumn} |
| 1651 | % The |\multicolumn| command works correctly in matrices, although you should |
| 1652 | % bear in mind that you should give \env{matrix} column types, not |
| 1653 | % \env{array} ones. |
| 1654 | % |
| 1655 | % \DescribeEnv{pmatrix} |
| 1656 | % The standard \env{matrix} environment doesn't put any delimiters around the |
| 1657 | % matrix. You can use the standard |\left| and |\right| commands, although |
| 1658 | % this is a bit nasty. The \env{pmatrix} environment will put parentheses |
| 1659 | % around the matrix it creates; it's otherwise exactly the same as |
| 1660 | % \env{matrix}. |
| 1661 | % |
| 1662 | % \DescribeEnv{dmatrix} |
| 1663 | % A \env{dmatrix} environment is also provided. It takes two extra |
| 1664 | % arguments: the left and right delimiter characters (without |\left| or |
| 1665 | % |\right|). |
| 1666 | % |
| 1667 | % \begin{figure} |
| 1668 | % \begin{demo}[w]{Various \env{matrix} environments} |
| 1669 | %\[ \begin{matrix} 1 & 0 \\ 0 & -1 \end{matrix} \quad |
| 1670 | % \begin{pmatrix} |
| 1671 | % \cos\theta & \sin\theta \\ |
| 1672 | % -\sin\theta & \cos\theta |
| 1673 | % \end{pmatrix} \quad |
| 1674 | % \begin{dmatrix}[] 0 & -i \\ i & 0 \end{dmatrix} |
| 1675 | %\] |
| 1676 | % \end{demo} |
| 1677 | % \end{figure} |
| 1678 | % |
| 1679 | % \DescribeEnv{smatrix} |
| 1680 | % Normal matrices always come out the same size; they don't change size |
| 1681 | % according to the surrounding context (unfortunately). However, it can be |
| 1682 | % occasionally useful to put matrices in running text, so you can talk about |
| 1683 | % $A$ being $\bigl( \begin{smatrix} a & b \\ b & c \end{smatrix} \bigr)$ |
| 1684 | % being its own transpose (i.e., $A = A^T$). This is accomplished using the |
| 1685 | % \env{smatrix} (the `s' stands for `small' -- I thought that `smallmatrix' |
| 1686 | % was too big to type inline). As well as inline text, the \env{smatrix} |
| 1687 | % can be useful in displays, if the matrix is deep in a subformula. I can't |
| 1688 | % think of any examples offhand, though. |
| 1689 | % |
| 1690 | % \DescribeEnv{spmatrix} |
| 1691 | % \DescribeEnv{sdmatrix} |
| 1692 | % The \env{smatrix} environment doesn't supply any delimiters, like |
| 1693 | % \env{matrix}. There are \env{spmatrix} and \env{sdmatrix} environments |
| 1694 | % which do, though. Note that delimiters have a tendency to get too big and |
| 1695 | % mess up the line spacing -- I had to use explicitly |\big| delimiters |
| 1696 | % in the above example. |
| 1697 | % |
| 1698 | % \DescribeEnv{pmatrix*} |
| 1699 | % \DescribeEnv{spmatrix*} |
| 1700 | % \DescribeEnv{sdmatrix*} |
| 1701 | % All the small matrix environments have starred versions, which are more |
| 1702 | % suitable for use in displays, since they have more space between the rows. |
| 1703 | % They're intended for typesetting really big matrices in displays. |
| 1704 | % |
| 1705 | % \DescribeMacro{\ddots} |
| 1706 | % \DescribeMacro{\vdots} |
| 1707 | % The standard |\vdots| and |\ddots| commands don't produce anything at all |
| 1708 | % nice in small matrices, so this package redefines them so that they scale |
| 1709 | % properly to smaller sizes. |
| 1710 | % |
| 1711 | % \DescribeEnv{genmatrix} |
| 1712 | % Actually, all these environments are special cases of one: \env{genmatrix}. |
| 1713 | % This takes oodles of arguments: |
| 1714 | % \begin{quote} \synshorts |
| 1715 | % "\\begin{genmatrix}{"<matrix-style>"}{"<outer-style>"}" \\ |
| 1716 | % \null \qquad "{"<spacing>"}{"<left-delim>"}{"<right-delim>"}" \\ |
| 1717 | % \null \quad\vdots \\ |
| 1718 | % "\\end{genmatrix}" |
| 1719 | % \end{quote} |
| 1720 | % The two `style' arguments should be things like |\textstyle| or |
| 1721 | % |\scriptstyle|; the first, \<matrix-style>, is the style to use for the |
| 1722 | % matrix elements, and the second, \<outer-style>, is the style to assume |
| 1723 | % for the surrounding text (this affects the spacing within the matrix; it |
| 1724 | % should usually be the same as \<matrix-style>). The \<spacing> is inserted |
| 1725 | % between the matrix and the delimiters, on each side of the matrix. It's |
| 1726 | % usually `|\,|' in full-size matrices, and blank for small ones. The |
| 1727 | % delimiters are inserted around the matrices, and sized appropriately. |
| 1728 | % |
| 1729 | % \DescribeEnv{newmatrix} |
| 1730 | % You can create your own matrix environments if you like, using the |
| 1731 | % |\newmatrix| command. It takes two arguments, although they're a bit |
| 1732 | % odd. The first is the name of the environment, and the second contains |
| 1733 | % the arguments to pass to \env{genmatrix}. For example, the \env{pmatrix} |
| 1734 | % environment was defined by saying |
| 1735 | % |
| 1736 | % \begin{listing} |
| 1737 | %\newmatrix{pmatrix}{{\textstyle}{\textstyle}{\,}{(}{)}} |
| 1738 | % \end{listing} |
| 1739 | % |
| 1740 | % If you don't pass all three arguments, then you end up requiring the |
| 1741 | % user to specify the remaining ones. This is how \env{dmatrix} works. |
| 1742 | % |
| 1743 | % \DescribeEnv{script} |
| 1744 | % Finally, although it's not really a matrix, stacked super- and subscripts |
| 1745 | % follow much the same sorts of spacing rules. The \env{script} environment |
| 1746 | % allows you to do this sort of thing very easily. It essentially provides |
| 1747 | % a `matrix' with the right sort of spacing. The default preamble string is |
| 1748 | % \lit{c}, giving you centred scripts, although you can say |
| 1749 | % |\begin{script}[l]| for left-aligned scripts, which is better if the |
| 1750 | % script is being placed to the right of its operator. If you're really |
| 1751 | % odd, you can have more than one column. |
| 1752 | % |
| 1753 | % \begin{demo}{Example of \env{script}} |
| 1754 | %\[ \mathop{{\sum}'}_{x \in A} |
| 1755 | % f(x) |
| 1756 | % \stackrel{\mathrm{def}}{=} |
| 1757 | % \sum_{\begin{script} |
| 1758 | % x \in A \\ x \ne 0 |
| 1759 | % \end{script}} f(x) |
| 1760 | %\] |
| 1761 | % \end{demo} |
| 1762 | % |
| 1763 | % |
| 1764 | % \subsection{Other \package{mathenv} environments} |
| 1765 | % |
| 1766 | % The \package{mathenv} package contains some other environments which may |
| 1767 | % be useful, based on the enhanced \env{tabular} and \env{array} |
| 1768 | % environments. |
| 1769 | % |
| 1770 | % \DescribeEnv{cases} |
| 1771 | % The \env{cases} environment lets you say things like the following: |
| 1772 | % |
| 1773 | % \begin{demo}[w]{Example of \env{cases}} |
| 1774 | %\[ P_{r-j} = \begin{cases} |
| 1775 | % 0 & if $r-j$ is odd \\ |
| 1776 | % r!\,(-1)^{(r-j)/2} & if $r-j$ is even |
| 1777 | % \end{cases} |
| 1778 | %\] |
| 1779 | % \end{demo} |
| 1780 | % |
| 1781 | % The spacing required for this is a bit messy, so providing an environment |
| 1782 | % for it is quite handy. |
| 1783 | % |
| 1784 | % \DescribeEnv{smcases} |
| 1785 | % The \env{smcases} environment works the same way as \env{cases}, but with |
| 1786 | % scriptsize lettering. |
| 1787 | % |
| 1788 | % \implementation |
| 1789 | % |
| 1790 | % |
| 1791 | %^^A------------------------------------------------------------------------- |
| 1792 | % \section{Implementation of table handling} |
| 1793 | % |
| 1794 | % |
| 1795 | % Here we go. It starts horrid and gets worse. However, it does stay nicer |
| 1796 | % than the original, IMHO. |
| 1797 | % |
| 1798 | % \begin{macrocode} |
| 1799 | %<*mdwtab> |
| 1800 | % \end{macrocode} |
| 1801 | % |
| 1802 | % |
| 1803 | % \subsection{Registers, switches and things} |
| 1804 | % |
| 1805 | % We need lots of these. It's great fun. |
| 1806 | % |
| 1807 | % The two count registers are simple enough: |
| 1808 | % |
| 1809 | % \begin{description} |
| 1810 | % \item [\cs{tab@state}] contains the current parser state. Since we |
| 1811 | % probably won't be parsing preambles recursively, this is a global |
| 1812 | % variable. |
| 1813 | % \item [\cs{tab@columns}] contains the number of the current column. |
| 1814 | % \item [\cs{tab@hlstate}] contains the state required for hline management. |
| 1815 | % \item [\textsf{tabrow}] contains the row number in the table. It's a |
| 1816 | % proper \LaTeX\ counter. |
| 1817 | % \end{description} |
| 1818 | % |
| 1819 | % \begin{macrocode} |
| 1820 | \newcount\tab@state |
| 1821 | \newcount\tab@columns |
| 1822 | \newcounter{tabrow} |
| 1823 | % \end{macrocode} |
| 1824 | % |
| 1825 | % We need \emph{lots} of token registers. Fortunately, most of them are only |
| 1826 | % used during parsing. We'll use \PlainTeX's scratch tokens for this. Note |
| 1827 | % that |\toks\tw@| isn't used here. It, and |\toks@|, are free for use by |
| 1828 | % column commands. |
| 1829 | % |
| 1830 | % \begin{macrocode} |
| 1831 | \newtoks\tab@preamble |
| 1832 | \newtoks\tab@shortline |
| 1833 | \toksdef\tab@pretext 4 |
| 1834 | \toksdef\tab@posttext 6 |
| 1835 | \toksdef\tab@userpretext 8 |
| 1836 | % \end{macrocode} |
| 1837 | % |
| 1838 | % The dimens are fairly straightforward. The inclusion of |\col@sep| is a |
| 1839 | % sacrifice to compatibility -- judicious use of |\let| in \package{array} |
| 1840 | % would have saved a register. |
| 1841 | % |
| 1842 | % \begin{macrocode} |
| 1843 | \newdimen\extrarowheight |
| 1844 | \newdimen\tabextrasep |
| 1845 | \newdimen\arrayextrasep |
| 1846 | \newdimen\smarraycolsep |
| 1847 | \newdimen\smarrayextrasep |
| 1848 | \newdimen\tab@width |
| 1849 | \newdimen\col@sep |
| 1850 | \newdimen\tab@endheight |
| 1851 | \newdimen\arraythickrulewidth |
| 1852 | \newdimen\tab@rulewidth |
| 1853 | % \end{macrocode} |
| 1854 | % |
| 1855 | % Some skip registers too. Phew. |
| 1856 | % |
| 1857 | % \begin{macrocode} |
| 1858 | \newskip\tab@leftskip |
| 1859 | \newskip\tab@rightskip |
| 1860 | % \end{macrocode} |
| 1861 | % |
| 1862 | % And some switches. The first three are for the parser. |
| 1863 | % |
| 1864 | % \begin{macrocode} |
| 1865 | \newif\iftab@firstcol |
| 1866 | \newif\iftab@initrule |
| 1867 | \newif\iftab@rule |
| 1868 | \newif\iftab@vgap |
| 1869 | \newif\iftab@colour |
| 1870 | % \end{macrocode} |
| 1871 | % |
| 1872 | % Now assign some default values to new dimen parameters. These definitions |
| 1873 | % are essentially the equivalent of an |\openup 1\jot| in \env{array}, but |
| 1874 | % not in \env{tabular}. This looks nice, I think. |
| 1875 | % |
| 1876 | % \begin{macrocode} |
| 1877 | \tabextrasep\z@ |
| 1878 | \arrayextrasep\jot |
| 1879 | \smarraycolsep\thr@@\p@ |
| 1880 | \smarrayextrasep\z@ |
| 1881 | \arraythickrulewidth\p@ |
| 1882 | % \end{macrocode} |
| 1883 | % |
| 1884 | % Set some things up for alien table environments. |
| 1885 | % |
| 1886 | % \begin{macrocode} |
| 1887 | \let\tab@extrasep\tabextrasep |
| 1888 | \let\tab@penalty\relax |
| 1889 | % \end{macrocode} |
| 1890 | % |
| 1891 | % |
| 1892 | % \subsection{Options processing} |
| 1893 | % |
| 1894 | % Notice options, load package. |
| 1895 | % |
| 1896 | % \begin{macrocode} |
| 1897 | \DeclareOption{colour}{\tab@colourtrue} |
| 1898 | \DeclareOption{color}{\tab@colourtrue} |
| 1899 | \ProcessOptions |
| 1900 | \RequirePackage{mdwkey} |
| 1901 | % \end{macrocode} |
| 1902 | % |
| 1903 | % \subsection{Some little details} |
| 1904 | % |
| 1905 | % \begin{macro}{\@maybe@unskip} |
| 1906 | % |
| 1907 | % This macro solves a little problem. In an alignment (and in other places) |
| 1908 | % it's desirable to suppress trailing space. The usual method, to say |
| 1909 | % |\unskip|, is a little hamfisted, because it removes perfectly reasonable |
| 1910 | % aligning spaces like |\hfil|s. While as a package writer I can deal with |
| 1911 | % this sort of thing by saying |\kern\z@| in appropriate places, it can |
| 1912 | % annoy users who are trying to use |\hfill| to override alignment in funny |
| 1913 | % places. |
| 1914 | % |
| 1915 | % My current solution seems to be acceptable. I'll remove the natural width |
| 1916 | % of the last glue item, so that it can still stretch and shrink if |
| 1917 | % necessary. The implementation makes use of the fact that multiplying |
| 1918 | % a \<skip> by a \<number> kills off the stretch. (Bug fix: don't do this |
| 1919 | % when we're in vertical mode.) |
| 1920 | % |
| 1921 | % \begin{macrocode} |
| 1922 | \def\@maybe@unskip{\ifhmode\hskip\m@ne\lastskip\relax\fi} |
| 1923 | % \end{macrocode} |
| 1924 | % |
| 1925 | % \end{macro} |
| 1926 | % |
| 1927 | % \begin{macro}{\q@delim} |
| 1928 | % |
| 1929 | % Finally, for the sake of niceness, here's a delimiter token I can use |
| 1930 | % for various things. It's a `quark', for what it's worth (i.e., it expands |
| 1931 | % to itself) although I'm not really sure why this is a good thing. As far |
| 1932 | % as I'm concerned, it's important that it has a unique meaning (i.e., that |
| 1933 | % it won't be |\ifx|-equal to other things, or something undefined) and that |
| 1934 | % it won't be used where I don't expect it to be used. \TeX\ will loop |
| 1935 | % horridly if it tries to expand this, so I don't think that quarks are |
| 1936 | % wonderfully clever thing to use. (Maybe it should really expand to |
| 1937 | % something like `\syntax{<quark>"."}', which will rapidly fill \TeX's memory |
| 1938 | % if it gets accidentally expanded. Still, I'll leave it as it is until |
| 1939 | % such time as I understand the idea more.) |
| 1940 | % |
| 1941 | % \begin{macrocode} |
| 1942 | \def\q@delim{\q@delim} |
| 1943 | % \end{macrocode} |
| 1944 | % |
| 1945 | % \end{macro} |
| 1946 | % |
| 1947 | % |
| 1948 | % \subsection{Parser states} |
| 1949 | % |
| 1950 | % Now we start on the parser. It's really simple, deep down. We progress |
| 1951 | % from state to state, extracting tokens from the preamble and building |
| 1952 | % command names from them. Each command calls one of the element-building |
| 1953 | % routines, which works out which state it should be in. We go through each |
| 1954 | % of the states in between (see later) doing default things for the ones we |
| 1955 | % missed out. |
| 1956 | % |
| 1957 | % Anyway, here's some symbolic names for the states. It makes my life |
| 1958 | % easier. |
| 1959 | % |
| 1960 | % \begin{macrocode} |
| 1961 | \chardef\tab@startstate 0 |
| 1962 | \chardef\tab@loopstate 1 |
| 1963 | \chardef\tab@rulestate 1 |
| 1964 | \chardef\tab@prespcstate 2 |
| 1965 | \chardef\tab@prestate 3 |
| 1966 | \chardef\tab@colstate 4 |
| 1967 | \chardef\tab@poststate 5 |
| 1968 | \chardef\tab@postspcstate 6 |
| 1969 | \chardef\tab@limitstate 7 |
| 1970 | % \end{macrocode} |
| 1971 | % |
| 1972 | % |
| 1973 | % \subsection{Adding things to token lists} |
| 1974 | % |
| 1975 | % Define some macros for adding stuff to the beginning and end of token |
| 1976 | % lists. This is really easy, actually. Here we go. |
| 1977 | % |
| 1978 | % \begin{macrocode} |
| 1979 | \def\tab@append#1#2{#1\expandafter{\the#1#2}} |
| 1980 | \def\tab@prepend#1#2{% |
| 1981 | \toks@{#2}#1\expandafter{\the\expandafter\toks@\the#1}% |
| 1982 | } |
| 1983 | % \end{macrocode}% |
| 1984 | % |
| 1985 | % |
| 1986 | % \subsection{Committing a column to the preamble} |
| 1987 | % |
| 1988 | % Each time we pass the `rule' state, we `commit' the tokens we've gathered |
| 1989 | % so far to the main preamble token list. This is how we do it. Note the |
| 1990 | % icky use of |\expandafter|. |
| 1991 | % |
| 1992 | % \begin{macrocode} |
| 1993 | \def\tab@commit{% |
| 1994 | % \end{macrocode} |
| 1995 | % |
| 1996 | % If this isn't the first column, then we need to put in a column separator. |
| 1997 | % |
| 1998 | % \begin{macrocode} |
| 1999 | \iftab@firstcol\else% |
| 2000 | \expandafter\tab@append\expandafter\tab@preamble% |
| 2001 | \expandafter{\tab@tabtext}% |
| 2002 | \fi% |
| 2003 | % \end{macrocode} |
| 2004 | % |
| 2005 | % Now dump in the |\tab@lefttext| material. |
| 2006 | % |
| 2007 | % \begin{macrocode} |
| 2008 | \expandafter\tab@append\expandafter\tab@preamble% |
| 2009 | \expandafter{\tab@lefttext}% |
| 2010 | % \end{macrocode} |
| 2011 | % |
| 2012 | % Now we spill the token registers into the main list in a funny order (which |
| 2013 | % is why we're doing it in this strange way in the first place. |
| 2014 | % |
| 2015 | % \begin{macrocode} |
| 2016 | \toks@\expandafter{\tab@midtext}% |
| 2017 | \tab@preamble\expandafter{% |
| 2018 | \the\expandafter\tab@preamble% |
| 2019 | \the\expandafter\tab@pretext% |
| 2020 | \the\expandafter\tab@userpretext% |
| 2021 | \the\expandafter\toks@% |
| 2022 | \the\expandafter\tab@posttext% |
| 2023 | \tab@righttext% |
| 2024 | }% |
| 2025 | % \end{macrocode} |
| 2026 | % |
| 2027 | % Now reset token lists and things for the next go round. |
| 2028 | % |
| 2029 | % \begin{macrocode} |
| 2030 | \tab@firstcolfalse% |
| 2031 | \tab@pretext{}% |
| 2032 | \tab@userpretext{}% |
| 2033 | \tab@posttext{}% |
| 2034 | } |
| 2035 | % \end{macrocode} |
| 2036 | % |
| 2037 | % |
| 2038 | % \subsection{Playing with parser states} |
| 2039 | % |
| 2040 | % \begin{macro}{\tab@setstate} |
| 2041 | % |
| 2042 | % This is how we set new states. The algorithm is fairly simple, really. |
| 2043 | % |
| 2044 | % ^^A Let's see how good my TeX really is... ;-) |
| 2045 | % ^^A Actually, it doesn't seem to have worked out too badly. Maybe I should |
| 2046 | % ^^A write a package to do this automatically. It's rather tricky, though. |
| 2047 | % |
| 2048 | % \def\qq{\mbox{\quad}} |
| 2049 | % \sbox{0}{\itshape\textunderscore}\def\_{\usebox{0}} |
| 2050 | % |
| 2051 | % \begin{quote} |
| 2052 | % {\bf while} $\it tab\_state \ne s$ {\bf do} \\ |
| 2053 | % \qq $\mathit{tab\_state = tab\_state}+1$; \\ |
| 2054 | % \qq {\bf if} $\it tab\_state = tab\_limitState$ {\bf then} |
| 2055 | % $\it tab\_state=tab\_loopState$; \\ |
| 2056 | % \qq {\bf if} $\it tab\_state = tab\_preSpcState$ {\bf then} \\ |
| 2057 | % \qq \qq {\bf if} $\it tab\_initRule$ {\bf then} \\ |
| 2058 | % \qq \qq \qq $\it tab\_initRule = {\bf false}$; \\ |
| 2059 | % \qq \qq {\bf else} \\ |
| 2060 | % \qq \qq \qq {\bf if} $\it tab\_inMultiCol$ {\bf then moan}; \\ |
| 2061 | % \qq \qq \qq $\it commit$; \\ |
| 2062 | % \qq \qq \qq $\it append(tab\_shortLine,\hbox{`|&\omit|')}$; \\ |
| 2063 | % \qq \qq {\bf end\,if}; \\ |
| 2064 | % \qq {\bf end\,if}; \\ |
| 2065 | % \qq {\bf if} $\it tab\_state \ne s$ {\bf then} |
| 2066 | % $\it do\_default(tab\_state)$; \\ |
| 2067 | % {\bf end\,while}; |
| 2068 | % \end{quote} |
| 2069 | % |
| 2070 | % First we decide if there's anything to do. If so, we call another macro to |
| 2071 | % do it for us. |
| 2072 | % |
| 2073 | % \begin{macrocode} |
| 2074 | \def\tab@setstate#1{% |
| 2075 | \ifnum#1=\tab@state\else% |
| 2076 | \def\@tempa{\tab@setstate@i{#1}}% |
| 2077 | \@tempa% |
| 2078 | \fi% |
| 2079 | } |
| 2080 | % \end{macrocode} |
| 2081 | % |
| 2082 | % This is where the fun is. First we bump the state by one, and loop back |
| 2083 | % if we fall off the end. |
| 2084 | % |
| 2085 | % \begin{macrocode} |
| 2086 | \def\tab@setstate@i#1{% |
| 2087 | \global\advance\tab@state\@ne% |
| 2088 | \ifnum\tab@state>\tab@limitstate% |
| 2089 | \global\tab@state\tab@loopstate% |
| 2090 | \fi% |
| 2091 | % \end{macrocode} |
| 2092 | % |
| 2093 | % Now, if we've just passed the ruleoff state, we commit the current text |
| 2094 | % \emph{unless} this was the strange initial rule at the very beginning. We |
| 2095 | % provide a little hook here so that |\multicolumn| can moan if you try and |
| 2096 | % give more than one column there. We also add another tab/omit pair to the |
| 2097 | % list we use for |\vgap|. |
| 2098 | % |
| 2099 | % \begin{macrocode} |
| 2100 | \ifnum\tab@state=\tab@prespcstate% |
| 2101 | \iftab@initrule% |
| 2102 | \tab@initrulefalse% |
| 2103 | \else% |
| 2104 | \tab@looped% |
| 2105 | \tab@commit% |
| 2106 | \expandafter\tab@append\expandafter\tab@shortline% |
| 2107 | \expandafter{\tab@rightruletext}% |
| 2108 | \tab@append\tab@shortline{&\omit}% |
| 2109 | \expandafter\tab@append\expandafter\tab@shortline% |
| 2110 | \expandafter{\tab@leftruletext}% |
| 2111 | \fi% |
| 2112 | \fi% |
| 2113 | % \end{macrocode} |
| 2114 | % |
| 2115 | % Now we decide whether to go round again. If not, we do the default thing |
| 2116 | % for this state. This is mainly here so that we can put the |\tabcolsep| or |
| 2117 | % whatever in if the user didn't give an \lit{@} expression. |
| 2118 | % |
| 2119 | % \begin{macrocode} |
| 2120 | \ifnum#1=\tab@state% |
| 2121 | \let\@tempa\relax% |
| 2122 | \else% |
| 2123 | \csname tab@default@\number\tab@state\endcsname% |
| 2124 | \fi% |
| 2125 | \@tempa% |
| 2126 | } |
| 2127 | % \end{macrocode} |
| 2128 | % |
| 2129 | % \end{macro} |
| 2130 | % |
| 2131 | % Now we set up the default actions for the various states. |
| 2132 | % |
| 2133 | % In state~2 (pre-space) we add in the default gap if either we didn't have |
| 2134 | % an \lit{@} expression in the post-space state or there was an explicit |
| 2135 | % intervening rule. |
| 2136 | % |
| 2137 | % \begin{macrocode} |
| 2138 | \@namedef{tab@default@2}{% |
| 2139 | \iftab@rule% |
| 2140 | \tab@append\tab@pretext{\hskip\col@sep}% |
| 2141 | \fi% |
| 2142 | } |
| 2143 | % \end{macrocode} |
| 2144 | % |
| 2145 | % If the user omits the column type, we insert an `l'-type column and moan |
| 2146 | % a lot. |
| 2147 | % |
| 2148 | % \begin{macrocode} |
| 2149 | \@namedef{tab@default@4}{% |
| 2150 | \tab@err@misscol% |
| 2151 | \tab@append\tab@pretext{\tab@bgroup\relax}% |
| 2152 | \tab@append\tab@posttext{\relax\tab@egroup\hfil}% |
| 2153 | \tab@append\tab@shortline{\hfil}% |
| 2154 | \advance\tab@columns\@ne% |
| 2155 | } |
| 2156 | % \end{macrocode} |
| 2157 | % |
| 2158 | % Finally we deal with the post-space state. We set a marker so that we |
| 2159 | % put in the default space in the pre-space state later too. |
| 2160 | % |
| 2161 | % \begin{macrocode} |
| 2162 | \@namedef{tab@default@6}{% |
| 2163 | \tab@append\tab@posttext{\hskip\col@sep}% |
| 2164 | \tab@ruletrue% |
| 2165 | } |
| 2166 | % \end{macrocode} |
| 2167 | % |
| 2168 | % |
| 2169 | % \subsection{Declaring token types} |
| 2170 | % |
| 2171 | % \begin{macro}{\tab@extracol} |
| 2172 | % |
| 2173 | % Before we start, we need to handle |\extracolsep|. This is a right pain, |
| 2174 | % because the original version of \env{tabular} worked on total expansion, |
| 2175 | % which is a Bad Thing. On the other hand, turning |\extracolsep| into a |
| 2176 | % |\tabskip| is also a major pain. |
| 2177 | % |
| 2178 | % \begin{macrocode} |
| 2179 | \def\tab@extracol#1#2{\tab@extracol@i#1#2\extracolsep{}\extracolsep\end} |
| 2180 | \def\tab@extracol@i#1#2\extracolsep#3#4\extracolsep#5\end{% |
| 2181 | \ifx @#3@% |
| 2182 | \def\@tempa{#1{#2}}% |
| 2183 | \else% |
| 2184 | \def\@tempa{#1{#2\tabskip#3\relax#4}}% |
| 2185 | \fi% |
| 2186 | \@tempa% |
| 2187 | } |
| 2188 | % \end{macrocode} |
| 2189 | % |
| 2190 | % \end{macro} |
| 2191 | % |
| 2192 | % This is where we do the work for inserting preamble elements. |
| 2193 | % |
| 2194 | % \begin{macro}{\tabruletype} |
| 2195 | % |
| 2196 | % Inserting rules is interesting, because we have to decide where to put |
| 2197 | % them. If this is the funny initial rule, it goes in the pre-text list, |
| 2198 | % otherwise it goes in the post-text list. We work out what to do first |
| 2199 | % thing: |
| 2200 | % |
| 2201 | % \begin{macrocode} |
| 2202 | \def\tabruletype#1{\tab@extracol\tabruletype@i{#1}}% |
| 2203 | \def\tabruletype@i#1{% |
| 2204 | \iftab@initrule% |
| 2205 | \let\tab@tok\tab@pretext% |
| 2206 | \else% |
| 2207 | \let\tab@tok\tab@posttext% |
| 2208 | \fi% |
| 2209 | % \end{macrocode} |
| 2210 | % |
| 2211 | % Now if we're already in the rule state, we must have just done a rule. |
| 2212 | % This means we must put in the |\doublerulesep| space, both here and in the |
| 2213 | % shortline list. Otherwise we just stick the rule in. |
| 2214 | % |
| 2215 | % This is complicated, because |\vgap| needs to be able to remove some bits |
| 2216 | % of rule. We pass each one to a macro |\tab@ckr|, together with the column |
| 2217 | % number, which is carefully bumped at the right times, and this macro will |
| 2218 | % vet the rules and output the appropriate ones. There's lots of extreme |
| 2219 | % |\expandafter| nastiness as a result. Amazingly, this actually works. |
| 2220 | % |
| 2221 | % \begin{macrocode} |
| 2222 | \ifnum\tab@state=\tab@rulestate% |
| 2223 | \tab@append\tab@tok{\hskip\doublerulesep\begingroup#1\endgroup}% |
| 2224 | \expandafter\tab@append\expandafter\tab@shortline\expandafter{% |
| 2225 | \expandafter\hskip\expandafter\doublerulesep% |
| 2226 | \expandafter\tab@ckr\expandafter{\the\tab@columns}% |
| 2227 | {\begingroup#1\endgroup}% |
| 2228 | }% |
| 2229 | \else% |
| 2230 | \tab@setstate\tab@rulestate% |
| 2231 | \tab@append\tab@tok{\begingroup#1\endgroup}% |
| 2232 | \expandafter\tab@append\expandafter\tab@shortline\expandafter{% |
| 2233 | \expandafter\tab@ckr\expandafter{\the\tab@columns}% |
| 2234 | {\begingroup#1\endgroup}% |
| 2235 | }% |
| 2236 | \fi% |
| 2237 | % \end{macrocode} |
| 2238 | % |
| 2239 | % Finally, we say there was a rule here, so that default space gets put in |
| 2240 | % after this. Otherwise we lose lots of generality. |
| 2241 | % |
| 2242 | % \begin{macrocode} |
| 2243 | \tab@ruletrue% |
| 2244 | } |
| 2245 | % \end{macrocode} |
| 2246 | % |
| 2247 | % \end{macro} |
| 2248 | % |
| 2249 | % \begin{macro}{\tabspctype} |
| 2250 | % |
| 2251 | % We need to work out which space-state we should be in. Then we just put |
| 2252 | % the text in. Easy, really. |
| 2253 | % |
| 2254 | % \begin{macrocode} |
| 2255 | \def\tabspctype#1{\tab@extracol\tabspctype@i{#1}}% |
| 2256 | \def\tabspctype@i#1{% |
| 2257 | \tab@rulefalse% |
| 2258 | \ifnum\tab@state>\tab@prespcstate% |
| 2259 | \tab@setstate\tab@postspcstate% |
| 2260 | \let\tab@tok\tab@posttext% |
| 2261 | \else% |
| 2262 | \tab@setstate\tab@prespcstate% |
| 2263 | \let\tab@tok\tab@pretext% |
| 2264 | \fi% |
| 2265 | \tab@append\tab@tok{\begingroup#1\endgroup}% |
| 2266 | } |
| 2267 | % \end{macrocode} |
| 2268 | % |
| 2269 | % \end{macro} |
| 2270 | % |
| 2271 | % \begin{macro}{\tabcoltype} |
| 2272 | % |
| 2273 | % If we're already in the column state, we bump the state and loop round |
| 2274 | % again, to get all the appropriate default behaviour. We bump the column |
| 2275 | % counter, and add the bits of text we were given to appropriate token lists. |
| 2276 | % We also add the |\hfil| glue to the shortline list, to space out the rules |
| 2277 | % properly. |
| 2278 | % |
| 2279 | % \begin{macrocode} |
| 2280 | \def\tabcoltype#1#2{% |
| 2281 | \ifnum\tab@state=\tab@colstate% |
| 2282 | \global\advance\tab@state\@ne% |
| 2283 | \fi% |
| 2284 | \advance\tab@columns\@ne% |
| 2285 | \tab@setstate\tab@colstate% |
| 2286 | \tab@append\tab@pretext{#1}% |
| 2287 | \tab@append\tab@posttext{#2}% |
| 2288 | \tab@append\tab@shortline{\hfil}% |
| 2289 | } |
| 2290 | % \end{macrocode} |
| 2291 | % |
| 2292 | % \end{macro} |
| 2293 | % |
| 2294 | % \begin{macro}{\tabuserpretype} |
| 2295 | % \begin{macro}{\tabuserposttype} |
| 2296 | % |
| 2297 | % These are both utterly trivial. |
| 2298 | % |
| 2299 | % \begin{macrocode} |
| 2300 | \def\tabuserpretype#1{% |
| 2301 | \tab@setstate\tab@prestate% |
| 2302 | \tab@prepend\tab@userpretext{#1}% |
| 2303 | } |
| 2304 | % \end{macrocode} |
| 2305 | % |
| 2306 | % \begin{macrocode} |
| 2307 | \def\tabuserposttype#1{% |
| 2308 | \tab@setstate\tab@poststate% |
| 2309 | \tab@prepend\tab@posttext{#1}% |
| 2310 | } |
| 2311 | % \end{macrocode} |
| 2312 | % |
| 2313 | % \end{macro} |
| 2314 | % \end{macro} |
| 2315 | % |
| 2316 | % |
| 2317 | % \subsection{The colset stack} |
| 2318 | % |
| 2319 | % Let's start with something fairly easy. We'll keep a stack of column sets |
| 2320 | % so that users don't get confused by package authors changing the current |
| 2321 | % column set. This is fairly easy, really. |
| 2322 | % |
| 2323 | % \begin{macro}{\tab@push} |
| 2324 | % \begin{macro}{\tab@pop} |
| 2325 | % \begin{macro}{\tab@head} |
| 2326 | % |
| 2327 | % These are the stack management routines. The only important thing to note |
| 2328 | % is that |\tab@head| must take place \emph{only} in \TeX's mouth, so we can |
| 2329 | % use it in |\csname|\dots|\endcsname| constructions. |
| 2330 | % |
| 2331 | % \begin{macrocode} |
| 2332 | \def\tab@push#1#2{% |
| 2333 | \toks@{{#2}}% |
| 2334 | \expandafter\def\expandafter#1\expandafter{\the\expandafter\toks@#1}% |
| 2335 | } |
| 2336 | \def\tab@pop#1{\edef#1{\expandafter\@gobble#1}} |
| 2337 | \def\tab@head#1{\expandafter\tab@head@i#1\relax} |
| 2338 | \def\tab@head@i#1#2\relax{#1} |
| 2339 | % \end{macrocode} |
| 2340 | % |
| 2341 | % \end{macro} |
| 2342 | % \end{macro} |
| 2343 | % \end{macro} |
| 2344 | % |
| 2345 | % \begin{macro}{\colset} |
| 2346 | % \begin{macro}{\colpush} |
| 2347 | % \begin{macro}{\colpop} |
| 2348 | % |
| 2349 | % Now we can define the user macros. |
| 2350 | % |
| 2351 | % \begin{macrocode} |
| 2352 | \def\tab@colstack{{tabular}} |
| 2353 | \def\colset{\colpop\colpush} |
| 2354 | \def\colpush{\tab@push\tab@colstack} |
| 2355 | \def\colpop{\tab@pop\tab@colstack} |
| 2356 | % \end{macrocode} |
| 2357 | % |
| 2358 | % \end{macro} |
| 2359 | % \end{macro} |
| 2360 | % \end{macro} |
| 2361 | % |
| 2362 | % \begin{macro}{\tab@colset} |
| 2363 | % |
| 2364 | % Now we define a shortcut for reading the top item off the stack. |
| 2365 | % |
| 2366 | % \begin{macrocode} |
| 2367 | \def\tab@colset{\tab@head\tab@colstack} |
| 2368 | % \end{macrocode} |
| 2369 | % |
| 2370 | % \end{macro} |
| 2371 | % |
| 2372 | % |
| 2373 | % \subsection{The main parser routine} |
| 2374 | % |
| 2375 | % \begin{macro}{\tab@initread} |
| 2376 | % |
| 2377 | % This macro sets up lots of variables to their normal states prior to |
| 2378 | % parsing a preamble. Some things may need changing, but not many. This |
| 2379 | % version just sets the major hooks, and then does a subread. The midtext |
| 2380 | % macro contains what to put in the very middle of each template -- |
| 2381 | % |\multicolumn| will insert its argument here. |
| 2382 | % |
| 2383 | % \begin{macrocode} |
| 2384 | \def\tab@initread{% |
| 2385 | \def\tab@lefttext{}% |
| 2386 | \def\tab@leftruletext{}% |
| 2387 | \def\tab@righttext{}% |
| 2388 | \def\tab@rightruletext{}% |
| 2389 | \def\tab@tabtext{&}% |
| 2390 | \def\tab@midtext{\ignorespaces####\@maybe@unskip}% |
| 2391 | \tab@initsubread% |
| 2392 | } |
| 2393 | % \end{macrocode} |
| 2394 | % |
| 2395 | % \end{macro} |
| 2396 | % |
| 2397 | % \begin{macro}{\tab@initsubread} |
| 2398 | % |
| 2399 | % This is where most of the activity is. We don't replace the left and right |
| 2400 | % texts, so that we effectively inherit them rfrom the enclosing |
| 2401 | % environment. |
| 2402 | % |
| 2403 | % \begin{macrocode} |
| 2404 | \def\tab@initsubread{% |
| 2405 | % \end{macrocode} |
| 2406 | % |
| 2407 | % First, reset the parser state to the start state. |
| 2408 | % |
| 2409 | % \begin{macrocode} |
| 2410 | \global\tab@state\tab@startstate% |
| 2411 | % \end{macrocode} |
| 2412 | % |
| 2413 | % We clear the token lists to sensible values, mostly. |
| 2414 | % |
| 2415 | % \begin{macrocode} |
| 2416 | \tab@preamble{}% |
| 2417 | \tab@shortline{}% |
| 2418 | \tab@pretext{}% |
| 2419 | \tab@userpretext{}% |
| 2420 | \tab@posttext{}% |
| 2421 | \let\tab@multicol\@empty% |
| 2422 | \def\tab@startpause{\penalty\postdisplaypenalty\medskip}% |
| 2423 | \def\tab@endpause{\penalty\predisplaypenalty\medskip}% |
| 2424 | % \end{macrocode} |
| 2425 | % |
| 2426 | % Finally, reset the column counter, don't raise errors when we loop, and set |
| 2427 | % some parser flags to their appropriate values. |
| 2428 | % |
| 2429 | % \begin{macrocode} |
| 2430 | \tab@columns\z@% |
| 2431 | \let\tab@looped\relax% |
| 2432 | \tab@ruletrue% |
| 2433 | \tab@initruletrue% |
| 2434 | \tab@firstcoltrue% |
| 2435 | } |
| 2436 | |
| 2437 | % \end{macrocode} |
| 2438 | % |
| 2439 | % \end{macro} |
| 2440 | % |
| 2441 | % \begin{macro}{\tab@readpreamble} |
| 2442 | % |
| 2443 | % This is the main macro for preamble handling. Actually, all it does is |
| 2444 | % gobble its argument's leading brace and call another macro, but it does it |
| 2445 | % with style. |
| 2446 | % |
| 2447 | % \begin{macrocode} |
| 2448 | \def\tab@readpreamble#1{% |
| 2449 | \expandafter\tab@append\expandafter\tab@shortline% |
| 2450 | \expandafter{\tab@leftruletext}% |
| 2451 | \tab@doreadpream{#1}% |
| 2452 | \iftab@initrule\global\tab@state\tab@prespcstate\fi% |
| 2453 | \tab@setstate\tab@rulestate% |
| 2454 | \expandafter\tab@append\expandafter\tab@shortline% |
| 2455 | \expandafter{\tab@rightruletext}% |
| 2456 | \tab@commit% |
| 2457 | } |
| 2458 | % \end{macrocode} |
| 2459 | % |
| 2460 | % \end{macro} |
| 2461 | % |
| 2462 | % \begin{macro}{\tab@doreadpream} |
| 2463 | % |
| 2464 | % The preamble is in an argument. Previous versions used a nasty trick using |
| 2465 | % |\let| and |\afterassignment|. Now we use an explicit end token, to allow |
| 2466 | % dodgy column type handlers to scoop up the remaining preamble tokens |
| 2467 | % and process them. Not that anyone would want to do that, oh no (see |
| 2468 | % the \lit{[} type in the \env{eqnarray} environment |;-)|). |
| 2469 | % |
| 2470 | % \begin{macrocode} |
| 2471 | \def\tab@doreadpream#1{\tab@mkpreamble#1\q@delim} |
| 2472 | % \end{macrocode} |
| 2473 | % |
| 2474 | % \end{macro} |
| 2475 | % |
| 2476 | % \begin{macro}{\tab@mkpreamble} |
| 2477 | % |
| 2478 | % This is the main parser routine. It takes each token in turn, scrutinises |
| 2479 | % it carefully, and does the appropriate thing with it. |
| 2480 | % |
| 2481 | % The preamble was given as an argument to |\tab@doreadpream|, and that has |
| 2482 | % helpfully stripped off the initial |{| character. We need to pick off the |
| 2483 | % next token (whatever it is) so we can examine it. We'll use |\futurelet| |
| 2484 | % so we can detect groups and things in funny places. |
| 2485 | % |
| 2486 | % \begin{macrocode} |
| 2487 | \def\tab@mkpreamble{\futurelet\@let@token\tab@mkpreamble@i} |
| 2488 | % \end{macrocode} |
| 2489 | % |
| 2490 | % If we find a space token, we'll go off and do something a bit special, |
| 2491 | % since spaces are sort of hard to handle. Otherwise we'll do it in the old |
| 2492 | % fashioned way. |
| 2493 | % |
| 2494 | % \begin{macrocode} |
| 2495 | \def\tab@mkpreamble@i{% |
| 2496 | \ifx\@let@token\@sptoken% |
| 2497 | \expandafter\tab@mkpreamble@spc% |
| 2498 | \else% |
| 2499 | \expandafter\tab@mkpreamble@ii% |
| 2500 | \fi% |
| 2501 | } |
| 2502 | % \end{macrocode} |
| 2503 | % |
| 2504 | % If we find a |\@@endpreamble| token, that's it and we're finished. We just |
| 2505 | % gobble it and return. Otherwise, if it's an open group character, we'll |
| 2506 | % complain because someone's probably tried to put an argument in the wrong |
| 2507 | % place. Finally, if none of the other things apply, we'll deal with the |
| 2508 | % character below. |
| 2509 | % |
| 2510 | % \begin{macrocode} |
| 2511 | \def\tab@mkpreamble@ii{% |
| 2512 | \ifx\@let@token\q@delim% |
| 2513 | \def\@tempa{\let\@let@token}% |
| 2514 | \else% |
| 2515 | \ifcat\bgroup\noexpand\@let@token% |
| 2516 | \tab@err@oddgroup% |
| 2517 | \def\@tempa##1{\tab@mkpreamble}% |
| 2518 | \else% |
| 2519 | \let\@tempa\tab@mkpreamble@iii% |
| 2520 | \fi% |
| 2521 | \fi% |
| 2522 | \@tempa% |
| 2523 | } |
| 2524 | % \end{macrocode} |
| 2525 | % |
| 2526 | % Handle a character. This involves checking to see if it's actually |
| 2527 | % defined, and then doing it. Doing things this way means we won't get |
| 2528 | % stranded in mid-preamble unless a package author has blown it. |
| 2529 | % |
| 2530 | % \begin{macrocode} |
| 2531 | \def\tab@mkpreamble@iii#1{% |
| 2532 | \@ifundefined{\tab@colset!col.\string#1}{% |
| 2533 | \tab@err@undef{#1}\tab@mkpreamble% |
| 2534 | }{% |
| 2535 | \@nameuse{\tab@colset!col.\string#1}% |
| 2536 | }% |
| 2537 | } |
| 2538 | % \end{macrocode} |
| 2539 | % |
| 2540 | % If we get given a space character, we'll look up the command name as |
| 2541 | % before. If no-one's defined the column type we'll just skip it silently, |
| 2542 | % which lets users do pretty formatting if they like. |
| 2543 | % |
| 2544 | % \begin{macrocode} |
| 2545 | \@namedef{tab@mkpreamble@spc} {% |
| 2546 | \@ifundefined{\tab@colset!col. }{% |
| 2547 | \tab@mkpreamble% |
| 2548 | }{% |
| 2549 | \@nameuse{\tab@colset!col. }% |
| 2550 | }% |
| 2551 | } |
| 2552 | % \end{macrocode} |
| 2553 | % |
| 2554 | % \end{macro} |
| 2555 | % |
| 2556 | % \begin{macro}{\coldef} |
| 2557 | % |
| 2558 | % Here's how to define column types the nice way. Some dexterity is required |
| 2559 | % to make everything work right, but it's simple really. |
| 2560 | % |
| 2561 | % \begin{macrocode} |
| 2562 | \def\coldef{\@testopt\coldef@i\tab@colset} |
| 2563 | \def\coldef@i[#1]#2#3#{\coldef@ii[#1]{#2}{#3}} |
| 2564 | \def\coldef@ii[#1]#2#3#4{% |
| 2565 | \expandafter\def\csname#1!col.\string#2\endcsname#3{% |
| 2566 | #4\tab@mkpreamble% |
| 2567 | }% |
| 2568 | } |
| 2569 | % \end{macrocode} |
| 2570 | % |
| 2571 | % \end{macro} |
| 2572 | % |
| 2573 | % \begin{macro}{\collet} |
| 2574 | % |
| 2575 | % We'd like to let people copy column types from other places. This is how |
| 2576 | % to do it. |
| 2577 | % |
| 2578 | % \begin{macrocode} |
| 2579 | \def\collet{\@testopt\collet@i\tab@colset} |
| 2580 | \def\collet@i[#1]#2{% |
| 2581 | \@ifnextchar=% |
| 2582 | {\collet@ii[#1]{#2}}% |
| 2583 | {\collet@ii[#1]{#2}=}% |
| 2584 | } |
| 2585 | \def\collet@ii[#1]#2={\@testopt{\collet@iii[#1]{#2}}\tab@colset} |
| 2586 | \def\collet@iii[#1]#2[#3]#4{% |
| 2587 | \expandafter\let\csname#1!col.\string#2\expandafter\endcsname% |
| 2588 | \csname#3!col.\string#4\endcsname% |
| 2589 | } |
| 2590 | % \end{macrocode} |
| 2591 | % |
| 2592 | % \end{macro} |
| 2593 | % |
| 2594 | % \begin{macro}{\newcolumntype} |
| 2595 | % |
| 2596 | % We just bundle the text off to |\newcommand| and expect it to cope. It |
| 2597 | % ought to. The column type code inserts the user's tokens directly, rather |
| 2598 | % than calling |\tab@doreadpream| recursively. The magic control sequence |
| 2599 | % is the one looked up by the parser. |
| 2600 | % |
| 2601 | % There's some additional magic here for compatibility with the obscure way |
| 2602 | % that \package{array} works. |
| 2603 | % |
| 2604 | % \begin{macrocode} |
| 2605 | \def\newcolumntype#1{\@testopt{\nct@i{#1}}0} |
| 2606 | \def\nct@i#1[#2]{\@ifnextchar[{\nct@ii{#1}[#2]}{\nct@iii{#1}{[#2]}}} |
| 2607 | \def\nct@ii#1[#2][#3]{\nct@iii{#1}{[#2][#3]}} |
| 2608 | \def\nct@iii#1#2#3{% |
| 2609 | \expandafter\let\csname\tab@colset!col.\string#1\endcsname\relax% |
| 2610 | \expandafter\newcommand\csname\tab@colset!col.\string#1\endcsname#2{% |
| 2611 | \tab@deepmagic{#1}% |
| 2612 | \tab@mkpreamble% |
| 2613 | #3% |
| 2614 | }% |
| 2615 | } |
| 2616 | % \end{macrocode} |
| 2617 | % |
| 2618 | % Now for some hacking for compatibility with \package{tabularx}. |
| 2619 | % |
| 2620 | % \begin{macrocode} |
| 2621 | \def\newcol@#1[#2]{\nct@iii{#1}{[#2]}} |
| 2622 | % \end{macrocode} |
| 2623 | % |
| 2624 | % And now some more. This is seriously deep magic. Hence the name. |
| 2625 | % |
| 2626 | % \begin{macrocode} |
| 2627 | \def\tab@deepmagic#1{% |
| 2628 | \csname NC@rewrite@\string#1\endcsname\NC@find\tab@@magic@@% |
| 2629 | } |
| 2630 | \def\NC@find#1\tab@@magic@@{} |
| 2631 | % \end{macrocode} |
| 2632 | % |
| 2633 | % \end{macro} |
| 2634 | % |
| 2635 | % |
| 2636 | % \subsection{Standard column types} |
| 2637 | % |
| 2638 | % First, make sure we're setting up the right columns. This also sets the |
| 2639 | % default for the user. Other packages must not use the |\colset| command |
| 2640 | % for defining columns -- they should use the stack operations defined above. |
| 2641 | % For colour support, we ensure that the total stretch in a table cell is |
| 2642 | % 1\,fil. |
| 2643 | % |
| 2644 | % \begin{macrocode} |
| 2645 | \def\tab@halfhfil{\hskip\z@\@plus.5fil\relax} |
| 2646 | % \end{macrocode} |
| 2647 | % |
| 2648 | % And now on with the show. |
| 2649 | % |
| 2650 | % \begin{macrocode} |
| 2651 | \colset{tabular} |
| 2652 | % \end{macrocode} |
| 2653 | % |
| 2654 | % Now do the simple alignment types. These are fairly simple. The |
| 2655 | % mysterious kern in the \lit{l} type is to stop the |\col@sep| glue from |
| 2656 | % vanishing due to the |\unskip| inserted by the standard |\tab@midtext| if |
| 2657 | % the column contains no text. (Thanks for spotting this bug go to that |
| 2658 | % nice Mr~Carlisle.) |
| 2659 | % |
| 2660 | % \begin{macrocode} |
| 2661 | \coldef l{\tabcoltype{\kern\z@\tab@bgroup}{\tab@egroup\hfil}} |
| 2662 | \coldef c{\tabcoltype{\tab@halfhfil\tab@bgroup}{\tab@egroup\tab@halfhfil}} |
| 2663 | \coldef r{\tabcoltype{\hfil\tab@bgroup}{\tab@egroup}} |
| 2664 | % \end{macrocode} |
| 2665 | % |
| 2666 | % Some extensions now. These are explicitly textual or mathematical |
| 2667 | % columns. Can be useful if you're providing column types for other people. |
| 2668 | % I've inserted a kern here for exactly the same reason as for the \lit{l} |
| 2669 | % column type above. |
| 2670 | % |
| 2671 | % \begin{macrocode} |
| 2672 | \coldef T#1{\tab@aligncol{#1}{\tab@btext}{\tab@etext}} |
| 2673 | \coldef M#1{\tab@aligncol{#1}{\tab@bmaths}{\tab@emaths}} |
| 2674 | \def\tab@aligncol#1#2#3{% |
| 2675 | \if#1l\tabcoltype{\kern\z@#2}{#3\hfil}\fi% |
| 2676 | \if#1c\tabcoltype{\tab@halfhfil#2}{#3\tab@halfhfil}\fi% |
| 2677 | \if#1r\tabcoltype{\hfil#2}{#3}\fi% |
| 2678 | } |
| 2679 | % \end{macrocode} |
| 2680 | % |
| 2681 | % Now for the default rules. |
| 2682 | % |
| 2683 | % \begin{macrocode} |
| 2684 | \coldef ${\@firstoftwo{\tab@withrp\tab@vrule}} |
| 2685 | \coldef |{\@firstoftwo{\tab@withrp\tab@vrule[]}} |
| 2686 | \def\tab@vrule#1{\tabruletype{#1\vrule\@width\dimen@}\tab@mkpreamble} |
| 2687 | \coldef !#1{\tabruletype{#1}} |
| 2688 | % \end{macrocode} |
| 2689 | % |
| 2690 | % Deal with \lit{@} expressions. |
| 2691 | % |
| 2692 | % \begin{macrocode} |
| 2693 | \coldef @#1{\tabspctype{#1}} |
| 2694 | % \end{macrocode} |
| 2695 | % |
| 2696 | % And the paragraph types. I've added things to handle footnotes here. |
| 2697 | % |
| 2698 | % \begin{macrocode} |
| 2699 | \coldef p#1{\tabcoltype% |
| 2700 | {\savenotes\vtop\tab@bpar{#1}}% |
| 2701 | {\tab@epar\spewnotes\hfil}} |
| 2702 | \coldef m#1{\tabcoltype% |
| 2703 | {\savenotes$\vcenter\tab@bpar{#1}}% |
| 2704 | {\tab@epar$\spewnotes\hfil}} |
| 2705 | \coldef b#1{\tabcoltype% |
| 2706 | {\savenotes\vbox\tab@bpar{#1}}% |
| 2707 | {\tab@epar\spewnotes\hfil}} |
| 2708 | % \end{macrocode} |
| 2709 | % |
| 2710 | % Phew. Only a few more left now. The user text ones. |
| 2711 | % |
| 2712 | % \begin{macrocode} |
| 2713 | \coldef >#1{\tabuserpretype{#1}} |
| 2714 | \coldef <#1{\tabuserposttype{#1}} |
| 2715 | \coldef ?#1#2{% |
| 2716 | \ifx>#1\expandafter\tabuserpretype% |
| 2717 | \else\expandafter\tabuserposttype\fi% |
| 2718 | {#2}% |
| 2719 | \tab@append\tab@shortline{#2}% |
| 2720 | } |
| 2721 | % \end{macrocode} |
| 2722 | % |
| 2723 | % The strange column type. |
| 2724 | % |
| 2725 | % \begin{macrocode} |
| 2726 | \coldef ##1#2{\tabcoltype{#1}{#2}} |
| 2727 | % \end{macrocode} |
| 2728 | % |
| 2729 | % And \lit{*}, which repeats a preamble spec. This is really easy, and not |
| 2730 | % at all like the original one. |
| 2731 | % |
| 2732 | % \begin{macrocode} |
| 2733 | \coldef *#1#2{% |
| 2734 | \count@#1% |
| 2735 | \loop\ifnum\count@>0\relax% |
| 2736 | \tab@doreadpream{#2}% |
| 2737 | \advance\count@\m@ne% |
| 2738 | \repeat% |
| 2739 | } |
| 2740 | % \end{macrocode} |
| 2741 | % |
| 2742 | % |
| 2743 | % \subsection{Paragraph handling} |
| 2744 | % |
| 2745 | % First of all, starting new paragraphs: the vbox token is already there, and |
| 2746 | % we have the width as an argument. |
| 2747 | % |
| 2748 | % \begin{macro}{\tab@bpar} |
| 2749 | % |
| 2750 | % There are some gymnastics to do here to support lists which form the |
| 2751 | % complete text of the parbox. One of the odd things I'll do here is to |
| 2752 | % not insert a strut on the first line: instead, I'll put the text into a |
| 2753 | % box register so that I can inspect it later. So that I have access to |
| 2754 | % the height of the first line, I'll use a |\vtop| -- I can get at the |
| 2755 | % final depth by using |\prevdepth|, so this seems to be the most general |
| 2756 | % solution. |
| 2757 | % |
| 2758 | % \begin{macrocode} |
| 2759 | \def\tab@bpar#1{% |
| 2760 | \bgroup% |
| 2761 | \setlength\hsize{#1}% |
| 2762 | \@arrayparboxrestore% |
| 2763 | \setbox\z@\vtop\bgroup% |
| 2764 | \global\@minipagetrue% |
| 2765 | \global\@noskipsecfalse% |
| 2766 | \everypar\expandafter{\the\everypar% |
| 2767 | \global\@minipagefalse% |
| 2768 | \everypar{}% |
| 2769 | }% |
| 2770 | } |
| 2771 | % \end{macrocode} |
| 2772 | % |
| 2773 | % \end{macro} |
| 2774 | % |
| 2775 | % \begin{macro}{\tab@epar} |
| 2776 | % |
| 2777 | % To end the paragraph, close the box. That sounds easy, doesn't it? |
| 2778 | % I need to space out the top and bottom of the box so that it looks as if |
| 2779 | % struts have been applied. |
| 2780 | % |
| 2781 | % \begin{macrocode} |
| 2782 | \def\tab@epar{% |
| 2783 | % \end{macrocode} |
| 2784 | % |
| 2785 | % Anyway, I should end the current paragraph if I'm still in horizontal |
| 2786 | % mode. A simple |\par| will do this nicely. I'll also remove any trailing |
| 2787 | % vertical glue (which may be left there by a list environment), because |
| 2788 | % things will look very strange otherwise. |
| 2789 | % |
| 2790 | % \begin{macrocode} |
| 2791 | \ifhmode\@maybe@unskip\par\fi% |
| 2792 | \unskip% |
| 2793 | % \end{macrocode} |
| 2794 | % |
| 2795 | % Now I'll look at the depth of the last box: if it's less deep than my |
| 2796 | % special strut, I'll cunningly backpedal by a bit, and add a box with the |
| 2797 | % appropriate depth. Since this will lie on the previous baseline, it won't |
| 2798 | % alter the effective height of the box. There's a snag here. |\prevdepth| |
| 2799 | % may be wrong for example if the last thing inserted was a rule, or the |
| 2800 | % box is just empty. Check for this specially. (Thanks to Rowland McDonnell |
| 2801 | % for spotting this.) |
| 2802 | % |
| 2803 | % \begin{macrocode} |
| 2804 | \ifdim\prevdepth>-\@m\p@\ifdim\prevdepth<\dp\@arstrutbox% |
| 2805 | \kern-\prevdepth% |
| 2806 | \nointerlineskip% |
| 2807 | \vtop to\dp\@arstrutbox{}% |
| 2808 | \fi\fi% |
| 2809 | % \end{macrocode} |
| 2810 | % |
| 2811 | % I've finished the bottom of the box now: I'll close it, and start work on |
| 2812 | % the top again. |
| 2813 | % |
| 2814 | % \begin{macrocode} |
| 2815 | \egroup% |
| 2816 | % \end{macrocode} |
| 2817 | % |
| 2818 | % For top-alignment to work, the first item in the box must be another box. |
| 2819 | % (This is why I couldn't just set |\prevdepth| at the beginning.) If the |
| 2820 | % box isn't high enough, I'll add a box of the right height and then kern |
| 2821 | % backwards so that the `real' first box ends up in the right place. |
| 2822 | % |
| 2823 | % \begin{macrocode} |
| 2824 | \ifdim\ht\z@<\ht\@arstrutbox% |
| 2825 | \vbox to\ht\@arstrutbox{}% |
| 2826 | \kern-\ht\z@% |
| 2827 | \fi% |
| 2828 | \unvbox\z@% |
| 2829 | \egroup% |
| 2830 | } |
| 2831 | % \end{macrocode} |
| 2832 | % |
| 2833 | % \end{macro} |
| 2834 | % |
| 2835 | % |
| 2836 | % \subsection{Gentle persuasion} |
| 2837 | % |
| 2838 | % To persuade \package{longtable} to work, we emulate some features of |
| 2839 | % the \package{array} way of doing things. It's a shame, but we have to do |
| 2840 | % it, because \package{longtable} came first. |
| 2841 | % |
| 2842 | % Note the horribleness with the grouping here. In order to get everything |
| 2843 | % expanded at the right time, |\@preamble| just replaces itself with the (not |
| 2844 | % expanded!) preamble string, using |\the|. This means that the preamble |
| 2845 | % string must be visible in the group just above us. Now, |
| 2846 | % \package{longtable} (and \package{array} for that matter) does |
| 2847 | % |\@mkpreamble| immediately after opening a new group. So all we need to do |
| 2848 | % is close that group, do our stuff, and reopen the group again. (Evil |
| 2849 | % laughter\dots) |
| 2850 | % |
| 2851 | % \begin{macrocode} |
| 2852 | \def\@mkpream#1{% |
| 2853 | \endgroup% |
| 2854 | \colset{tabular}% |
| 2855 | \tab@initread% |
| 2856 | \def\tab@multicol{\@arstrut}% |
| 2857 | \tab@preamble{\tab@multicol}% |
| 2858 | \let\tab@lefttext\tab@lefttexthook% |
| 2859 | \let\tab@righttext\tab@righttexthook% |
| 2860 | \let\tab@leftruletext\tab@leftruletexthook% |
| 2861 | \let\tab@rightruletext\tab@rightruletexthook% |
| 2862 | \def\tab@midtext{\tab@setcr\ignorespaces\@sharp\@sharp\@maybe@unskip}% |
| 2863 | \tab@readpreamble{#1}% |
| 2864 | \gdef\@preamble{\the\tab@preamble}% |
| 2865 | \let\tab@bgroup\begingroup% |
| 2866 | \let\tab@egroup\endgroup% |
| 2867 | \begingroup% |
| 2868 | } |
| 2869 | % \end{macrocode} |
| 2870 | % |
| 2871 | % \subsection{Debugging} |
| 2872 | % |
| 2873 | % This macro just parses a preamble and displays it on the terminal. It |
| 2874 | % means I can see whether the thing's working. |
| 2875 | % |
| 2876 | % \begin{macrocode} |
| 2877 | \def\showpream#1{% |
| 2878 | \tab@initread% |
| 2879 | \tab@readpreamble{#1}% |
| 2880 | \showthe\tab@preamble% |
| 2881 | \showthe\tab@shortline% |
| 2882 | } |
| 2883 | % \end{macrocode} |
| 2884 | % |
| 2885 | % A quick macro for showing column types. |
| 2886 | % |
| 2887 | % \begin{macrocode} |
| 2888 | \def\showcol#1{% |
| 2889 | \expandafter\show\csname\tab@colset!col.\string#1\endcsname% |
| 2890 | } |
| 2891 | % \end{macrocode} |
| 2892 | % |
| 2893 | % |
| 2894 | % \subsection{The \env{tabular} and \env{array} environments} |
| 2895 | % |
| 2896 | % This is where we define the actual environments which users play with. |
| 2897 | % |
| 2898 | % \subsubsection{The environment routines} |
| 2899 | % |
| 2900 | % The real work is done in the |\@array| macro later. We just set up lots |
| 2901 | % (and I mean \emph{lots}) of parameters first, and then call |\@array|. |
| 2902 | % |
| 2903 | % \begin{macro}{\tab@array} |
| 2904 | % |
| 2905 | % The |\tab@array| macro does most of the common array things. |
| 2906 | % |
| 2907 | % \begin{macrocode} |
| 2908 | \def\tab@array{% |
| 2909 | \tab@width\z@% |
| 2910 | \let\tab@bgroup\tab@bmaths% |
| 2911 | \let\tab@egroup\tab@emaths% |
| 2912 | \@tabarray% |
| 2913 | } |
| 2914 | % \end{macrocode} |
| 2915 | % |
| 2916 | % \end{macro} |
| 2917 | % |
| 2918 | % \begin{macro}{\tab@btext} |
| 2919 | % \begin{macro}{\tab@bmaths} |
| 2920 | % \begin{macro}{\tab@etext} |
| 2921 | % \begin{macro}{\tab@emaths} |
| 2922 | % |
| 2923 | % These macros contain appropriate things to use when typesetting |
| 2924 | % text or maths macros. They're all trivial. They're here only for |
| 2925 | % later modification by funny things like the \env{smarray} environment. |
| 2926 | % |
| 2927 | % \begin{macrocode} |
| 2928 | \def\tab@btext{\begingroup} |
| 2929 | \def\tab@bmaths{\color@begingroup$} |
| 2930 | \def\tab@etext{\endgroup} |
| 2931 | \def\tab@emaths{\m@th$\color@endgroup} |
| 2932 | % \end{macrocode} |
| 2933 | % |
| 2934 | % \end{macro} |
| 2935 | % \end{macro} |
| 2936 | % \end{macro} |
| 2937 | % \end{macro} |
| 2938 | % |
| 2939 | % \begin{environment}{array} |
| 2940 | % |
| 2941 | % Now for the \env{array} environment. The `|$|' signs act as a group, so we |
| 2942 | % don't need to do extra grouping this time. Closing the environment is |
| 2943 | % easy. |
| 2944 | % |
| 2945 | % \begin{macrocode} |
| 2946 | \def\array{% |
| 2947 | \col@sep\arraycolsep% |
| 2948 | \let\tab@extrasep\arrayextrasep% |
| 2949 | \tab@normalstrut% |
| 2950 | \tab@array% |
| 2951 | } |
| 2952 | \def\endarray{% |
| 2953 | \crcr% |
| 2954 | \egroup% |
| 2955 | \tab@right% |
| 2956 | \endgroup% |
| 2957 | \tab@restorehlstate% |
| 2958 | \global\c@tabrow\count@% |
| 2959 | \def\@currentlabel{\p@tabrow\thetabrow}% |
| 2960 | \tab@endhook% |
| 2961 | } |
| 2962 | % \end{macrocode} |
| 2963 | % |
| 2964 | % \end{environment} |
| 2965 | % |
| 2966 | % \begin{environment}{smarray} |
| 2967 | % |
| 2968 | % Now for something a little different. The \env{smarray} environment |
| 2969 | % gives you an array with lots of small text. |
| 2970 | % |
| 2971 | % \begin{macrocode} |
| 2972 | \def\smarray{% |
| 2973 | \extrarowheight\z@% |
| 2974 | \col@sep\smarraycolsep% |
| 2975 | \let\tab@extrasep\smarrayextrasep% |
| 2976 | \def\tab@bmaths{\color@begingroup$\scriptstyle}% |
| 2977 | \def\tab@btext{\begingroup\scriptsize}% |
| 2978 | \setbox\z@\hbox{\scriptsize\strut}% |
| 2979 | \dimen@\ht\z@\dimen@ii\dp\z@\tab@setstrut% |
| 2980 | \tab@array% |
| 2981 | } |
| 2982 | \let\endsmarray\endarray |
| 2983 | % \end{macrocode} |
| 2984 | % |
| 2985 | % \end{environment} |
| 2986 | % |
| 2987 | % \begin{macro}{\tabstyle} |
| 2988 | % |
| 2989 | % This is a little hook that document designers can use to modify the |
| 2990 | % appearance of tables throughout a document. For example, I've set it to |
| 2991 | % make the text size |\small| in all tables in this document. Macro writers |
| 2992 | % shouldn't try to use it as a hook for their own evilness, though. I've |
| 2993 | % used |\providecommand| to avoid nobbling an existing definition. |
| 2994 | % |
| 2995 | % \begin{macrocode} |
| 2996 | \providecommand\tabstyle{} |
| 2997 | % \end{macrocode} |
| 2998 | % |
| 2999 | % \end{macro} |
| 3000 | % |
| 3001 | % \begin{macro}{\@tabular} |
| 3002 | % |
| 3003 | % The two \env{tabular} environments share lots of common code, so we |
| 3004 | % separate that out. (This needs to be done better.) All we really do here |
| 3005 | % is set up the |\tab@bgroup| and |\tab@egroup| to localise things properly, |
| 3006 | % and then go. |
| 3007 | % |
| 3008 | % \begin{macrocode} |
| 3009 | \def\@tabular#1{% |
| 3010 | \tabstyle% |
| 3011 | \setlength\tab@width{#1}% |
| 3012 | \let\tab@bgroup\tab@btext% |
| 3013 | \let\tab@egroup\tab@etext% |
| 3014 | \col@sep\tabcolsep% |
| 3015 | \let\tab@extrasep\tabextrasep% |
| 3016 | \tab@normalstrut% |
| 3017 | \@tabarray% |
| 3018 | } |
| 3019 | % \end{macrocode} |
| 3020 | % |
| 3021 | % \end{macro} |
| 3022 | % |
| 3023 | % \begin{environment}{tabular} |
| 3024 | % \begin{environment}{tabular*} |
| 3025 | % |
| 3026 | % These environments just call a macro which does all the common stuff. |
| 3027 | % |
| 3028 | % \begin{macrocode} |
| 3029 | \def\tabular{\@tabular\z@} |
| 3030 | \expandafter\let\csname tabular*\endcsname\@tabular |
| 3031 | \let\endtabular\endarray |
| 3032 | \expandafter\let\csname endtabular*\endcsname\endarray |
| 3033 | % \end{macrocode} |
| 3034 | % |
| 3035 | % \end{environment} |
| 3036 | % \end{environment} |
| 3037 | % |
| 3038 | % \subsubsection{Setting the strut height} |
| 3039 | % |
| 3040 | % \begin{macro}{\tab@setstrut} |
| 3041 | % |
| 3042 | % We use a magical strut, called |\@arstrut|, which keeps the table from |
| 3043 | % collapsing around our heads. This is where we set it up. |
| 3044 | % |
| 3045 | % It bases the array strut size on the given values of |\dimen@| and |
| 3046 | % |\dimen@ii|, amended by various appropriate fiddle values added in by |
| 3047 | % various people. |
| 3048 | % |
| 3049 | % \begin{macrocode} |
| 3050 | \def\tab@setstrut{% |
| 3051 | \setbox\@arstrutbox\hbox{% |
| 3052 | \vrule% |
| 3053 | \@height\arraystretch\dimen@% |
| 3054 | \@depth\arraystretch\dimen@ii% |
| 3055 | \@width\z@% |
| 3056 | }% |
| 3057 | } |
| 3058 | % \end{macrocode} |
| 3059 | % |
| 3060 | % \end{macro} |
| 3061 | % |
| 3062 | % \begin{macro}{\tab@normalstrut} |
| 3063 | % |
| 3064 | % This sets the strut the normal way, from the size of |\strutbox|. |
| 3065 | % |
| 3066 | % \begin{macrocode} |
| 3067 | \def\tab@normalstrut{% |
| 3068 | \dimen@\ht\strutbox\advance\dimen@\extrarowheight% |
| 3069 | \dimen\tw@\dp\strutbox% |
| 3070 | \tab@setstrut% |
| 3071 | } |
| 3072 | % \end{macrocode} |
| 3073 | % |
| 3074 | % \end{macro} |
| 3075 | % |
| 3076 | % \subsubsection{Setting up the alignment} |
| 3077 | % |
| 3078 | % The following bits are mainly for other packages to hook themselves onto. |
| 3079 | % |
| 3080 | % \begin{macrocode} |
| 3081 | \let\@arrayleft\relax |
| 3082 | \let\@arrayright\relax |
| 3083 | \let\tab@beginhook\@empty |
| 3084 | \let\tab@lefttexthook\@empty |
| 3085 | \let\tab@righttexthook\@empty |
| 3086 | \let\tab@leftruletexthook\@empty |
| 3087 | \let\tab@rightruletexthook\@empty |
| 3088 | \let\tab@endhook\@empty |
| 3089 | % \end{macrocode} |
| 3090 | % |
| 3091 | % For setting these hooks, we provide some handy commands. |
| 3092 | % |
| 3093 | % \begin{macrocode} |
| 3094 | \def\tab@addhookbefore#1#2{% |
| 3095 | \toks@{#2}\toks@\expandafter{\the\expandafter\toks@#1}% |
| 3096 | \edef#1{\the\toks@}% |
| 3097 | } |
| 3098 | \def\tab@addhookafter#1#2{% |
| 3099 | \toks@\expandafter{#1#2}% |
| 3100 | \edef#1{\the\toks@}% |
| 3101 | } |
| 3102 | % \end{macrocode} |
| 3103 | % |
| 3104 | % And now we get on with the real thing. |
| 3105 | % |
| 3106 | % \begin{macrocode} |
| 3107 | \def\@tabarray{% |
| 3108 | \let\@arrayleft\relax% |
| 3109 | \let\@arrayright\relax% |
| 3110 | \@testopt\@array c% |
| 3111 | } |
| 3112 | % \end{macrocode} |
| 3113 | % |
| 3114 | % \begin{macro}{\@array} |
| 3115 | % |
| 3116 | % The |\@array| macro does most of the real work for the environments. The |
| 3117 | % first job is to set up the row strut, which keeps the table rows at the |
| 3118 | % right height. We just take the normal strut box, and extend its height by |
| 3119 | % the |\extrarowheight| length parameter. |
| 3120 | % |
| 3121 | % \begin{macrocode} |
| 3122 | \def\@array[#1]#2{% |
| 3123 | % \end{macrocode} |
| 3124 | % |
| 3125 | % Sort out the hline state variable. We'll store the old value in a |
| 3126 | % control sequence to avoid wasting any more count registers. |
| 3127 | % |
| 3128 | % \begin{macrocode} |
| 3129 | \tab@beginhook% |
| 3130 | \count@\c@tabrow% |
| 3131 | \global\c@tabrow\z@% |
| 3132 | \edef\tab@restorehlstate{% |
| 3133 | \global\tab@endheight\the\tab@endheight% |
| 3134 | \gdef\noexpand\tab@hlstate{\tab@hlstate}% |
| 3135 | }% |
| 3136 | \begingroup% |
| 3137 | \def\tab@hlstate{n}% |
| 3138 | % \end{macrocode} |
| 3139 | % |
| 3140 | % Now we read the preamble. All the clever things we've already done are |
| 3141 | % terribly useful here. |
| 3142 | % |
| 3143 | % The |\tab@setcr| sets up |\\| to be a newline even if users have changed it |
| 3144 | % using something like |\raggedright|. |
| 3145 | % |
| 3146 | % \begin{macrocode} |
| 3147 | \colset{tabular}% |
| 3148 | \tab@initread% |
| 3149 | \let\tab@lefttext\tab@lefttexthook% |
| 3150 | \let\tab@righttext\tab@righttexthook% |
| 3151 | \let\tab@leftruletext\tab@leftruletexthook% |
| 3152 | \let\tab@rightruletext\tab@rightruletexthook% |
| 3153 | \def\tab@midtext{\tab@setcr\ignorespaces####\@maybe@unskip}% |
| 3154 | \def\tab@multicol{\@arstrut\tab@startrow}% |
| 3155 | \tab@preamble{\tab@multicol\tabskip\z@skip}% |
| 3156 | \tab@readpreamble{#2}% |
| 3157 | % \end{macrocode} |
| 3158 | % |
| 3159 | % Set up the default tabskip glue. This is easy: there isn't any. |
| 3160 | % |
| 3161 | % \begin{macrocode} |
| 3162 | \tab@leftskip\z@skip% |
| 3163 | \tab@rightskip\z@skip% |
| 3164 | % \end{macrocode} |
| 3165 | % |
| 3166 | % Now set up the positioning of the table. This is put into a separate macro |
| 3167 | % because it's rather complicated. |
| 3168 | % |
| 3169 | % \begin{macrocode} |
| 3170 | \tab@setposn{#1}% |
| 3171 | % \end{macrocode} |
| 3172 | % |
| 3173 | % Now work out how to start the alignment. |
| 3174 | % |
| 3175 | % \begin{macrocode} |
| 3176 | \ifdim\tab@width=\z@% |
| 3177 | \def\tab@halign{}% |
| 3178 | \else% |
| 3179 | \def\tab@halign{to\tab@width}% |
| 3180 | \fi% |
| 3181 | % \end{macrocode} |
| 3182 | % |
| 3183 | % Finally, do all the normal things we need to do before an alignment. Note |
| 3184 | % that we define |\tabularnewline| first, then set |\\| from that (using |
| 3185 | % |\tab@setcr|). Since |\\| is reset in the |\tab@midtext| of every table |
| 3186 | % cell, it becomes secondary to |\tabularnewline|. Doing things this way |
| 3187 | % avoids the problems with declarations like |\raggedright| which redefine |
| 3188 | % |\\| in their own (usually rather strange) way, so you don't need to mess |
| 3189 | % about with things like the |\PreserveBackslash| command given in the |
| 3190 | % \textit{\LaTeX\ Companion}. |
| 3191 | % |
| 3192 | % \begin{macrocode} |
| 3193 | \lineskip\z@\baselineskip\z@% |
| 3194 | \m@th% |
| 3195 | \def\tabularnewline{\tab@arraycr\tab@penalty}% |
| 3196 | \tab@setcr% |
| 3197 | \let\par\@empty% |
| 3198 | \everycr{}\tabskip\tab@leftskip% |
| 3199 | \tab@left\halign\tab@halign\expandafter\bgroup% |
| 3200 | \the\tab@preamble\tabskip\tab@rightskip\cr% |
| 3201 | } |
| 3202 | % \end{macrocode} |
| 3203 | % |
| 3204 | % \end{macro} |
| 3205 | % |
| 3206 | % You've no doubt noticed the |\tab@left| and |\tab@right| macros above. |
| 3207 | % These are set up here and elsewhere to allow other things to gain control |
| 3208 | % at various points of the table (they include and take the place of the |
| 3209 | % |\@arrayleft| and |\@arrayright| hooks in \package{array}, put in for |
| 3210 | % \package{delarray}'s use. |
| 3211 | % |
| 3212 | % \subsubsection{Positioning the table} |
| 3213 | % |
| 3214 | % \begin{macro}{\tab@setposn} |
| 3215 | % |
| 3216 | % This macro sets everything up for the table's positioning. It's rather |
| 3217 | % long, but not all that complicated. Honest. |
| 3218 | % |
| 3219 | % First, we set up some defaults (for centring). If anything goes wrong, we |
| 3220 | % just do the centring things. |
| 3221 | % |
| 3222 | % \begin{macrocode} |
| 3223 | \def\tab@setposn#1{% |
| 3224 | \def\tab@left{% |
| 3225 | \savenotes% |
| 3226 | \leavevmode\hbox\bgroup$\@arrayleft\vcenter\bgroup% |
| 3227 | }% |
| 3228 | \def\tab@right{% |
| 3229 | \egroup% |
| 3230 | \m@th\@arrayright$\egroup% |
| 3231 | \spewnotes% |
| 3232 | }% |
| 3233 | \global\tab@endheight\z@% |
| 3234 | % \end{macrocode} |
| 3235 | % |
| 3236 | % For the standard positioning things, we just do appropriate boxing things. |
| 3237 | % Note that the dollar signs are important, since \package{delarray} might |
| 3238 | % want to put its delimiters in here. |
| 3239 | % |
| 3240 | % The |\if@tempswa| switch it used to decide if we're doing an unboxed |
| 3241 | % tabular. We'll set it if we find an unbox-type position code, and then |
| 3242 | % check that everything's OK for this. |
| 3243 | % |
| 3244 | % \begin{macrocode} |
| 3245 | \@tempswafalse% |
| 3246 | \let\tab@penalty\relax% |
| 3247 | \if#1t% |
| 3248 | \def\tab@left{% |
| 3249 | \savenotes% |
| 3250 | \leavevmode\setbox\z@\hbox\bgroup$\@arrayleft\vtop\bgroup% |
| 3251 | }% |
| 3252 | \def\tab@right{% |
| 3253 | \egroup% |
| 3254 | \m@th\@arrayright$\egroup% |
| 3255 | \tab@raisebase% |
| 3256 | \spewnotes% |
| 3257 | }% |
| 3258 | \gdef\tab@hlstate{t}% |
| 3259 | \global\tab@endheight\ht\@arstrutbox% |
| 3260 | \else\if#1b% |
| 3261 | \def\tab@left{% |
| 3262 | \savenotes% |
| 3263 | \leavevmode\setbox\z@\hbox\bgroup$\@arrayleft\vbox\bgroup% |
| 3264 | }% |
| 3265 | \def\tab@right{% |
| 3266 | \egroup% |
| 3267 | \m@th\@arrayright$\egroup% |
| 3268 | \tab@lowerbase% |
| 3269 | \spewnotes% |
| 3270 | }% |
| 3271 | \gdef\tab@hlstate{b}% |
| 3272 | \else% |
| 3273 | \if#1L\@tempswatrue\fi% |
| 3274 | \if#1C\@tempswatrue\fi% |
| 3275 | \if#1R\@tempswatrue\fi% |
| 3276 | \fi\fi% |
| 3277 | % \end{macrocode} |
| 3278 | % |
| 3279 | % Now for some tests to make sure we're allowed to do the unboxing. We text |
| 3280 | % for |\@arrayleft| being defined, because people trying to hook us won't |
| 3281 | % understand unboxed tabulars. |
| 3282 | % |
| 3283 | % \begin{macrocode} |
| 3284 | \if@tempswa\ifhmode% |
| 3285 | \ifinner\tab@err@unbrh\@tempswafalse\else\par\fi% |
| 3286 | \fi\fi% |
| 3287 | \if@tempswa\ifmmode\tab@err@unbmm\@tempswafalse\fi\fi% |
| 3288 | \if@tempswa\ifx\@arrayleft\relax\else% |
| 3289 | \tab@err@unbext\@tempswafalse% |
| 3290 | \fi\fi% |
| 3291 | % \end{macrocode} |
| 3292 | % |
| 3293 | % Finally, if we're still doing an unboxed alignment, we need to sort out the |
| 3294 | % spacing. We know that no-one's tried to hook on to the environment, so we |
| 3295 | % clear |\tab@left| and |\tab@right|. |
| 3296 | % |
| 3297 | % \begin{macrocode} |
| 3298 | \if@tempswa% |
| 3299 | \def\tab@left{\vskip\parskip\medskip}% |
| 3300 | \def\tab@right{\par\@endpetrue\global\@ignoretrue}% |
| 3301 | % \end{macrocode} |
| 3302 | % |
| 3303 | % Now we need to sort out the alignment. The only way we can do this is by |
| 3304 | % playing with tabskip glue. There are two possibilities: |
| 3305 | % |
| 3306 | % \begin{itemize} |
| 3307 | % |
| 3308 | % \item If this is a straight \env{tabular} or an \env{array}, we just use |
| 3309 | % infinite glue. This is reasonable, I think. |
| 3310 | % |
| 3311 | % \item If we have a width for the table, we calculate the fixed values of |
| 3312 | % glue on either side. This is fairly easy, and forces the table to |
| 3313 | % the required width. |
| 3314 | % |
| 3315 | % \end{itemize} |
| 3316 | % |
| 3317 | % First, set up the left and right glues to represent the prevailing |
| 3318 | % margins set up by \env{list} environments. I think this is the right |
| 3319 | % thing to do. |
| 3320 | % |
| 3321 | % \begin{macrocode} |
| 3322 | \tab@leftskip\@totalleftmargin% |
| 3323 | \tab@rightskip\hsize% |
| 3324 | \advance\tab@rightskip-\linewidth% |
| 3325 | \advance\tab@rightskip-\@totalleftmargin% |
| 3326 | % \end{macrocode} |
| 3327 | % |
| 3328 | % First of all, deal with the simple case. I'm using 10000\,fill glue here, |
| 3329 | % in an attempt to suppress |\extracolsep| glue from making the table the |
| 3330 | % wrong width. It can always use filll glue if it really needs to, though. |
| 3331 | % |
| 3332 | % \begin{macrocode} |
| 3333 | \ifdim\tab@width=\z@% |
| 3334 | \if#1L\else\advance\tab@leftskip\z@\@plus10000fill\fi% |
| 3335 | \if#1R\else\advance\tab@rightskip\z@\@plus10000fill\fi% |
| 3336 | % \end{macrocode} |
| 3337 | % |
| 3338 | % Now for the fun bit. This isn't too hard really. The extra space I must |
| 3339 | % add around the table adds up to $|\linewidth| - |\tab@width|$. I just |
| 3340 | % need to add this onto the appropriate sides of the table. |
| 3341 | % |
| 3342 | % \begin{macrocode} |
| 3343 | \else% |
| 3344 | \dimen@\linewidth% |
| 3345 | \advance\dimen@-\tab@width% |
| 3346 | \if#1L\advance\tab@rightskip\dimen@\fi% |
| 3347 | \if#1R\advance\tab@leftskip\dimen@\fi% |
| 3348 | \if#1C% |
| 3349 | \advance\tab@leftskip.5\dimen@% |
| 3350 | \advance\tab@rightskip.5\dimen@% |
| 3351 | \fi% |
| 3352 | \fi% |
| 3353 | % \end{macrocode} |
| 3354 | % |
| 3355 | % Don't allow page breaks. David Carlisle's wonderful \env{longtable} |
| 3356 | % package does page breaks far better than I could possibly do here, and |
| 3357 | % we're compatible with it (wahey!). |
| 3358 | % |
| 3359 | % \begin{macrocode} |
| 3360 | \def\tab@penalty{\penalty\@M}% |
| 3361 | % \end{macrocode} |
| 3362 | % |
| 3363 | % Finally, set the new width of the table, and leave. |
| 3364 | % |
| 3365 | % \begin{macrocode} |
| 3366 | \tab@width\hsize% |
| 3367 | \fi% |
| 3368 | } |
| 3369 | % \end{macrocode} |
| 3370 | % |
| 3371 | % \end{macro} |
| 3372 | % |
| 3373 | % \subsubsection{Handling tops and bottoms} |
| 3374 | % |
| 3375 | % This is how the tops and bottoms of tables are made to line up with the |
| 3376 | % text on the same line, in the presence of arbitrary rules and space. The |
| 3377 | % old method, based on the way the \package{array} package worked, wasn't |
| 3378 | % terribly good. This new version copes much better with almost anything |
| 3379 | % that gets thrown at it. |
| 3380 | % |
| 3381 | % I'll keep a state in a macro (|\tab@hlstate|), which tells me what I'm |
| 3382 | % meant to be doing. The possible values are \lit{n}, which means I don't |
| 3383 | % have to do anything, \lit{t}, which means that I'm meant to be handling |
| 3384 | % top-aligned tables, and \lit{b}, which means that I'm meant to be lining |
| 3385 | % up the bottom. There are several other `substates' which have various |
| 3386 | % magic meanings. |
| 3387 | % |
| 3388 | % \begin{macrocode} |
| 3389 | \def\tab@hlstate{n} |
| 3390 | % \end{macrocode} |
| 3391 | % |
| 3392 | % When all's said and done, I extract the box containing the table, and |
| 3393 | % play with the height and depth to try and make it correct. |
| 3394 | % |
| 3395 | % \begin{macro}{\tab@addruleheight} |
| 3396 | % |
| 3397 | % This macro is called by `inter-row' things to add their height to our |
| 3398 | % dimen register. |
| 3399 | % |
| 3400 | % Only do this if the state indicates that it's sensible. |
| 3401 | % |
| 3402 | % \begin{macrocode} |
| 3403 | \def\tab@addruleheight#1{% |
| 3404 | \if\tab@hlstate n\else% |
| 3405 | \global\advance\tab@endheight#1\relax% |
| 3406 | \fi% |
| 3407 | } |
| 3408 | % \end{macrocode} |
| 3409 | % |
| 3410 | % \end{macro} |
| 3411 | % |
| 3412 | % \begin{macro}{\tab@startrow} |
| 3413 | % |
| 3414 | % This is called at the start of a row, from within the array preamble. |
| 3415 | % Currently, this assumes that the rows aren't bigger than their struts: |
| 3416 | % this is reasonable, although slightly limiting, and it could be done better |
| 3417 | % if I was willing to rip the alignment apart and put it back together |
| 3418 | % again. |
| 3419 | % |
| 3420 | % \begin{macrocode} |
| 3421 | \def\tab@startrow{% |
| 3422 | \if\tab@hlstate t% |
| 3423 | \gdef\tab@hlstate{n}% |
| 3424 | \else\if\tab@hlstate b% |
| 3425 | \global\tab@endheight\dp\@arstrutbox% |
| 3426 | \fi\fi% |
| 3427 | } |
| 3428 | % \end{macrocode} |
| 3429 | % |
| 3430 | % \end{macro} |
| 3431 | % |
| 3432 | % \begin{macro}{\tab@raisebase} |
| 3433 | % |
| 3434 | % This macro is called at the end of it all, to set the height and depth |
| 3435 | % of the box correctly. It sets the height to |\tab@endheight|, and the |
| 3436 | % depth to everything else. The box is in |\box|~0 currently. |
| 3437 | % |
| 3438 | % \begin{macrocode} |
| 3439 | \def\tab@raisebase{% |
| 3440 | \global\advance\tab@endheight-\ht\z@% |
| 3441 | \raise\tab@endheight\box\z@% |
| 3442 | } |
| 3443 | % \end{macrocode} |
| 3444 | % |
| 3445 | % \end{macro} |
| 3446 | % |
| 3447 | % \begin{macro}{\tab@lowerbase} |
| 3448 | % |
| 3449 | % And, for symmetry's sake, here's how to set the bottom properly instead. |
| 3450 | % |
| 3451 | % \begin{macrocode} |
| 3452 | \def\tab@lowerbase{% |
| 3453 | \global\advance\tab@endheight-\dp\z@% |
| 3454 | \lower\tab@endheight\box\z@% |
| 3455 | } |
| 3456 | % \end{macrocode} |
| 3457 | % |
| 3458 | % \end{macro} |
| 3459 | % |
| 3460 | % |
| 3461 | % \subsection{Breaking tables into bits} |
| 3462 | % |
| 3463 | % Unboxed tables have a wonderful advantage over boxed ones: you can stop |
| 3464 | % halfway through and do something else for a bit. Here's how: |
| 3465 | % |
| 3466 | % \begin{macro}{\tabpause} |
| 3467 | % |
| 3468 | % I'd like to avoid forbidding catcode changes here. I'll use |\doafter| |
| 3469 | % now I've got it, to ensure that colour handling and things occur |
| 3470 | % \emph{inside} the |\noalign| (otherwise they'll mess up the alignment |
| 3471 | % very seriously). We selectively include lots of stuff from |
| 3472 | % |\arrayparboxrestore|. |
| 3473 | % |
| 3474 | % We have to be careful here to ensure that everything works correctly within |
| 3475 | % lists. (The \package{amsmath} package had this problem in its |
| 3476 | % |\intertext| macro, so I'm not alone here.) |
| 3477 | % |
| 3478 | % \begin{macrocode} |
| 3479 | \def\tabpause#{% |
| 3480 | \noalign{\ifnum0=`}\fi% |
| 3481 | \let\if@nobreak\iffalse |
| 3482 | \let\if@noskipsec\iffalse |
| 3483 | \let\par\@@par |
| 3484 | \let\-\@dischyph |
| 3485 | \let\'\@acci\let\`\@accii\let\=\@acciii |
| 3486 | \everypar{}% |
| 3487 | \lineskip\normallineskip% |
| 3488 | \let\\\@normalcr% |
| 3489 | \color@begingroup% |
| 3490 | \tab@startpause% |
| 3491 | \vskip-\parskip% |
| 3492 | \parshape\@ne\@totalleftmargin\linewidth% |
| 3493 | \noindent% |
| 3494 | \doafter\tabpause@i% |
| 3495 | } |
| 3496 | \def\tabpause@i{% |
| 3497 | \nobreak% |
| 3498 | \tab@endpause% |
| 3499 | \color@endgroup% |
| 3500 | \ifnum0=`{\fi}% |
| 3501 | } |
| 3502 | % \end{macrocode} |
| 3503 | % |
| 3504 | % \end{macro} |
| 3505 | % |
| 3506 | % |
| 3507 | % \subsection{The wonderful world of \cmd\multicolumn} |
| 3508 | % |
| 3509 | % \begin{macro}{\multicolumn} |
| 3510 | % |
| 3511 | % This is actually fantastically easy. Watch and learn. Make sure you |
| 3512 | % notice the |\long|s here: remember that some table cells can contain |
| 3513 | % paragraphs, so it seems sensible to allow |\par| into the argument. |
| 3514 | % (As far as I know, most other |\multicolumn| commands don't do this, |
| 3515 | % which seems a little silly. Then again, I forgot to do it the first |
| 3516 | % time around.) |
| 3517 | % |
| 3518 | % \begin{macrocode} |
| 3519 | \long\def\multicolumn#1#2#3{% |
| 3520 | \multispan{#1}% |
| 3521 | \begingroup% |
| 3522 | \tab@multicol% |
| 3523 | \tab@initsubread% |
| 3524 | \long\def\tab@midtext{#3}% |
| 3525 | \let\tab@looped\tab@err@multi% |
| 3526 | \tab@readpreamble{#2}% |
| 3527 | \the\tab@preamble% |
| 3528 | \endgroup% |
| 3529 | \ignorespaces% |
| 3530 | } |
| 3531 | % \end{macrocode} |
| 3532 | % |
| 3533 | % \end{macro} |
| 3534 | % |
| 3535 | % |
| 3536 | % \subsection{Interlude: range lists} |
| 3537 | % |
| 3538 | % For processing arguments to |\vgap| and |\cline|, we need to be able to |
| 3539 | % do things with lists of column ranges. To save space, and to make my |
| 3540 | % fingers do less typing, here's some routines which do range handling. |
| 3541 | % |
| 3542 | % \begin{macro}{\ranges} |
| 3543 | % |
| 3544 | % Given a macro name and a comma separated list of ranges and simple numbers, |
| 3545 | % this macro will call the macro giving it each range in the list in turn. |
| 3546 | % Single numbers~$n$ will be turned into ranges $n$--$n$. |
| 3547 | % |
| 3548 | % The first job is to read the macro to do (which may already have some |
| 3549 | % arguments attached to it). We'll also start a group to make sure that |
| 3550 | % our changes to temp registers don't affect anyone else. |
| 3551 | % |
| 3552 | % There's a space before the delimiting |\q@delim| to stop numbers being |
| 3553 | % parsed to far and expanding our quark (which will stop \TeX\ dead in its |
| 3554 | % tracks). Since we use |\@ifnextchar| to look ahead, spaces in range lists |
| 3555 | % are perfectly all right. |
| 3556 | % |
| 3557 | % \begin{macrocode} |
| 3558 | \def\ranges#1#2{% |
| 3559 | \gdef\ranges@temp{#1}% |
| 3560 | \begingroup% |
| 3561 | \ranges@i#2 \q@delim% |
| 3562 | } |
| 3563 | % \end{macrocode} |
| 3564 | % |
| 3565 | % |
| 3566 | % We're at the beginning of the list. We expect either the closing marker |
| 3567 | % (if this is an empty list) or a number, which we can scoop up into a |
| 3568 | % scratch register. |
| 3569 | % |
| 3570 | % \begin{macrocode} |
| 3571 | \def\ranges@i{% |
| 3572 | \@ifnextchar\q@delim\ranges@done{\afterassignment\ranges@ii\count@}% |
| 3573 | } |
| 3574 | % \end{macrocode} |
| 3575 | % |
| 3576 | % We've read the first number in the range. If there's another number, we'll |
| 3577 | % expect a `|-|' sign to be next. If there is no `|-|', call the user's code |
| 3578 | % with the number duplicated and then do the rest of the list. |
| 3579 | % |
| 3580 | % \begin{macrocode} |
| 3581 | \def\ranges@ii{% |
| 3582 | \@ifnextchar-\ranges@iii{\ranges@do\count@\count@\ranges@v}% |
| 3583 | } |
| 3584 | % \end{macrocode} |
| 3585 | % |
| 3586 | % Now we strip the `|-|' off and read the other number into a temporary |
| 3587 | % register. |
| 3588 | % |
| 3589 | % \begin{macrocode} |
| 3590 | \def\ranges@iii-{\afterassignment\ranges@iv\@tempcnta} |
| 3591 | % \end{macrocode} |
| 3592 | % |
| 3593 | % We have both ends of the range now, so call the user's code, passing it |
| 3594 | % both ends of the range. |
| 3595 | % |
| 3596 | % \begin{macrocode} |
| 3597 | \def\ranges@iv{\ranges@do\count@\@tempcnta\ranges@v} |
| 3598 | % \end{macrocode} |
| 3599 | % |
| 3600 | % We've finished doing an item now. If we have a `|,|' next, then start |
| 3601 | % over with the next item. Otherwise, if we're at the end of the list, |
| 3602 | % we can end happily. Finally, if we're totally confused, raise an |
| 3603 | % error. |
| 3604 | % |
| 3605 | % \begin{macrocode} |
| 3606 | \def\ranges@v{% |
| 3607 | \@ifnextchar,% |
| 3608 | \ranges@vi% |
| 3609 | {% |
| 3610 | \@ifnextchar\q@delim% |
| 3611 | \ranges@done% |
| 3612 | {\tab@err@range\ranges@vi,}% |
| 3613 | }% |
| 3614 | } |
| 3615 | % \end{macrocode} |
| 3616 | % |
| 3617 | % We had a comma, so gobble it, read the next number, and go round again. |
| 3618 | % |
| 3619 | % \begin{macrocode} |
| 3620 | \def\ranges@vi,{\afterassignment\ranges@ii\count@} |
| 3621 | % \end{macrocode} |
| 3622 | % |
| 3623 | % Here's how we call the user's code, now. We close the group, so that the |
| 3624 | % user's code doesn't have to do global things to remember its results, and |
| 3625 | % we expand the two range ends from their count registers. We also ensure |
| 3626 | % that the range is the right way round. |
| 3627 | % |
| 3628 | % \begin{macrocode} |
| 3629 | \def\ranges@do#1#2{% |
| 3630 | \ifnum#1>#2\else% |
| 3631 | \expandafter\endgroup% |
| 3632 | \expandafter\ranges@temp% |
| 3633 | \expandafter{% |
| 3634 | \the\expandafter#1% |
| 3635 | \expandafter}% |
| 3636 | \expandafter{% |
| 3637 | \the#2% |
| 3638 | }% |
| 3639 | \begingroup% |
| 3640 | \fi% |
| 3641 | } |
| 3642 | % \end{macrocode} |
| 3643 | % |
| 3644 | % And finishing the scan is really easy. We close the group after gobbling |
| 3645 | % the close token. |
| 3646 | % |
| 3647 | % \begin{macrocode} |
| 3648 | \def\ranges@done\q@delim{\endgroup} |
| 3649 | % \end{macrocode} |
| 3650 | % |
| 3651 | % \end{macro} |
| 3652 | % |
| 3653 | % \begin{macro}{\ifinrange} |
| 3654 | % |
| 3655 | % Something a little more useful, now. |\ifinrange| takes four arguments: |
| 3656 | % a number, a range list (as above), and two token lists which I'll call |
| 3657 | % \emph{then} and \emph{else}. If the number is in the list, I'll do |
| 3658 | % \emph{then}, otherwise I'll do \emph{else}. |
| 3659 | % |
| 3660 | % \begin{macrocode} |
| 3661 | \def\ifinrange#1#2{% |
| 3662 | \@tempswafalse% |
| 3663 | \count@#1% |
| 3664 | \ranges\ifinrange@i{#2}% |
| 3665 | \if@tempswa% |
| 3666 | \expandafter\@firstoftwo% |
| 3667 | \else% |
| 3668 | \expandafter\@secondoftwo% |
| 3669 | \fi% |
| 3670 | } |
| 3671 | \def\ifinrange@i#1#2{% |
| 3672 | \ifnum\count@<#1 \else\ifnum\count@>#2 \else\@tempswatrue\fi\fi% |
| 3673 | } |
| 3674 | % \end{macrocode} |
| 3675 | % |
| 3676 | % \end{macro} |
| 3677 | % |
| 3678 | % |
| 3679 | % \subsection{Horizontal rules OK} |
| 3680 | % |
| 3681 | % This is where all the gubbins for |\vgap| and friends is kept, lest it |
| 3682 | % contaminate fairly clean bits of code found elsewhere. |
| 3683 | % |
| 3684 | % \subsubsection{Common parsing for rule parameters twiddling} |
| 3685 | % |
| 3686 | % \begin{macro}{\tab@ruleparams} |
| 3687 | % |
| 3688 | % Given a macro name, make a (global) macro |\tab@ruledecls|, which sets |
| 3689 | % |\dimen0| to be the chosen rule thickness, and sets up colours and whatnot, |
| 3690 | % and then and calls the macro. We parse a `|*|' to mean |
| 3691 | % |\arraythickrulewidth|, an optional argument which should be something |
| 3692 | % |\setlength| can understand, or nothing, which gives the default |
| 3693 | % |\arrayrulewidth|. |
| 3694 | % |
| 3695 | % To make this properly hookable, we need to make a list of properties and |
| 3696 | % gather them together. |
| 3697 | % |
| 3698 | % \begin{macrocode} |
| 3699 | \let\tab@rp@inithook\@empty |
| 3700 | \let\tab@rp@sethook\@empty |
| 3701 | \let\tab@rp@donehook\@empty |
| 3702 | \let\tab@rp@default\@empty |
| 3703 | \def\tab@ruleparams#1{% |
| 3704 | {\ifnum0=`}\fi% |
| 3705 | \tab@rp@inithook% |
| 3706 | \def\tab@rp@next{\ifnum0=`{\fi}#1}% |
| 3707 | \expandafter\tab@rp@keys\expandafter{\tab@rp@default}% |
| 3708 | \@ifstar\tab@rp@star\tab@rp@what% |
| 3709 | } |
| 3710 | \def\tab@rp@star{\dimen@\arraythickrulewidth\tab@rp@what} |
| 3711 | \def\tab@rp@what{\@ifnextchar[\tab@rp@opt\tab@rp@done} |
| 3712 | \def\tab@rp@opt[#1]{\tab@rp@keys{#1}\tab@rp@done} |
| 3713 | \def\tab@rp@keys{\mkparse{mdwtab:rule}} |
| 3714 | \def\tab@rp@done{% |
| 3715 | \protected@xdef\tab@rp@{\tab@rp@sethook}% |
| 3716 | \tab@rp@donehook% |
| 3717 | \tab@rp@next% |
| 3718 | } |
| 3719 | \def\tab@withrp#1{\tab@ruleparams{\tab@withrp@i{#1}}} |
| 3720 | \def\tab@withrp@i#1{% |
| 3721 | \toks@{#1}% |
| 3722 | \toks@\expandafter{\the\expandafter\toks@\expandafter{\tab@rp@}}% |
| 3723 | \the\toks@% |
| 3724 | } |
| 3725 | % \end{macrocode} |
| 3726 | % |
| 3727 | % And now to define the width parameters. |
| 3728 | % |
| 3729 | % \begin{macrocode} |
| 3730 | \tab@addhookafter\tab@rp@inithook{\dimen@\arrayrulewidth} |
| 3731 | \tab@addhookafter\tab@rp@sethook{\dimen@\the\dimen@} |
| 3732 | \tab@addhookafter\tab@rp@donehook{\global\tab@rulewidth\dimen@} |
| 3733 | \mkdef{mdwtab:rule}{width}{\setlength\dimen@{#1}} |
| 3734 | \mkdef{mdwtab:rule}{thin}*{\dimen@\arrayrulewidth} |
| 3735 | \mkdef{mdwtab:rule}{thick}*{\dimen@\arraythickrulewidth} |
| 3736 | \mkdef*{mdwtab:rule}*{\setlength\dimen@{#1}} |
| 3737 | % \end{macrocode} |
| 3738 | % |
| 3739 | % \end{macro} |
| 3740 | % |
| 3741 | % \begin{macro}{\tabsetruleparams} |
| 3742 | % |
| 3743 | % And the user default-parameter list. |
| 3744 | % |
| 3745 | % \begin{macrocode} |
| 3746 | \def\tabsetruleparams{\def\tab@rp@default} |
| 3747 | % \end{macrocode} |
| 3748 | % |
| 3749 | % \end{macro} |
| 3750 | % |
| 3751 | % \subsubsection{Drawing horizontal rules} |
| 3752 | % |
| 3753 | % \begin{macro}{\hline} |
| 3754 | % |
| 3755 | % Note the funny use of |\noalign| to allow \TeX\ stomach ops like |
| 3756 | % |\futurelet| without starting a new table row. This lets us see if there's |
| 3757 | % another |\hline| coming up, so we can see if we need to insert extra |
| 3758 | % vertical space. |
| 3759 | % |
| 3760 | % \begin{macrocode} |
| 3761 | \def\hline{\noalign\tab@ruleparams\hline@prep} |
| 3762 | \def\hline@prep{% |
| 3763 | \tab@dohline% |
| 3764 | \noalign{\ifnum0=`}\fi% |
| 3765 | \tab@penalty% |
| 3766 | \futurelet\@let@token\hline@i% |
| 3767 | } |
| 3768 | % \end{macrocode} |
| 3769 | % |
| 3770 | % We check here for another |\hline| command, and insert glue if there is. |
| 3771 | % This looks terrible, though, and |\hlx{hvh}| is much nicer. Still\dots |
| 3772 | % |
| 3773 | % \begin{macrocode} |
| 3774 | \def\hline@i{% |
| 3775 | \@tempswafalse% |
| 3776 | \ifx\@let@token\hline\@tempswatrue\fi% |
| 3777 | \ifx\@let@token\hline@prep\@tempswatrue\fi% |
| 3778 | \if@tempswa% |
| 3779 | \vskip\doublerulesep% |
| 3780 | \tab@addruleheight\doublerulesep% |
| 3781 | \fi% |
| 3782 | \ifnum0=`{\fi}% |
| 3783 | } |
| 3784 | % \end{macrocode} |
| 3785 | % |
| 3786 | % \end{macro} |
| 3787 | % |
| 3788 | % \begin{macro}{\tab@dohline} |
| 3789 | % |
| 3790 | % This is where hlines actually get drawn. |
| 3791 | % Drawing lines is more awkward than it used to be, particularly in unboxed |
| 3792 | % tables. It used to be a case simply of saying |\noalign{\hrule}|. |
| 3793 | % However, since unboxed tables are actually much wider than they look, this |
| 3794 | % would make the rules stretch right across the page and look generally |
| 3795 | % horrible. |
| 3796 | % |
| 3797 | % The solution is simple: we basically do a dirty big |\cline|. |
| 3798 | % |
| 3799 | % \begin{macrocode} |
| 3800 | \def\tab@dohline{% |
| 3801 | \multispan\tab@columns% |
| 3802 | \color@begingroup% |
| 3803 | \tab@rp@\leaders\hrule\@height\dimen@\hfil% |
| 3804 | \tab@addruleheight\dimen@% |
| 3805 | \color@endgroup% |
| 3806 | \cr% |
| 3807 | } |
| 3808 | % \end{macrocode} |
| 3809 | % |
| 3810 | % \end{macro} |
| 3811 | % |
| 3812 | % \subsubsection{Vertical rules} |
| 3813 | % |
| 3814 | % I couldn't fit these in anywhere else, so they'll have to go here. I'll |
| 3815 | % provide a new optional argument which specifies the width of the rule; this |
| 3816 | % gets rid of the problem described in the \emph{Companion}, where to get |
| 3817 | % an unusually wide vertical rule, you have to play with things like |
| 3818 | % \syntax{"\\vrule width" <dimen>} which really isn't too nice. |
| 3819 | % |
| 3820 | % \begin{macro}{\vline} |
| 3821 | % |
| 3822 | % Now uses the general |\tab@ruleparams| parser. We save and restore the |
| 3823 | % global |\tab@rulewidth| parameter here. |
| 3824 | % |
| 3825 | % \begin{macrocode} |
| 3826 | \def\vline{% |
| 3827 | \begingroup% |
| 3828 | \@tempdima\tab@rulewidth\let\safe@\tab@rp@% |
| 3829 | \tab@ruleparams\tab@vline% |
| 3830 | } |
| 3831 | \def\tab@vline{% |
| 3832 | \tab@rp@\vrule\@width\dimen@% |
| 3833 | \global\tab@rulewidth\@tempdima\global\let\tab@rp@\safe@% |
| 3834 | \endgroup% |
| 3835 | } |
| 3836 | % \end{macrocode} |
| 3837 | % |
| 3838 | % \end{macro} |
| 3839 | % |
| 3840 | % \subsubsection{Drawing bits of lines} |
| 3841 | % |
| 3842 | % Just for a bit of fun, here's an extended version of |\cline| which takes |
| 3843 | % a list of columns to draw lines under, rather than just a single range. |
| 3844 | % |
| 3845 | % \begin{macro}{\cline} |
| 3846 | % |
| 3847 | % Not a single line of code written yet, and we already have a dilemma on |
| 3848 | % our hands. Multiple consecutive |\cline| commands are meant to draw |
| 3849 | % on the same vertical bit of table. But horizontal lines are meant to have |
| 3850 | % thickness now. Worse, if the lines have real thickness then we leave gaps |
| 3851 | % in the vertical rules which aren't covered by our line. But if we |
| 3852 | % backspace over the line, then we overwrite it with coloured blobs. |
| 3853 | % |
| 3854 | % We give up on doing the job properly -- that's just doomed. Backspace over |
| 3855 | % the previous row, and provide a hack for doing the spacing right elsewhere. |
| 3856 | % |
| 3857 | % Now the problem remains how best to do the job. The way I see it, there |
| 3858 | % are three possibilities: |
| 3859 | % |
| 3860 | % \begin{itemize} |
| 3861 | % |
| 3862 | % \item We can start a table row, and then for each column of the table |
| 3863 | % (as recorded in |\tab@columns|) we look to see if that column is |
| 3864 | % listed in the range list and if so draw the rule. This requires |
| 3865 | % lots of scanning of the range list. |
| 3866 | % |
| 3867 | % \item We can take each range in the list, and draw rules appropriately, |
| 3868 | % just like the old |\cline| used to do, and starting a new table row |
| 3869 | % for each. |
| 3870 | % |
| 3871 | % \item We can start a table row, and then for each range remember where we |
| 3872 | % stopped drawing the last row, move to the start of the new one, and |
| 3873 | % draw it. If we start moving backwards, we close the current row |
| 3874 | % and open a new one. |
| 3875 | % |
| 3876 | % \end{itemize} |
| 3877 | % |
| 3878 | % The last option looks the most efficient, and the most difficult. This |
| 3879 | % is therefore what I shall do |;-)|. |
| 3880 | % |
| 3881 | % The first thing to do is to add in a little negative space, and start a |
| 3882 | % table row (omitting the first item). Then scan the range list, and finally |
| 3883 | % close the table row and add some negative space again. |
| 3884 | % |
| 3885 | % We need a global count register to keep track of where we are. Mixing |
| 3886 | % local and global assignments causes all sorts of tragedy, so I shall hijack |
| 3887 | % |\tab@state|. |
| 3888 | % |
| 3889 | % \begin{macrocode} |
| 3890 | \def\cline{\noalign\tab@ruleparams\cline@do} |
| 3891 | % \end{macrocode} |
| 3892 | % |
| 3893 | % Now for the tricky bit. When we're given a range, we look to see if the |
| 3894 | % first number is less than |\tab@state|. If so, we quickly close the |
| 3895 | % current row, kern backwards and start again with an |\omit| and reset |
| 3896 | % |\tab@state| to 1, and try again. This is hardly perfect, but gets the job |
| 3897 | % done in many cases. Correct |\vgap| insertion fixes the remaining bugs. |
| 3898 | % |
| 3899 | % \begin{macrocode} |
| 3900 | \def\cline@do#1{% |
| 3901 | \noalign{\kern-\tab@rulewidth}% |
| 3902 | \omit% |
| 3903 | \global\tab@state\@ne% |
| 3904 | \ranges\cline@do@i{#1}\cr% |
| 3905 | } |
| 3906 | \def\cline@do@i#1#2{% |
| 3907 | \ifnum#1<\tab@state\relax% |
| 3908 | \tab@@cr% |
| 3909 | \noalign{\kern-\tab@rulewidth\tab@penalty}% |
| 3910 | \omit% |
| 3911 | \global\tab@state\@ne% |
| 3912 | \fi% |
| 3913 | % \end{macrocode} |
| 3914 | % |
| 3915 | % We are now either at or in front of the column position required. If |
| 3916 | % we're too far back, we must |\hfil&\omit| our way over to the correct |
| 3917 | % column. |
| 3918 | % |
| 3919 | % \begin{macrocode} |
| 3920 | \@whilenum\tab@state<#1\do{% |
| 3921 | \hfil\tab@@tab@omit% |
| 3922 | \global\advance\tab@state\@ne% |
| 3923 | }% |
| 3924 | % \end{macrocode} |
| 3925 | % |
| 3926 | % We've found the start correctly. We must deal with a tiny problem now: |
| 3927 | % if this is not the first table cell, the left hand vertical rule is in the |
| 3928 | % column to the left, so our horizontal rule won't match up properly. So |
| 3929 | % we skip back by a bit to compensate. If there isn't actually a vertical |
| 3930 | % rule to line up with, no-one will notice, because the rules are so thin. |
| 3931 | % This adds a little touch of quality to the whole thing, which is after all |
| 3932 | % the point of this whole exercise. |
| 3933 | % |
| 3934 | % \begin{macrocode} |
| 3935 | \ifnum\tab@state>\@ne% |
| 3936 | \kern-\arrayrulewidth% |
| 3937 | \fi% |
| 3938 | % \end{macrocode} |
| 3939 | % |
| 3940 | % Now we must stretch this table cell to the correct width. |
| 3941 | % |
| 3942 | % \begin{macrocode} |
| 3943 | \@whilenum\tab@state<#2\do{% |
| 3944 | \tab@@span@omit% |
| 3945 | \global\advance\tab@state\@ne% |
| 3946 | }% |
| 3947 | % \end{macrocode} |
| 3948 | % |
| 3949 | % We're ready. Draw the rule. Note that this is |\hfill| glue, just in case |
| 3950 | % we start putting in |\hfil| glue when we step onto the next cell. |
| 3951 | % |
| 3952 | % \begin{macrocode} |
| 3953 | \color@begingroup% |
| 3954 | \tab@rp@% |
| 3955 | \leaders\hrule\@height\tab@rulewidth\hfill% |
| 3956 | \color@endgroup% |
| 3957 | } |
| 3958 | % \end{macrocode} |
| 3959 | % |
| 3960 | % Some alignment primitives are hidden inside macros so they don't get seen |
| 3961 | % at the wrong time. This is what they look like: |
| 3962 | % |
| 3963 | % \begin{macrocode} |
| 3964 | \def\tab@@cr{\cr} |
| 3965 | \def\tab@@tab@omit{&\omit} |
| 3966 | \def\tab@@span@omit{\span\omit} |
| 3967 | % \end{macrocode} |
| 3968 | % |
| 3969 | % \end{macro} |
| 3970 | % |
| 3971 | % \subsubsection{Drawing short table rows} |
| 3972 | % |
| 3973 | % Before I start on a description of more code, I think I'll briefly discuss |
| 3974 | % my reasons for leaving the |\vgap| command in its current state. There's a |
| 3975 | % reasonable case for introducing an interface between |\vgap| and |
| 3976 | % |\multicolumn|, to avoid all the tedious messing about with column |
| 3977 | % ranges. There are good reasons why I'm not going to do this: |
| 3978 | % |
| 3979 | % \begin{itemize} |
| 3980 | % |
| 3981 | % \item It's very difficult to do: it requires either postprocessing of |
| 3982 | % the table or delaying processing of each row until I know exactly |
| 3983 | % what's in it; a |\multicolumn| in a row should be able to affect |
| 3984 | % a |\vgap| before the row, which gets very nasty. This package is |
| 3985 | % probably far too large already, and adding more complexity and |
| 3986 | % running the risk of exhausting \TeX's frustratingly finite capacity |
| 3987 | % for the sake of relieving the user of a fairly trivial job doesn't |
| 3988 | % seem worthwhile. |
| 3989 | % |
| 3990 | % \item Perhaps more importantly, there are perfectly valid occasions when |
| 3991 | % it's useful to have the current vgap behaviour. For example, the |
| 3992 | % \texttt{MIX} word layout diagrams found in \emph{The Art of |
| 3993 | % Computer Programming} use the little `stub lines' to show where |
| 3994 | % data items cross byte boundaries: |
| 3995 | % |
| 3996 | % ^^A This actually looks terrifyingly similar to the original. |
| 3997 | % ^^A The leading @{} is there to stop the table looking off-centre, |
| 3998 | % ^^A because there's no left hand rule telling you where the table |
| 3999 | % ^^A starts, like there is on the right, just the \tabcolsep glue. |
| 4000 | % |
| 4001 | % \begingroup |
| 4002 | % \newcommand{\wide}[2]{\multicolumn{#1}{c|}{\ttfamily #2}} |
| 4003 | % \begin{tabular}[C]{@{} r @{\qquad} | Mc | *{5}{c|}} \hlx{c{2-7} v} |
| 4004 | % empty & - & 1 & 0 & 0 & 0 & 0 \\ \hlx{v c{2-7} v} |
| 4005 | % occupied & + & \wide{2}{LINK} & \wide{3}{KEY} \\ \hlx{v c{2-7}} |
| 4006 | % \end{tabular} |
| 4007 | % \endgroup |
| 4008 | % |
| 4009 | % \end{itemize} |
| 4010 | % |
| 4011 | % That's my excuses out of the way; now I'll press on with the actual |
| 4012 | % programming. |
| 4013 | % |
| 4014 | % \begin{macro}{\tab@checkrule} |
| 4015 | % |
| 4016 | % We have a range list in |\tab@xcols| and a number as an argument. If we |
| 4017 | % find the number in the list, we just space out the following group, |
| 4018 | % otherwise we let it be. |
| 4019 | % |
| 4020 | % \begin{macrocode} |
| 4021 | \def\tab@checkrule#1{% |
| 4022 | \count@#1\relax% |
| 4023 | \expandafter\ifinrange% |
| 4024 | \expandafter\count@% |
| 4025 | \expandafter{\tab@xcols}% |
| 4026 | {\tab@checkrule@i}% |
| 4027 | {}% |
| 4028 | } |
| 4029 | \def\tab@checkrule@i#1{\setbox\z@\hbox{#1}\hb@xt@\wd\z@{}} |
| 4030 | % \end{macrocode} |
| 4031 | % |
| 4032 | % \end{macro} |
| 4033 | % |
| 4034 | % \begin{macro}{\vgap} |
| 4035 | % |
| 4036 | % We must tread carefully here. A single misplaced stomach operation can |
| 4037 | % cause error messages. We therefore start with an |\omit| so we can search |
| 4038 | % for optional arguments. |
| 4039 | % |
| 4040 | % So that |\hlx| can get control after |\vgap| has finished, we provide a |
| 4041 | % hook called |\vgap@after| which is expanded after |\vgap| has finished. |
| 4042 | % Here we make it work like |\@empty|, which expands to nothing. (Note that |
| 4043 | % |\relax| will start a new table row, so we can't use that.) There are |
| 4044 | % some penalty items here to stick the |\vgap| row to the text row and |
| 4045 | % |\hline| that are adjacent to it. The \package{longtable} package will |
| 4046 | % split an |\hline| in half, so this is the correct thing to do. |
| 4047 | % |
| 4048 | % \begin{macrocode} |
| 4049 | \def\vgap{% |
| 4050 | \noalign{\nobreak}% |
| 4051 | \omit% |
| 4052 | \global\let\vgap@after\@empty% |
| 4053 | \iffalse{\fi\ifnum0=`}\fi% |
| 4054 | \@ifnextchar[\vgap@i\vgap@simple% |
| 4055 | } |
| 4056 | % \end{macrocode} |
| 4057 | % |
| 4058 | % We set up two different sorts of |\vgap| -- a simple one which allows all |
| 4059 | % rules to be passed through, and a specific one which carefully vets each |
| 4060 | % one (and is therefore slower). We decide which to so based on the presence |
| 4061 | % of an optional argument. |
| 4062 | % |
| 4063 | % The optional argument handler just passes its argument to an interface |
| 4064 | % routine which is used by |\hlx|. |
| 4065 | % |
| 4066 | % \begin{macrocode} |
| 4067 | \def\vgap@i[#1]{\vgap@spec{#1}} |
| 4068 | % \end{macrocode} |
| 4069 | % |
| 4070 | % Now we handle specified columns. Since we're in an omitted table cell, we |
| 4071 | % must set things up globally. Assign the column spec to a macro, and set up |
| 4072 | % vetting by the routine above. Then just go and do the job. |
| 4073 | % |
| 4074 | % \begin{macrocode} |
| 4075 | \def\vgap@spec#1#2{% |
| 4076 | \gdef\tab@xcols{#1}% |
| 4077 | \global\let\tab@ckr\tab@checkrule% |
| 4078 | \vgap@do{#2}% |
| 4079 | } |
| 4080 | % \end{macrocode} |
| 4081 | % |
| 4082 | % Handle all columns. Just gobble the column number for each rule, and let |
| 4083 | % the drawing pass unharmed. Easy. |
| 4084 | % |
| 4085 | % \begin{macrocode} |
| 4086 | \def\vgap@simple#1{% |
| 4087 | \global\let\tab@ckr\@gobble% |
| 4088 | \vgap@do{#1}% |
| 4089 | } |
| 4090 | % \end{macrocode} |
| 4091 | % |
| 4092 | % This is where stuff actually gets done. We set the |\vgap| flag on while |
| 4093 | % we do the short row. Then just expand the token list we built while |
| 4094 | % scanning the preamble. |
| 4095 | % |
| 4096 | % Note that the flag is cleared at the end of the last column, to allow other |
| 4097 | % funny things like |\noalign| and |\omit| before a new row is started. |
| 4098 | % |
| 4099 | % \begin{macrocode} |
| 4100 | \def\vgap@do#1{% |
| 4101 | \ifnum0=`{}\fi% |
| 4102 | \global\tab@vgaptrue% |
| 4103 | \the\tab@shortline% |
| 4104 | \vrule\@height#1\@width\z@% |
| 4105 | \global\tab@vgapfalse |
| 4106 | \tab@addruleheight{#1}% |
| 4107 | \cr% |
| 4108 | \noalign{\nobreak}% |
| 4109 | \vgap@after% |
| 4110 | } |
| 4111 | % \end{macrocode} |
| 4112 | % |
| 4113 | % \end{macro} |
| 4114 | % |
| 4115 | % \subsubsection{Prettifying syntax} |
| 4116 | % |
| 4117 | % \begin{macro}{\hlx} |
| 4118 | % |
| 4119 | % This is like a poor cousin to the preamble parser. The whole loop is |
| 4120 | % carefully written to take place \emph{only} in \TeX's mouth, so the |
| 4121 | % alignment handling bits half way down the gullet don't see any of this. |
| 4122 | % |
| 4123 | % First, pass the string to another routine. |
| 4124 | % |
| 4125 | % \begin{macrocode} |
| 4126 | \def\hlx{\noalign\tab@ruleparams\hlx@prep} |
| 4127 | \def\hlx@prep#1{\hlx@loop#1\q@delim} |
| 4128 | % \end{macrocode} |
| 4129 | % |
| 4130 | % Now peel off a token, and dispatch using |\csname|. We handle |
| 4131 | % undefinedness of the command in a fairly messy way, although it probably |
| 4132 | % works. Maybe. |
| 4133 | % |
| 4134 | % \begin{macrocode} |
| 4135 | \def\hlx@loop#1{% |
| 4136 | \ifx#1\q@delim\else% |
| 4137 | \@ifundefined{hlx@cmd@\string#1}{% |
| 4138 | \expandafter\hlx@loop% |
| 4139 | }{% |
| 4140 | \csname hlx@cmd@\string#1\expandafter\endcsname% |
| 4141 | }% |
| 4142 | \fi% |
| 4143 | } |
| 4144 | % \end{macrocode} |
| 4145 | % |
| 4146 | % \end{macro} |
| 4147 | % |
| 4148 | % \begin{macro}{\hlxdef} |
| 4149 | % |
| 4150 | % New |\hlx| commands can be defined using |\hlxdef|. This is a simple |
| 4151 | % abbreviation. |
| 4152 | % |
| 4153 | % \begin{macrocode} |
| 4154 | \def\hlxdef#1{\@namedef{hlx@cmd@#1}} |
| 4155 | % \end{macrocode} |
| 4156 | % |
| 4157 | % \end{macro} |
| 4158 | % |
| 4159 | % \begin{macro}{\hlx h} |
| 4160 | % |
| 4161 | % Handle an \lit{h} character. Just do an |\hline| and return to the loop. |
| 4162 | % We look ahead to see if there's another \lit{h} coming up, and if so |
| 4163 | % insert two |\hline| commands. This strange (and inefficient) behaviour |
| 4164 | % keeps packages which redefine |\hline| happy. |
| 4165 | % |
| 4166 | % \begin{macrocode} |
| 4167 | \hlxdef h#1{% |
| 4168 | \noalign{% |
| 4169 | \ifx#1h\def\@tempa{\hline@prep\hline@prep\hlx@loop}% |
| 4170 | \else\def\@tempa{\hline@prep\hlx@loop#1}% |
| 4171 | \fi\expandafter |
| 4172 | }% |
| 4173 | \@tempa% |
| 4174 | } |
| 4175 | % \end{macrocode} |
| 4176 | % |
| 4177 | % \end{macro} |
| 4178 | % |
| 4179 | % \begin{macro}{\hlx b} |
| 4180 | % |
| 4181 | % The \lit{b} character does a nifty backspace, for \package{longtable}'s |
| 4182 | % benefit. |
| 4183 | % |
| 4184 | % \begin{macrocode} |
| 4185 | \hlxdef b{\noalign{\kern-\arrayrulewidth}\hlx@loop} |
| 4186 | % \end{macrocode} |
| 4187 | % |
| 4188 | % \end{macro} |
| 4189 | % |
| 4190 | % \begin{macro}{\hlx /} |
| 4191 | % |
| 4192 | % The `"/"' character allows a page break at the current position. |
| 4193 | % |
| 4194 | % \begin{macrocode} |
| 4195 | \hlxdef /{\noalign{\ifnum0=`}\fi\@testopt\hlx@cmd@break@i0} |
| 4196 | \def\hlx@cmd@break@i[#1]{\ifnum0=`{\fi}\pagebreak[#1]\hlx@loop} |
| 4197 | % \end{macrocode} |
| 4198 | % |
| 4199 | % \end{macro} |
| 4200 | % |
| 4201 | % \begin{macro}{\hlx v} |
| 4202 | % \begin{macro}{\hlx z} |
| 4203 | % |
| 4204 | % Handle a \lit{v} or \lit{z} character. This is rather like the |\vgap| |
| 4205 | % code above, although there are syntactic differences. |
| 4206 | % |
| 4207 | % \begin{macrocode} |
| 4208 | \hlxdef v{\hlx@vgap\doublerulesep} |
| 4209 | \hlxdef z{\hlx@vgap\tab@rulewidth} |
| 4210 | \def\hlx@vgap#1{% |
| 4211 | \noalign{\nobreak}% |
| 4212 | \omit% |
| 4213 | \iffalse{\fi\ifnum0=`}\fi% |
| 4214 | \global\let\vgap@after\hlx@loop% |
| 4215 | \@ifnextchar[{\hlx@vgap@i{#1}}{\hlx@vgap@ii\vgap@simple{#1}}% |
| 4216 | } |
| 4217 | \def\hlx@vgap@i#1[#2]{% |
| 4218 | \ifx!#2!\def\@tempa{\hlx@vgap@ii\vgap@simple{#1}}% |
| 4219 | \else\def\@tempa{\hlx@vgap@ii{\vgap@spec{#2}}{#1}}\fi% |
| 4220 | \@tempa% |
| 4221 | } |
| 4222 | \def\hlx@vgap@ii#1#2{\@testopt{\hlx@vgap@iii{#1}}{#2}} |
| 4223 | \def\hlx@vgap@iii#1[#2]{#1{#2}} |
| 4224 | % \end{macrocode} |
| 4225 | % |
| 4226 | % \end{macro} |
| 4227 | % \end{macro} |
| 4228 | % |
| 4229 | % \begin{macro}{\hlx s} |
| 4230 | % |
| 4231 | % Allow the user to leave a small gap using the \lit{s} command. |
| 4232 | % |
| 4233 | % \begin{macrocode} |
| 4234 | \hlxdef s{% |
| 4235 | \noalign{\ifnum0=`}\fi% |
| 4236 | \nobreak% |
| 4237 | \@testopt\hlx@space@i\doublerulesep% |
| 4238 | } |
| 4239 | \def\hlx@space@i[#1]{% |
| 4240 | \vskip#1% |
| 4241 | \tab@addruleheight{#1}% |
| 4242 | \ifnum0=`{\fi}% |
| 4243 | \hlx@loop% |
| 4244 | } |
| 4245 | % \end{macrocode} |
| 4246 | % |
| 4247 | % \end{macro} |
| 4248 | % |
| 4249 | % \begin{macro}{\hlx c} |
| 4250 | % |
| 4251 | % We might as well allow a \lit{c} command to do a |\cline|. The fix to |
| 4252 | % |\cline| permeates here. |
| 4253 | % |
| 4254 | % \begin{macrocode} |
| 4255 | \hlxdef c#1{\cline@do{#1}\hlx@loop} |
| 4256 | % \end{macrocode} |
| 4257 | % |
| 4258 | % \end{macro} |
| 4259 | % |
| 4260 | % \begin{macro}{\hlx ?} |
| 4261 | % |
| 4262 | % Do some arbitrary stuff which won't typeset. Put the stuff in a box which |
| 4263 | % is discarded, just in case. |
| 4264 | % |
| 4265 | % \begin{macrocode} |
| 4266 | \hlxdef ?#1{% |
| 4267 | \noalign{\setbox\z@\hbox{\color@begingroup#1\color@endgroup}}\hlx@loop% |
| 4268 | } |
| 4269 | % \end{macrocode} |
| 4270 | % |
| 4271 | % \end{macro} |
| 4272 | % |
| 4273 | % \begin{macro}{\hlx !} |
| 4274 | % |
| 4275 | % Change parameters in mid-flow. |
| 4276 | % |
| 4277 | % \begin{macrocode} |
| 4278 | \hlxdef !#1{\noalign\tab@ruleparams\hlx@loop[{#1}]} |
| 4279 | % \end{macrocode} |
| 4280 | % |
| 4281 | % \end{macro} |
| 4282 | % |
| 4283 | % \begin{macro}{\hlx .} |
| 4284 | % |
| 4285 | % The \lit{.} character forces a start of the new column. There's a little |
| 4286 | % problem here. Since the \lit{.} character starts the next column, we need |
| 4287 | % to gobble any spaces following the |\hlx| command before the cell contents |
| 4288 | % actually starts. Unfortunately, |\ignorespaces| will start the column for |
| 4289 | % us, so we can't put it in always. We'll handle it here, then. We'll take |
| 4290 | % the rest of the `preamble' string, and warn if it's not empty. Then we'll |
| 4291 | % |\ignorespaces| -- this will start the column for us, so we don't need to |
| 4292 | % |\relax| any more. |
| 4293 | % |
| 4294 | % \begin{macrocode} |
| 4295 | \hlxdef .#1\q@delim{% |
| 4296 | \ifx @#1@\else% |
| 4297 | \PackageWarning{mdwtab}{% |
| 4298 | Ignoring \protect\hlx\space command characters following a |
| 4299 | `.'\MessageBreak command% |
| 4300 | }% |
| 4301 | \fi% |
| 4302 | \ignorespaces% |
| 4303 | } |
| 4304 | % \end{macrocode} |
| 4305 | % |
| 4306 | % \end{macro} |
| 4307 | % |
| 4308 | % \begin{macro}{\hlx +} |
| 4309 | % \begin{macro}{\nextrow} |
| 4310 | % |
| 4311 | % The \lit{+} subcommand just steps the table-row counter. |
| 4312 | % |
| 4313 | % \begin{macrocode} |
| 4314 | \hlxdef +{\nextrow\hlx@loop} |
| 4315 | \def\nextrow{\noalign{\ifnum0=`}\fi\@testopt\nextrow@i\@ne} |
| 4316 | \def\nextrow@i[#1]{\global\advance\c@tabrow#1\ifnum0=`{\fi}} |
| 4317 | % \end{macrocode} |
| 4318 | % |
| 4319 | % \end{macro} |
| 4320 | % \end{macro} |
| 4321 | % |
| 4322 | % |
| 4323 | % \subsection{Starting new table rows} |
| 4324 | % |
| 4325 | % We take a break from careful mouthery at last, and start playing with |
| 4326 | % newlines. The standard one allows pagebreaks in unboxed tables, which |
| 4327 | % isn't really too desirable. |
| 4328 | % |
| 4329 | % Anyway, we'll try to make this macro rather more reusable than the standard |
| 4330 | % one. Here goes. |
| 4331 | % |
| 4332 | % \begin{macro}{\@arraycr} |
| 4333 | % |
| 4334 | % We pass lots of information to a main parser macro, and expect it to cope. |
| 4335 | % |
| 4336 | % \begin{macrocode} |
| 4337 | \def\@arraycr{\tab@arraycr{}} |
| 4338 | \def\tab@arraycr#1{\tab@cr{\tab@tabcr{#1}}{}{}} |
| 4339 | % \end{macrocode} |
| 4340 | % |
| 4341 | % Now to actually do the work. |\tab@cr| passes us the skip size, and the |
| 4342 | % appropriate one of the two arguments given above (both of which are empty) |
| 4343 | % depending on the presence of the $*$. |
| 4344 | % |
| 4345 | % \begin{macrocode} |
| 4346 | \def\tab@tabcr#1#2{% |
| 4347 | % \end{macrocode} |
| 4348 | % |
| 4349 | % If the total height I need to add between rows (from the optional argument |
| 4350 | % and the `extrasep' parameter) is greater than zero, I'll handle this by |
| 4351 | % extending the strut slightly. I'm not actually sure whether this is the |
| 4352 | % right thing to do, to be honest, although it's easier than trying to |
| 4353 | % to an automatic |\vgap|, because I need to know which columns to skip. |
| 4354 | % If the space is less than zero, I'll just insert the vertical space with |
| 4355 | % in a |\noalign|. |
| 4356 | % |
| 4357 | % First, to calculate how much space needs adding. |
| 4358 | % |
| 4359 | % \begin{macrocode} |
| 4360 | \setlength\dimen@{#2}% |
| 4361 | \advance\dimen@\tab@extrasep% |
| 4362 | % \end{macrocode} |
| 4363 | % |
| 4364 | % If the height is greater than zero, I need to play with the strut. I must |
| 4365 | % bear in mind that the current table cell (which I'm still in, remember) |
| 4366 | % may be in vertical mode, and I may or may not be in a paragraph. |
| 4367 | % |
| 4368 | % If I am in vertical mode, I'll backpedal to the previous box and put the |
| 4369 | % strut in an hbox superimposed on the previous baseline. Otherwise, I can |
| 4370 | % just put the strut at the end of the text. (This works in either LR |
| 4371 | % or paragraph mode as long as I'm not between paragraphs.) Again, Rowland's |
| 4372 | % empty cell bug strikes. (See |\tab@epar| for details.) |
| 4373 | % |
| 4374 | % \begin{macrocode} |
| 4375 | \ifdim\dimen@>\z@% |
| 4376 | \ifvmode% |
| 4377 | \unskip\ifdim\prevdepth>-\@m\p@\kern-\prevdepth\fi% |
| 4378 | \nointerlineskip\expandafter\hbox% |
| 4379 | \else% |
| 4380 | \@maybe@unskip\expandafter\@firstofone% |
| 4381 | \fi% |
| 4382 | {\advance\dimen@\dp\@arstrutbox\vrule\@depth\dimen@\@width\z@}% |
| 4383 | \fi% |
| 4384 | % \end{macrocode} |
| 4385 | % |
| 4386 | % This table cell works as a group (which is annoying here). I'll copy the |
| 4387 | % interrow gap into a global register so that I can use it in the |\noalign|. |
| 4388 | % |
| 4389 | % \begin{macrocode} |
| 4390 | \global\dimen\@ne\dimen@% |
| 4391 | \cr% |
| 4392 | \noalign{% |
| 4393 | #1% |
| 4394 | \ifdim\dimen\@ne<\z@\vskip\dimen\@ne\relax\fi% |
| 4395 | }% |
| 4396 | \@gobble% |
| 4397 | } |
| 4398 | % \end{macrocode} |
| 4399 | % |
| 4400 | % \end{macro} |
| 4401 | % |
| 4402 | % \begin{macro}{\tab@setcr} |
| 4403 | % |
| 4404 | % To set the |\\| command correctly in each table cell, we make it a part of |
| 4405 | % the preamble (in |\tab@midtext|) to call this routine. It's easy -- just |
| 4406 | % saves the preamble from being huge. |
| 4407 | % |
| 4408 | % \begin{macrocode} |
| 4409 | \def\tab@setcr{\let\\\tabularnewline} |
| 4410 | % \end{macrocode} |
| 4411 | % |
| 4412 | % \end{macro} |
| 4413 | % |
| 4414 | % \begin{macro}{\tab@cr} |
| 4415 | % |
| 4416 | % Now we do the parsing work. This is fun. Note the revenge of the funny |
| 4417 | % braces here. Nothing to worry about, honest. The tricky bit is to keep |
| 4418 | % track of which arguments are which. (Thanks to David Carlisle for pointing |
| 4419 | % out that I'd missed out the |\relax| here.) |
| 4420 | % |
| 4421 | % \begin{macrocode} |
| 4422 | \def\tab@cr#1#2#3{% |
| 4423 | \relax% |
| 4424 | \iffalse{\fi\ifnum0=`}\fi% |
| 4425 | \@ifstar{\tab@cr@i{#1}{#3}}{\tab@cr@i{#1}{#2}}% |
| 4426 | } |
| 4427 | \def\tab@cr@i#1#2{\@testopt{\tab@cr@ii{#1}{#2}}\z@} |
| 4428 | \def\tab@cr@ii#1#2[#3]{\ifnum0=`{}\fi#1{#3}{#2}} |
| 4429 | % \end{macrocode} |
| 4430 | % |
| 4431 | % \end{macro} |
| 4432 | % |
| 4433 | % |
| 4434 | % \subsection{Gratuitous grotesquery} |
| 4435 | % |
| 4436 | % So far we've had an easy-ish ride (or should that be \emph{queasy}?). Now |
| 4437 | % for something unexplainably evil. We convince \LaTeX\ that it's loaded the |
| 4438 | % \package{array} package, so that packages which need it think they've got |
| 4439 | % it. |
| 4440 | % |
| 4441 | % The bogus date is the same as the date for the \package{array} package I've |
| 4442 | % got here -- this will raise a warning if Frank updates his package which |
| 4443 | % should filter back to me telling me that there's something I need to |
| 4444 | % know about. |
| 4445 | % |
| 4446 | % The messing with |\xdef| and the funny parsing ought to insert the current |
| 4447 | % \package{mdwtab} version and date into the fake \package{array} version |
| 4448 | % string, giving a visible clue to the user that this isn't the real |
| 4449 | % \package{array} package. |
| 4450 | % |
| 4451 | % \begin{macrocode} |
| 4452 | \begingroup |
| 4453 | \catcode`.=11 |
| 4454 | \def\@tempa#1 #2 #3\@@{#1 #2} |
| 4455 | \xdef\ver@array.sty |
| 4456 | {1995/11/19 [mdwtab.sty \expandafter\@tempa\ver@mdwtab.sty\@@]} |
| 4457 | \endgroup |
| 4458 | % \end{macrocode} |
| 4459 | % |
| 4460 | % |
| 4461 | % \subsection{Error messages} |
| 4462 | % |
| 4463 | % I've put all the error messages together, where I can find them, translate |
| 4464 | % them or whatever. |
| 4465 | % |
| 4466 | % First, some token-space saving (which also saves my fingers): |
| 4467 | % |
| 4468 | % \begin{macrocode} |
| 4469 | \def\tab@error{\PackageError{mdwtab}} |
| 4470 | % \end{macrocode} |
| 4471 | % |
| 4472 | % Now do the error messages. |
| 4473 | % |
| 4474 | % \begin{macrocode} |
| 4475 | \def\tab@err@misscol{% |
| 4476 | \tab@error{Missing column type}{% |
| 4477 | I'm lost. I was expecting something describing^^J% |
| 4478 | the type of the current column, but you seem to^^J% |
| 4479 | have missed it out. I've inserted a type `l'^^J% |
| 4480 | column here in the hope that this makes sense.% |
| 4481 | }% |
| 4482 | } |
| 4483 | % \end{macrocode} |
| 4484 | % |
| 4485 | % \begin{macrocode} |
| 4486 | \def\tab@err@oddgroup{% |
| 4487 | \tab@error{Misplaced group in table preamble}{% |
| 4488 | I've found an open brace character in your preamble^^J% |
| 4489 | when I was expecting a specifier character. I'm^^J% |
| 4490 | going to gobble the whole group and carry on as if^^J% |
| 4491 | I'd never seen it.% |
| 4492 | }% |
| 4493 | } |
| 4494 | % \end{macrocode} |
| 4495 | % |
| 4496 | % \begin{macrocode} |
| 4497 | \def\tab@err@undef#1{% |
| 4498 | \tab@error{Unknown `\tab@colset' preamble character `\string#1'}{% |
| 4499 | I don't understand what you meant by typing this^^J% |
| 4500 | character. Anyway, I'll ignore it this time around.^^J% |
| 4501 | Just don't you do it again.% |
| 4502 | }% |
| 4503 | } |
| 4504 | % \end{macrocode} |
| 4505 | % |
| 4506 | % \begin{macrocode} |
| 4507 | \def\tab@err@unbrh{% |
| 4508 | \tab@error{Can't use unboxed tabular in LR mode}{% |
| 4509 | You've asked for a tabular or array environment with^^J% |
| 4510 | `L', `C' or `R' as the position specifier, but you're^^J% |
| 4511 | in LR (restricted horizontal) mode, so it won't work.^^J% |
| 4512 | I'll assume you really meant `c' and soldier on.% |
| 4513 | }% |
| 4514 | } |
| 4515 | % \end{macrocode} |
| 4516 | % |
| 4517 | % \begin{macrocode} |
| 4518 | \def\tab@err@unbmm{% |
| 4519 | \tab@error{Can't use unboxed tabular in maths mode}{% |
| 4520 | You've asked for a tabular or array environment with^^J% |
| 4521 | `L', `C' or `R' as the position specifier, but you're^^J% |
| 4522 | in maths mode, so it won't work. I'll pretend that^^J% |
| 4523 | you really typed `c', and that this is all a bad dream.% |
| 4524 | }% |
| 4525 | } |
| 4526 | % \end{macrocode} |
| 4527 | % |
| 4528 | % \begin{macrocode} |
| 4529 | \def\tab@err@unbext{% |
| 4530 | \tab@error{Can't extend unboxed tabulars}{% |
| 4531 | You're trying to use kludgy extensions (e.g.,^^J% |
| 4532 | `delarray') on an array or tabular with `L', `C'^^J% |
| 4533 | or `R' as the position specifier. I'll assume you^^J% |
| 4534 | subconsciously wanted a `c' type all along.% |
| 4535 | }% |
| 4536 | } |
| 4537 | % \end{macrocode} |
| 4538 | % |
| 4539 | % \begin{macrocode} |
| 4540 | \def\tab@err@multi{% |
| 4541 | \tab@error{More than one column in a \protect\multicolumn}{% |
| 4542 | You've put more than one column into a \string\multicolumn^^J% |
| 4543 | descriptor. It won't work. I have no idea what^^J% |
| 4544 | will happen, although it won't be pleasant. Hold^^J% |
| 4545 | on tight now...% |
| 4546 | }% |
| 4547 | } |
| 4548 | % \end{macrocode} |
| 4549 | % |
| 4550 | % \begin{macrocode} |
| 4551 | \def\tab@err@range{% |
| 4552 | \tab@error{Expected `,' or `<end>' in range list}{% |
| 4553 | I was expecting either the end of the range list,^^J% |
| 4554 | or a comma, followed by another range. I've^^J% |
| 4555 | inserted a comma to try and get me back on track.^^J% |
| 4556 | Good luck.% |
| 4557 | }% |
| 4558 | } |
| 4559 | % \end{macrocode} |
| 4560 | % |
| 4561 | % |
| 4562 | % \subsection{Loading the colour package} |
| 4563 | % |
| 4564 | % If requested, we load the \package{mtcolour} package here. This ensures |
| 4565 | % that it can patch this code if it needs to. |
| 4566 | % |
| 4567 | % \begin{macrocode} |
| 4568 | \iftab@colour |
| 4569 | \RequirePackage{mtcolour} |
| 4570 | \fi |
| 4571 | % \end{macrocode} |
| 4572 | % |
| 4573 | % That's it. No more. Move along please. |
| 4574 | % |
| 4575 | % \begin{macrocode} |
| 4576 | %</mdwtab> |
| 4577 | % \end{macrocode} |
| 4578 | % |
| 4579 | % |
| 4580 | %^^A------------------------------------------------------------------------- |
| 4581 | % \section{Implementation of \package{mtcolour}} |
| 4582 | % |
| 4583 | % |
| 4584 | % This is in a separate package to avoid dragging in the \package{color} |
| 4585 | % package if it's unwanted. |
| 4586 | % |
| 4587 | % I prefer English spellings. Here's a trivial redirection for Americans. |
| 4588 | % |
| 4589 | % \begin{macrocode} |
| 4590 | %<*color> |
| 4591 | \DeclareOption*{\PassOptionsToPackage{\CurrentOption}{mtcolour}} |
| 4592 | \ProcessOptions |
| 4593 | \RequirePackage{mtcolour} |
| 4594 | %</color> |
| 4595 | % \end{macrocode} |
| 4596 | % |
| 4597 | % And now we can start the thing properly. |
| 4598 | % |
| 4599 | % \begin{macrocode} |
| 4600 | %<*colour> |
| 4601 | \RequirePackage{color} |
| 4602 | % \end{macrocode} |
| 4603 | % |
| 4604 | % |
| 4605 | % \subsection{Cell background colours} |
| 4606 | % |
| 4607 | % First, some simple preliminaries. The |\iftab@colour| switch is set if the |
| 4608 | % current cell is meant to have a colour. |
| 4609 | % |
| 4610 | % \begin{macrocode} |
| 4611 | \newif\iftab@colour |
| 4612 | \tab@colourfalse |
| 4613 | % \end{macrocode} |
| 4614 | % |
| 4615 | % We shall store the cell colour information in |\tab@cellcolour|, and the |
| 4616 | % row colour information as |\tab@rowcolour|. Because of the structure of |
| 4617 | % tables, we need to make global assignments; so we must copy the current |
| 4618 | % value away at the start of a table and put the value back at the end. In |
| 4619 | % order to transfer the overhang information reliably, we use a separate |
| 4620 | % control sequence |\tab@colouroverhangs| for that -- otherwise |\color| can |
| 4621 | % corrupt it. |
| 4622 | % |
| 4623 | % \begin{macrocode} |
| 4624 | \tab@addhookbefore\tab@beginhook{% |
| 4625 | \let\tab@saverowcolour\tab@rowcolour% |
| 4626 | \let\tab@savecolouroverhangs\tab@colouroverhangs% |
| 4627 | \let\tab@savecellcolour\tab@cellcolour% |
| 4628 | } |
| 4629 | \tab@addhookafter\tab@endhook{% |
| 4630 | \global\let\tab@rowcolour\tab@saverowcolour% |
| 4631 | \global\let\tab@colouroverhangs\tab@savecolouroverhangs% |
| 4632 | \global\let\tab@cellcolour\tab@savecellcolour% |
| 4633 | } |
| 4634 | % \end{macrocode} |
| 4635 | % |
| 4636 | % Initially, there are no colours. |
| 4637 | % |
| 4638 | % \begin{macrocode} |
| 4639 | \let\tab@rowcolour\@empty% |
| 4640 | \let\tab@cellcolour\@empty% |
| 4641 | \let\tab@colouroverhangs\@empty% |
| 4642 | % \end{macrocode} |
| 4643 | % |
| 4644 | % \begin{macro}{\@snarfcolour} |
| 4645 | % |
| 4646 | % Reading a colour specification is something we'll need to do a few times, |
| 4647 | % so an abstraction is useful. Its single argument is a continuation to |
| 4648 | % which we pass a colour-spec acceptable to the |\color| command. (This is |
| 4649 | % the same code as found in the \package{sverb} package. Remember to keep |
| 4650 | % them in step.) |
| 4651 | % |
| 4652 | % \begin{macrocode} |
| 4653 | \def\@snarfcolour#1{% |
| 4654 | \@ifnextchar[{\@snarfcolour@i{#1}}{\@snarfcolour@ii{#1}{}}% |
| 4655 | } |
| 4656 | \def\@snarfcolour@i#1[#2]{\@snarfcolour@ii{#1}{[#2]}} |
| 4657 | \def\@snarfcolour@ii#1#2#3{#1{#2{#3}}} |
| 4658 | % \end{macrocode} |
| 4659 | % |
| 4660 | % \end{macro} |
| 4661 | % |
| 4662 | % \begin{macro}{\cellcolour} |
| 4663 | % |
| 4664 | % Setting a cell colour is a matter of stashing the right declarations in |
| 4665 | % |\tab@cellcolour| and |\tab@colouroverhangs|. Note that the overhangs end |
| 4666 | % up in |\dimen0| and |\dimen2|. |
| 4667 | % |
| 4668 | % \begin{macrocode} |
| 4669 | \def\cellcolour{% |
| 4670 | \@ifstar{\tab@ccol@i{\let\tab@rowcolour\@empty}}{\tab@ccol@i{}}% |
| 4671 | } |
| 4672 | \def\tab@ccol@i#1{\@snarfcolour{\tab@ccol@ii{#1}}} |
| 4673 | \def\tab@ccol@ii#1#2{\@testopt{\tab@ccol@iii{#2#1}}\z@} |
| 4674 | \def\tab@ccol@iii#1[#2]{\@testopt{\tab@ccol@iv{#1}{#2}}{#2}} |
| 4675 | \def\tab@ccol@iv#1#2[#3]{% |
| 4676 | \gdef\tab@cellcolour{\color#1\tab@colourtrue}% |
| 4677 | \gdef\tab@colouroverhangs{% |
| 4678 | \setlength\dimen@{#2}% |
| 4679 | \setlength{\dimen\tw@}{#3}% |
| 4680 | }% |
| 4681 | } |
| 4682 | % \end{macrocode} |
| 4683 | % |
| 4684 | % \end{macro} |
| 4685 | % |
| 4686 | % \begin{macro}{\rowcolour} |
| 4687 | % |
| 4688 | % Setting the global row colour is simpler, because we don't mess with |
| 4689 | % overhangs. |
| 4690 | % |
| 4691 | % \begin{macrocode} |
| 4692 | \def\rowcolour{\@snarfcolour\tab@setrowcolour} |
| 4693 | \let\rowcolor\rowcolour |
| 4694 | \def\tab@setrowcolour#1{% |
| 4695 | \gdef\tab@rowcolour{\color#1\tab@colourtrue}% |
| 4696 | } |
| 4697 | % \end{macrocode} |
| 4698 | % |
| 4699 | % \end{macro} |
| 4700 | % |
| 4701 | % \begin{macro}{\rowcolouroff} |
| 4702 | % |
| 4703 | % And turning the global colouring off is easy. |
| 4704 | % |
| 4705 | % \begin{macrocode} |
| 4706 | \def\rowcolouroff{\global\let\tab@rowcolour\@empty} |
| 4707 | \let\rowcoloroff\rowcolouroff |
| 4708 | % \end{macrocode} |
| 4709 | % |
| 4710 | % \end{macro} |
| 4711 | % |
| 4712 | % \begin{macro}{\tab@colourleft} |
| 4713 | % |
| 4714 | % Now we start on the table-cell hooks. The left hook starts a box which |
| 4715 | % will capture the cell's text and natural width. We add the hook to the |
| 4716 | % rule list as well, so that we can colour the bits in |\vgap|s correctly. |
| 4717 | % |
| 4718 | % \begin{macrocode} |
| 4719 | \tab@addhookbefore\tab@lefttexthook\tab@colourleft |
| 4720 | \tab@addhookbefore\tab@leftruletexthook\tab@colourleft |
| 4721 | \def\tab@colourleft{% |
| 4722 | \global\let\tab@cellcolour\@empty% |
| 4723 | \global\let\tab@colouroverhangs\@empty% |
| 4724 | \setbox\z@\hbox\bgroup\color@begingroup% |
| 4725 | } |
| 4726 | % \end{macrocode} |
| 4727 | % |
| 4728 | % \end{macro} |
| 4729 | % |
| 4730 | % \begin{macro}{\tab@colourright} |
| 4731 | % |
| 4732 | % The right hook will insert an appropriate rule behind the cell and |
| 4733 | % retypeset the cell contents over the top. Note that the stretch in a table |
| 4734 | % cell is exactly 1\,fil. Because we add (leaders) and subtract (negative |
| 4735 | % |\hskip|) 1\,fil, we retain this stretch exactly. Don't bother unless |
| 4736 | % there's actually some colouring. |
| 4737 | % |
| 4738 | % \begin{macrocode} |
| 4739 | \tab@addhookafter\tab@righttexthook\tab@colourright |
| 4740 | \tab@addhookafter\tab@rightruletexthook\tab@colourright |
| 4741 | \def\tab@colourright{% |
| 4742 | \color@endgroup\egroup% |
| 4743 | \color@begingroup% |
| 4744 | \global\tab@colourfalse% |
| 4745 | \tab@cellcolour\tab@rowcolour% |
| 4746 | \dimen@\z@\dimen\tw@\z@\tab@colouroverhangs% |
| 4747 | \iftab@colour% |
| 4748 | \skip@\wd\z@\advance\skip@\z@\@plus1fil% |
| 4749 | \skip\tw@\skip@% |
| 4750 | \kern-\dimen@% |
| 4751 | \advance\skip\tw@\dimen@% |
| 4752 | \advance\skip\tw@\dimen\tw@% |
| 4753 | \leaders\vrule\hskip\skip\tw@% |
| 4754 | \kern-\dimen\tw@% |
| 4755 | \hskip-\skip@% |
| 4756 | \fi% |
| 4757 | \color@endgroup% |
| 4758 | \unhbox\z@% |
| 4759 | } |
| 4760 | % \end{macrocode} |
| 4761 | % |
| 4762 | % \end{macro} |
| 4763 | % |
| 4764 | % |
| 4765 | % \subsection{Coloured rules} |
| 4766 | % |
| 4767 | % We hook ourselves onto the rule-parameters edifice. This is rather |
| 4768 | % straightforward. |
| 4769 | % |
| 4770 | % \begin{macrocode} |
| 4771 | \tab@addhookafter\tab@rp@inithook{% |
| 4772 | \let\tab@rulecolour\@empty% |
| 4773 | \let\tab@rulecolourmodel\@empty% |
| 4774 | } |
| 4775 | \mkdef{mdwtab:rule}{colour}{\tab@setrulecolour{#1}} |
| 4776 | \mkdef{mdwtab:rule}{colourmodel}{\tab@setrulecolourmodel{#1}} |
| 4777 | \mkdef{mdwtab:rule}{color}{\tab@setrulecolour{#1}} |
| 4778 | \mkdef{mdwtab:rule}{colormodel}{\tab@setrulecolourmodel{#1}} |
| 4779 | \mkdef{mdwtab:rule}{nocolour}*{\let\tab@rulecolour\@empty} |
| 4780 | \mkdef{mdwtab:rule}{nocolor}*{\let\tab@rulecolour\@empty} |
| 4781 | \mkdef{mdwtab:rule}{nocolourmodel}*{\let\tab@rulecolourmodel\@empty} |
| 4782 | \mkdef{mdwtab:rule}{nocolormodel}*{\let\tab@rulecolourmodel\@empty} |
| 4783 | \def\tab@setrulecolour#1{% |
| 4784 | \def\tab@rulecolour{\color\tab@rulecolourmodel{#1}}% |
| 4785 | } |
| 4786 | \def\tab@setrulecolourmodel#1{\def\tab@rulecolourmodel{[#1]}} |
| 4787 | \tab@addhookafter\tab@rp@sethook{\tab@rulecolour} |
| 4788 | % \end{macrocode} |
| 4789 | % |
| 4790 | % |
| 4791 | % \subsection{Other stuff} |
| 4792 | % |
| 4793 | % \begin{macro}{\ifmod} |
| 4794 | % |
| 4795 | % \syntax{"\\ifmod{"$x$"}{"$m$"}{"y"}{"<yes>"}{"<no>"}"} -- if $x \bmod m = |
| 4796 | % y$ then do \<yes>; otherwise do \<no>. |
| 4797 | % |
| 4798 | % \begin{macrocode} |
| 4799 | \def\ifmod#1#2#3{% |
| 4800 | \begingroup% |
| 4801 | \@tempcnta#1% |
| 4802 | \@tempcntb#2% |
| 4803 | \count@\@tempcnta% |
| 4804 | \divide\count@\@tempcntb% |
| 4805 | \multiply\count@\@tempcntb% |
| 4806 | \advance\@tempcnta-\count@% |
| 4807 | \count@#3\relax% |
| 4808 | \ifnum\@tempcnta=\count@\endgroup\expandafter\@firstoftwo% |
| 4809 | \else\endgroup\expandafter\@secondoftwo\fi% |
| 4810 | } |
| 4811 | % \end{macrocode} |
| 4812 | % |
| 4813 | % \end{macro} |
| 4814 | % |
| 4815 | % Done. |
| 4816 | % |
| 4817 | % \begin{macrocode} |
| 4818 | %</colour> |
| 4819 | % \end{macrocode} |
| 4820 | % |
| 4821 | %^^A------------------------------------------------------------------------- |
| 4822 | % \section{Implementation of \package{mathenv}} |
| 4823 | % |
| 4824 | % |
| 4825 | % This is in a separate package, mainly to avoid wasting people's memory. |
| 4826 | % |
| 4827 | % \begin{macrocode} |
| 4828 | %<*mathenv> |
| 4829 | % \end{macrocode} |
| 4830 | % |
| 4831 | % |
| 4832 | % \subsection{Options handling} |
| 4833 | % |
| 4834 | % We need to be able to cope with \textsf{fleqn} and \textsf{leqno} options. |
| 4835 | % This will adjust our magic modified \env{eqnarray} environment |
| 4836 | % appropriately. |
| 4837 | % |
| 4838 | % \begin{macrocode} |
| 4839 | \newif\if@fleqn |
| 4840 | \newif\if@leqno |
| 4841 | \DeclareOption{fleqn}{\@fleqntrue} |
| 4842 | \DeclareOption{leqno}{\@leqnotrue} |
| 4843 | \ProcessOptions |
| 4844 | % \end{macrocode} |
| 4845 | % |
| 4846 | % We use the \package{mdwtab} package for all its nice table handling things. |
| 4847 | % (Oh, and to inflict it on users who want to do nice equations and don't |
| 4848 | % care about our tables.) |
| 4849 | % |
| 4850 | % \begin{macrocode} |
| 4851 | \RequirePackage{mdwtab} |
| 4852 | % \end{macrocode} |
| 4853 | % |
| 4854 | % |
| 4855 | % \subsection{Some useful registers} |
| 4856 | % |
| 4857 | % The old \LaTeX\ version puts the equation numbers in by keeping a count of |
| 4858 | % where it is in the alignment. Since I don't know how may columns there are |
| 4859 | % going to be, I'll just use a switch in the preamble to tell me to stop |
| 4860 | % tabbing. |
| 4861 | % |
| 4862 | % \begin{macrocode} |
| 4863 | \newif\if@eqalast |
| 4864 | % \end{macrocode} |
| 4865 | % |
| 4866 | % Now define some useful length parameters. First allocate them: |
| 4867 | % |
| 4868 | % \begin{macrocode} |
| 4869 | \newskip\eqaopenskip |
| 4870 | \newskip\eqacloseskip |
| 4871 | \newskip\eqacolskip |
| 4872 | \newskip\eqainskip |
| 4873 | \newskip\splitleft |
| 4874 | \newskip\splitright |
| 4875 | % \end{macrocode} |
| 4876 | % |
| 4877 | % Now assign some default values. Users can play with these if they really |
| 4878 | % want although I can't see the point myself. |
| 4879 | % |
| 4880 | % \begin{macrocode} |
| 4881 | \AtBeginDocument{% |
| 4882 | \eqacloseskip\@centering% |
| 4883 | \eqacolskip1.5em\@plus\@m\p@ |
| 4884 | \eqainskip\z@% |
| 4885 | \if@fleqn% |
| 4886 | \eqaopenskip\mathindent% |
| 4887 | \splitleft\mathindent\relax% |
| 4888 | \splitright\mathindent\@minus\mathindent\relax% |
| 4889 | \else% |
| 4890 | \eqaopenskip\@centering% |
| 4891 | \splitleft2.5em\@minus2.5em% |
| 4892 | \splitright\splitleft% |
| 4893 | \fi% |
| 4894 | \relax% |
| 4895 | } |
| 4896 | % \end{macrocode} |
| 4897 | % |
| 4898 | % |
| 4899 | % \subsection{A little display handling} |
| 4900 | % |
| 4901 | % I'm probably going a little far here, and invading territory already |
| 4902 | % claimed by the \package{amsmath} stuff (and done a good deal better than |
| 4903 | % I can be bothered to do), but just for completeness, this is how we handle |
| 4904 | % attempts to put displays inside other displays without screwing up the |
| 4905 | % spacing. |
| 4906 | % |
| 4907 | % \begin{macro}{\dsp@startouter} |
| 4908 | % |
| 4909 | % This is how we start an outermost display. It's fairly easy really. We |
| 4910 | % make |\dsp@start| start an inner display, and make |\dsp@end| close the |
| 4911 | % outer display. |
| 4912 | % |
| 4913 | % \begin{macrocode} |
| 4914 | \def\dsp@startouter{% |
| 4915 | \let\dsp@end\dsp@endouter% |
| 4916 | $$% |
| 4917 | } |
| 4918 | % \end{macrocode} |
| 4919 | % |
| 4920 | % \end{macro} |
| 4921 | % |
| 4922 | % \begin{macro}{\dsp@endouter} |
| 4923 | % |
| 4924 | % Ending the outer display is utterly trivial. |
| 4925 | % |
| 4926 | % \begin{macrocode} |
| 4927 | \def\dsp@endouter{$$} |
| 4928 | % \end{macrocode} |
| 4929 | % |
| 4930 | % \end{macro} |
| 4931 | % |
| 4932 | % \begin{macro}{\dsp@startinner} |
| 4933 | % |
| 4934 | % Starting inner displays is done in a vbox (actually I choose |\vbox| or |
| 4935 | % |\vtop| depending on the setting of \textsf{leqno} to put the equation |
| 4936 | % number the right way round). |
| 4937 | % |
| 4938 | % \begin{macrocode} |
| 4939 | \def\dsp@startinner{% |
| 4940 | \let\dsp@end\dsp@endinner% |
| 4941 | \if@fleqn\kern-\mathindent\fi% |
| 4942 | \if@leqno\vtop\else\vtop\fi\bgroup% |
| 4943 | } |
| 4944 | % \end{macrocode} |
| 4945 | % |
| 4946 | % \end{macro} |
| 4947 | % |
| 4948 | % \begin{macro}{\dsp@endinner} |
| 4949 | % |
| 4950 | % Ending an inner display is also really easy. |
| 4951 | % |
| 4952 | % \begin{macrocode} |
| 4953 | \def\dsp@endinner{\egroup} |
| 4954 | % \end{macrocode} |
| 4955 | % |
| 4956 | % \end{macro} |
| 4957 | % |
| 4958 | % \begin{macro}{\dsp@start} |
| 4959 | % |
| 4960 | % This is what other bits of code uses to start displays. It's one of the |
| 4961 | % start macros up above, and outer by default. |
| 4962 | % |
| 4963 | % \begin{macrocode} |
| 4964 | \def\dsp@start{% |
| 4965 | \ifmmode% |
| 4966 | %\ifinner\mth@err@mdsp\fi% |
| 4967 | \expandafter\dsp@startinner% |
| 4968 | \else% |
| 4969 | \ifhmode\ifinner\mth@err@hdsp\fi\fi% |
| 4970 | \expandafter\dsp@startouter% |
| 4971 | \fi% |
| 4972 | } |
| 4973 | % \end{macrocode} |
| 4974 | % |
| 4975 | % \end{macro} |
| 4976 | % |
| 4977 | % \begin{macro}{\dsp@tabpause} |
| 4978 | % |
| 4979 | % This sets up the correct pre- and postambles for the |\tabpause| macro in |
| 4980 | % maths displays. This is fairly simple stuff. |
| 4981 | % |
| 4982 | % \begin{macrocode} |
| 4983 | \def\dsp@tabpause{% |
| 4984 | \def\tab@startpause% |
| 4985 | {\penalty\postdisplaypenalty\vskip\belowdisplayskip}% |
| 4986 | \def\tab@endpause% |
| 4987 | {\penalty\predisplaypenalty\vskip\abovedisplayskip}% |
| 4988 | } |
| 4989 | % \end{macrocode} |
| 4990 | % |
| 4991 | % \end{macro} |
| 4992 | % |
| 4993 | % |
| 4994 | % \subsection{The \env{eqnarray} environment} |
| 4995 | % |
| 4996 | % We allow the user to play with the style if this is really wanted. I dunno |
| 4997 | % why, really. Maybe someone wants very small alignments. |
| 4998 | % |
| 4999 | % \begin{macrocode} |
| 5000 | \let\eqastyle\displaystyle |
| 5001 | % \end{macrocode} |
| 5002 | % |
| 5003 | % \subsubsection{The main environments} |
| 5004 | % |
| 5005 | % \begin{environment}{eqnarray} |
| 5006 | % \begin{environment}{eqnarray*} |
| 5007 | % |
| 5008 | % We define the toplevel commands here. They just add in default arguments |
| 5009 | % and then call |\@eqnarray| with a preamble string. We handle equation |
| 5010 | % numbers by setting up a default (|\eqa@defnumber|) which is put into |
| 5011 | % the final column. At the beginning of each row, we globally |\let| |
| 5012 | % |\eqa@number| equal to |\eqa@defnumber|. The |\eqnumber| macro just |
| 5013 | % changes |\eqa@number| as required. Since |\eqa@number| is changed globally |
| 5014 | % we must save it in this environment. |
| 5015 | % |
| 5016 | % First, we must sort out the optional arguments and things. This is really |
| 5017 | % easy. The only difference between the starred and non-starred environments |
| 5018 | % is the default definition of |\eqa@defnumber|. |
| 5019 | % |
| 5020 | % \begin{macrocode} |
| 5021 | \def\eqnarray{% |
| 5022 | \eqnarray@i\eqa@eqcount% |
| 5023 | } |
| 5024 | \@namedef{eqnarray*}{\eqnarray@i{}} |
| 5025 | \def\eqnarray@i#1{\@testopt{\eqnarray@ii{#1}}{rcl}} |
| 5026 | % \end{macrocode} |
| 5027 | % |
| 5028 | % Right. Now for the real work. The first argument is the default numbering |
| 5029 | % tokens; the second is the preamble string. |
| 5030 | % |
| 5031 | % \begin{macrocode} |
| 5032 | \def\eqnarray@ii#1[#2]{% |
| 5033 | % \end{macrocode} |
| 5034 | % |
| 5035 | % Set up the equation counter and labels correctly. |
| 5036 | % |
| 5037 | % \medskip\par\noindent|\begin{rant}|\par |
| 5038 | % The hacking with |\@currentlabel| is here because (in the author's opinion) |
| 5039 | % \LaTeX's |\refstepcounter| macro is broken. It's currently defined as |
| 5040 | % \begin{listing} |
| 5041 | %\def\refstepcounter#1{% |
| 5042 | % \stepcounter{#1}% |
| 5043 | % \protected@edef\@currentlabel% |
| 5044 | % {\csname p@#1\endcsname\csname the#1\endcsname}% |
| 5045 | %} |
| 5046 | % \end{listing} |
| 5047 | % which means that the current label gets `frozen' as soon as you do the |
| 5048 | % counter step. By redefining the macro as |
| 5049 | % \begin{listing} |
| 5050 | %\def\refstepcounter#1{% |
| 5051 | % \stepcounter{#1}% |
| 5052 | % \edef\@currentlabel{% |
| 5053 | % \expandafter\noexpand\csname p@#1\endcsname% |
| 5054 | % \expandafter\noexpand\csname the#1\endcsname% |
| 5055 | % }% |
| 5056 | %} |
| 5057 | % \end{listing} |
| 5058 | % these sorts of problems would be avoided, without any loss of functionality |
| 5059 | % or compatibility that I can see. |
| 5060 | % \par\noindent|\end{rant}|\par |
| 5061 | % |
| 5062 | % \begin{macrocode} |
| 5063 | \stepcounter{equation}% |
| 5064 | \def\@currentlabel{\p@equation\theequation}% |
| 5065 | % \end{macrocode} |
| 5066 | % |
| 5067 | % The next step is to set up the numbering. I must save the old numbering |
| 5068 | % so I can restore it later (once in the alignment, I must assign these |
| 5069 | % things globally). |
| 5070 | % |
| 5071 | % \begin{macrocode} |
| 5072 | \let\eqa@oldnumber\eqa@number% |
| 5073 | \def\eqa@defnumber{#1}% |
| 5074 | \global\let\eqa@number\eqa@defnumber% |
| 5075 | % \end{macrocode} |
| 5076 | % |
| 5077 | % The |\if@eqalastfalse| switch is false everywhere except when we're in the |
| 5078 | % final column. |
| 5079 | % |
| 5080 | % \begin{macrocode} |
| 5081 | \@eqalastfalse% |
| 5082 | % \end{macrocode} |
| 5083 | % |
| 5084 | % Remove the |\mathsurround| kerning, since it will look very odd inside |
| 5085 | % the display. We have our own spacing parameters for configuring these |
| 5086 | % things, so |\mathsurround| is unnecessary. |
| 5087 | % |
| 5088 | % \begin{macrocode} |
| 5089 | \m@th% |
| 5090 | % \end{macrocode} |
| 5091 | % |
| 5092 | % Time to parse the preamble string now. I must choose the correct column |
| 5093 | % set, initialise the preamble parser and set up the various macros. The% |
| 5094 | % extra `|@{\tabskip\eqacloseskip}|' item sets up the tabskip glue to centre |
| 5095 | % the alignment properly. |
| 5096 | % |
| 5097 | % \begin{macrocode} |
| 5098 | \colset{eqnarray}% |
| 5099 | \tab@initread% |
| 5100 | \def\tab@tabtext{&\tabskip\z@skip}% |
| 5101 | \tab@preamble{\tabskip\z@skip}% |
| 5102 | \tab@readpreamble{#2@{\tabskip\eqacloseskip}}% |
| 5103 | \dsp@tabpause% |
| 5104 | % \end{macrocode} |
| 5105 | % |
| 5106 | % Now for some final setting up. The column separation is set from the |
| 5107 | % user's parameter, the |\everycr| tokens are cleared, and I set up the |
| 5108 | % newline command appropriately. |
| 5109 | % |
| 5110 | % \begin{macrocode} |
| 5111 | \col@sep.5\eqainskip% |
| 5112 | \everycr{}% |
| 5113 | \let\\\@eqncr% |
| 5114 | % \end{macrocode} |
| 5115 | % |
| 5116 | % Now start a maths display and do the alignment. Set up the left hand |
| 5117 | % tabskip glue to centre the alignment, and do the actual alignment. |
| 5118 | % The preamble used is mainly that generated from the user's string, although |
| 5119 | % the stuff at the end is how we set up the equation number -- it repeats |
| 5120 | % appropriately so we can always find it. |
| 5121 | % |
| 5122 | % \begin{macrocode} |
| 5123 | \dsp@start% |
| 5124 | \tabskip\eqaopenskip% |
| 5125 | \halign to\displaywidth\expandafter\bgroup% |
| 5126 | \the\tab@preamble% |
| 5127 | &&\eqa@lastcol\hb@xt@\z@{\hss##}\tabskip\z@\cr% |
| 5128 | } |
| 5129 | % \end{macrocode} |
| 5130 | % |
| 5131 | % Now for the end of the environment. This is really easy. Set the final |
| 5132 | % equation number, close the |\halign|, tidy up the equation counter (it's |
| 5133 | % been stepped once too many times) and close the display. |
| 5134 | % |
| 5135 | % \begin{macrocode} |
| 5136 | \def\endeqnarray{% |
| 5137 | \eqa@eqnum% |
| 5138 | \egroup% |
| 5139 | \dsp@end% |
| 5140 | \global\let\eqa@number\eqa@oldnumber% |
| 5141 | \global\@ignoretrue% |
| 5142 | \global\advance\c@equation\m@ne% |
| 5143 | } |
| 5144 | \expandafter\let\csname endeqnarray*\endcsname\endeqnarray |
| 5145 | % \end{macrocode} |
| 5146 | % |
| 5147 | % \end{environment} |
| 5148 | % \end{environment} |
| 5149 | % |
| 5150 | % Now we can define the column types. |
| 5151 | % |
| 5152 | % \begin{macrocode} |
| 5153 | \colpush{eqnarray} |
| 5154 | % \end{macrocode} |
| 5155 | % |
| 5156 | % Note the positioning of ord atoms in the stuff below. This will space out |
| 5157 | % relations and binops correctly when they occur at the edges of columns, and |
| 5158 | % won't affect ord atoms at the edges, because ords pack closely. |
| 5159 | % |
| 5160 | % First the easy ones. Just stick |\hfil| in the right places and |
| 5161 | % everything will be all right. |
| 5162 | % |
| 5163 | % \begin{macrocode} |
| 5164 | \coldef r{\tabcoltype{\hfil$\eqastyle}{{}$}} |
| 5165 | \coldef c{\tabcoltype{\hfil$\eqastyle{}}{{}$\hfil}} |
| 5166 | \coldef l{\tabcoltype{$\eqastyle{}}{$\hfil}} |
| 5167 | \coldef x{\tabcoltype{\if@fleqn\else\hfil\fi$\eqastyle}{$\hfil}} |
| 5168 | % \end{macrocode} |
| 5169 | % |
| 5170 | % Now for the textual ones. This is also fairly easy. |
| 5171 | % |
| 5172 | % \begin{macrocode} |
| 5173 | \collet T [tabular]T |
| 5174 | % \end{macrocode} |
| 5175 | % |
| 5176 | % Sort of split types of equations. I mustn't use |\rlap| here, or |
| 5177 | % everything goes wrong -- |\\| doesn't get noticed by \TeX\ in the same way |
| 5178 | % as |\cr| does. |
| 5179 | % |
| 5180 | % \begin{macrocode} |
| 5181 | \coldef L{\tabcoltype{\hb@xt@2em\bgroup$\eqastyle}{$\hss\egroup}} |
| 5182 | % \end{macrocode} |
| 5183 | % |
| 5184 | % The \lit{:} column type is fairly simple. |
| 5185 | % |
| 5186 | % \begin{macrocode} |
| 5187 | \coldef :{\tabspctype{\tabskip\eqacolskip}} |
| 5188 | \coldef q{\tabspctype{\quad}} |
| 5189 | % \end{macrocode} |
| 5190 | % |
| 5191 | % The other column types just insert given text in an appropriate way. |
| 5192 | % |
| 5193 | % \begin{macrocode} |
| 5194 | \collet > [tabular]> |
| 5195 | \collet < [tabular]< |
| 5196 | \collet * [tabular]* |
| 5197 | \collet @ [tabular]@ |
| 5198 | % \end{macrocode} |
| 5199 | % |
| 5200 | % Finally, the magical `|\magic|' column type, which sets the equation |
| 5201 | % number. We set up the |\tabskip| glue properly, tab on, and set the flag |
| 5202 | % which marks the final column. The |\eqa@lastcol| command is there to |
| 5203 | % raise an error if the user tabs over to this column. I'll temporarily |
| 5204 | % redefine it to |\@eqalasttrue| when I enter this column legitimately. |
| 5205 | % The extra magical bits here will make the final column repeat, so that we |
| 5206 | % can find it if necessary. Well is this column type named. |
| 5207 | % |
| 5208 | % That's it. We can return to normal now. |
| 5209 | % |
| 5210 | % \begin{macrocode} |
| 5211 | \colpop |
| 5212 | % \end{macrocode} |
| 5213 | % |
| 5214 | % \subsubsection{Newline codes} |
| 5215 | % |
| 5216 | % Newline sequences (|\\|) get turned into calls of |\@eqncr|. The job is |
| 5217 | % fairly simple, really. |
| 5218 | % |
| 5219 | % \begin{macrocode} |
| 5220 | \def\@eqncr{\tab@cr\eqacr@i\interdisplaylinepenalty\@M}% |
| 5221 | \def\eqacr@i#1#2{% |
| 5222 | \eqa@eqnum% |
| 5223 | \noalign{\penalty#2\vskip\jot\vskip#1}% |
| 5224 | } |
| 5225 | % \end{macrocode} |
| 5226 | % |
| 5227 | % \subsubsection{Setting equation numbers} |
| 5228 | % |
| 5229 | % \begin{macro}{\eqa@eqpos} |
| 5230 | % |
| 5231 | % Before we start, we need to generalise the flush-left number handling bits. |
| 5232 | % The macro |\eqa@eqpos| will put its argument in the right place. |
| 5233 | % |
| 5234 | % \begin{macrocode} |
| 5235 | \def\eqa@eqpos#1{% |
| 5236 | \if@leqno% |
| 5237 | \hb@xt@.01\p@{}\rlap{\normalfont\normalcolor\hskip-\displaywidth#1}% |
| 5238 | \else% |
| 5239 | \normalfont\normalcolor#1% |
| 5240 | \fi% |
| 5241 | } |
| 5242 | % \end{macrocode} |
| 5243 | % |
| 5244 | % \end{macro} |
| 5245 | % |
| 5246 | % \begin{macro}{\eqa@eqnum} |
| 5247 | % |
| 5248 | % Here we typeset an equation number in roughly the right place. First I'll |
| 5249 | % redefine |\eqa@lastcol| so that it tells me I'm in the right place, and |
| 5250 | % start a loop to find that place. |
| 5251 | % |
| 5252 | % \begin{macrocode} |
| 5253 | \def\eqa@eqnum{% |
| 5254 | \global\let\eqa@lastcol\@eqalasttrue% |
| 5255 | \eqa@eqnum@i% |
| 5256 | } |
| 5257 | % \end{macrocode} |
| 5258 | % |
| 5259 | % Now for the loop. The |\relax| here is absolutely vital -- it starts the |
| 5260 | % table column, inserting useful tokens like `|\eqa@lastcol|' which tell |
| 5261 | % me where I am in the alignment. Then, if I've reached the end, I can |
| 5262 | % typeset the equation number; otherwise I go off into another macro and |
| 5263 | % step on to the next column. |
| 5264 | % |
| 5265 | % \begin{macrocode} |
| 5266 | \def\eqa@eqnum@i{% |
| 5267 | \relax% |
| 5268 | \if@eqalast% |
| 5269 | \expandafter\eqa@eqnum@ii% |
| 5270 | \else% |
| 5271 | \expandafter\eqa@eqnum@iii% |
| 5272 | \fi% |
| 5273 | } |
| 5274 | \def\eqa@eqnum@ii{% |
| 5275 | \eqa@eqpos\eqa@number% |
| 5276 | \global\let\eqa@number\eqa@defnumber% |
| 5277 | \global\let\eqa@lastcol\eqa@@lastcol% |
| 5278 | \cr% |
| 5279 | } |
| 5280 | \def\eqa@eqnum@iii{&\eqa@eqnum@i} |
| 5281 | % \end{macrocode} |
| 5282 | % |
| 5283 | % \end{macro} |
| 5284 | % |
| 5285 | % \begin{macro}{\eqa@lastcol} |
| 5286 | % |
| 5287 | % This is used as a marker for the final column in an \env{eqnarray} |
| 5288 | % environment. By default it informs the user that they've been very |
| 5289 | % silly and swallows the contents of the column. I'll redefine it to |
| 5290 | % something more useful at appropriate times, and then turn it back again. |
| 5291 | % |
| 5292 | % \begin{macrocode} |
| 5293 | \def\eqa@@lastcol{\mth@err@number\setbox\z@} |
| 5294 | \let\eqa@lastcol\eqa@@lastcol |
| 5295 | % \end{macrocode} |
| 5296 | % |
| 5297 | % \end{macro} |
| 5298 | % |
| 5299 | % \subsubsection{Numbering control} |
| 5300 | % |
| 5301 | % \begin{macro}{\eqnumber} |
| 5302 | % |
| 5303 | % The |\eqnumber| command sets the equation number on the current equation. |
| 5304 | % This is really easy, actually. |
| 5305 | % |
| 5306 | % \begin{macrocode} |
| 5307 | \newcommand\eqnumber[1][\eqa@eqcount]{\gdef\eqa@number{#1}} |
| 5308 | % \end{macrocode} |
| 5309 | % |
| 5310 | % \end{macro} |
| 5311 | % |
| 5312 | % \begin{macro}{\eqa@eqcount} |
| 5313 | % |
| 5314 | % This is how a standard equation number is set, stepping the counter and |
| 5315 | % all. It's really easy and obvious. |
| 5316 | % |
| 5317 | % \begin{macrocode} |
| 5318 | \def\eqa@eqcount{(\theequation)\global\advance\c@equation\@ne} |
| 5319 | % \end{macrocode} |
| 5320 | % |
| 5321 | % \end{macro} |
| 5322 | % |
| 5323 | % \begin{macro}{\nonumber} |
| 5324 | % |
| 5325 | % The \LaTeX\ |\nonumber| command could be defined by saying |
| 5326 | % \begin{listing} |
| 5327 | %\renewcommand{\nonumber}{\eqnumber[]} |
| 5328 | % \end{listing} |
| 5329 | % but I'll be slightly more efficient and redefine |\eqa@number| directly. |
| 5330 | % |
| 5331 | % \begin{macrocode} |
| 5332 | \def\nonumber{\global\let\eqa@number\@empty} |
| 5333 | % \end{macrocode} |
| 5334 | % |
| 5335 | % \end{macro} |
| 5336 | % |
| 5337 | % \subsubsection{The \env{eqnalign} environment} |
| 5338 | % |
| 5339 | % As a sort of companion to \env{eqnarray}, here's an environment which does |
| 5340 | % similar things inside a box, rather than taking up the whole display width. |
| 5341 | % It uses the same column types that we've already created, so there should |
| 5342 | % be no problems. |
| 5343 | % |
| 5344 | % \begin{environment}{eqnalign} |
| 5345 | % |
| 5346 | % First, sort out some simple things like optional arguments. |
| 5347 | % |
| 5348 | % \begin{macrocode} |
| 5349 | \def\eqnalign{\@testopt\eqnalign@i{rcl}} |
| 5350 | \def\eqnalign@i[#1]{\@testopt{\eqnalign@ii{#1}}c} |
| 5351 | % \end{macrocode} |
| 5352 | % |
| 5353 | % Now we actually do the environment. This is fairly easy, actually. |
| 5354 | % |
| 5355 | % \begin{macrocode} |
| 5356 | \def\eqnalign@ii#1[#2]{% |
| 5357 | \let\\\eqn@cr% |
| 5358 | \colset{eqnarray}% |
| 5359 | \tab@initread% |
| 5360 | \def\tab@tabtext{&\tabskip\z@skip}% |
| 5361 | \tabskip\z@skip% |
| 5362 | \col@sep.5\eqainskip% |
| 5363 | \tab@readpreamble{#1}% |
| 5364 | \everycr{}% |
| 5365 | \if#2t\vtop\else% |
| 5366 | \if#2b\vbox\else% |
| 5367 | \vcenter% |
| 5368 | \fi% |
| 5369 | \fi% |
| 5370 | \bgroup% |
| 5371 | \halign\expandafter\bgroup\the\tab@preamble\cr% |
| 5372 | } |
| 5373 | % \end{macrocode} |
| 5374 | % |
| 5375 | % Finishing the environment is even simpler. |
| 5376 | % |
| 5377 | % \begin{macrocode} |
| 5378 | \def\endeqnalign{% |
| 5379 | \crcr% |
| 5380 | \egroup% |
| 5381 | \egroup% |
| 5382 | } |
| 5383 | % \end{macrocode} |
| 5384 | % |
| 5385 | % \end{environment} |
| 5386 | % |
| 5387 | % \begin{macro}{\eqn@cr} |
| 5388 | % |
| 5389 | % Newlines are really easy here. |
| 5390 | % |
| 5391 | % \begin{macrocode} |
| 5392 | \def\eqn@cr{\tab@cr\eqn@cr@i{}{}} |
| 5393 | \def\eqn@cr@i#1{\cr\noalign{\vskip\jot\vskip#1}\@gobble} |
| 5394 | % \end{macrocode} |
| 5395 | % |
| 5396 | % \end{macro} |
| 5397 | % |
| 5398 | % |
| 5399 | % \subsection{Simple multiline equations} |
| 5400 | % |
| 5401 | % As a sort of example and abbreviation, here's a multiline display |
| 5402 | % environment which just centres everything. |
| 5403 | % |
| 5404 | % \begin{environment}{eqlines} |
| 5405 | % |
| 5406 | % We just get |\eqnarray| to do everything for us. This is really easy. |
| 5407 | % |
| 5408 | % \begin{macrocode} |
| 5409 | \def\eqlines{\eqnarray[x]} |
| 5410 | \let\endeqlines\endeqnarray |
| 5411 | % \end{macrocode} |
| 5412 | % |
| 5413 | % \end{environment} |
| 5414 | % |
| 5415 | % \begin{environment}{eqlines*} |
| 5416 | % |
| 5417 | % There's a $*$ version which omits numbers. This is easy too. Lots of |
| 5418 | % hacking with expansion here to try and reduce the number of tokens being |
| 5419 | % used. Is it worth it? |
| 5420 | % |
| 5421 | % \begin{macrocode} |
| 5422 | \expandafter\edef\csname eqlines*\endcsname{% |
| 5423 | \expandafter\noexpand\csname eqnarray*\endcsname[x]% |
| 5424 | } |
| 5425 | \expandafter\let\csname endeqlines*\expandafter\endcsname |
| 5426 | \csname endeqnarray*\endcsname |
| 5427 | % \end{macrocode} |
| 5428 | % |
| 5429 | % \end{environment} |
| 5430 | % |
| 5431 | % |
| 5432 | % \subsection{Split equations} |
| 5433 | % |
| 5434 | % Based on an idea from \textit{The \TeX book}, we provide some simple |
| 5435 | % environments for doing split equations. There's plenty of scope for |
| 5436 | % improvement here, though. |
| 5437 | % |
| 5438 | % \begin{environment}{spliteqn} |
| 5439 | % \begin{environment}{spliteqn*} |
| 5440 | % |
| 5441 | % The only difference between these two is that the $*$-version doesn't put |
| 5442 | % in an equation number by default (although this behaviour can be |
| 5443 | % changed by |\eqnumber|). |
| 5444 | % |
| 5445 | % The fun here mainly concerns putting in the equation number at the right |
| 5446 | % place -- for |leqno| users, we need to put the number on the first line; |
| 5447 | % otherwise we put it on the last line. |
| 5448 | % |
| 5449 | % The way we handle this is to have two macros, |\\| (which clearly does |
| 5450 | % all the user line breaks) and |\seq@lastcr| which is used at the end of |
| 5451 | % the environment to wrap everything up. The |\seq@eqnocr| macro puts an |
| 5452 | % equation number on the current line and then does a normal |\\|. It also |
| 5453 | % resets |\\| and |\seq@lastcr| so that they don't try to put another |
| 5454 | % equation number in. This must be done globally, although anyone who tries |
| 5455 | % to nest maths displays will get what they deserve. |
| 5456 | % |
| 5457 | % For the non-$*$ environment, then, we need to step the equation counter, |
| 5458 | % and set |\\| to |\seq@cr| or |\seq@eqnocr| as appropriate for the setting |
| 5459 | % of the |leqno| flag -- |\seq@lastcr| always gets set to put an equation |
| 5460 | % number in (because it will be reset if the number actually gets done |
| 5461 | % earlier -- this catches stupid users trying to put a single row into |
| 5462 | % a split environment). |
| 5463 | % |
| 5464 | % \begin{macrocode} |
| 5465 | \def\spliteqn{% |
| 5466 | \let\eqa@oldnumber\eqa@number% |
| 5467 | \global\let\eqa@number\eqa@eqcount% |
| 5468 | \spliteqn@i% |
| 5469 | } |
| 5470 | % \end{macrocode} |
| 5471 | % |
| 5472 | % For the $*$ variant, we don't need to bother with equation numbering, so |
| 5473 | % this is really easy. |
| 5474 | % |
| 5475 | % \begin{macrocode} |
| 5476 | \@namedef{spliteqn*}{% |
| 5477 | \let\eqa@oldnumber\eqa@number% |
| 5478 | \gdef\eqa@number{}% |
| 5479 | \spliteqn@i% |
| 5480 | } |
| 5481 | % \end{macrocode} |
| 5482 | % |
| 5483 | % Ending the environments is easy. Most of the stuff here will be described |
| 5484 | % later. |
| 5485 | % |
| 5486 | % \begin{macrocode} |
| 5487 | \def\endspliteqn{% |
| 5488 | \hfilneg\seq@lastcr% |
| 5489 | \egroup% |
| 5490 | \dsp@end% |
| 5491 | \global\let\eqa@number\eqa@oldnumber% |
| 5492 | \global\advance\c@equation\m@ne% |
| 5493 | \global\@ignoretrue% |
| 5494 | } |
| 5495 | \expandafter\let\csname endspliteqn*\endcsname\endspliteqn |
| 5496 | % \end{macrocode} |
| 5497 | % |
| 5498 | % \end{environment} |
| 5499 | % \end{environment} |
| 5500 | % |
| 5501 | % \begin{macro}{\spliteqn@i} |
| 5502 | % |
| 5503 | % Here we handle the full display splits. Start a maths display, and make |
| 5504 | % each row of the alignment take up the full display width. |
| 5505 | % |
| 5506 | % The macro |\seq@dosplit| does most of the real work for us -- setting up |
| 5507 | % the alignment and so forth. The template column is interesting. There |
| 5508 | % are two items glue on both sides of the actual text: |
| 5509 | % |
| 5510 | % \begin{itemize} |
| 5511 | % |
| 5512 | % \item Some glue which can shrink. This keeps the display from the edges |
| 5513 | % of the page unless we get a really wide item. |
| 5514 | % |
| 5515 | % \item An |\hfil| to do the alignment. By default, this centres the |
| 5516 | % equations. On the first line, however, we put a leading |\hfilneg| |
| 5517 | % which cancels the first |\hfil|, making the first row left aligned. |
| 5518 | % Similarly, at the end, we put an |\hfilneg| after the last equation |
| 5519 | % to right align the last line. |
| 5520 | % |
| 5521 | % \end{itemize} |
| 5522 | % |
| 5523 | % We pass this information on as an argument. It's easy really. |
| 5524 | % |
| 5525 | % \begin{macrocode} |
| 5526 | \def\spliteqn@i{% |
| 5527 | % \end{macrocode} |
| 5528 | % |
| 5529 | % First, set up equation numbering properly. See my rant about |
| 5530 | % |\refstepcounter| above. |
| 5531 | % |
| 5532 | % \begin{macrocode} |
| 5533 | \stepcounter{equation}% |
| 5534 | \def\@currentlabel{\p@equation\theequation}% |
| 5535 | % \end{macrocode} |
| 5536 | % |
| 5537 | % Right; now to sort out the numbering and newline handling. If the number's |
| 5538 | % meant to be on the first line (for \textsf{leqno} users), then it gets |
| 5539 | % typeset on the first like; otherwise we just do a normal newline on |
| 5540 | % all lines except the first. Once |\seq@eqnocr| has done its stuff, it |
| 5541 | % redefines all the newline handling not to insert another number. |
| 5542 | % |
| 5543 | % \begin{macrocode} |
| 5544 | \if@leqno% |
| 5545 | \global\let\seq@docr\seq@eqnocr% |
| 5546 | \else% |
| 5547 | \global\let\seq@docr\seq@cr% |
| 5548 | \fi% |
| 5549 | \global\let\seq@lastcr\seq@eqnocr% |
| 5550 | % \end{macrocode} |
| 5551 | % |
| 5552 | % For my next trick, I'll do some display handling -- start a (possibly |
| 5553 | % nested) maths display, set up the |\tabpause| macro appropriately, and |
| 5554 | % set the newline command to do the right thing. |
| 5555 | % |
| 5556 | % \begin{macrocode} |
| 5557 | \dsp@start% |
| 5558 | \dsp@tabpause% |
| 5559 | \def\\{\seq@docr}% |
| 5560 | % \end{macrocode} |
| 5561 | % |
| 5562 | % Finally, call another macro to do the remaining bits of setting up. |
| 5563 | % |
| 5564 | % \begin{macrocode} |
| 5565 | \seq@dosplit% |
| 5566 | {\hb@xt@\displaywidth{% |
| 5567 | \hskip\splitleft\hfil$\displaystyle##$% |
| 5568 | \hfil\hskip\splitright}}% |
| 5569 | {\hfilneg}% |
| 5570 | } |
| 5571 | % \end{macrocode} |
| 5572 | % |
| 5573 | % \end{macro} |
| 5574 | % |
| 5575 | % \begin{environment}{subsplit} |
| 5576 | % |
| 5577 | % For doing splits in the middle of equations, we provide a similar |
| 5578 | % environment. Here, we make |\\| just start a new line. We also use |
| 5579 | % a |\vcenter| rather than a full maths display. The glue items are also |
| 5580 | % a bit different: we use plain double-quads on each side of the item, and |
| 5581 | % we need to remove them by hand at the extremities of the environment. |
| 5582 | % |
| 5583 | % \begin{macrocode} |
| 5584 | \def\subsplit{\@ifnextchar[\subsplit@i{\subsplit@i[c]}} |
| 5585 | \def\subsplit@i[#1]{% |
| 5586 | \let\@tempa\vcenter% |
| 5587 | \if#1t\let\@tempa\vtop\fi% |
| 5588 | \if#1b\let\@tempa\vbox\fi% |
| 5589 | \let\\\seq@cr% |
| 5590 | \@tempa\bgroup% |
| 5591 | \seq@dosplit{\hfil\qquad$##$\qquad\hfil}{\hfilneg\hskip-2em}% |
| 5592 | } |
| 5593 | % \end{macrocode} |
| 5594 | % |
| 5595 | % Ending the environment is fairly easy. We remove the final glue item, |
| 5596 | % and close the alignment and the vbox. |
| 5597 | % |
| 5598 | % \begin{macrocode} |
| 5599 | \def\endsubsplit{% |
| 5600 | \hfilneg\hskip-2em\cr% |
| 5601 | \egroup\egroup% |
| 5602 | } |
| 5603 | % \end{macrocode} |
| 5604 | % |
| 5605 | % \end{environment} |
| 5606 | % |
| 5607 | % \begin{macro}{\seq@dosplit} |
| 5608 | % |
| 5609 | % Here we do most of the real work. Actually, since the preamble is passed |
| 5610 | % in as an argument, most of the work is already done. The only thing to |
| 5611 | % really note is the template for subsequent columns. To stop users putting |
| 5612 | % in extra columns (which is where we put the equation number) we raise an |
| 5613 | % error and discard the input in a scratch box register. This template is |
| 5614 | % repeated infinitely so as to allow us to put the equation number in nicely. |
| 5615 | % However, the final negative glue item won't work properly, so the equation |
| 5616 | % will look awful. |
| 5617 | % |
| 5618 | % \begin{macrocode} |
| 5619 | \def\seq@dosplit#1#2{% |
| 5620 | \halign\bgroup% |
| 5621 | #1&&\mth@err@number\setbox\z@\hbox{##}\cr% |
| 5622 | #2\relax% |
| 5623 | } |
| 5624 | % \end{macrocode} |
| 5625 | % |
| 5626 | % \end{macro} |
| 5627 | % |
| 5628 | % \begin{macro}{\seq@eqnocr} |
| 5629 | % |
| 5630 | % Here's how we set equation numbers. Since the column provided raises |
| 5631 | % errors as soon as a token finds its way into it, we start with a |&\omit|. |
| 5632 | % Then we just put the equation number in a zero-width box. Finally, we |
| 5633 | % reset the newline commands to avoid putting in more than one equation |
| 5634 | % number, and do normal newline things. |
| 5635 | % |
| 5636 | % \begin{macrocode} |
| 5637 | \def\seq@eqnocr{% |
| 5638 | &\omit% |
| 5639 | \hb@xt@\z@{\hss\eqa@eqpos\eqa@number}% |
| 5640 | \global\let\seq@docr\seq@cr% |
| 5641 | \global\let\seq@lastcr\seq@cr% |
| 5642 | \seq@cr% |
| 5643 | } |
| 5644 | % \end{macrocode} |
| 5645 | % |
| 5646 | % \end{macro} |
| 5647 | % |
| 5648 | % \begin{macro}{\seq@cr} |
| 5649 | % |
| 5650 | % Newlines are very easy. We add a |\jot| of extra space, since this is |
| 5651 | % a nice thing to do. |
| 5652 | % |
| 5653 | % \begin{macrocode} |
| 5654 | \def\seq@cr{\tab@cr\seq@cr@i\interdisplaylinepenalty\@M} |
| 5655 | \def\seq@cr@i#1#2{\cr\noalign{\penalty#2\vskip\jot\vskip#1}} |
| 5656 | % \end{macrocode} |
| 5657 | % |
| 5658 | % \end{macro} |
| 5659 | % |
| 5660 | % |
| 5661 | % \subsection{Matrix handling} |
| 5662 | % |
| 5663 | % There's been a complete and total overhaul of the spacing calculations |
| 5664 | % for matrices here. The vertical spacing now bears an uncanny similarity |
| 5665 | % to the rules \TeX\ uses to space out |\atop|-like fractions, the difference |
| 5666 | % being that you can have more than one column in a matrix. This has the |
| 5667 | % interesting side-effect that we get an \package{amsmath}-style |
| 5668 | % sub/superscript environment almost free of charge with the matrix handling |
| 5669 | % (it just ends up being a script-size single-column matrix). |
| 5670 | % |
| 5671 | % What is rather gratifying is that our \env{matrix} environment looks |
| 5672 | % rather nicer than \package{amsmath}'s (which is based directly on |
| 5673 | % \env{array}, giving it nasty restrictions on the numbers of columns and |
| 5674 | % so on); in particular, the version here gives the `correct' result for |
| 5675 | % Knuth's exercise~18.42 (which states categorically that a |\smallskip| |
| 5676 | % should be placed between the rows of the big matrix). |
| 5677 | % |
| 5678 | % The reason the interrow space doesn't come out in the AMS version is |
| 5679 | % that \env{array} inserts extra vertical space by extending the depth of |
| 5680 | % the final row using a strut: the big matrix already extends deeper than |
| 5681 | % this, so the strut doesn't make any difference. If the space was added |
| 5682 | % by |\hlx{s[\smallskipamount]}| instead of the |\\| command, things would |
| 5683 | % be different. |
| 5684 | % |
| 5685 | % \begin{figure} |
| 5686 | % |
| 5687 | % ^^A This is essentially what amsmath (version 1.2b) does. The real |
| 5688 | % ^^A implementation requires a counter MaxMatrixCols, and has fewer braces: |
| 5689 | % ^^A that's all the difference. Oh, and I turn off \arrayextrasep here, |
| 5690 | % ^^A since amsmath doesn't expect it to be there (accurate emulation, see?) |
| 5691 | % ^^A and I've used \hspace instead of \hskip since everything else is |
| 5692 | % ^^A `proper' LaTeX stuff. |
| 5693 | % |
| 5694 | % \newenvironment{ams-pmatrix}{^^A |
| 5695 | % \setlength{\arrayextrasep}{0pt}^^A |
| 5696 | % \left(^^A |
| 5697 | % \hspace{-\arraycolsep}^^A |
| 5698 | % \begin{array}{*{10}{c}}^^A |
| 5699 | % }{^^A |
| 5700 | % \end{array}^^A |
| 5701 | % \hspace{-\arraycolsep}^^A |
| 5702 | % \right)^^A |
| 5703 | % } |
| 5704 | % |
| 5705 | % \begin{demo}{Exercise 18.42 from \emph{The \TeX book}} |
| 5706 | %\newcommand{\domatrix}[1]{ |
| 5707 | % \def\mat##1 |
| 5708 | % {\begin{#1}##1\end{#1}} |
| 5709 | % \[ \begin{#1} |
| 5710 | % \mat{a & b \\ c & d} & |
| 5711 | % \mat{e & f \\ g & h} |
| 5712 | % \\[\smallskipamount] |
| 5713 | % 0 & |
| 5714 | % \mat{i & j \\ k & l} |
| 5715 | % \end{#1} |
| 5716 | % \] |
| 5717 | %} |
| 5718 | %\domatrix{pmatrix} |
| 5719 | %\domatrix{ams-pmatrix} |
| 5720 | % \end{demo} |
| 5721 | % |
| 5722 | % \end{figure} |
| 5723 | % |
| 5724 | % \begin{environment}{genmatrix} |
| 5725 | % |
| 5726 | % The first job is to store my maths style and font away, because I'll be |
| 5727 | % needing it lots later. |
| 5728 | % |
| 5729 | % \begin{macrocode} |
| 5730 | \def\genmatrix#1#2#3#4#5{% |
| 5731 | \let\mat@style#1% |
| 5732 | \ifx#2\scriptstyle% |
| 5733 | \let\mat@font\scriptfont% |
| 5734 | \else\ifx#2\scriptscriptstyle% |
| 5735 | \let\mat@font\scriptscriptfont% |
| 5736 | \else% |
| 5737 | \let\mat@font\textfont% |
| 5738 | \fi\fi% |
| 5739 | % \end{macrocode} |
| 5740 | % |
| 5741 | % Now to cope with inserted text. This is easy. |
| 5742 | % |
| 5743 | % \begin{macrocode} |
| 5744 | \ifx\mat@style\scriptstyle% |
| 5745 | \let\mat@textsize\scriptsize% |
| 5746 | \else\ifx\mat@style\scriptscriptstyle% |
| 5747 | \let\mat@textsize\scriptscriptsize% |
| 5748 | \else% |
| 5749 | \let\mat@textsize\relax% |
| 5750 | \fi\fi% |
| 5751 | % \end{macrocode} |
| 5752 | % |
| 5753 | % Now for some fun. I'll remember how to start and end the matrix in a |
| 5754 | % couple of macros |\mat@left| and |\mat@right|. I haven't yet worked out |
| 5755 | % exactly what needs to be in |\mat@right| yet, though, so I'll build that |
| 5756 | % up in a scratch token list while I'm making my mind up. |
| 5757 | % |
| 5758 | % Initially, I want to open a group (to trap the style changes), set the |
| 5759 | % maths style (to get the right spacing), insert the left delimiter, insert |
| 5760 | % some spacing around the matrix, and start a centred box. The ending just |
| 5761 | % closes all the groups and delimiters I opened. |
| 5762 | % |
| 5763 | % \begin{macrocode} |
| 5764 | \def\mat@left{\bgroup\mat@style\left#4#3\vcenter\bgroup}% |
| 5765 | \toks@{\egroup#3\right#5\egroup}% |
| 5766 | % \end{macrocode} |
| 5767 | % |
| 5768 | % Now comes a slightly trickier bit. If the maths style is script or |
| 5769 | % scriptscript, then I need to raise the box by a little bit to make it look |
| 5770 | % really good. The right amount is somewhere around \smallf 3/4\,pt, I |
| 5771 | % think, so that's what I'll use. |
| 5772 | % |
| 5773 | % \begin{macrocode} |
| 5774 | \@tempswatrue% |
| 5775 | \ifx\mat@style\displaystyle\else\ifx\mat@style\textstyle\else% |
| 5776 | \@tempswafalse% |
| 5777 | \setbox\z@\hbox\bgroup$% |
| 5778 | \toks@\expandafter{\the\toks@$\m@th\egroup\raise.75\p@\box\z@}% |
| 5779 | \fi\fi% |
| 5780 | % \end{macrocode} |
| 5781 | % |
| 5782 | % If I'm not in maths mode right now, then I should enter maths mode, and |
| 5783 | % remember to leave it later. |
| 5784 | % |
| 5785 | % \begin{macrocode} |
| 5786 | \if@tempswa\ifmmode\else% |
| 5787 | $\m@th% |
| 5788 | \toks@\expandafter{\the\toks@$}% |
| 5789 | \fi\fi% |
| 5790 | % \end{macrocode} |
| 5791 | % |
| 5792 | % Now I've sorted out how to end the environment properly, so I can set up |
| 5793 | % the macro, using |\edef|. |
| 5794 | % |
| 5795 | % \begin{macrocode} |
| 5796 | \edef\mat@right{\the\toks@}% |
| 5797 | % \end{macrocode} |
| 5798 | % |
| 5799 | % Now see if there's an optional argument. If not, create lots of centred |
| 5800 | % columns. |
| 5801 | % |
| 5802 | % \begin{macrocode} |
| 5803 | \@testopt\genmatrix@i{[c}% |
| 5804 | } |
| 5805 | % \end{macrocode} |
| 5806 | % |
| 5807 | % Now to sort out everything else. |
| 5808 | % |
| 5809 | % \begin{macrocode} |
| 5810 | \def\genmatrix@i[#1]{% |
| 5811 | % \end{macrocode} |
| 5812 | % |
| 5813 | % Some initial setting up: choose the correct column set, and set up some |
| 5814 | % variables for reading the preamble. |
| 5815 | % |
| 5816 | % \begin{macrocode} |
| 5817 | \colset{matrix}% |
| 5818 | \tab@initread% |
| 5819 | % \end{macrocode} |
| 5820 | % |
| 5821 | % Now comes some of the tricky stuff. The space between columns should be |
| 5822 | % 12\,mu (by trial and error). We put the space in a box so we can measure |
| 5823 | % it in the correct mathstyle. |
| 5824 | % |
| 5825 | % \begin{macrocode} |
| 5826 | \setbox\z@\hbox{$\mat@style\mskip12mu$}% |
| 5827 | \edef\tab@tabtext{&\kern\the\wd\z@}% |
| 5828 | \tab@readpreamble{#1}% |
| 5829 | % \end{macrocode} |
| 5830 | % |
| 5831 | % Now we need to decide how to space out the rows. The code here is based |
| 5832 | % on the information in appendix~G of \emph{The \TeX book}: I think it'd be |
| 5833 | % nice if my matrices were spaced out in the same way as normal fractions |
| 5834 | % (particularly |\choose|y things). The standard |\baselineskip| and |
| 5835 | % |\lineskip| parameters come in really handy here. |
| 5836 | % |
| 5837 | % The parameters vary according to the size of the text, so I need to see |
| 5838 | % if we have scriptsize or less, or not. The tricky |\if| sorts this out. |
| 5839 | % |
| 5840 | % \begin{macrocode} |
| 5841 | \if1\ifx\mat@style\scriptstyle1\else% |
| 5842 | \ifx\mat@style\scriptscriptstyle1\else0\fi\fi% |
| 5843 | \baselineskip\fontdimen10\mat@font\tw@% |
| 5844 | \advance\baselineskip\fontdimen12\mat@font\tw@% |
| 5845 | \lineskip\thr@@\fontdimen8\mat@font\thr@@% |
| 5846 | \else% |
| 5847 | \baselineskip\fontdimen8\mat@font\tw@% |
| 5848 | \advance\baselineskip\fontdimen11\mat@font\tw@% |
| 5849 | \lineskip7\fontdimen8\mat@font\thr@@% |
| 5850 | \fi% |
| 5851 | \lineskiplimit\lineskip% |
| 5852 | % \end{macrocode} |
| 5853 | % |
| 5854 | % Now actually set up for the alignment. Assign |\\| to the correct value. |
| 5855 | % Set up the |\tabskip|. Do the appropriate |\mat@left| thing set up above. |
| 5856 | % And then start the alignment. |
| 5857 | % |
| 5858 | % \begin{macrocode} |
| 5859 | \let\\\mat@cr% |
| 5860 | \tabskip\z@skip% |
| 5861 | \col@sep\z@% |
| 5862 | \mat@left% |
| 5863 | \halign\expandafter\bgroup\the\tab@preamble\tabskip\z@skip\cr% |
| 5864 | % \end{macrocode} |
| 5865 | % |
| 5866 | % Now for a little hack to make the spacing consistent between matrices of |
| 5867 | % the same height. This comes directly from \PlainTeX. This appears to |
| 5868 | % make the spacing \emph{exactly} the same as the \TeX\ primitives, oddly |
| 5869 | % enough. |
| 5870 | % |
| 5871 | % \begin{macrocode} |
| 5872 | \ifx\mat@font\textfont% |
| 5873 | \omit$\mat@style\mathstrut$\cr\noalign{\kern-\baselineskip}% |
| 5874 | \fi% |
| 5875 | } |
| 5876 | % \end{macrocode} |
| 5877 | % |
| 5878 | % Finishing the environment is really easy. We do the spacing hack again |
| 5879 | % at the bottom, close the alignment and then tidy whatever we started in |
| 5880 | % |\mat@left|. |
| 5881 | % |
| 5882 | % \begin{macrocode} |
| 5883 | \def\endgenmatrix{% |
| 5884 | \crcr% |
| 5885 | \ifx\mat@font\textfont% |
| 5886 | \omit$\mat@style\mathstrut$\cr\noalign{\kern-\baselineskip}% |
| 5887 | \fi% |
| 5888 | \egroup% |
| 5889 | \mat@right% |
| 5890 | } |
| 5891 | % \end{macrocode} |
| 5892 | % |
| 5893 | % \end{environment} |
| 5894 | % |
| 5895 | % \begin{macro}{\mat@cr} |
| 5896 | % |
| 5897 | % Newlines are really easy. The $*$-form means nothing here, so we ignore |
| 5898 | % it. |
| 5899 | % |
| 5900 | % \begin{macrocode} |
| 5901 | \def\mat@cr{\tab@cr\mat@cr@i{}{}} |
| 5902 | \def\mat@cr@i#1{\cr\noalign{\vskip#1}\@gobble} |
| 5903 | % \end{macrocode} |
| 5904 | % |
| 5905 | % \end{macro} |
| 5906 | % |
| 5907 | % \begin{macro}{\newmatrix} |
| 5908 | % |
| 5909 | % This is how we define new matrix environments. It's simple fun with |
| 5910 | % |\csname| and |\expandafter|. |
| 5911 | % |
| 5912 | % \begin{macrocode} |
| 5913 | \def\newmatrix#1#2{% |
| 5914 | \@namedef{#1}{\genmatrix#2}% |
| 5915 | \expandafter\let\csname end#1\endcsname\endgenmatrix% |
| 5916 | } |
| 5917 | % \end{macrocode} |
| 5918 | % |
| 5919 | % \end{macro} |
| 5920 | % |
| 5921 | % \begin{environment}{matrix} |
| 5922 | % \begin{environment}{pmatrix} |
| 5923 | % \begin{environment}{dmatrix} |
| 5924 | % \begin{environment}{smatrix} |
| 5925 | % \begin{environment}{spmatrix} |
| 5926 | % \begin{environment}{sdmatrix} |
| 5927 | % \begin{environment}{smatrix*} |
| 5928 | % \begin{environment}{spmatrix*} |
| 5929 | % \begin{environment}{sdmatrix*} |
| 5930 | % |
| 5931 | % Now we define all the other environments we promised. This is easy. |
| 5932 | % |
| 5933 | % \begin{macrocode} |
| 5934 | \newmatrix{matrix}{{\textstyle}{\textstyle}{\,}{.}{.}} |
| 5935 | \newmatrix{pmatrix}{{\textstyle}{\textstyle}{\,}{(}{)}} |
| 5936 | \newmatrix{dmatrix}{{\textstyle}{\textstyle}{\,}} |
| 5937 | \newmatrix{smatrix}{{\scriptstyle}{\scriptstyle}{}{.}{.}} |
| 5938 | \newmatrix{spmatrix}{{\scriptstyle}{\scriptstyle}{}{(}{)}} |
| 5939 | \newmatrix{sdmatrix}{{\scriptstyle}{\scriptstyle}{}} |
| 5940 | \newmatrix{smatrix*}{{\scriptstyle}{\textstyle}{}{.}{.}} |
| 5941 | \newmatrix{spmatrix*}{{\scriptstyle}{\textstyle}{}{(}{)}} |
| 5942 | \newmatrix{sdmatrix*}{{\scriptstyle}{\textstyle}{}} |
| 5943 | % \end{macrocode} |
| 5944 | % |
| 5945 | % \end{environment} |
| 5946 | % \end{environment} |
| 5947 | % \end{environment} |
| 5948 | % \end{environment} |
| 5949 | % \end{environment} |
| 5950 | % \end{environment} |
| 5951 | % \end{environment} |
| 5952 | % \end{environment} |
| 5953 | % \end{environment} |
| 5954 | % |
| 5955 | % \begin{environment}{script} |
| 5956 | % |
| 5957 | % Now for superscripts and subscripts. This is fairly easy, because I |
| 5958 | % took so much care over the matrix handling. |
| 5959 | % |
| 5960 | % \begin{macrocode} |
| 5961 | \def\script{% |
| 5962 | \let\mat@style\scriptstyle% |
| 5963 | \def\mat@left{\vcenter\bgroup}% |
| 5964 | \def\mat@right{\egroup}% |
| 5965 | \let\mat@font\scriptfont% |
| 5966 | \let\mat@textsize\scriptsize% |
| 5967 | \@testopt\genmatrix@i c% |
| 5968 | } |
| 5969 | \let\endscript\endgenmatrix |
| 5970 | % \end{macrocode} |
| 5971 | % |
| 5972 | % \end{environment} |
| 5973 | % |
| 5974 | % Now define the column types. |
| 5975 | % |
| 5976 | % \begin{macrocode} |
| 5977 | \colpush{matrix} |
| 5978 | \coldef l{\tabcoltype{\kern\z@$\mat@style}{\m@th$\hfil}} |
| 5979 | \coldef c{\tabcoltype{\hfil$\mat@style}{\m@th$\hfil}} |
| 5980 | \coldef r{\tabcoltype{\hfil$\mat@style}{\m@th$}} |
| 5981 | \coldef T#1{\tab@aligncol{#1}{\begingroup\mat@textsize}{\endgroup}} |
| 5982 | \collet > [tabular]> |
| 5983 | \collet < [tabular]< |
| 5984 | \collet * [tabular]* |
| 5985 | \collet @ [tabular]@ |
| 5986 | % \end{macrocode} |
| 5987 | % |
| 5988 | % The repeating type is more awkward. Things will go wrong if this is |
| 5989 | % given before the first column, so we must do a whole repeat by hand. We |
| 5990 | % can tell if we haven't contributed a column yet, since |\tab@column| will |
| 5991 | % be zero. Otherwise, we fiddle the parser state to start a new column, and |
| 5992 | % insert the |&| character to make \TeX\ repeat the preamble. |
| 5993 | % |
| 5994 | % \begin{macrocode} |
| 5995 | \coldef {[}{% |
| 5996 | \@firstoftwo{% |
| 5997 | \ifnum\tab@columns=\z@% |
| 5998 | \def\@tempa##1\q@delim{% |
| 5999 | \tab@mkpreamble##1[##1\q@delim% |
| 6000 | }% |
| 6001 | \expandafter\@tempa% |
| 6002 | \else% |
| 6003 | \tab@setstate\tab@prestate% |
| 6004 | \tab@append\tab@preamble{&}% |
| 6005 | \expandafter\tab@mkpreamble% |
| 6006 | \fi% |
| 6007 | }% |
| 6008 | } |
| 6009 | % \end{macrocode} |
| 6010 | % |
| 6011 | % We're done defining columns now. |
| 6012 | % |
| 6013 | % \begin{macrocode} |
| 6014 | \colpop |
| 6015 | % \end{macrocode} |
| 6016 | % |
| 6017 | % |
| 6018 | % \subsection{Dots\dots} |
| 6019 | % |
| 6020 | % Nothing whatsoever to do with alignments, although vertical and diagonal |
| 6021 | % dots in small matrices look really silly. The following hacky definitions |
| 6022 | % work rather better. |
| 6023 | % |
| 6024 | % \begin{macro}{\mdw@dots} |
| 6025 | % |
| 6026 | % First of all, here's some definitions common to both of the dots macros. |
| 6027 | % The macro takes as an argument the actual code to draw the dots, passing |
| 6028 | % it the scaled size of a point in the scratch register |\dimen@|; the |
| 6029 | % register |\box 0| is set to contain a dot of the appropriate size. |
| 6030 | % |
| 6031 | % \begin{macrocode} |
| 6032 | \def\mdw@dots#1{\ensuremath{\mathpalette\mdw@dots@i{#1}}} |
| 6033 | \def\mdw@dots@i#1#2{% |
| 6034 | \setbox\z@\hbox{$#1\mskip1.8mu$}% |
| 6035 | \dimen@\wd\z@% |
| 6036 | \setbox\z@\hbox{$#1.$}% |
| 6037 | #2% |
| 6038 | } |
| 6039 | % \end{macrocode} |
| 6040 | % |
| 6041 | % \end{macro} |
| 6042 | % |
| 6043 | % \begin{macro}{\vdots} |
| 6044 | % |
| 6045 | % I'll start with the easy one. This is a simple translation of the original |
| 6046 | % implementation. |
| 6047 | % |
| 6048 | % \begin{macrocode} |
| 6049 | \def\vdots{% |
| 6050 | \mdw@dots{\vbox{% |
| 6051 | \baselineskip4\dimen@% |
| 6052 | \lineskiplimit\z@% |
| 6053 | \kern6\dimen@% |
| 6054 | \copy\z@\copy\z@\box\z@% |
| 6055 | }}% |
| 6056 | } |
| 6057 | % \end{macrocode} |
| 6058 | % |
| 6059 | % \end{macro} |
| 6060 | % |
| 6061 | % \begin{macro}{\ddots} |
| 6062 | % |
| 6063 | % And I'll end with the other easy one\dots |
| 6064 | % |
| 6065 | % \begin{macrocode} |
| 6066 | \def\ddots{% |
| 6067 | \mdw@dots{\mathinner{% |
| 6068 | \mkern1mu% |
| 6069 | \raise7\dimen@\vbox{\kern7\dimen@\copy\z@}% |
| 6070 | \mkern2mu% |
| 6071 | \raise4\dimen@\copy\z@% |
| 6072 | \mkern2mu% |
| 6073 | \raise\dimen@\box\z@% |
| 6074 | \mkern1mu% |
| 6075 | }}% |
| 6076 | } |
| 6077 | % \end{macrocode} |
| 6078 | % |
| 6079 | % \end{macro} |
| 6080 | % |
| 6081 | % |
| 6082 | % \subsection{Lucky dip} |
| 6083 | % |
| 6084 | % Time to round off with some trivial environments, just to show how easy |
| 6085 | % this stuff is. |
| 6086 | % |
| 6087 | % \begin{environment}{cases} |
| 6088 | % \begin{environment}{smcases} |
| 6089 | % |
| 6090 | % These are totally and utterly trivial. |
| 6091 | % |
| 6092 | % \begin{macrocode} |
| 6093 | \def\cases{\left\{\,\array{@{}lTl@{}}} |
| 6094 | \def\endcases{\endarray\,\right.} |
| 6095 | \def\smcases{\left\{\smarray{@{}lTl@{}}} |
| 6096 | \def\endsmcases{\endsmarray\,\right.} |
| 6097 | % \end{macrocode} |
| 6098 | % |
| 6099 | % \end{environment} |
| 6100 | % \end{environment} |
| 6101 | % |
| 6102 | % \subsection{Error messages} |
| 6103 | % |
| 6104 | % Some token saving: |
| 6105 | % |
| 6106 | % \begin{macrocode} |
| 6107 | \def\mth@error{\PackageError{mathenv}} |
| 6108 | % \end{macrocode} |
| 6109 | % |
| 6110 | % Now for the error messages. |
| 6111 | % |
| 6112 | % \begin{macrocode} |
| 6113 | \def\mth@err@number{% |
| 6114 | \mth@error{Too many `&' characters found}{% |
| 6115 | You've put too many `&' characters in an alignment^^J% |
| 6116 | environment (like `eqnarray' or `spliteqn') and wandered^^J% |
| 6117 | into trouble. I've gobbled the contents of that column^^J% |
| 6118 | and hopefully I can recover fairly easily.% |
| 6119 | }% |
| 6120 | } |
| 6121 | % \end{macrocode} |
| 6122 | % |
| 6123 | % \begin{macrocode} |
| 6124 | \def\mth@err@mdsp{% |
| 6125 | \mth@error{Can't do displays in nondisplay maths mode}{% |
| 6126 | You're trying to start a display environment, but you're^^J% |
| 6127 | in nondisplay maths mode. The display will appear but^^J% |
| 6128 | don't blame me when it looks horrible.% |
| 6129 | }% |
| 6130 | } |
| 6131 | % \end{macrocode} |
| 6132 | % |
| 6133 | % \begin{macrocode} |
| 6134 | \def\mth@err@hdsp{% |
| 6135 | \mth@error{Can't do displays in LR mode}{% |
| 6136 | You're trying to start a display environment, but you're^^J% |
| 6137 | in LR (restricted horizontal) mode. Everything will go^^J% |
| 6138 | totally wrong, so your best bet is to type `X', fix the^^J% |
| 6139 | mistake and start again.% |
| 6140 | }% |
| 6141 | } |
| 6142 | % \end{macrocode} |
| 6143 | % |
| 6144 | % \vskip\parskip\vbox{ ^^A The best way I could find of keeping this lot |
| 6145 | % ^^A together, I'm afraid. |
| 6146 | % That's all there is. Byebye. |
| 6147 | % |
| 6148 | % \begin{macrocode} |
| 6149 | %</mathenv> |
| 6150 | % \end{macrocode} |
| 6151 | % \nopagebreak |
| 6152 | % |
| 6153 | % \hfill Mark Wooding, \today |
| 6154 | % } |
| 6155 | % |
| 6156 | % \Finale |
| 6157 | % |
| 6158 | \endinput |