Debian Codegraph @

texmf-dist/source/latex/arydshln/arydshln.dtx

raw
1 % \iffalse
2 %% Style file `arydshln'
3 %% Copyright (C) 1993,1998-2018
4 %% Hiroshi Nakashima <h.nakasima'AT'media'DOT'kyoto-u'DOT'ac'DOT'jp>
5 %% (Kyoto Univ.)
6 %% This program can be redistributed and/or modified under the terms
7 %% of the LaTeX Project Public License Distributed from CTAN
8 %% archives in directory macros/latex/base/lppl.txt; either
9 %% version 1 of the License, or any later version.
10 % \fi
11 %
12 %\CheckSum{3697}
13 %% \CharacterTable
14 %% {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
15 %% 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
16 %% Digits \0\1\2\3\4\5\6\7\8\9
17 %% Exclamation \! Double quote \" Hash (number) \#
18 %% Dollar \$ Percent \% Ampersand \&
19 %% Acute accent \' Left paren \( Right paren \)
20 %% Asterisk \* Plus \+ Comma \,
21 %% Minus \- Point \. Solidus \/
22 %% Colon \: Semicolon \; Less than \<
23 %% Equals \= Greater than \> Question mark \?
24 %% Commercial at \@ Left bracket \[ Backslash \\
25 %% Right bracket \] Circumflex \^ Underscore \_
26 %% Grave accent \` Left brace \{ Vertical bar \|
27 %% Right brace \} Tilde \~}
28 %%
29 % \iffalse
30 %%
31 %<*dtx>
32 \ProvidesFile{arydshln.dtx}
33 %</dtx>
34 %<package>\def\next{LaTeX2e}
35 %<package>\ifx\fmtname\next\def\next{
36 %<package>\NeedsTeXFormat{LaTeX2e}[1994/12/01]
37 %<package>\ProvidesPackage{arydshln}}
38 %<package>\else\def\next[#1]{}\fi\next
39 %<driver>\ProvidesFile{arydshln-man.tex}
40 % \fi
41 % \ProvidesFile{arydshln.dtx}
42 [2019/02/21 v1.76 ]
43 %
44 % \iffalse
45 %<*driver>
46 \documentclass{ltxdoc}
47 \usepackage{arydshln}
48 \DisableCrossrefs
49 \PageIndex
50 \CodelineNumbered
51 \RecordChanges
52 %</driver>
53 %<driver>\OnlyDescription
54 %<driver>\def\ONLYDESCRIPTION{}
55 %<*driver>
56 \advance\textwidth4em
57 \begin{document}
58 \DocInput{arydshln.dtx}
59 \end{document}
60 %</driver>
61 % \fi
62 % \MakeShortVerb{\|}
63 % \GetFileInfo{arydshln.dtx}
64 % \IndexPrologue{\section*{Index}}
65 % \setcounter{IndexColumns}{2}
66 % \def\th{^{\it th}}
67 % \def\LOR{{\ {\lor}\ }}
68 % \def\LAND{{\ {\land}\ }}
69 % \let\togets\leftrightarrow
70 % \let\epsilon\varepsilon
71 % \def\cs#1{\hbox{\tt\char\escapechar#1}}
72 %
73 % \makeatletter
74 % \def\Meta#1{\penalty\@highpenalty \hbox{{\tt\char`\{}\meta{#1}{\tt\char`\}}}}
75 %
76 % \def\eqnarray{\stepcounter{equation}\let\@currentlabel=\theequation
77 % \global\@eqnswtrue
78 % \global\@eqcnt\z@\tabskip\@centering\let\\=\@eqncr
79 % $$\halign to \displaywidth\bgroup\hskip\@centering
80 % $\displaystyle\tabskip\z@{##}$\@eqnsel&\global\@eqcnt\@ne
81 % \hfil$\;{##}\;$\hfil
82 % &\global\@eqcnt\tw@ $\displaystyle\tabskip\z@{##}$\hfil
83 % \tabskip\@centering&\llap{##}\tabskip\z@\cr}
84 %
85 % \let\latex@addmarginpar\@addmarginpar
86 % \let\latex@latex@warning@no@line\@latex@warning@no@line
87 % \def\@addmarginpar{\let\@latex@warning@no@line\@gobble
88 % \latex@addmarginpar
89 % \let\@latex@warning@no@line\latex@latex@warning@no@line}
90 % \let\@font@warning\@gobble
91 %
92 % \long\def\m@cro@#1#2{\endgroup \topsep\MacroTopsep \trivlist
93 % \edef\saved@macroname{\string#2}
94 % \def\makelabel##1{\llap{##1\hskip\@totalleftmargin}}
95 % \if@inlabel
96 % \let\@tempa\@empty \count@\macro@cnt
97 % \loop \ifnum\count@>\z@
98 % \edef\@tempa{\@tempa\hbox{\strut}}\advance\count@\m@ne \repeat
99 % \edef\makelabel##1{\llap{\vtop to\baselineskip
100 % {\@tempa\hbox{##1}\vss}
101 % \hskip\@totalleftmargin}}
102 % \advance \macro@cnt \@ne
103 % \else \macro@cnt\@ne \fi
104 % \edef\@tempa{\noexpand\item[
105 % #1
106 % \noexpand\PrintMacroName
107 % \else
108 % \noexpand\PrintEnvName
109 % \fi
110 % {\string#2}]}
111 % \@tempa
112 % \global\advance\c@CodelineNo\@ne
113 % #1
114 % \SpecialMainIndex{#2}\nobreak
115 % \DoNotIndex{#2}
116 % \else
117 % \SpecialMainEnvIndex{#2}\nobreak
118 % \fi
119 % \global\advance\c@CodelineNo\m@ne
120 % \ignorespaces}
121 %
122 % \def\DescribeOpt{\leavevmode\@bsphack\begingroup\MakePrivateLetters
123 % \Describe@Opt}
124 % \def\Describe@Opt#1{\endgroup \Describe@pt{#1}}
125 % \def\Describe@pt#1#2{
126 % \marginpar{\raggedleft\PrintDescribeMacro{#1}}
127 % \SpecialOptUsageIndex{#1}{#2}}
128 % \def\SpecialOptUsageIndex#1#2{
129 % \index{#1\actualchar{\protect\ttfamily#1} (#2)\encapchar usage}
130 % \index{#2s:\levelchar{\protect\ttfamily#1}\encapchar usage}
131 % \@esphack\ignorespaces}
132 % \def\SpecialOptMainIndex#1#2{
133 % \index{#1\actualchar{\protect\ttfamily#1} (#2)\encapchar main}
134 % \index{#2s:\levelchar{\protect\ttfamily#1}\encapchar main}
135 % \@esphack\ignorespaces}
136 % \def\SpecialOptIndex#1#2{
137 % \index{#1\actualchar{\protect\ttfamily#1} (#2)}
138 % \index{#2s:\levelchar{\protect\ttfamily#1}}
139 % \@esphack\ignorespaces}
140 % \def\SpecialOptIndexS#1#2{
141 % \index{#1\actualchar{\protect\ttfamily#1} (#2)}
142 % \index{#2es:\levelchar{\protect\ttfamily#1}}
143 % \@esphack\ignorespaces}
144 %
145 % \def\SpecialMainIndex#1{\@bsphack
146 % \RemoveAtFromCS{#1}
147 % \protected@edef\@tempa{\noexpand\special@index{\RAFCtempa
148 % \actualchar \string\verb
149 % \quotechar*\verbatimchar\string#1\verbatimchar
150 % \encapchar main}}\@tempa
151 % \@esphack}
152 %
153 % \def\SpecialIndex#1{\@bsphack
154 % \RemoveAtFromCS{#1}
155 % \protected@edef\@tempa{\noexpand\special@index{\RAFCtempa
156 % \actualchar \string\verb
157 % \quotechar*\verbatimchar\string#1\verbatimchar}}\@tempa
158 % \@esphack}
159 %
160 % \def\Macro{\begingroup \@sanitize \iMacro}
161 % \def\iMacro#1{\endgroup
162 % \def\SpecialMainIndex##1{\SWiIndex{##1}{\encapchar main}}
163 % \macro{#1}}
164 % \let\endMacro\endmacro
165 %
166 % \def\SpecialWeirdIndex{\@bsphack \begingroup \@sanitize \SWIndex{}}
167 % \def\SWIndex#1#2{\endgroup \SWiIndex{#2}{#1}}
168 % \def\SWiIndex#1#2{%
169 % \protected@edef\@tempa{\noexpand\special@index{
170 % #1\actualchar \string\verb
171 % \quotechar*\verbatimchar#1\verbatimchar#2}}\@tempa
172 % \@esphack}
173 %
174 % \let\RAFCempty\@empty
175 % \let\RAFCgobble\@gobble
176 % \makeatother
177 % \def\RemoveAtFromCS#1{\edef\RAFCtempa{\expandafter\RAFCgobble\string#1}
178 % \edef\RAFCtempa{\expandafter\string\csname\RAFCtempa\endcsname}
179 % \expandafter\iRemoveAtFromCS\RAFCtempa@\NIL}
180 % \def\iRemoveAtFromCS#1#2@#3\NIL{\def\RAFCtempb{#3}
181 % \ifx\RAFCtempb\RAFCempty \def\RAFCtempa{#2}\let\next\relax
182 % \else \def\RAFCtempa{#2@}\def\next{\iiRemoveAtFromCS#3\NIL}
183 % \def\RAFCtempb{#2}
184 % \ifx\RAFCtempb\RAFCempty
185 % \else\ifx\RAFCtempb\CommonCSPrefix
186 % \else\ifx\RAFCtempb\CommonIfPrefix
187 % \else \def\RAFCtempa{}\def\next{\xRemoveAtFromCS#2@#3\NIL}\fi\fi\fi\fi
188 % \next}
189 % \def\iiRemoveAtFromCS#1{\ifx#1@\edef\RAFCtempa{\RAFCtempa@}
190 % \let\next\iiRemoveAtFromCS
191 % \else \def\next{\xRemoveAtFromCS#1}\fi
192 % \next}
193 % \def\xRemoveAtFromCS#1@\NIL{\edef\RAFCtempa{#1\RAFCtempa}}
194 % \makeatletter
195 %
196 % \begingroup
197 % \gdef\@sverbcat{\catcode`\@\active\catcode`\#\active}
198 % \catcode`\@\active \gdef@{\char`\@\penalty\exhyphenpenalty}
199 % \catcode`\#\active \gdef#{\penalty\exhyphenpenalty}
200 % \endgroup
201 % \let\latex@sverb\@sverb
202 % \def\@sverb{\@sverbcat\latex@sverb}
203 %
204 % \newcount\ChangesCounter \ChangesCounter10000
205 % \def\changes@#1#2#3{\global\advance\ChangesCounter\@ne
206 % \protected@edef\@tempa{\noexpand\glossary{#1\levelchar
207 % x\number\ChangesCounter
208 % \ifx\saved@macroname\@empty
209 % \actualchar\generalname
210 % \else
211 % \expandafter\@gobble\saved@macroname
212 % \actualchar \string\verb
213 % \quotechar*\verbatimchar \saved@macroname
214 % \verbatimchar
215 % \fi
216 % :\levelchar #3}}\@tempa\endgroup\@esphack}
217 % \def\theglossary{\onecolumn \glossary@prologue
218 % \GlossaryParms \let\item\@idxitem \ignorespaces}
219 % \def\endtheglossary{\relax}
220 %
221 % \let\EQ\relax \let\GT\relax \let\BAR\relax \let\NEQ\relax
222 %
223 % \makeatother
224 % \errorcontextlines1000
225 %
226 %
227 %
228 % \changes{v1.0}{1993/04/01}
229 % {The style was born on a good day \ldots\ (1993/04/01)}
230 % \changes{v1.05}{1993/06/18}
231 % {Cope with \texttt{\char`\\\char`\\} with negative optional
232 % vertical space. (1993/06/18)}
233 % \changes{v1.1}{1993/06/24}
234 % {Save and restore the \texttt{\char`\\catcode} for
235 % `\texttt{@}'. (1993/06/24)}
236 %
237 % \changes{v1.2}{1998/07/16}
238 % {Various changes shown below. (1998/07/16)}
239 % \changes{v1.2-1}{1998/07/16}
240 % {Add this document.}
241 % \changes{v1.2-2}{1998/07/16}
242 % {Cope with \protect\LaTeXe.}
243 % \changes{v1.2-3}{1998/07/16}
244 % {Allow mixture of vertical solid- and dash-lines.}
245 % \changes{v1.2-4}{1998/07/16}
246 % {Add the feature of explicit dash/gap specification.}
247 % \changes{v1.2-5}{1998/07/16}
248 % {Fix some bugs and change codes.}
249 %
250 % \changes{v1.3}{1998/10/08}
251 % {Fix one bug shown below. (1998/10/08)}
252 %
253 % \changes{v1.4}{1999/06/25}
254 % {Make compatible with array package and add new features. (1999/06/25)}
255 %
256 % \changes{v1.4-1}{1999/06/25}
257 % {The following are changes of this document.}
258 %
259 % \changes{v1.4-2-1}{1999/06/25}
260 % {The following are for the general compatibility with \textsf{array}.}
261 % \changes{v1.4-2-2}{1999/06/25}
262 % {The following are to control the effective width of vertical lines.}
263 % \changes{v1.4-2-3}{1999/06/25}
264 % {The following are for inactivation of dash-line functions.}
265 % \changes{v1.4-2-4}{1999/06/25}
266 % {The following are for drawing mode to cope with the bug of
267 % \cs{xlearders}.}
268 % \changes{v1.4-2-5}{1999/06/25}
269 % {The following are to implement dashed version of \cs{firsthline}
270 % and \cs{lasthline} of \textsf{array}.}
271 % \changes{v1.4-2-6}{1999/06/25}
272 % {The following are to fix the bug by which the depth of
273 % \texttt{array}\slash\texttt{tabular} was always zero.}
274 % \changes{v1.4-2-7}{1999/06/25}
275 % {The following are to rename macros for \cs{cdashline}.}
276 % \changes{v1.4-2-8}{1999/06/25}
277 % {The following are to cope with very narrow or negative wide columns.}
278 % \changes{v1.4-3}{2000/07/04}
279 % {Released to CTAN on 2000/07/04.}
280 %
281 % \changes{v1.5}{2000/07/12}
282 % {Make compatible with \textsf{colortab}, and fix bugs. (2000/07/12)}
283 % \changes{v1.5-1}{2000/07/12}
284 % {The following are for the compatibility with \textsf{colortab}.}
285 % \changes{v1.5-2}{2002/06/26}
286 % {The following are for bug fix of \cs{adl@putlrc}.}
287 % \changes{v1.5-3}{2003/07/11}
288 % {The following are for bug fix of \cs{adl@inactivate}.}
289 %
290 % \changes{v1.54}{2003/08/25}
291 % {Bug fixes. (2003/08/25)}
292 % \changes{v1.54-1}{2003/08/25}
293 % {The following are for bug fix of \cs{adl@@vl}.}
294 % \changes{v1.54-2}{2003/08/25}
295 % {The following are for bug fix of activation.}
296 % \changes{v1.54-3}{2003/08/25}
297 % {The following are miscellaneous modifications.}
298 %
299 % \changes{v1.6}{2003/08/25}
300 % {The following are for the compatibility with
301 % \textsf{longtable}. (2003/08/25)}
302 %
303 % \changes{v1.7}{2004/05/21}
304 % {The following are for the compatibility with
305 % \textsf{colortbl}. (2004/05/21)}
306 % \changes{v1.7-1}{2004/5/21}
307 % {The following are for null-wide horizontal lines.}
308 % \changes{v1.7-2}{2004/5/21}
309 % {The following are to fix the bug of \cs{arrayrulecolor} etc. in
310 % \textsf{colortbl}.}
311 % \changes{v1.7-3}{2004/5/21}
312 % {The following are for vertical line coloring.}
313 %
314 % \changes{v1.71}{2004/7/31}
315 % {The following are for bug fix for \textsf{array}'s
316 % \texttt{m}-columns. (2004/7/31)}
317 %
318 % \changes{v1.72}{2016/03/19}
319 % {Bug fix and revision of \S2.4. (2016/03/19)}
320 % \changes{v1.72-1}{2016/03/19}
321 % {The following are for bug fix for footnotes in \textsf{longtable}'s
322 % \texttt{m}-columns.}
323 %
324 % \changes{v1.73}{2016/04/28}
325 % {Bug fix. (2016/04/28)}
326 % \changes{v1.73-1}{2016/04/28}
327 % {The following are to fix the problem that the top edge a vertical
328 % (dash-)line is at the bottom of a horizontal line rather than it
329 % top.}
330 % \changes{v1.73-2}{2016/04/28}
331 % {The following are to fix the bug that \cs{hdashline}
332 % is not properly processed in a \texttt{array}\slash
333 % \texttt{tabular} environment if \textsf{longtable} is loaded.}
334 % \changes{v1.74}{2018/09/26}
335 % {The following are to fix the bug in the \textsf{array}-compatible
336 % mechanism by which \textsf{delarray} did not work well.}
337 % \changes{v1.75}{2018/12/26}
338 % {The following are to cope with the change in \textsf{array} v2.4i
339 % or later in which \cs{@startpbox} and \cs{@endpbox} have
340 % \cs{color@begingroup} and \cs{color@endgroup}, respectively.}
341 % \changes{v1.76}{2019/02/21}
342 % {The following are to cope with the change in \textsf{longtable}
343 % bundled in \textsf{latex-tools} 2019-01-05 or later in which
344 % \cs{LT@startpbox} and \cs{LT@endpbox} have
345 % \cs{color@begingroup} and \cs{color@endgroup}, respectively.}
346 %
347 %
348 %
349 % \def\EXCLAMATION{\texttt{!}}
350 % \def\VERTBAR{\texttt{\char`\|}}
351 % \edef\CommonCSPrefix{\expandafter\RAFCgobble\string\adl}
352 % \edef\CommonIfPrefix{\expandafter\RAFCgobble\string\ifadl}
353 %
354 % \title{The \textsf{arydshln} package\thanks{
355 % This file has version number \fileversion, last revised \filedate.}}
356 % \author{Hiroshi Nakashima\\(Kyoto University)}
357 % \date{\filedate}
358 % \maketitle
359 %
360 % \begin{abstract}
361 % This file gives \LaTeX's |array| and |tabular| environments the capability
362 % to draw horizontal\slash vertical dash-lines.
363 % \end{abstract}
364 %
365 % \tableofcontents
366 % \newpage
367 %
368 %
369 %
370 % \section{Introduction}
371 % \label{sec:intro}
372 %
373 % In January 1993, Weimin Zhang kindly posted a style \textsf{hvdashln}
374 % written by the author, which draws horizontal\slash vertical dash-lines in
375 % \LaTeX's |array| and |tabular| environments, to the news group
376 % \textsf{comp.text.tex}. The style, unfortunately, has a known problem
377 % that vertical lines are broken when an array contains tall rows.
378 %
379 % In March of the year, Monty Hayes complained of this problem encouraging
380 % the author to make a new version \textsf{arydshln} to solve the problem.
381 % The new style also has new features, such as allowing `|:|' to specify a
382 % vertical dash-line in preamble, and |\cdashline| being a counterpart of
383 % |\cline|.
384 %
385 % \changes{v1.4-1}{1999/06/25}
386 % {The history on the compatibility with \textsf{array} package.}
387 % In March 1999, Sebastian Rahtz kindly invited the style, which had been
388 % improved following the bug report from Takahiro Kubota, to be included in
389 % \TeX{} CTAN and also in the online catalogue compiled by Graham Williams.
390 % This invitation gave the style new users including Peter Ehrbar who wished
391 % to use it with \textsf{array} style in Standard \LaTeX{} Tools Bundle and
392 % had trouble because these styles were incompatible with each other.
393 % Therefore, the style became compatible with \textsf{array} and got
394 % additional new features.
395 %
396 % \changes{v1.5-1}{2000/7/12}
397 % {The history on the compatibility with \textsf{colortab} package.}
398 % In February 2000, Zsuzsanna Nagy reported that \textsf{arydshln} is not
399 % compatible with \textsf{colortab} style to let the author work on the
400 % compatibility issue again.
401 %
402 % \changes{v1.6}{2003/8/25}
403 % {The history on the compatibility with \textsf{longtable} package.}
404 % In Feburary 2001, Craig Leech reported another compatibility problem with
405 % \textsf{longtable}. Although the author promised that the problem would
406 % be attacked some day, the issue had left long time\footnote{
407 %
408 % Two years and a half! Sorry Craig.}
409 %
410 % until three other complaints were made. Then the author attacked the
411 % problem hoping it is the last compatibility issue\footnote{
412 %
413 % But his hope was dashed as described below.}.
414 %
415 % \changes{v1.7}{2004/5/21}
416 % {The history on the compatibility with \textsf{colortbl} package.}
417 % In May 2004, Klaus Dalinghaus found another incompatibility with
418 % \textsf{colortbl}. Although he was satisfied by a quick hack for cell
419 % painting, the author attacked a harder problem for line
420 % coloring to solve the problem\footnote{
421 %
422 % Without dreaming it is the last compatibility issue.}.
423 %
424 %
425 %
426 % \section{Usage}
427 % \label{sec:usage}
428 % \subsection{Loading Package}
429 % \label{sec:usage-load}
430 %
431 % \changes{v1.4-1}{1999/06/25}
432 % {Explanation of package loading is added.}
433 %
434 % The package is usable to both \LaTeXe{} and \LaTeX-2.09 users with their
435 % standard package loading declaration. If you use \LaTeXe, simply do the
436 % following.
437 % \begin{quote}
438 % \SpecialUsageIndex{\usepackage}
439 % |\usepackage{arydshln}|
440 % \end{quote}
441 % If you still love \LaTeX-2.09, the following is what you have to do.
442 % \begin{quote}
443 % \SpecialUsageIndex{\documentstyle}
444 % |\documentstyle[..,arydshln,...]|\Meta{style}
445 % \end{quote}
446 %
447 % \changes{v1.5-1}{2000/7/12}
448 % {Caution about loading order of \textsf{colortab} is added.}
449 % \changes{v1.6}{2003/8/25}
450 % {Caution about loading order of \textsf{longtable} is added.}
451 % \changes{v1.7}{2004/5/21}
452 % {Caution about loading order of \textsf{colortbl} is added.}
453 % \SpecialOptIndex{array}{package}
454 % \SpecialOptIndex{colortab}{package}
455 % \SpecialOptIndex{longtable}{package}
456 % \SpecialOptIndex{colortbl}{package}
457 %
458 % Only one caution given to users of \textsf{array} (v2.3m or later) and
459 % \textsf{longtable} (v4.10 or later) packages, included in Standard \LaTeX{}
460 % Tools Bundle, and \textsf{colortab} and \textsf{colortbl} package is that
461 % \textsf{arydshln} has to be loaded {\em after} \textsf{array},
462 % \textsf{longtable}, \textsf{colortab} and/or \textsf{colortbl} done. That
463 % is, the following is correct but reversing the order of |\usepackage| will
464 % cause some mysterious error.
465 %
466 % \begin{quote}
467 % |\usepackage{array} % and/or|\\
468 % |\usepackage{longtable} % and/or|\\
469 % |\usepackage{colortab} % or|\\
470 % |\usepackage{colortbl}|\\
471 % |\usepackage{arydshln}|
472 % \end{quote}
473 %
474 %
475 %
476 % \subsection{Basic Usage}
477 % \label{sec:usage-basic}
478 %
479 % \DescribeEnv{array}\DescribeEnv{tabular}
480 % \SpecialIndex{\hline}
481 % \SpecialIndex{\cline}
482 % \SpecialIndex{\multicolumn}
483 % You can simply use |array| or |tabular(*)| environments with standard
484 % preamble, such as \verb!{r|c|ll}!, and standard commands |\\|, |\hline|,
485 % |\cline| and |\multicolumn|.
486 %
487 % \leavevmode\marginpar{\raggedleft\PrintDescribeMacro{:}}
488 % \index{:={\protect\ttfamily:}|usage}
489 % Drawing a vertical dash-line is quite simple. Use `|:|' in the preamble
490 % as the separator of columns separated by the dash-line, just like using
491 % `\verb!|!' to draw a vertical solid-line. The {\em preamble} means not
492 % only that of the environment, but also the first argument of
493 % |\multicolumn|.
494 %
495 % \DescribeMacro{\hdashline}\DescribeMacro{\cdashline}
496 % \changes{v1.4-1}{1999/06/25}
497 % {Description of \cs{first}\slash\texttt{lasthdashline} is added.}
498 % It is also simple to draw a horizontal dash-line. Use |\hdashline| and
499 % |\cdashline| as the counterparts of |\hline| and |\cline|.
500 %
501 % For example;
502 %
503 %\begin{quote}
504 %\begin{verbatim}
505 %\begin{tabular}{|l::c:r|}\hline
506 %A&B&C\\\hdashline
507 %AAA&BBB&CCC\\\cdashline{1-2}
508 %\multicolumn{2}{|l:}{AB}&C\\\hdashline\hdashline
509 %\end{tabular}
510 %\end{verbatim}
511 %\end{quote}
512 %
513 % will produce the following result.
514 %
515 %\begin{quote}
516 %\catcode`\|12
517 %\begin{tabular}{|l::c:r|}\hline
518 %A&B&C\\\hdashline
519 %AAA&BBB&CCC\\\cdashline{1-2}
520 %\multicolumn{2}{|l:}{AB}&C\\\hdashline\hdashline
521 %\end{tabular}
522 %\end{quote}
523 %
524 % \changes{v1.73-1}{2016/04/28}
525 % {Add a paragraph describing the perfect contacts of vertical and
526 % horizontal lines.}
527 %
528 % Note that the intersections of leftmost/rightmost vertical lines and
529 % horizontal dash-lines are little bit different from those produced by
530 % ordinary |array|\slash|tabular|. That is, with very careful examination
531 % you will find that vertical lines of ordinary ones are {\em broken\/} with
532 % small white specks at intersections, while in the example above they have
533 % no specks. In addition, the four corners of outermost rectangular also
534 % have specks in ordinary ones, while those in the example above have
535 % perfect contacts of L-shape\footnote{
536 % 
537 % The top-left/right corners had specks before v1.73, the fix in which made
538 % the topmost dash segment of a vertical dash-line a little bit shorter.}.
539 %
540 % \DescribeMacro{\firsthdashline}\DescribeMacro{\lasthdashline}
541 % If you use \textsf{array}, the dashed version of |\firsthline| and
542 % |\lasthline| named |\first#hdash#line| and |\lasthdashline| are available.
543 %
544 %
545 %
546 % \subsection{Style Parameters}
547 % \label{sec:usage-par}
548 %
549 % \DescribeMacro{\dashlinedash}\DescribeMacro{\dashlinegap}
550 % You have two style parameters to control the shape of dash-lines:
551 % |\dashlinedash| is for the length of each dash segment in a dash line;
552 % |\dashlinegap| controls the amount of each gap between dash segments.
553 % Both parameters have a common default value, 4\,pt.
554 %
555 %
556 %
557 % \subsection{Fine Tuning}
558 % \label{sec:usage-fine}
559 %
560 % \leavevmode\marginpar{\raggedleft\PrintDescribeMacro{;}}
561 % \index{;={\protect\ttfamily;}|usage}
562 % Although you can control the shape of dash-lines in an
563 % |array|\slash|tabular| environment as described in \S\ref{sec:usage-par},
564 % you might want to draw a dash-line of a shape different from others. To
565 % specify the shape of a vertical dash-line explicitly, you may use;
566 %
567 % \begin{quote}
568 % |;{|\meta{dash}|/|\meta{gap}|}|
569 % \end{quote}
570 %
571 % instead of ordinary `|:|' and will have a dash-line with dash segments of
572 % \meta{dash} long separated by spaces of \meta{gap}.
573 %
574 % \DescribeMacro{\hdashline}\DescribeMacro{\cdashline}
575 % As for horizontal dash-lines, explicit shape specifications may be given
576 % through optional arguments of |\hdashline| and |\cdashline| as follows.
577 %
578 % \begin{quote}
579 % |\hdashline[|\meta{dash}|/|\meta{gap}|]|\\
580 % |\cdashline{|\meta{col1}|-|\meta{col2}|}[|\meta{dash}|/|\meta{gap}|]|
581 % \end{quote}
582 %
583 %
584 % For example;
585 %
586 %\begin{quote}
587 %\begin{verbatim}
588 %\begin{tabular}{|l::c;{2pt/2pt}r|}\hline
589 %A&B&C\\\hdashline[1pt/1pt]
590 %AAA&BBB&CCC\\\cdashline{1-2}[.4pt/1pt]
591 %\multicolumn{2}{|l;{2pt/2pt}}{AB}&C\\\hdashline\hdashline
592 %\end{tabular}
593 %\end{verbatim}
594 %\end{quote}
595 %
596 % will produce the following result.
597 %
598 %\begin{quote}
599 %\catcode`\|12
600 %\begin{tabular}{|l::c;{2pt/2pt}r|}\hline
601 %A&B&C\\\hdashline[1pt/1pt]
602 %AAA&BBB&CCC\\\cdashline{1-2}[.4pt/1pt]
603 %\multicolumn{2}{|l;{2pt/2pt}}{AB}&C\\\hdashline\hdashline
604 %\end{tabular}
605 %\end{quote}
606 %
607 % \changes{v1.4-1}{1999/06/25}
608 % {Description of the real width of vertical lines is added.}
609 %
610 % \DescribeMacro{\ADLnullwide}
611 % \DescribeMacro{\ADLsomewide}
612 % \SpecialIndex{\arrayrulewidth}
613 %
614 % The vertical solid and dashed lines are drawn as if their width is zero, as
615 % standard \LaTeX's |array| and |tabular| do, if you don't use
616 % \textsf{array} package. Otherwise, they have {\em real} width of
617 % |\arrayrulewidth| as the authors of \textsf{array} prefers. However, you
618 % may explicitly tell \textsf{arydshln} to follow your own preference by
619 % |\ADLnullwide| if you love \LaTeX{} standard, or |\ADLsomewide| if
620 % you second the preference of \textsf{array} authors.
621 %
622 %
623 %
624 % \subsection{Finer Tuning}
625 % \label{sec:usage-finer}
626 %
627 % \changes{v1.4-1}{1999/06/25}
628 % {Description of drawing mode is added.}
629 % \changes{v1.72-2}{1999/06/25}
630 % {Revise \S2.4 reflecting the fix of \cs{xleaders}.}
631 % 
632 % To draw dash-lines, we use a powerful primitive of \TeX{} called |\xleaders|.
633 % It replicates a segment that consist of a dash and gap so that a dash-line
634 % has as many segments as possible and distributes {\it remainder} space to
635 % make the spaces between adjacent dash segments (almost) equal to each
636 % other. Therefore, you will have dash-lines with consistent steps of gaps
637 % and spaces the lines in Figure~\ref{fig:mode}(1) are.
638 %
639 % \begin{figure}
640 % \catcode`\|12
641 % \tabcolsep0pt \setbox\strutbox\hbox{\vrule height14pt depth6pt width0pt}
642 % \def\mb{\makebox[20pt]}
643 % \dashlinedash3.01pt \dashlinegap3.01pt
644 % \def\example#1{\begin{tabular}[b]{|c:c;{2pt/2pt}c|}
645 % \hline
646 % \mb{A}&\mb{A}&\mb{A}\\[-\arrayrulewidth]\hdashline
647 % \mb{B}&\mb{B}&\mb{B}\\[-\arrayrulewidth]\hdashline[2pt/2pt]
648 % \mb{C}&\mb{C}&\mb{C}\\[-\arrayrulewidth]\hline
649 % \multicolumn3c{(#1)}
650 % \end{tabular}}
651 %
652 % \centerline{\example1\qquad
653 % \ADLdrawingmode{2}\example2\qquad
654 % \ADLdrawingmode{3}\example3}
655 % \caption{Drawing mode controlled by \cs{ADLdrawingmode}}
656 % \label{fig:mode}
657 % \end{figure}
658 %
659 % However, because of a bug (or buggy feature) of |\xleaders|, there {\em
660 % had been} a small possibility that a dash segment near the right\slash
661 % bottom end drops, until it was fixed in the version of 3.141592\footnote{
662 % 
663 % By pointing out this problem, the author got a check of \$327.68 plus a
664 % significantly large amount of interest from DEK\@. Wow!!}.
665 % 
666 % Though the fix ultimately made any effort to cope with the problem
667 % unnecessary, the pacakge still gives you alternative {\em drawing modes}
668 % which you may specify by |\ADLdrawingmode|\Meta{$m$} as follows.
669 %
670 % \begin{itemize}
671 % \item $m=1$
672 % \DescribeMacro{\ADLdrawingmode}\\
673 % As shown in Figure~\ref{fig:mode}(1), it gives most beautiful result by
674 % |\xleaders|\footnote{
675 % 
676 % Until the fix of \cs{xleaders}, the second bottom\slash rightmost segments
677 % of right\slash lower lines were dropped.}.
678 % 
679 % This is default.
680 %
681 % \item $m=2$\\
682 % As shown in (2) of the figure, beautiful if dash-lines are not so sparse
683 % as right/lower lines, but dash segments near the both ends may be a little
684 % bit too long as left/upper lines, because in this mode the second
685 % first\slash last segments are drawn by a special mechanism.
686 %
687 % \item $m=3$\\
688 % As shown in (3) of the figure, beautiful if dash-lines are not so sparse
689 % as right/lower lines, but gaps near the both ends may be considerably too
690 % large as left/upper lines, because in this mode the lines are drawn by
691 % |\cleaders|.
692 % \end{itemize}
693 %
694 % It is strongly recommended to use default mode 1 unless you want to have
695 % some special effect.
696 %
697 %
698 %
699 % \subsection{Performance Tuning}
700 % \label{sec:usage-performance}
701 %
702 % \changes{v1.4-1}{1999/06/25}
703 % {Description of (in)activation is added.}
704 %
705 % Since drawing dash-lines is a hard job, you have to be patient with the
706 % fact that the performance of typesetting |array|\slash|tabular| with
707 % dash-lines is poorer than that of ordinary ones. In fact, according to
708 % author's small performance evaluation with a |tabular| having nine
709 % vertical and ten horizontal dash-lines, typesetting the |tabular| is
710 % approximately ten times as slow as its ordinary counterpart with solid
711 % lines.
712 %
713 % However, this is not a really bad news, unfortunately. The real one is
714 % that loading \textsf{arydshln} makes typesetting |array|\slash|tabular|
715 % slower even if they only have solid lines which the package treats as
716 % special ones of dash-lines. The evaluation result shows the degradation
717 % factor is about nine. Therefore, if your document has many
718 % |array|\slash|tabular| with solid lines, \LaTeX{} will run slowly even
719 % with quite few (or no) |array|\slash|tabular| with dash-lines,
720 %
721 % \DescribeMacro{\ADLinactivate}
722 % To cope with this problem, you may inactivate dash-line functions by the
723 % command |\ADLinactivate| that replaces dash-lines with solid
724 % lines drawn by a faster (i.e. ordinary) mechanism. Although the
725 % inactivation does not completely solve the performance problem, the
726 % degradation factor will become much smaller and acceptable, approximately
727 % 1.5 in the author's evaluation. For example, the draft version of your
728 % document will have the command in its preamble, which you will remove from
729 % your final version.
730 %
731 % \DescribeMacro{\ADLactivate}
732 % Alternatively, you may do |\ADLinactivate| in the preamble, switch on by
733 % |\ADLactivate| before you really need dash-lines, and switch off again
734 % afterword. A wiser way could be surrounding |array|\slash|tabular| by
735 % |\begin{ADLactivate}| and |\end{ADLactivate}|.
736 %
737 % \DescribeEnv{Array}
738 % \DescribeEnv{Tabular}
739 % If you feel it tiresome to type the long command\slash environment name for
740 % the activation, you may use |Array| and |Tabular(*)| environment in which
741 % dash-line functions are always active. Note that, however, since these
742 % environment names are too natural to keep them from being used by authors
743 % of other packages or yourself, name conflict could occur. If |Array|
744 % and/or |Tabular| have already been defined when \textsf{arydshln} is
745 % loaded, you will get a warning to show you have to define new
746 % environments, say |dlarray| and |dltabular|, as follows.
747 %
748 % \begin{itemize}\item[]
749 % |\newenvironment{dlarray}{\ADLactivate\begin{array}}%|\\
750 % | {\end{array}}|\\
751 % |\newenvironment{dltabular}{\ADLactivate\begin{tabular}}%|\\
752 % | {\end{tabular}}|\\
753 % |\newenvironment{dltabular*}{\ADLactivate\begin{tabular*}}%|\\
754 % | {\end{tabular*}}|
755 % \end{itemize}
756 %
757 % \begingroup
758 % \let\PDM\PrintDescribeMacro
759 % \def\PrintDescribeMacro#1{\leavevmode\llap{\PDM{#1}}}
760 % \DescribeMacro{\ADLnoshorthanded}
761 % On the other hand, if they are defined after \textsf{arydshln} is loaded,
762 % their definitions are silently replaced or \LaTeX{} complains of multiple
763 % definitions. The error in the latter case will be avoided by putting
764 % |\ADLnoshorthanded| just after |\usepackage{arydshln}|.
765 % \par
766 % \endgroup
767 %
768 %
769 %
770 % \subsection{Compatibility with Other Packages}
771 % \label{sec:usage-compat}
772 %
773 % \changes{v1.5-1}{2000/07/12}
774 % {Section 2.7 is added.}
775 % \changes{v1.4-1}{1999/06/25}
776 % {Description of characters and commands of \textsf{array} package is
777 % added.}
778 % \SpecialOptIndex{array}{package}
779 % \SpecialIndex{\extrarowheight}
780 % \SpecialIndex{\firsthline}
781 % \SpecialIndex{\lasthline}
782 % \SpecialIndex{\newcolumntype}
783 %
784 % Users of \textsf{array} package may use all of newly introduced
785 % preamble characters, such as `|>|', `|<|', `|m|', `|b|', and all the
786 % commands such as |\extrarowheight|, |\firsthilne| and |\lasthline|. The
787 % preamble characters given by \textsf{arydshln} may be included in the
788 % second argument of |\newcolumntype|.
789 %
790 % \changes{v1.5-1}{2000/07/12}
791 % {Description of \textsf{colortab} commands is added.}
792 % \SpecialOptIndex{colortab}{package}
793 % \SpecialIndex{\LCC}
794 % \SpecialIndex{\NAC}
795 % \SpecialIndex{\ENAC}
796 % \SpecialIndex{\AC}
797 % \SpecialIndex{\EAC}
798 %
799 % Also users of \textsf{colortab} package may use |\LCC|\slash |\ECC|
800 % construct to color columns. A horizontal solid\slash dash line may be
801 % colored by, e.g.\ |\NAC\hdashline\ENAC|\@. The pair of |\AC| and |\EAC| may
802 % be used to color everything between them {\em but}, unfortunately,
803 % vertical lines are not. There are no ways to color vertical lines in a
804 % table having dash lines. You may color vertical lines of a ordinary table
805 % inactivating dash line functions by |\ADLinactivate|.
806 %
807 % \changes{v1.7}{2004/5/21}
808 % {Description of \textsf{colortbl} and related commands is added.}
809 % \SpecialOptIndex{colortbl}{package}
810 % \SpecialIndex{\rowcolor}
811 % \SpecialIndex{\columncolor}
812 % \SpecialIndex{\cellcolor}
813 % \SpecialIndex{\arrayrulecolor}
814 % \SpecialIndex{\doublerulesepcolor}
815 %
816 % Another (and more convenient) table coloring tool \textsf{colortbl} may be
817 % also used simply by loading it before \textsf{arydshln}. Not only the
818 % painting commands |\rowcolor|, |\columncolor| and |\cellcolor| work well,
819 % but both solid and dash lines are also colored by the command
820 % |\arrayrulecolor| of \textsf{colortbl}\footnote{
821 %
822 % The \textsf{colortbl} manual says \cs{arrayrulecolor} and
823 % \cs{doublerulesepcolor} may be in \texttt{>\char`\{\ldots\char`\}}
824 % in a preamble but they cause an error with the original implementation.
825 % This bug is fixed in \textsf{arydshln} and they are now usable to specify
826 % the color of the vertical (dash) lines whose specifications occur after
827 % the commands.}.
828 %
829 % One caution is that |\arrayrulecolor| defines the color of the dash-part
830 % of dash lines and thus gap-part has no color (i.e. color of the paper on
831 % which the line drawn). Therefore, if you have a \textsf{tabular} like;
832 %
833 %\begin{quote}
834 %\begin{verbatim}
835 %\begin{tabular}{|>{\columncolor{red}}l:>{\colomncolor{green}}r|}
836 %...
837 %\end{tabular}
838 %\end{verbatim}
839 %\end{quote}
840 %
841 % you will find the vertical dash line is a sequence of black (or the color of
842 % |\arrayrulecolor|) and white segments. This problem is partly solved by
843 % declaring |\ADLnullwide|\footnote{
844 %
845 % Since \textsf{colortbl} automatically loads \textsf{array}, the default is
846 % \cs{ADLsomewide}}
847 %
848 % to conjunct the red and blue columns and to draw the dash line on their
849 % border.
850 %
851 % \DescribeMacro{\ADLnullwidehline}
852 % \DescribeMacro{\ADLsomewidehline}
853 % Unfortunately, however, |\ADLnullwide| does not affect the real
854 % width of horizontal (dash) lines and thus you will still see white gaps
855 % in |\hdashline| and |\cdashline|. A solution is to put
856 % |\ADLnullwidehline| before you start a
857 % \texttt{array}\slash\texttt{tabular}\footnote{
858 %
859 % This command also makes \cs{cline} and \cs{cdashline} visisble even if the
860 % row below is painted.}.
861 %
862 % With this command, a horizontal (dash) line is drawn adjusting its bottom
863 % edge to that of the row above. The command |\ADLsomewidehline| turns the
864 % switch to default and the top edge of a horizontal (dash) line will be
865 % adjusted to the bottom edge of the row above.
866 %
867 % \DescribeMacro{\dashgapcolor}
868 % \DescribeMacro{\nodashgapcolor}
869 % \SpecialIndex{\color}
870 % Another method to avoid white gaps is to give a color to gaps by
871 % |\dashgapcolor| with arguments same as |\color|. For example;
872 %
873 % \begin{quote}
874 % |\arrayrulecolor{green}\dashgapcolor[rgb]{1,1,0}|
875 % \end{quote}
876 %
877 % makes colorful dash lines with green dashes and yellow gaps. The command
878 % can be placed outside of \texttt{array}\slash\texttt{tabular} for dash
879 % lines in the environment, in the argument of preamble character |>|
880 % for vertical dash lines following them, or at the beginning of a row for
881 % horizontal dash lines following the command. The commmand
882 % |\nodashgapcolor| (no arguments) nullifies the effect of |\dashgapcolor|.
883 % Note that |\nodashgapcolor| is different from |\dashgapcolor{white}|
884 % because the former makes gaps {\it transparent} while the later whiten them.
885 %
886 % \DescribeEnv{longtable}
887 % \DescribeEnv{Longtable}
888 % \changes{v1.6}{2003/8/25}
889 % {Description of \textsf{longtable} is added.}
890 % \SpecialOptIndex{longtable}{package}
891 % Usage of \textsf{longtable} with \textsf{arydshln} is quite simple. Just
892 % loading \textsf{arydshln} after \textsf{longtable} is enough to make the
893 % |longtable| environment able to draw dash-lines. A shorthand activation
894 % of dash-line functions is also available by |Longtable| environment. One
895 % caution to \textsf{longtable} users is that the temporary results before
896 % the {\em convergence} of the column widths may be different from those
897 % without \textsf{arydshln}. For example, the following is the first pass
898 % result of the example shown in Table~3 of the \textsf{longtable} manual.
899 %
900 % \begin{quote}
901 % \catcode`\|12
902 % \leavevmode\vtop{\offinterlineskip
903 % \hbox{\begin{tabular}{|c|c|c|}\hline 1&2&3\end{tabular}}
904 % \hbox{\begin{tabular}{|c|}wide multicolumn spanning 1--3\end{tabular}}
905 % \hbox{\begin{tabular}{|c|c|}\omit&
906 % \omit\vrule height0pt width116.66695pt\cr
907 % multicolumn 1--2&3\end{tabular}}
908 % \hbox{\begin{tabular}{|c|c|c|}\omit&
909 % \omit\vrule height0pt width68.05573pt&
910 % \omit\vrule height0pt width116.66695pt\cr
911 % wide 1&2&3\\\hline\end{tabular}}}
912 % \end{quote}
913 %
914 % \SpecialOptIndex{LTchunksize}{counter}
915 % Since |LTchunksize| is one in the example, columns of each row has their own
916 % widths and thus has vertical lines drawn at the edges of the columns. On
917 % the other hand, you will have the following as the first pass result with
918 % \textsf{arydshln}.
919 %
920 % \begin{quote}
921 % \catcode`\|12
922 % \settowidth\unitlength{\begin{tabular}{|c|c|c|}\hline 1&2&3\end{tabular}}
923 % \begin{tabular}[t]{|l|c|c|}
924 % \multispan3\relax
925 % \leaders\hrule height\arrayrulewidth\hskip\unitlength\cr
926 % \hbox to0pt{1\hskip2\tabcolsep 2\hskip2\tabcolsep 3\hss}&&\\
927 % \multicolumn3{|l|}{wide multicolumn spanning 1--3}\\
928 % \multicolumn2{|l|}{\hbox to0pt{multicolumn 1--2\hskip\tabcolsep
929 % \hbox to116.66695pt{\hss3\hss}\hss}}&\\
930 % wide 1&2&3\\\hline
931 % \omit&
932 % \omit\vrule height0pt width68.05573pt&
933 % \omit\vrule height0pt width116.66695pt\cr
934 % \end{tabular}
935 % \end{quote}
936 %
937 % As you see, the vertical lines are drawn at the column edges of the last
938 % row\footnote{
939 %
940 % More precisely, drawn according to the column widths established by all
941 % the chunks preceding page output.}
942 %
943 % because \textsf{arydshln} draws them when it see the last row. Anyway,
944 % you may ignore temporary results and will have a compatible result when
945 % the column widths are converged like the following.
946 %
947 % \begin{quote}
948 % \catcode`\|12
949 % \begin{tabular}[t]{|c|c|c|}\hline
950 % 1&2&3\\
951 % \multicolumn{3}{|c|}{wide multicolumn spanning 1--3}\\
952 % \multicolumn{2}{|c|}{multicolumn 1--2}&3\\
953 % wide 1&2&3\\\hline
954 % \end{tabular}
955 % \end{quote}
956 %
957 %
958 %
959 % \section{Known Problems}
960 % \label{sec:prob}
961 %
962 % There are following known problems.
963 %
964 % \begin{enumerate}
965 % \item
966 % The new preamble specifiers `|:|' and `|;{|\meta{dash}|/|\meta{gap}|}|'
967 % cannot be followed or preceded by `|@{|\meta{text}|}|', or you will have
968 % an ugly result. More specifically, a specifier to draw a dash-line at the
969 % left edge of a column cannot be preceded by `|@{|\meta{text}|}|', while
970 % that to draw at the right edge cannot be followed by `|@{|\meta{text}|}|'.
971 %
972 % \item\leavevmode
973 % \changes{v1.4-1}{1999/06/25}
974 % {Description about `\protect\EXCLAMATION' of array package is added.}
975 %
976 % If you use \textsf{array} package, the restriction of `|@|' shown above is
977 % also applied to `|!|'.
978 %
979 % \item
980 % In order to make it sure that a dash-line always {\em touches} its both
981 % end, i.e. a dash-line always begins and ends with a dash segment, the
982 % amount of a gap will slightly vary depending on the dash-line length.
983 %
984 % \changes{v1.4-1}{1999/06/25}
985 % {Reference to the section for drawing mode is added.}
986 % \changes{v1.72-2}{1999/06/25}
987 % {Remove the caution about the dash segment dropping.}
988 %
989 % \item
990 % \changes{v1.4-1}{1999/06/25}
991 % {Description on minimum length is added.}
992 %
993 % If a dash-line is too short, you will have an ugly result without overfull
994 % message. More specifically, in mode 1 or 3, a line will look to protrude
995 % beyond its column\slash row borders if it is shorter than a half of
996 % |\dashlinedash|. In mode 2, the minimum length to avoid the protrusion is
997 % $1.5\times|\dashlinedash|+ |\dashlinegap|$.
998 %
999 % \item
1000 % \changes{v1.4-1}{1999/06/25}
1001 % {Reference to the performance tuning section is added.}
1002 %
1003 % As described in \S\ref{sec:usage-performance},
1004 % the processing speed for |array| and |tabular| environment will become
1005 % slower even if dash-lines are not included.
1006 %
1007 % \item
1008 % \changes{v1.5-1}{2000/07/12}
1009 % {Caution about \cs{AC}/\cs{EAC} pair for vertical line coloring is
1010 % added.}
1011 % \changes{v1.7}{2004/5/21}
1012 % {Comment on vertical line coloring with \textsf{colortbl} is added.}
1013 %
1014 % As described in \S\ref{sec:usage-compat}, |\AC| and |\EAC| pair of
1015 % \textsf{colortab} such as |\AC&\EAC| cannot color the vertical line at
1016 % |&|. Use |\ADLinactivate| if you want to have a ordinary table with
1017 % colored vertical lines. Note that you may color vertical lines with
1018 % \textsf{colortbl} package.
1019 %
1020 % \item
1021 % \changes{v1.74}{2018/09/26}
1022 % {Comment on \texttt{plextarydshln} is added.}
1023 % There should be a number of packages whose own |array|/|tabular|
1024 % implementations are not compatible with \textsf{arydshln}, though the
1025 % author has made efforts at the compatibility. One of them is
1026 % \textsf{plext} package for Japanese typesetting but it has a style file
1027 % named |plextarydshln.sty| to solve the compatibility issue. So if you use
1028 % the functionality of \textsf{arydshln} with \textsf{plext}, do
1029 % |\usepackage{plextarydshln}| instead of |\usepackage{arydshln}|.
1030 % \end{enumerate}
1031 %
1032 %
1033 %
1034 % \IndexPrologue{\newpage\section*{Index}
1035 % Italicized number refers to the page where the specification of
1036 % corresponding entry are described.}
1037 % \StopEventually{
1038 % \section*{Acknowledgments}
1039 %
1040 % \changes{v1.4-1}{1999/06/25}
1041 % {Thank to more people.}
1042 % \changes{v1.5-3}{2003/07/11}
1043 % {Thank to Yaxin Liu.}
1044 % \changes{v1.6}{2003/08/25}
1045 % {Thank to people for \textsf{longtable}.}
1046 % \changes{v1.7}{2004/5/21}
1047 % {Thank to Klaus Dalinghaus and refer orignal \textsf{colortbl}.}
1048 % \changes{v1.71}{2004/8/31}
1049 % {Thank to Morten H\o gholm.}
1050 % \changes{v1.72-1}{2016/03/19}
1051 % {Thank to Ma\"ieul Rouquette.}
1052 % \changes{v1.73}{2016/03/19}
1053 % {Thank to Ma\"ieul Rouquette again.}
1054 % \changes{v1.74}{2018/09/26}
1055 % {Thank to Hironobu Yamashita.}
1056 % \changes{v1.74}{2018/09/26}
1057 % {Thank to Hironobu Yamashita for coloring problem.}
1058 %
1059 % The author thanks to Monty Hayes who gave the author the opportunity to
1060 % make this style, and Weimin Zhang and Takahiro Kubota who pointed out bugs
1061 % in early versions. He also thanks to the following people; Sebastian
1062 % Rahtz and Graham Williams who kindly invited the style to \TeX{} CTAN and
1063 % online catalogue compiled by Graham; Peter Ehrbar who showed the style was
1064 % incompatible with \textsf{array} and kindly accepted the offer to be an
1065 % alpha-user of v1.4 alone; Zsuzsanna Nagy who reported another
1066 % incompatibility problem with \textsf{colortab}; Ralf Heydenreich who
1067 % reported the bug causing that glues in a column have no effect; Yaxin
1068 % Liu who reported the incompatibility bug of \textsf{array} and
1069 % \cs{ADLinactivate}; Craig Leech who reported the incompatibility
1070 % problem with \textsf{longtable}, which was also reported by Uwe Jehmlich,
1071 % Torge Thielemann and Florian Weig, and had waited for two years and a
1072 % half (!)\ for the solution; Klaus Dalinghaus who reported yet another
1073 % incompatibility with \textsf{colortbl}; Morten H\o gholm who reported the
1074 % bug of |m|-type columns of \textsf{array} which had not manifested in five
1075 % (!!)\ years since the author realesed the first \textsf{array}-compatible
1076 % version; Ma\"ieul Rouquette who reported another bug of |m|-type
1077 % columns of \textsf{longtable} with \textsf{array} which had peacefully
1078 % hidden in the package for eleven years and a half (!!!)\ since the author
1079 % made the bug fix shown above carelessly, yet another bug related to
1080 % \textsf{longtable}, and most surprisingly a problem on intersections of
1081 % horizontal and vertical (dash-)lines which has hidden for 23 years (!!!!)
1082 % since the very first version of the package; and Hironobu Yamashita who
1083 % pointed out bugs hidden for 19 years (!!!!!) by which \textsf{delarray}
1084 % did not work, and compatibility problems with \textsf{array} v2.4i and
1085 % \textsf{longtable} in \textsf{latex-tools} 2019-01-05.
1086 %
1087 % The base implementation of \texttt{array} and \texttt{tabular}
1088 % environments, part of which the author gives new definitions referring
1089 % original ones, are written by Leslie Lamport as a part of \LaTeX-2.09 and
1090 % \LaTeXe{} (1997/12/01) to which Johannes Braams and other authors also
1091 % contributed. The author also refers \textsf{array} package (v2.4j)
1092 % written by Frank Mittelbach and David Carlisle; \textsf{colortab} package
1093 % (v0.9) written by Timothy van Zandt; and \textsf{longtable} (v4.11) and
1094 % \textsf{colortbl} (v0.1j) packages
1095 % written by David Carlisle; to make the style compatible with those
1096 % packages.
1097 %
1098 % \PrintIndex}
1099 %
1100 %
1101 %
1102 % \newpage
1103 % \section{Implementation}
1104 % \label{sec:imp}
1105 % \subsection{Problems and Solutions}
1106 % \label{sec:imp-ps1}
1107 %
1108 % \changes{v1.4-1}{1999/06/25}
1109 % {The title of section 4.1 is changed.}
1110 % \changes{v1.4-1}{1999/06/25}
1111 % {\cs{hfil} is replaced with \cs{hss} taking the possibility of
1112 % negative wide columns into account.}
1113 % \changes{v1.7}{2004/5/21}
1114 % {Add notes for dash line coloring.}
1115 %
1116 % We have two different problems to solve; how to draw horizontal dash-lines
1117 % and how to draw vertical dash-lines. The former problem is relatively
1118 % easy because the technique for drawing |\cline|-s can be used. That is, if
1119 % we know the number of columns, we can draw a dash-line across the
1120 % |\multispan|-ed columns by |\xleaders| of dash. Modifying a preamble of
1121 % |array|\slash|tabular| to count the number of columns is not hard. Since
1122 % |\cdashline| is given beginning and ending columns, its implementation is
1123 % also easy.
1124 %
1125 % \SpecialIndex{\@arstrutbox}
1126 %
1127 % The latter problem, however, is much harder. Remember that
1128 % |array|\slash|tabular| draws vertical solid lines by |\vrule|-s in each row
1129 % without height\slash depth specification exploiting \TeX's sophisticated
1130 % mechanism of the rule extension in the surrounding box. Since \TeX{} does
1131 % not have such a mechanism for |\xleaders| unfortunately, we at least have
1132 % to know the height and depth of a row which includes vertical dash-lines.
1133 % Although the height and depth are often same as those of |\@arstrutbox|, we
1134 % will have an exceptionally tall and/or deep row that makes dash-lines {\em
1135 % broken} if we assume every row has the standard height and depth.
1136 %
1137 % Moreover, even if we can measure the height\slash depth of each row (in
1138 % fact we will do as described later), drawing dash-lines in each row will
1139 % not produce a good result. Look at the following two examples closely.
1140 %
1141 %\begin{quote}
1142 %\makeatletter
1143 %\begin{tabular}{c:c}\hline
1144 %A&B\\\hline
1145 %\noalign{\xdef\adl@rowsL{\adl@rowsL(\relax/0);}
1146 % \xdef\adl@rowsR{\adl@rowsR(\relax/0);}}
1147 %{\vrule height2\ht\strutbox depth2\dp\strutbox width\z@\LARGE A}&B\\\hline
1148 %\end{tabular}\qquad
1149 %\begin{tabular}{c:c}\hline
1150 %A&B\\\hline
1151 %{\vrule height2\ht\strutbox depth2\dp\strutbox width\z@\LARGE A}&B\\\hline
1152 %\end{tabular}
1153 %\end{quote}
1154 %
1155 % In the left example, two dash-lines are individually drawn in two rows.
1156 % Since the first row is not so tall and deep (8.4\,pt\slash
1157 % 3.6\,pt) as to contain enough number of default dash segments (4\,pt dash and
1158 % 4\,pt gap) to keep |\xleaders| from inserting a large space, the dash-line
1159 % in the first row is {\em sparse}. On the other hand, the second row is
1160 % enough tall and deep (16.8\,pt\slash 7.2\,pt) and thus the dash-line in the
1161 % row looks better. Thus the resulting dash-line is awful because it does
1162 % not have a continuous dash/gap sequence.
1163 %
1164 % The right example, which we wish to produce, is much better than the left.
1165 % In this example, the dash line is drawn across two rows keeping continuous
1166 % steps of dashes and gaps. In order to have this result, we have to draw
1167 % the dash-line {\em after} two rows are built because it is necessary to
1168 % know the total height and depth of two rows. In general, if we know the
1169 % total height and depth of rows and whether a column has a dash-line, we can
1170 % draw dash-lines by adding an extra row containing dash-lines. For
1171 % example, the result shown above is obtained by the following row.
1172 %
1173 % \begin{quote}
1174 % |\omit\hss|\meta{dash-line of 36\,pt high}|&\omit\cr|
1175 % \end{quote}
1176 %
1177 % Note that \meta{dash-line of 36\,pt high} have to be |\smash|-ed.
1178 %
1179 % In addition to this basic scheme, we have to take the following points
1180 % into account.
1181 %
1182 % \begin{itemize}
1183 % \item
1184 % A dash-line drawn by the preamble character `|;|' will have non-default
1185 % dash\slash gap specification.
1186 %
1187 % \item
1188 % A column may have two or more dash-lines separated by spaces of
1189 % |\doublerulesep|. Mixed sequence of solid- and dash-lines also have to be
1190 % allowed.
1191 %
1192 % \item
1193 % The first column may have dash-lines at both ends, while those of others
1194 % will appear at right ends only. An exception of this rule is brought by
1195 % |\multicolumn| that may have leading sequence of solid- and/or dash-line
1196 % specifiers in its preamble.
1197 %
1198 % \item
1199 % A |\multicolumn| may break or add a dash-line, or may change the
1200 % dash\slash gap specification of a dash-line. A sequence of
1201 % |\h(dash)line|-s also break dash-lines.
1202 %
1203 % \item
1204 % If \textsf{colortbl} is in use, coloring dash\slash gap by
1205 % |\arrayrulecolor| and |\dashgapcolor| gives another possibility of the
1206 % variation of dash\slash gap specification.
1207 % \end{itemize}
1208 %
1209 % In order to cope with them, the following data structure is constructed
1210 % during rows are built.
1211 %
1212 % \begin{enumerate}
1213 % \item
1214 % The list of row information $R=\langle r_1,r_2,\ldots,r_N\rangle$.
1215 %
1216 % \item
1217 % The $i\th$ element of $R$, $r_i$, is one of the following\footnote{
1218 %
1219 % In the real implementation, the structure of $r_i$ is slightly different.}.
1220 %
1221 % \begin{enumerate}
1222 % \item
1223 % A triple $\langle C^L_i,C^R_i,h_i\rangle$, where $C^L_i$ and $C^R_i$ are
1224 % the lists of solid- or dash-line segments drawn at the left and right edge
1225 % of columns respectively, and $h_i$ is the height plus depth of the $i\th$
1226 % row.
1227 %
1228 % \item
1229 % ${\it connect}(h_i)$ for a |\h(dash)line| of $h_i$ wide meaning that $r_i$
1230 % is an empty pseudo row of $h_i$ high and dash-lines are not broken at the
1231 % row.
1232 %
1233 % \item
1234 % \changes{v1.6}{2003/8/25}
1235 % {Description of \textit{discard} is added.}
1236 % In |longtable| environment, ${\it discard}(h_i)$ for a negative vertical
1237 % space inserted by |\\[|\meta{$h_i$}|]| or |\h(dash)line| meaning $r_i$
1238 % is an empty pseudo row of $h_i$ high and dash-lines are not broken but may
1239 % be discarded by the page break at the row.
1240 %
1241 % \item
1242 % ${\it disconnect}(h_i)$ for a vertical gap generated by a sequence of
1243 % |\h(dash)line| meaning that $r_i$ is an empty pseudo row of $h_i$ high and
1244 % dash-lines are broken at the row.
1245 % \end{enumerate}
1246 %
1247 % \item
1248 % $C^L_i=\langle e^i_1,e^i_2,\ldots,e^i_m\rangle$ where $e^i_j$ corresponds
1249 % to the $j\th$ (leftmost is first) solid- or dash-line segment. $C^R_i$ is
1250 % similar but its elements are ordered in reverse, i.e. the rightmost segment
1251 % is the first element.
1252 %
1253 % \item
1254 % \changes{v1.7}{2004/5/21}
1255 % {A dash/gap specification $d^i_j/g^i_j$ now has color.}
1256 % The $j\th$ element of $C^L_i$ or $C^R_i$, $e^i_j$, is a triple $\langle
1257 % c^i_j,d^i_j,g^i_j\rangle$ where $c^i_j$ is the column number in which the
1258 % segment appears, and $d^i_j$ and $g^i_j$ are dash\slash gap specification,
1259 % length and color, of the segment. For a solid line segment, the length
1260 % attributes of both $d^i_j$ and $g^i_j$ are 0.
1261 % \end{enumerate}
1262 %
1263 % Then this data structure is processed to draw solid- and dash-lines at the
1264 % end of the |array|\slash|tabular| as follows. Let $e^i_j=\langle c^i_j,
1265 % d^i_j, g^i_j\rangle$ be the $j\th$ element of $C^L_i$ of $r_i$. The {\em
1266 % position} $p^i_j$ of $e^i_j$ in the column $c^i_j$ is defined as follows.
1267 %$$
1268 % p^i_j = \cases{1& if $j=1 \LOR c^i_j\neq c^i_{j-1}$\cr
1269 % p^i_{j-1}+1& otherwise}.
1270 %$$
1271 % The following defines whether two elements $e^i_j$ and $e^{i'}_{j'}$ are
1272 % {\em connected}, or $e^i_j\sim e^{i'}_{j'}$.
1273 %
1274 % \begin{eqnarray*}
1275 % e^i_j\sim e^{i'}_{j'}&
1276 % \togets&i<i'\LAND\\
1277 % && c^i_j=c^{i'}_{j'}\LAND d^i_j=d^{i'}_{j'}\LAND g^i_j=g^{i'}_{j'}\LAND
1278 % p^i_j=p^{i'}_{j'}\LAND\\
1279 % && ^\forall k(i<k<i'\to r_k\in
1280 % \{\mathit{connect}(h_k),\mathit{discard}(h_k)\}).
1281 % \end{eqnarray*}
1282 %
1283 % With these definitions, we can classify all $e^i_j$ into ordered sets
1284 % $S_1,S_2,\ldots S_n$ as follows.
1285 %
1286 % \def\PH{\phantom{^\forall e^i_j\in S_k,\forall e^{i'}_{j'}\in S_{k'}:{}}}
1287 % \begin{eqnarray*}
1288 % k\neq k'&\togets&
1289 % S_k\cap S_{k'}=\emptyset\\
1290 % e^i_j\sim e^{i'}_{j'}&\togets&
1291 % ^\exists k:e^i_j,e^{i'}_{j'}\in S_k\LAND
1292 % S_k=\{\ldots,e^i_j,e^{i'}_{j'},\ldots\}\\
1293 % k<k'&\togets&
1294 % ^\forall e^i_j\in S_k,\forall e^{i'}_{j'}\in S_{k'}:
1295 % (c^i_j<c^{i'}_{j'})\LOR\\
1296 % && \PH (c^i_j=c^{i'}_{j'}\LAND p^i_j<p^{i'}_{j'})\LOR\\
1297 % && \PH (c^i_j=c^{i'}_{j'}\LAND p^i_j=p^{i'}_{j'}\LAND i<i').
1298 % \end{eqnarray*}
1299 %
1300 % Now we can draw a dash-line $L_k=\langle\gamma_k, \pi_k, \delta_k, \xi_k,
1301 % \tau_k, \beta_k\rangle$ corresponding to $S_k=\{e^i_j,\ldots,e^{i'}_{j'}\}$
1302 % as follows.
1303 %
1304 % \begin{itemize}
1305 % \item
1306 % $L_k$ is $\pi_k\th$ line in the $\gamma_k\th$ column where $\gamma_k=
1307 % c^i_j=\ldots=c^{i'}_{j'}$ and $\pi_k=p^i_j=\ldots=p^{i'}_{j'}$.
1308 %
1309 % \item
1310 % $L_k$ has the dash specification (size and color)
1311 % $\delta_k=d^i_j=\ldots=d^{i'}_{j'}$ and gap specification
1312 % $\xi_k=g^i_j=\ldots=g^{i'}_{j'}$.
1313 %
1314 % \item
1315 % \changes{v1.73-1}{2016/04/28}
1316 % {Add the definition of $\eta_l$ and addition\slash subtraction of it
1317 % for $\tau_k$ and $\beta_k$.}
1318 %
1319 % The top and bottom ends of $L_k$ are at $\tau_k$ and $\beta_k$ above the
1320 % bottom of the |array|\slash|tabular|, where;
1321 %$$
1322 % \eta_l=\cases{h_l&$r_l=\mathit{connect(h_l)}$\cr
1323 % 0&otherwise},\quad
1324 % \tau_k=\eta_{i-1}+\sum_{l=i}^N h_l,\quad
1325 % \beta_k=-\eta_{i'+1}+\sum_{l=i'+1}^N h_l.
1326 %$$
1327 % Note that $\eta_{i-1}$ and $\eta_{i'+1}$ are added\slash subtracted so that
1328 % the top\slash bottom of $L_k$ is at the top\slash bottom edge of the
1329 % horizontal lines above\slash below the set $S_k$.
1330 % \end{itemize}
1331 %
1332 % The row to draw $L_1,\ldots,L_n$ is;
1333 %$$
1334 % \sigma_1 L_1 \sigma_2 L_2 \ldots L_{n-1}\sigma_n L_n\sigma_{n+1}|\cr|
1335 %$$
1336 % where;
1337 %
1338 % \def\amp{\hbox{\tt\&}}
1339 % \begin{eqnarray*}
1340 % \sigma_1&=&
1341 % \cs{omit}[\cs{hss}\amp\cs{omit}]^{\gamma_1-1}\\[1ex]
1342 % \sigma_{1<k\leq n}&=&
1343 % \cases{ \cs{null}&
1344 % if $\gamma_{k-1}=\gamma_k\LAND \pi_{k-1}=\pi_k$\cr
1345 % \cs{hskip}\cs{doublerulesep}&
1346 % if $\gamma_{k-1}=\gamma_k\LAND \pi_{k-1}\neq\pi_k$\cr
1347 % [\cs{hss}\amp\cs{omit}]^{\gamma_k-\gamma_{k-1}}&
1348 % if $\gamma_{k-1}\neq\gamma_k$\cr}\\[1ex]
1349 % \sigma_{n+1}&=&
1350 % [\cs{hss}\amp\cs{omit}]^{{\mit\Gamma}-\gamma_n-1}\cs{hss}.
1351 % \end{eqnarray*}
1352 %
1353 % Note that $[x]^m$ means $m$-times iteration of $x$, and $\mit\Gamma$ is
1354 % the number of columns specified in the preamble.
1355 %
1356 % Dash-lines at the right edges of columns are similarly drawn by processing
1357 % $C^R_i$ with the following modifications.
1358 %
1359 % \begin{eqnarray*}
1360 % k<k'&\togets&
1361 % ^\forall e^i_j\in S_k,\forall e^{i'}_{j'}\in S_{k'}:
1362 % (c^i_j<c^{i'}_{j'})\LOR\\
1363 % && \PH (c^i_j=c^{i'}_{j'}\LAND p^i_j>p^{i'}_{j'})\LOR\\
1364 % && \PH (c^i_j=c^{i'}_{j'}\LAND p^i_j=p^{i'}_{j'}\LAND i<i')\\
1365 % \sigma_1&=&
1366 % \cs{omit}\cs{hss}[\amp\cs{omit}\cs{hss}]^{\gamma_1-1}\\[1ex]
1367 % \sigma_{k>1}&=&
1368 % \cases{ \cs{null}&
1369 % if $\gamma_{k-1}=\gamma_k\LAND \pi_{k-1}=\pi_k$\cr
1370 % \cs{hskip}\cs{doublerulesep}&
1371 % if $\gamma_{k-1}=\gamma_k\LAND \pi_{k-1}\neq\pi_k$\cr
1372 % [\amp\cs{omit}\cs{hss}]^{\gamma_k-\gamma_{k-1}}&
1373 % if $\gamma_{k-1}\neq\gamma_k$\cr}\\[1ex]
1374 % \sigma_{n+1}&=&
1375 % [\amp\cs{omit}\cs{hss}]^{{\mit\Gamma}-\gamma_n-1}
1376 % \end{eqnarray*}
1377 %
1378 %
1379 %
1380 % \subsection{Another Old Problem}
1381 % \label{sec:imp-ps2}
1382 % 
1383 % \changes{v1.72-2}{1999/06/25}
1384 % {Change the title of \S4.2 and rephrase sentences according to the
1385 % fix of \cs{xleader}'s problem.}
1386 %
1387 % In the default mode 1, we draw a dash line of dash size $d$ and gap size
1388 % $g$ as follows. Let $W$ be the length of the line plus 10\,|sp|\footnote{
1389 %
1390 % This small amount is added by \cs{xleaders} in order to, according to the
1391 % comment in \texttt{tex.web}, compensate floating point rounding error.},
1392 %
1393 % which is unknown for us if horizontal but known for \TeX, and assume
1394 % $W\geq d/2$ (or the line protrude to the column\slash row boarder.) At
1395 % the both ends of the columns, dashes of $d/2$ long are drawn to make the
1396 % dash-line {\em touched} to the ends. Then $n=\lfloor(W-d-g)/(d+g)\rfloor$
1397 % dashes are equally distributed in the remaining space. Thus we will have;
1398 % $$
1399 % D_0(d/2)G_0(g+\epsilon')D_1(d)G_1(g+\epsilon)\ldots
1400 % G_{n-1}(g+\epsilon)D_n(d)G_n(g+\epsilon')D_{n+1}(d/2)
1401 % $$
1402 % where $D_i(l)$ and $G_i(l)$ are dash and gap of $l$ long,
1403 % $\epsilon=(W-(n+1)(d+g))/(n+1)$ (rounded), and
1404 % $\epsilon'=(W-(n+1)(d+g)-(n-1)\epsilon)/2$ to compensate the rounding
1405 % error on the calculation of $\epsilon$. For a horizontal line, this
1406 % result will be obtained by |\xleaders| as follows where $G^m_i(\epsilon)$
1407 % and $G^m_i(\epsilon')$ are the spaces inserted by |\xleaders|.
1408 %
1409 % \begin{eqnarray*}
1410 % \rlap{$\displaystyle D_0(d/2)G^l_0(g/2)
1411 % \hbox{\cs{xleaders}\cs{hbox}{\tt\char`\{}}
1412 % G^r(g/2)D(d)G^l(g/2)\hbox{\tt\char`\}\cs{hss}\ }
1413 % G^r_n(g)D_{n+1}(d/2)$}\qquad\\
1414 % &=&D_0(d/2)G^l_0(g/2)G^m_0(\epsilon')
1415 % \left(G^r_0(g/2)D_1(d)G^l_1(g/2)\right)G^m_1(\epsilon)\\
1416 % && \phantom{D_0(d/2)G^l_0(g/2)G^m_0(\epsilon)}
1417 % \left(G^r_1(g/2)D_2(d)G^l_2(g/2)\right)G^m_2(\epsilon)\\
1418 % && \phantom{D_0(d/2)G^l_0(g/2)G^m_0(\epsilon)}\ldots\\
1419 % && \phantom{D_0(d/2)G^l_0(g/2)G^m_0(\epsilon)}
1420 % \llap{$\displaystyle G^m_{n-1}(\epsilon)$}
1421 % \left(G^r_{n-1}(g/2)D_n(d)G^l_n(g/2)\right)
1422 % G^m_n(\epsilon')G^r_n(g/2)D_{n+1}(d/2)\\
1423 % &=&D_0(d/2)G_0(g+\epsilon')D_1(d)G_1(g+\epsilon)\ldots
1424 % G_{n-1}(g+\epsilon)D_n(d)G_n(g+\epsilon')D_{n+1}(d/2)
1425 % \end{eqnarray*}
1426 %
1427 % The problem is that $\epsilon'$ could be negative and old \TeX{} mistakingly
1428 % ignored this possibility. That is, since the \TeX{} older than 3.141592
1429 % did not put |\hbox| beyond the right edge of |\xleaders|, the rightmost
1430 % |\hbox| was omitted if $\epsilon'$ is negative.
1431 %
1432 % Since it is (almost) impossible to know the length of a horizontal line,
1433 % we could not cope with this problem by adding or subtracting its length.
1434 % Thus we introduced {\em drawing mode} to have imperfect solutions. In the
1435 % mode 2, we draw a line by the following sequence.
1436 %
1437 % \begin{eqnarray*}
1438 % \rlap{$\displaystyle D_0(d/2)G^l_0(g/2)G^r_{0'}(g/2)D_{1'}(d)G^l_{1'}(g/2)
1439 % G(-d-g)$}\qquad\\
1440 % && \hbox{\cs{xleaders}\cs{hbox}{\tt\char`\{}}
1441 % G^r(g/2)D(d)G^l(g/2)\hbox{\tt\char`\}\cs{hss}\ }\\
1442 % && G(-d-g)G^r_{n'}(g/2)D_{n'}(d)G^l_{n'}(g/2)G^r_n(g)D_{n+1}(d/2)
1443 % \end{eqnarray*}
1444 %
1445 % That is, $n\th$ |\hbox| that could be disappeared is put twice and the
1446 % first one is also overlaid for symmetrization. Therefore the length of the
1447 % first and $n\th$ dashes is $d+\vert\epsilon'\vert$ and thus could be a
1448 % little bit longer than others.
1449 %
1450 % On the other hand, we replace |\xleaders| of mode 1 with |\cleaders| for
1451 % the drawing in mode 3. The result will be;
1452 % $$
1453 % D_0(d/2)G_0(g+R)D_1(d)G_1(g)\ldots
1454 % G_{n-1}(g)D_n(d)G_n(g+R)D_{n+1}(d/2)
1455 % $$
1456 % where $R=(W-(n+1)(d+g))/2$ to make the first and last gaps considerably wider
1457 % than others.
1458 %
1459 %
1460 %
1461 % \iffalse
1462 %<*package>
1463 % \fi
1464 %
1465 % \subsection{Register Declaration}
1466 % \label{sec:imp-decl}
1467 %
1468 % Here registers and switches are declared.
1469 %
1470 % \begin{macro}{\dashlinedash}
1471 % \begin{macro}{\dashlinegap}
1472 % \begin{macro}{\hdashlinewidth}
1473 % \begin{macro}{\hdashlinegap}
1474 % First of all, two |\dimen| registers |\dashlinedash| and |\dashlinegap|
1475 % to control the shape of dash-lines are declared, and their default values,
1476 % 4\,pt for both, are assigned to them. They have aliases,
1477 % |\hdashlinewidth| and |\hdashlinegap| respectively, for the backward
1478 % compatibility.
1479 % \end{macro}\end{macro}\end{macro}\end{macro}
1480 %
1481 % \begin{macrocode}
1482 %% Register Declaration
1483
1484 \newdimen\dashlinedash \dashlinedash4pt %
1485 \newdimen\dashlinegap \dashlinegap4pt %
1486 \let\hdashlinewidth\dashlinedash
1487 \let\hdashlinegap\dashlinegap
1488
1489 % \end{macrocode}
1490 %
1491 % Next, the following six switches are declared.
1492 % \begin{itemize}
1493 % \item
1494 % \begin{macro}{\ifadl@leftrule}
1495 % |\ifadl@leftrule| is used in the preamble analysis macro |\@mkpream| and
1496 % is true during it processes leading characters for solid- and dash-lines,
1497 % i.e.\ `\verb!|!', `|:|', and `|;|'.
1498 % \end{macro}
1499 %
1500 % \item
1501 % \begin{macro}{\ifadl@connected}
1502 % |\ifadl@connected| is used to indicate the {\em connection} $e^i_j\sim
1503 % e^{i'}_{j'}$. When we process $e^{i'}_{j'}$, the switch is true iff
1504 % $^\exists e^i_j(e^i_j\sim e^{i'}_{j'})$.
1505 % \end{macro}
1506 %
1507 % \item
1508 % \begin{macro}{\ifadl@doublerule}
1509 % |\ifadl@doublerule| is used to make $\sigma_k$. When we are to make
1510 % $\sigma_k L_k$, it is true iff $\gamma_{k-1}=\gamma_k\LAND\pi_{k-1}\neq
1511 % \pi_k$.
1512 % \end{macro}
1513 %
1514 % \item
1515 % \begin{macro}{\ifadl@zwvrule}
1516 % \changes{v1.4-2-2}{1999/06/25}
1517 % {Introduced to indicate vertical lines have null width.}
1518 % |\ifadl@zwvrule| controls the {\em real} width of vertical lines. If it
1519 % is true, lines are drawn as if their width is zero following \LaTeX's
1520 % standard. Otherwise, their width |\arrayrulewidth| contribute to the
1521 % width of columns as \textsf{array} does.
1522 % \end{macro}
1523 %
1524 % \item
1525 % \begin{macro}{\ifadl@zwhrule}
1526 % \changes{v1.7-1}{2004/5/21}
1527 % {Introduced to indicate horizontal lines have null width.}
1528 % |\ifadl@zwhrule| controls the {\em real} width of horizontal lines. If it
1529 % is true, a line is drawn as if its width is zero and its bottom edge is
1530 % adjusted to that of the row above by inserting |\vskip-\arrayulewidth|
1531 % before the drawing. Thus a horizontal dash line is included in the row
1532 % above and its gaps look colored properly if the row is painted. If it is
1533 % false, the width |\arrayrulewidth| contribute to the height of
1534 % \texttt{array}\slash\texttt{tabular} as usual.
1535 % \end{macro}
1536 %
1537 % \item
1538 % \begin{macro}{\ifadl@usingarypkg}
1539 % \changes{v1.4-2-1}{1999/06/25}
1540 % {Introduced to know if \textsf{array} is loaded.}
1541 % |\ifadl@usingarypkg| is true iff \textsf{array} has been loaded prior to
1542 % \textsf{arydshln}. This switch shows us which definitions, by \LaTeX{} or
1543 % \textsf{array}, we have to modify. Its value is set by examining if
1544 % |\extrarowheight|, which is introduced by \textsf{array}, is defined.
1545 % \end{macro}
1546 %
1547 % \item
1548 % \begin{macro}{\ifadl@inactive}
1549 % \changes{v1.4-2-3}{1999/06/25}
1550 % {Introduced to indicate dash-line functions are inactive.}
1551 % |\ifadl@inactive| inactivates dash-line functions if it is true. Its
1552 % default value is false.
1553 % \end{macro}
1554 % \end{itemize}
1555 %
1556 % \SpecialIndex{\@tempswa}
1557 % We also use a working switch |\@tempswa|.
1558 %
1559 % \begin{macrocode}
1560 \newif\ifadl@leftrule
1561 \newif\ifadl@connected
1562 \newif\ifadl@doublerule
1563 \newif\ifadl@zwvrule
1564 \newif\ifadl@zwhrule
1565 \newif\ifadl@usingarypkg
1566 \ifx\extrarowheight\undefined \adl@usingarypkgfalse
1567 \else \adl@usingarypkgtrue \fi
1568 \newif\ifadl@inactive \adl@inactivefalse
1569
1570 % \end{macrocode}
1571 %
1572 % \begin{macro}{\ADLnullwide}
1573 % \changes{v1.7-1}{2004/5/21}
1574 % {Introduced to make horizontal lines null wide.}
1575 % \begin{macro}{\ADLsomewide}
1576 % \changes{v1.7-1}{2004/5/21}
1577 % {Introduced to make horizontal lines \cs{arraydashline} wide.}
1578 % The switch |\ifadl@hwvrule| is turned on/off by user interface macros
1579 % |\ADLnullwide| and |\ADLsomewide|. Its initial value is the complement of
1580 % |\adl@usingarypkg|.
1581 % \end{macro}\end{macro}
1582 %
1583 % \begin{macro}{\ADLnullwidehline}
1584 % \changes{v1.4-2-2}{1999/06/25}
1585 % {Introduced to make vertical lines null wide.}
1586 % \begin{macro}{\ADLsomewidehline}
1587 % \changes{v1.4-2-2}{1999/06/25}
1588 % {Introduced to make vertical lines \cs{arraydashline} wide.}
1589 % The switch |\ifadl@zwvrule| is turned on/off by user interface macros
1590 % |\ADLnullwidehline| and |\ADLsomewidehline|. Its initial value is false.
1591 % \end{macro}\end{macro}
1592 %
1593 % \begin{macro}{\ADLactivate}
1594 % \begin{macro}{\ADLinactivate}
1595 % The switch |\ifadl@inactive| is also turned on/off by user interface macros
1596 % |\ADL#inactivate| and |\ADLactivate|.
1597 % \end{macro}\end{macro}
1598 %
1599 % \begin{macrocode}
1600 \def\ADLnullwide{\adl@zwvruletrue}
1601 \def\ADLsomewide{\adl@zwvrulefalse}
1602 \ifadl@usingarypkg \ADLsomewide \else \ADLnullwide \fi
1603 \def\ADLnullwidehline{\adl@zwhruletrue}
1604 \def\ADLsomewidehline{\adl@zwhrulefalse}
1605 \ADLsomewidehline
1606
1607 \def\ADLactivate{\adl@inactivefalse}
1608 \def\ADLinactivate{\adl@inactivetrue}
1609
1610 % \end{macrocode}
1611 %
1612 % The following |\box| register and three |\dimen| registers are used to
1613 % measure the height and depth of a row.
1614 %
1615 % \begin{itemize}
1616 % \item
1617 % \begin{macro}{\adl@box}
1618 % The contents of a column is packed into the |\box| register |\adl@box| to
1619 % measure its height and depth.
1620 % \end{macro}
1621 %
1622 % \item
1623 % \begin{macro}{\adl@height}
1624 % \begin{macro}{\adl@depth}
1625 % The |\dimen| registers |\adl@height| and |\adl@depth| contain the
1626 % height\slash depth of the tallest\slash deepest column in a row. When a
1627 % column is processed, they are compared to the height and depth of
1628 % |\adl@box| and are updated if they are less.
1629 % \end{macro}\end{macro}
1630 %
1631 % \vskip-2ex\vskip0pt
1632 % \begin{macro}{\adl@heightsave}
1633 % \begin{macro}{\adl@depthsave}
1634 % Since we have to update these registers |\global|-ly to pass their values
1635 % across |&| and we may have a column containing |array|\slash|tabular|,
1636 % they are saved into |\adl@heightsave|\slash|\adl@depthsave| at the
1637 % beginning of the environment and are restored at its end.
1638 % \end{macro}\end{macro}
1639 %
1640 % \item
1641 % \begin{macro}{\adl@finaldepth}
1642 % \changes{v1.4-2-6}{1999/06/25}
1643 % {Introduced to measure the depth of the last row.}
1644 % The other |\dimen| register |\adl@finaldepth| is set to the depth of the
1645 % last row, or zero if the last vertical item is a horizontal line. This
1646 % value is used to shift |array|\slash|tabular| down because we add extra
1647 % two |\smash|-ed rows which make the depth of |array|\slash|tabular| zero.
1648 % \end{macro}
1649 % \end{itemize}
1650 %
1651 % \SpecialIndex{\@tempdima}
1652 % \SpecialIndex{\@tempdimb}
1653 % We also use working |\dimen| registers |\@tempdima| and |\@tempdimb|.
1654 %
1655 % \begin{macrocode}
1656 \newbox\adl@box
1657 \newdimen\adl@height \newdimen\adl@heightsave
1658 \newdimen\adl@depth \newdimen\adl@depthsave
1659 \newdimen\adl@finaldepth
1660 % \end{macrocode}
1661 %
1662 % Then the following |\count| registers are declared. Note that some of
1663 % them contain dimensions measured by the unit |sp|.
1664 %
1665 % \begin{itemize}
1666 % \item
1667 % \begin{macro}{\adl@columns}
1668 % \begin{macro}{\adl@ncol}
1669 % \changes{v1.4-2-1}{1999/06/25}
1670 % {Introduced for new column counting in preamble construction.}
1671 % |\adl@columns| has the number of columns specified in the preamble of the
1672 % environment. Because of a complicated reason related to the compatibility
1673 % with \textsf{array}, we cannot count up |\adl@columns| directly but
1674 % increment |\adl@ncol| when each column of preamble is built and move its
1675 % value to |\adl@columns| after the preamble is constructed.
1676 % \end{macro}\end{macro}
1677 %
1678 % \item
1679 % \begin{macro}{\adl@currentcolumn}
1680 % \begin{macro}{\adl@currentcolumnsave}
1681 % To process |\multicolumn|, we have to know the column number where it
1682 % appears. Thus we have a column counter |\adl@currentcolumn| which is
1683 % |\global|-ly incremented when each column is built. Because of the
1684 % |\global| assignment, the counter has to be saved\slash restored
1685 % into\slash from |\adl@currentcolumnsave|.
1686 % \end{macro}\end{macro}
1687 %
1688 % \item
1689 % \begin{macro}{\adl@totalheight}
1690 % In the real implementation, $\tau_k$ and $\beta_k$ are calculated by the
1691 % following equations rather than those shown in \S\ref{sec:imp-ps1}.
1692 %$$
1693 % H=\sum_{l=1}^N h_l,\quad
1694 % \tau_k=H+\eta_{i-1}-\sum_{l=1}^{i-1} h_l,\quad
1695 % \beta_k=\tau_k-\eta_{i-1}-\eta_{i'-1}-\sum_{l=i'}^i h_l.
1696 %$$
1697 % |\adl@totalheight| contains $\sum_{l=1}^i h_l$ when the $i\th$ row is
1698 % built and thus its final value is $H$. Since the data structure $R$ are
1699 % represented by a text, we have to pay attention to the precision of its
1700 % dimensional elements, such as $h_i$. That is, if we append $h_i$ to $R$
1701 % by expanding |\the\dimen|$n$ which has the height plus depth of $i\th$
1702 % row, $h_i$ will be an approximation of |\dimen|$n$ represented by a
1703 % decimal fraction with |pt|. Although the error of the approximation is
1704 % quite small and may be negligible, the error must be avoided because it is
1705 % avoidable by simply using |\number\dimen|$n$. Therefore, $h_i$ is an
1706 % integer and thus |\adl@totalheight| is too.
1707 % \end{macro}
1708 %
1709 % \begin{macro}{\adl@totalheightsave}
1710 % Because of the |\global| assignment to |\adl@totalheight| to pass its value
1711 % across rows, it has to be saved\slash restored into\slash from
1712 % |\adl@totalheightsave|.
1713 % \end{macro}
1714 %
1715 % \item
1716 % \begin{macro}{\adl@dash}
1717 % \begin{macro}{\adl@gap}
1718 % In order to check $e^i_j\sim e^{i'}_{j'}$, the size attributes of $d^i_j$
1719 % and $g^i_j$ are kept in
1720 % the registers |\adl@dash| and |\adl@gap| when we process $e^{i'}_{j'}$.
1721 % As explained above, $d^i_j$ and $g^i_j$ are integers and thus |\adl@dash|
1722 % and |\adl@gap| are |\count| registers.
1723 % \end{macro}\end{macro}
1724 %
1725 % \item
1726 % \begin{macro}{\adl@cla}
1727 % \begin{macro}{\adl@clb}
1728 % \SpecialIndex{\@cla}
1729 % \SpecialIndex{\@clb}
1730 % The coding of |\cdashline| is similar to that of |\cline| in \LaTeX-2.09
1731 % which uses two global |\count| registers |\@cla| and |\@clb|. These
1732 % registers are omitted from \LaTeXe{} because its |\cline| is completely
1733 % recoded. We could adopt new coding but it requires some other macro
1734 % definitions that \LaTeX-2.09 does not have. Thus we simply introduce new
1735 % global counters |\adl@cla| and |\adl@clb| for |\cdashline| in order to
1736 % make |\cdashline| work in both \LaTeX-2.09 and \LaTeXe.
1737 % \end{macro}\end{macro}
1738 % \end{itemize}
1739 %
1740 % \SpecialIndex{\@tempcnta}
1741 % \SpecialIndex{\@tempcntb}
1742 % We also use working |\count| registers |\@tempcnta| and |\@tempcntb|.
1743 %
1744 % \begin{macrocode}
1745 \newcount\adl@columns \newcount\adl@ncol
1746 \newcount\adl@currentcolumn \newcount\adl@currentcolumnsave
1747 \newcount\adl@totalheight \newcount\adl@totalheightsave
1748 \newcount\adl@dash \newcount\adl@gap
1749 \newcount\adl@cla \newcount\adl@clb
1750 % \end{macrocode}
1751 %
1752 % \begin{macro}{\adl@everyvbox}
1753 % \changes{v1.4-2-1}{1999/06/25}
1754 % {Introduced for a tricky modification of \cs{@array}.}
1755 % The last register declaration is for a |\toks| register named
1756 % |\adl@everyvbox|. In order to minimize the copy-and-modify of the codes
1757 % in \LaTeX{} and \textsf{array}, we need to use |\everyvbox| in our own
1758 % definition of |\@array|. The register is used to save the contents of
1759 % |\everyvbox|.
1760 % \end{macro}
1761 %
1762 % \begin{macrocode}
1763 \newtoks\adl@everyvbox
1764
1765 % \end{macrocode}
1766 %
1767 % \begin{macro}{\adl@org@arrayclassz}
1768 % \changes{v1.4-2-3}{1999/06/25}
1769 % {Introduced to restore \cs{@arrayclassz}.}
1770 % \begin{macro}{\adl@org@tabclassz}
1771 % \changes{v1.4-2-3}{1999/06/25}
1772 % {Introduced to restore \cs{@tabclassz}.}
1773 % \begin{macro}{\adl@org@classz}
1774 % \changes{v1.4-2-3}{1999/06/25}
1775 % {Introduced to restore \cs{@classz}.}
1776 % \begin{macro}{\adl@org@@startpbox}
1777 % \changes{v1.4-2-3}{1999/06/25}
1778 % {Introduced to restore \cs{@@startpbox}.}
1779 % \begin{macro}{\adl@org@@endpbox}
1780 % \changes{v1.4-2-3}{1999/06/25}
1781 % {Introduced to restore \cs{@@endpbox}.}
1782 % \begin{macro}{\adl@org@startpbox}
1783 % \changes{v1.75}{2018/12/26}
1784 % {Introduced to restore \cs{@startpbox}.}
1785 % \begin{macro}{\adl@org@endpbox}
1786 % \changes{v1.4-2-3}{1999/06/25}
1787 % {Introduced to restore \cs{@endpbox}.}
1788 % \begin{macro}{\adl@org@cline}
1789 % \changes{v1.4-2-3}{1999/06/25}
1790 % {Introduced to restore \cs{cline}.}
1791 % \changes{v1.4-2-6}{1999/06/25}
1792 % {Introduced to refer original version in modified \cs{cline}.}
1793 % The other declarative stuff consists of the sequence of |\let| to capture
1794 % the original definitions of macros that we will modify afterword. The
1795 % main purpose of them is to nullify the modification when dash-line
1796 % functions are inactive, while |\adl@org@cline| is also referred to in its
1797 % modified version.
1798 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
1799 % \end{macro}\end{macro}\end{macro}
1800 %
1801 % \begin{macrocode}
1802 \let\adl@org@arrayclassz\@arrayclassz
1803 \let\adl@org@tabclassz\@tabclassz
1804 \let\adl@org@classz\@classz
1805 \let\adl@org@@startpbox\@@startpbox
1806 \let\adl@org@@endpbox\@@endpbox
1807 \let\adl@org@startpbox\@startpbox
1808 \let\adl@org@endpbox\@endpbox
1809 \let\adl@org@cline\cline
1810
1811 %%^L
1812 % \end{macrocode}
1813 %
1814 %
1815 %
1816 % \subsection{Initialization}
1817 % \label{sec:imp-init}
1818 %
1819 % \begin{macro}{\adl@array}
1820 % \changes{v1.4-2-1}{1999/06/25}
1821 % {Introduced to save original definition of \cs{@array}.}
1822 % \begin{macro}{\@array}
1823 % \changes{v1.4-2-1}{1999/06/25}
1824 % {Drastically modified to avoid copy-and-modify.}
1825 % \begin{macro}{\adl@noalign}
1826 % \changes{v1.7-2}{2004/5/21}
1827 % {Introduced to fix a bug of \textsf{colortbl}.}
1828 %
1829 % \LaTeX's macro |\@array| is modified to save and initialize registers and
1830 % data structures which are |\global|-ly updated in order to allow nested
1831 % |array|\slash|tabular|. This saving and initializing are performed by
1832 % |\adl@arrayinit| as explained below. The problem in the modification is
1833 % that the code of |\@array| in \textsf{array} is completely different from
1834 % that of \LaTeX{} original.
1835 %
1836 % The main difference is that \LaTeX{} builds |\@preamble| locally, while
1837 % \textsf{array} does globally exploiting the fact that the lifetime of
1838 % |\@preamble| ends before another |array|\slash|tabular| appears in a
1839 % column. The latter implementation will work well unless the building
1840 % process in |\@mkpream| produces something referred to after |\@preamble|
1841 % is thrown into \TeX's {\em stomach}. In our implementation,
1842 % unfortunately, the number of columns has to be counted in
1843 % \break|\@mkpream| and will be referred to by |\hdashline| and the vertical
1844 % line drawing procedure.
1845 %
1846 % Thus we have to change the column counting mechanism depending on whether
1847 % or not \textsf{array} is in use. The simplest way could be to copy the codes
1848 % of \LaTeX{} and \textsf{array} and modify them appropriately examining the
1849 % value of |\ifadl@usingarypkg|. However this solution is vulnerable to the
1850 % modification of the original version and thus we wish to refuse it as far
1851 % as possible.
1852 %
1853 % \SpecialIndex{\everyvbox}
1854 % \SpecialIndex{\adl@everyvbox}
1855 % Therefore, we use a trick with |\everyvbox| in which |\adl@arrayinit| is
1856 % temporarily included to initialize registers and locally set
1857 % |\adl@columns| to the number of columns |\global|-ly counted by
1858 % |\adl@ncol|. This trick works well so far because;
1859 %
1860 % \begin{itemize}
1861 % \item
1862 % the first |\vbox|, |\vtop| or |\vcenter| made by |\@array| is the vertical
1863 % box surrounding |\halign|, and;
1864 %
1865 % \item
1866 % in |\@array| of \textsf{array} the box is opened {\em after} the preamble
1867 % is constructed;
1868 % \end{itemize}
1869 %
1870 % and will hopefully work in future.
1871 %
1872 % \changes{v1.5-3}{2003/07/11}
1873 % {Move \cs{adl@inactivate} to \cs{@array} from \cs{adl@arrayinit}.}
1874 % \changes{v1.54-2}{2003/08/25}
1875 % {Invoke \cs{adl@activate} if not \cs{ifadl@inactive}.}
1876 % \changes{v1.7-2}{2004/5/21}
1877 % {Make \cs{adl@noalign} \cs{let}-equal to \cs{noalign}.}
1878 % \SpecialIndex{\arrayrulecolor}
1879 % \SpecialIndex{\doublerulesepcolor}
1880 % \SpecialIndex{\dashgapcolor}
1881 % Next, if |\ifadl@inactive| is true, |\adl@inactivate| is invoked to
1882 % inactivate dash-line functions. Otherwise, |\adl@activate| is invoked to
1883 % activate them because an inactivated |array|\slash|tabular|
1884 % may have active children in it. Finally, |\adl@noalign| is made
1885 % |\let|-equal to |\noalign| so that |\arrayrulecolor|,
1886 % |\doublerulesepcolor| and |\dashgapcolor| are expanded with |\noalign| in
1887 % the environment.
1888 % \end{macro}\end{macro}\end{macro}
1889 %
1890 % \begin{macro}{\@@array}
1891 % \changes{v1.4-2-1}{1999/06/25}
1892 % {Introduced because \textsf{array} uses it.}
1893 % \changes{v1.74}{2018/09/26}
1894 % {Make \cs{@@array} \cs{let}-equal to \cs{@array} only when it is
1895 % made so by \textsf{array} and the equality is kept.}
1896 %
1897 % \SpecialIndex{\@tabarray}
1898 % Another stuff for the compatibility with \textsf{array} is to |\let| a
1899 % control sequence |\@@array| be equal to |\@array| if it is made so by
1900 % \textsf{array} and the equality is kept. That is, with \textsf{array}
1901 % |\@@array| is invoked by |\@tabarray| and it is |\let|-equal to |\@array|
1902 % by default, while\break |\@@array| can be made different from |\@array| by
1903 % some other package, e.g., \textsf{delarray}, to do some special operations
1904 % defined in the package. Therefore by the conditional equalization with
1905 % |\ifx|, our own |\@array| is directly invoked through |\@@array| if the
1906 % default equality is kept, while otherwise the package-dependent definition
1907 % of |\@@array| is respected.
1908 % \end{macro}
1909 %
1910 % \begin{macrocode}
1911
1912 %% Initialization
1913
1914 \let\adl@array\@array
1915 \def\@array{\adl@everyvbox\everyvbox
1916 \everyvbox{\adl@arrayinit \the\adl@everyvbox \everyvbox\adl@everyvbox}%
1917 \ifadl@inactive \adl@inactivate \else \adl@activate \fi
1918 \let\adl@noalign\noalign
1919 \adl@array}
1920 \ifx\@@array\adl@array \let\@@array\@array \fi
1921
1922 % \end{macrocode}
1923 %
1924 % \begin{macro}{\adl@arrayinit}
1925 % \changes{v1.4-2-9}{1999/06/25}
1926 % {The bug of saving \cs{adl@colsR} is fixed.}
1927 % \changes{v1.5-1}{2000/07/12}
1928 % {Use new macro \cs{adl@arraysave} to save registers/structures.}
1929 % \begin{macro}{\adl@arraysave}
1930 % \changes{v1.5-1}{2000/07/12}
1931 % {Introduced to use in modified \cs{CC@} of \textsf{colortab}.}
1932 % \SpecialIndex{\adl@height}
1933 % \SpecialIndex{\adl@depth}
1934 % \SpecialIndex{\adl@currentcolumn}
1935 % \SpecialIndex{\adl@totalheight}
1936 %
1937 % As described in \S\ref{sec:imp-decl}, registers updated |\global|-ly,
1938 % which are |\adl@height|, |\adl@depth|, |\adl@currentcolumn| and
1939 % |\adl@totalheight|, are saved in |\adl@arrayinit| by calling
1940 % |\adl@arraysave|, and also given initial values. The macro also saves the
1941 % following data structures and initializes them to empty lists.
1942 % \end{macro}
1943 %
1944 % \begin{macro}{\adl@rowsL}
1945 % \begin{macro}{\adl@rowsR}
1946 % \begin{macro}{\adl@rowsLsave}
1947 % \begin{macro}{\adl@rowsRsave}
1948 % \begin{macro}{\adl@colsL}
1949 % \begin{macro}{\adl@colsR}
1950 % \begin{macro}{\adl@colsLsave}
1951 % \begin{macro}{\adl@colsRsave}
1952 % \begin{itemize}
1953 % \item
1954 % In the real implementation, the data structure $R$ is split into two
1955 % lists;
1956 %
1957 % \begin{eqnarray*}
1958 % \cs{adl@rowsL}&=&R^L=\langle\langle C^L_1,h_1\rangle,\ldots\rangle\\
1959 % \cs{adl@rowsR}&=&R^R=\langle\langle C^R_1,h_1\rangle,\ldots\rangle
1960 % \end{eqnarray*}
1961 %
1962 % and they are saved into |\adl@rowsLsave| and |\adl@rowsRsave|.
1963 %
1964 % \item
1965 % When the $i\th$ row is building, $C^L_i$ and $C^R_i$ are constructed in
1966 % the macros |\adl@colsL| and |\adl@colsR|. They are saved
1967 % into |\adl@colsLsave| and |\adl@colsRsave|.
1968 % \end{itemize}
1969 % \end{macro}\end{macro}\end{macro}\end{macro}
1970 % \end{macro}\end{macro}\end{macro}\end{macro}
1971 %
1972 % \begin{macro}{\adl@connect}
1973 % \begin{macro}{\adl@discard}
1974 % \changes{v1.6}{2003/8/25}
1975 % {Add initializetion of \cs{adl@discard}.}
1976 % \SpecialIndex{\@elt}
1977 % \SpecialOptIndex{longtable}{environment}
1978 %
1979 % In the real implementation, $e^i_j$ is represented by a control sequence
1980 % |\@elt|, and ${\it connect}(i)$ by |\adl@connect|. They are made
1981 % |\let|-equal to |\relax| to keep them from expansion during $R$ is
1982 % constructed. In |longtable| environment, ${\it connect}(i)$ for negative
1983 % vertical space inserted by |\\[|\meta{h}|]| or a horizontal line has
1984 % another representation |\adl@discard| to indicate it corresponds to a
1985 % discardable item of page breaking. Since this representation, however, is
1986 % nonsense in usural |array|\slash|tabular| even if they are included in
1987 % |\longtable|, we define |\adl@discard| as |\adl@connect| so that it
1988 % transforms itself into |\adl@connect| when it is added to
1989 % |\adl@rowsL|\slash|R| by |\xdef|. Note that |\adl@discard| is made
1990 % |\let|-equal to |\relax| to inhibit the transformation at the beginning of
1991 % |longtable| environment.
1992 % \end{macro}\end{macro}
1993 %
1994 % \changes{v1.4-2-1}{1999/06/25}
1995 % {Modified for new column counting in preamble construction.}
1996 %
1997 % Then, we set to |\adl@columns| to the value of |\adl@ncol| {\em locally}.
1998 % As explained above, this has an effect with \textsf{array} because
1999 % |\adl@arrayinit| is called {\em after} the preamble is generated. Without
2000 % \textsf{array}, on the other hand, this assignment has no effect but safe
2001 % because it is included in a group of |\vbox| etc.
2002 %
2003 % \changes{v1.4-2-3}{1999/06/25}
2004 % {Modified to call \cs{adl@inactivate}.}
2005 % \changes{v1.5-3}{2003/07/11}
2006 % {Move \cs{adl@inactivate} from \cs{adl@arrayinit} to \cs{@array}.}
2007 %
2008 % \end{macro}
2009 % \SpecialIndex{\adl@height}
2010 % \SpecialIndex{\adl@depth}
2011 % \SpecialIndex{\adl@currentcolumn}
2012 % \SpecialIndex{\adl@totalheight}
2013 % \SpecialIndex{\adl@heightsave}
2014 % \SpecialIndex{\adl@depthsave}
2015 % \SpecialIndex{\adl@currentcolumnsave}
2016 % \SpecialIndex{\adl@totalheightsave}
2017 %
2018 % \begin{macrocode}
2019 \def\adl@arrayinit{%
2020 \adl@arraysave
2021 \global\adl@height\z@ \global\adl@depth\z@
2022 \global\adl@currentcolumn\@ne \global\adl@totalheight\z@
2023 \gdef\adl@rowsL{}\gdef\adl@rowsR{}\gdef\adl@colsL{}\gdef\adl@colsR{}%
2024 \let\@elt\relax \let\adl@connect\relax \def\adl@discard{\adl@connect}%
2025 \adl@columns\adl@ncol}
2026 \def\adl@arraysave{%
2027 \adl@heightsave\adl@height
2028 \adl@depthsave\adl@depth
2029 \adl@currentcolumnsave\adl@currentcolumn
2030 \adl@totalheightsave\adl@totalheight
2031 \let\adl@rowsLsave\adl@rowsL
2032 \let\adl@rowsRsave\adl@rowsR
2033 \let\adl@colsLsave\adl@colsL
2034 \let\adl@colsRsave\adl@colsR}
2035
2036 % \end{macrocode}
2037 %
2038 % \begin{macro}{\adl@inactivate}
2039 % \changes{v1.4-2-3}{1999/06/25}
2040 % {Introduced to inactivate \cs{@arrayclassz} etc.}
2041 % \changes{v1.5-3}{2003/07/11}
2042 % {Change \cs{adl@inactivate} caller to \cs{@array}.}
2043 % \changes{v1.7-1}{2004/5/21}
2044 % {Remove \cs{cline} because our own version is needed for null-wide.}
2045 % If |\ADLinactivate| has effect and thus |\ifadl@inactive| is true, the
2046 % macro |\adl@inactivate| is called from |\@array|\footnote{
2047 %
2048 % Before v1.53, \cs{adl@inactivate} was called from \cs{adl@arrayinit} and
2049 % thus invokded {\em after} the preamble of \textsf{array} is built. This
2050 % was incorrect of course and made inactive version of \texttt{p},
2051 % \texttt{m} and \texttt{b} produce nothing.}.
2052 %
2053 % This |\let|-s the following control sequences be equal to their
2054 % counterparts in \LaTeX{} and/or \textsf{array} package.
2055 %
2056 % \changes{v1.54-2}{2003/08/25}
2057 % {Add \cs{adl@argcr} to inactivation.}
2058 % \changes{v1.75}{2018/12/26}
2059 % {Add \cs{@startpbox} to inactivation.}
2060 %
2061 % \begin{itemize}\item[]
2062 % \SpecialIndex{\@arrayclassz}
2063 % \SpecialIndex{\@tabclassz}
2064 % \SpecialIndex{\@classz}
2065 % \SpecialIndex{\@@startpbox}
2066 % \SpecialIndex{\@@endpbox}
2067 % \SpecialIndex{\@startpbox}
2068 % \SpecialIndex{\@endpbox}
2069 % \SpecialIndex{\adl@cr}
2070 % \SpecialIndex{\adl@argcr}
2071 % \SpecialIndex{\adl@endarray}
2072 % |\@arrayclassz|\quad
2073 % |\@tabclassz|\quad
2074 % |\@classz|\quad
2075 % |\@@startpbox|\quad
2076 % |\@@endpbox|\\
2077 % |\@startpbox|\quad
2078 % |\@endpbox|\quad
2079 % |\adl@cr|\quad
2080 % |\adl@argcr|\quad
2081 % |\adl@endarray|
2082 % \end{itemize}
2083 %
2084 % Note that |\@classz| has to be |\let|-equal to |\adl@org@classz| only if
2085 % \textsf{array} is in use, because \LaTeX{} does not define |\@classz| but
2086 % refers to it which is either |\@arrayclassz| or |\@tabclassz|. Yet
2087 % another remark is that we have to conceal |\cr| for
2088 % |\adl@cr|\slash|\adl@argcr| and |\crcr| for |\adl@endarray| by bracing
2089 % them from \TeX's |\halign| mechanism that searches them when an
2090 % |array|\slash|tabular| has an nested |array|\slash|tabular|. This could
2091 % be done by a tricky |\let|-assignment such as;
2092 %
2093 % \begin{quote}
2094 % |\iffalse{\let\adl@cr\cr \iffalse}\fi|
2095 % \end{quote}
2096 %
2097 % but we simply use |\def| instead of |\let| because of clarity.
2098 %
2099 % We also |\let| the following be {\it no-operation} or their inactive
2100 % versions.
2101 %
2102 % \begin{itemize}\item[]
2103 % \SpecialIndex{\adl@hline}
2104 % \SpecialIndex{\adl@ihdashline}
2105 % \SpecialIndex{\adl@cdline}
2106 % \SpecialIndex{\adl@@vlineL}
2107 % \SpecialIndex{\adl@@vlineR}
2108 % \SpecialIndex{\adl@vlineL}
2109 % \SpecialIndex{\adl@vlineR}
2110 % |\adl@hline|\quad
2111 % |\adl@ihdashline|\quad
2112 % |\adl@cdline|\quad
2113 % |\adl@@vlineL|\quad
2114 % |\adl@@vlineR|\\
2115 % |\adl@vlineL|\quad
2116 % |\adl@vlineR|
2117 % \end{itemize}
2118 % Note that we have to inactivate both |\adl@@vlineL| and |\adl@vlineL|,
2119 % because the latter is referred to when \textsf{array} is in use while the
2120 % former is done otherwise. Their |R| relatives are also inactivated by
2121 % the same reason.
2122 % \end{macro}
2123 %
2124 % \SpecialIndex{\@arrayclassz}
2125 % \SpecialIndex{\@tabclassz}
2126 % \SpecialIndex{\@classz}
2127 % \SpecialIndex{\@@startpbox}
2128 % \SpecialIndex{\@@endpbox}
2129 % \SpecialIndex{\@startpbox}
2130 % \SpecialIndex{\@endpbox}
2131 % \SpecialIndex{\adl@cr}
2132 % \SpecialIndex{\adl@argcr}
2133 % \SpecialIndex{\adl@endarray}
2134 % \SpecialIndex{\adl@hline}
2135 % \SpecialIndex{\adl@ihdashline}
2136 % \SpecialIndex{\adl@cdline}
2137 % \SpecialIndex{\adl@@vlineL}
2138 % \SpecialIndex{\adl@@vlineR}
2139 % \SpecialIndex{\adl@vlineL}
2140 % \SpecialIndex{\adl@vlineR}
2141 % \SpecialIndex{\adl@org@arrayclassz}
2142 % \SpecialIndex{\adl@org@tabclassz}
2143 % \SpecialIndex{\adl@org@classz}
2144 % \SpecialIndex{\adl@org@@startpbox}
2145 % \SpecialIndex{\adl@org@@endpbox}
2146 % \SpecialIndex{\adl@org@startpbox}
2147 % \SpecialIndex{\adl@org@endpbox}
2148 % \SpecialIndex{\adl@cr}
2149 % \SpecialIndex{\adl@argcr}
2150 % \SpecialIndex{\adl@org@cline}
2151 % \SpecialIndex{\adl@inactivehdl}
2152 % \SpecialIndex{\adl@inactivecdl}
2153 % \SpecialIndex{\adl@inactivevl}
2154 % \begin{macrocode}
2155 \def\adl@inactivate{%
2156 \let\@arrayclassz\adl@org@arrayclassz
2157 \let\@tabclassz\adl@org@tabclassz
2158 \ifadl@usingarypkg \let\@classz\adl@org@classz \fi
2159 \let\@@startpbox\adl@org@@startpbox
2160 \let\@@endpbox\adl@org@@endpbox
2161 \let\@startpbox\adl@org@startpbox
2162 \let\@endpbox\adl@org@endpbox
2163 \def\adl@cr{\cr}%
2164 \def\adl@argcr##1{\cr}%
2165 \def\adl@endarray{\crcr}%
2166 \let\adl@hline\@gobbletwo
2167 \let\adl@ihdashline\adl@inactivehdl
2168 \let\adl@cdline\adl@inactivecdl
2169 \let\adl@@vlineL\adl@inactivevl
2170 \let\adl@@vlineR\adl@inactivevl
2171 \let\adl@vlineL\adl@inactivevl
2172 \let\adl@vlineR\adl@inactivevl}
2173 % \end{macrocode}
2174 %
2175 % \begin{macro}{\adl@activate}
2176 % \changes{v1.54-2}{2003/08/25}
2177 % {Introduced to activate \cs{@arrayclassz} etc.\ again.}
2178 % \changes{v1.75}{2018/12/26}
2179 % {Add \cs{@startpbox} to activation.}
2180 %
2181 % On the other hand, if |\ifadl@inactive| is false, the macro |\adl@activate|
2182 % is called from |\@array| to make inactivated macros active again in order
2183 % to cope with the case in which an inactive \textsf{array}\slash
2184 % \textsf{tabular} has active children in it\footnote{
2185 %
2186 % Before v1.54, an active \textsf{array}\slash\textsf{tabular} in an
2187 % inactive parent was not activated.}.
2188 %
2189 % To do that, |\adl@activate| makes |\@arrayclassz| etc.\ |\let|-equal to
2190 % their active version |\adl@act@arrayclassz| etc.\ which will be defined
2191 % (|\let|-equal to) as our own |\@arrayclassz| etc.\ in \S\ref{sec:imp-sh}.
2192 % \end{macro}
2193 %
2194 % \SpecialIndex{\@arrayclassz}
2195 % \SpecialIndex{\@tabclassz}
2196 % \SpecialIndex{\@classz}
2197 % \SpecialIndex{\@@startpbox}
2198 % \SpecialIndex{\@@endpbox}
2199 % \SpecialIndex{\@startpbox}
2200 % \SpecialIndex{\@endpbox}
2201 % \SpecialIndex{\adl@cr}
2202 % \SpecialIndex{\adl@argcr}
2203 % \SpecialIndex{\adl@endarray}
2204 % \SpecialIndex{\adl@hline}
2205 % \SpecialIndex{\adl@ihdashline}
2206 % \SpecialIndex{\adl@cdline}
2207 % \SpecialIndex{\adl@@vlineL}
2208 % \SpecialIndex{\adl@@vlineR}
2209 % \SpecialIndex{\adl@vlineL}
2210 % \SpecialIndex{\adl@vlineR}
2211 % \SpecialIndex{\adl@act@arrayclassz}
2212 % \SpecialIndex{\adl@act@tabclassz}
2213 % \SpecialIndex{\adl@act@classz}
2214 % \SpecialIndex{\adl@act@@startpbox}
2215 % \SpecialIndex{\adl@act@@endpbox}
2216 % \SpecialIndex{\adl@act@startpbox}
2217 % \SpecialIndex{\adl@act@endpbox}
2218 % \SpecialIndex{\adl@act@cr}
2219 % \SpecialIndex{\adl@act@argcr}
2220 % \SpecialIndex{\adl@act@cline}
2221 % \SpecialIndex{\adl@act@endarray}
2222 % \SpecialIndex{\adl@act@hline}
2223 % \SpecialIndex{\adl@act@ihdashline}
2224 % \SpecialIndex{\adl@act@cdline}
2225 % \SpecialIndex{\adl@act@@vlineL}
2226 % \SpecialIndex{\adl@act@@vlineR}
2227 % \begin{macrocode}
2228 \def\adl@activate{%
2229 \let\@arrayclassz\adl@act@arrayclassz
2230 \let\@tabclassz\adl@act@tabclassz
2231 \ifadl@usingarypkg \let\@classz\adl@act@classz \fi
2232 \let\@@startpbox\adl@act@@startpbox
2233 \let\@@endpbox\adl@act@@endpbox
2234 \let\@startpbox\adl@act@startpbox
2235 \let\@endpbox\adl@act@endpbox
2236 \let\adl@cr\adl@act@cr
2237 \let\adl@argcr\adl@act@argcr
2238 \let\adl@endarray\adl@act@endarray
2239 \let\adl@hline\adl@act@hline
2240 \let\adl@ihdashline\adl@act@ihdashline
2241 \let\adl@cdline\adl@act@cdline
2242 \let\adl@@vlineL\adl@act@@vlineL
2243 \let\adl@@vlineR\adl@act@@vlineR
2244 \let\adl@vlineL\adl@act@@vlineL
2245 \let\adl@vlineR\adl@act@@vlineR}
2246
2247 %%^L
2248 % \end{macrocode}
2249 %
2250 % \changes{v1.6}{2003/8/25}
2251 % {Add a summary of activation/inactivation.}
2252 % The summary of the activation and inactivation is shown in
2253 % Table~\ref{tab:act}.
2254 %
2255 %
2256 %
2257 % \begin{table}
2258 % \catcode`\|12
2259 % \caption{Active and Inactive Operations}\label{tab:act}
2260 % \def\!{\texttt{\char`\\\char`\\}}
2261 % \centerline{\begin{tabular}[t]{|l||l|l|}\hline
2262 % command&active&inactive\\\hline\hline
2263 % \tt l c r&&\\
2264 % \qquad with \textsf{array}& \cs{adl@act@classz}&
2265 % \cs{adl@org@classz}\\
2266 % \qquad without \textsf{array}&\cs{adl@act@tabclassz}&
2267 % \cs{adl@org@tabclassz}\\
2268 % & \cs{adl@act@arrayclassz}&
2269 % \cs{adl@org@arrayclassz}\\\hline
2270 % {\tt p m b} (open)&&\\
2271 % \qquad with \textsf{array}& \cs{adl@act@classz}&
2272 % \cs{adl@org@classz}\\
2273 % \qquad without \textsf{array}&\cs{adl@act@@startpbox}&
2274 % \cs{adl@org@@startpbox}\\\hline
2275 % {\tt p m b} (close)& \cs{adl@act@@endpbox}&
2276 % \cs{adl@org@@endpbox}\\\hline
2277 % \tt \char`\|/:/;& \cs{adl@act@@vlineL/R}&
2278 % \cs{adl@inactivevl}\\\hline
2279 % \!& $\to$\cs{adl@act@(arg)cr}&
2280 % $\to$\cs{cr}\\\hline
2281 % \cs{hline}& $\to$\cs{adl@act@hline}&
2282 % $\to$\cs{@gobbletwo}\\\hline
2283 % \cs{hdashline}& $\to$\cs{adl@act@ihdashline}&
2284 % $\to$\cs{adl@inactivehdl}\\\hline
2285 % \cs{cdashline}& $\to$\cs{adl@act@cdline}&
2286 % $\to$\cs{adl@inactivecdl}\\\hline
2287 % \end{tabular}}
2288 % \end{table}
2289 %
2290 %
2291 %
2292 % \subsection{Making Preamble}
2293 % \label{sec:imp-pream}
2294 %
2295 % Each preamble character is converted to a part of |\halign|'s preamble as
2296 % follows.
2297 %
2298 % \begin{itemize}
2299 % \item
2300 % \begin{macro}{\adl@colhtdp}
2301 % \changes{v1.5-2}{2002/06/26}
2302 % {The pseudo-formal description of $\langle\hbox{\it put-lrc}\rangle$
2303 % is modified.}
2304 % \SpecialIndex{\adl@box}
2305 % \SpecialIndex{\adl@height}
2306 % \SpecialIndex{\adl@depth}
2307 % `|l|', `|r|' and `|c|' are converted to the following \meta{lrc}.
2308 %
2309 % \begingroup \catcode`\|12
2310 % \begin{eqnarray*}
2311 % \meta{lrc}&::=&
2312 % [\cs{hfil}]\meta{put-lrc}[\cs{hfil}]\\
2313 % \meta{put-lrc}&::=&
2314 % \cs{setbox}\cs{adl@box}\cs{hbox}\Meta{lrc-contents}\\
2315 % && \cs{adl@colhtdp }\cs{unhbox}\cs{adl@box}\\
2316 % \meta{lrc-contents}&::=&
2317 % \hbox{\texttt{\$\cs{relax}\#\$}}\ |\\
2318 % && \hbox{\texttt{\#\cs{unskip}}}\\
2319 % \end{eqnarray*}
2320 % \endgroup
2321 %
2322 % That is, the content of a column is at first packed into the |\box|
2323 % register |\adl@box|, then its height and depth are compared to
2324 % |\adl@height| and |\adl@depth| by the macro |\adl@colhtdp|, and finally
2325 % the box is put with leading and/or trailing |\hfil|.
2326 % \end{macro}
2327 %
2328 % \item
2329 % \begin{macro}{\adl@vlineL}
2330 % \begin{macro}{\adl@vlineR}
2331 % \def\G{{\mit\Gamma}}
2332 % `\verb!|!', `|:|' and |;{|\meta{dash}|/|\meta{gap}|}| are converted to the
2333 % following \meta{vline}.
2334 %
2335 % \begingroup \catcode`\|12
2336 % \begin{eqnarray*}
2337 % \meta{vline}&::=&
2338 % [\cs{hskip}\cs{doublerulesep}]\meta{vline-LR}\\
2339 % \meta{vline-LR}&::=&
2340 % \cs{adl@vlineL}\Meta{$\G_d$}\Meta{$\G_g$}\Meta{c}
2341 % \hbox{{\tt\char`\{}\meta{d}/\meta{g}{\tt\char`\}}}\ |\\
2342 % && \cs{adl@vlineR}\Meta{$\G_d$}\Meta{$\G_g$}\Meta{c}
2343 % \hbox{{\tt\char`\{}\meta{d}/\meta{g}{\tt\char`\}}}\\
2344 % \meta{d}&::=&
2345 % 0\ |&\hbox to.3\textwidth{\ldots\ for `\texttt{|}'\hfil}\cr
2346 % && \cs{dashlinedash}\ |&
2347 % \hbox to.3\textwidth{\ldots\ for `\texttt{:}'\hfil}\cr
2348 % && \meta{dash}&
2349 % \hbox to.3\textwidth{\ldots\ for `\texttt{;}'\hfil}\cr
2350 % \meta{g}&::=&
2351 % 0\ |&\hbox to.3\textwidth{\ldots\ for `\texttt{|}'\hfil}\cr
2352 % && \cs{dashlinegap}\ |&
2353 % \hbox to.3\textwidth{\ldots\ for `\texttt{:}'\hfil}\cr
2354 % && \meta{gap}&
2355 % \hbox to.3\textwidth{\ldots\ for `\texttt{;}'\hfil}\cr
2356 % \end{eqnarray*}
2357 % \endgroup
2358 %
2359 % \SpecialIndex{\CT@arc@}
2360 % \SpecialIndex{\adl@dashgapcolor}
2361 % Note that \meta{c} is the column number (leftmost is 1) where the
2362 % character appears, and \meta{$\G_d$} and \meta{$\G_g$} is the
2363 % color of dashs and gaps specified in |\CT@arc@| and |\adl@dashgapcolor|.
2364 % \end{macro}\end{macro}
2365 % \end{itemize}
2366 %
2367 % \SpecialIndex{\adl@currentcolumn}
2368 % Additionally, each column except for the last one has;
2369 %
2370 % \begin{quote}
2371 % |\global\advance\adl@currentcolumn\@ne|
2372 % \end{quote}
2373 %
2374 % just before |&| to increment |\adl@currentcolumn|. Other features, such
2375 % as inserting spaces of |\arraycolsep|\slash|\tabcolsep|, are as same as
2376 % original scheme. This means that |@|\Meta{text} and |!|\Meta{text} of
2377 % \textsf{array} are {\em not} handled specially although it could interfere
2378 % with drawing vertical lines. Therefore, we have the problem 1 shown in
2379 % \S\ref{sec:prob}, which is very hard to solve. Note that the measurement
2380 % of the column of `|p|' of \LaTeX{} original is done by (modified)
2381 % |\@@startpbox| and |\@@endpbox| and thus the preamble for `|p|' is not
2382 % modified. In the case with \textsf{array}, however, the preambles for
2383 % `|p|' and its relatives `|m|' and `|b|' are modified to set |\adl@box| to
2384 % the box for them.
2385 %
2386 % \begin{macro}{\adl@mkpream}
2387 % \begin{macro}{\@mkpream}
2388 % \changes{v1.4-2-1}{1999/06/25}
2389 % {Modified for new column counting and control sequence redefinition.}
2390 % \changes{v1.71}{2004/08/31}
2391 % {Modified to nullify \cs{adl@startmbox} and \cs{adl@endmbox} for
2392 % \textsf{array's} \texttt{m}-columns.}
2393 % \SpecialIndex{\adl@colhtdp}
2394 % \SpecialIndex{\adl@vlineL}
2395 % \SpecialIndex{\adl@vlineR}
2396 % \SpecialIndex{\adl@startmbox}
2397 % \SpecialIndex{\adl@endmbox}
2398 % \SpecialIndex{\adl@preaminit}
2399 % \SpecialIndex{\afterassignment}
2400 %
2401 % To make the preamble shown above, |\@mkpream| is modified to |\let|
2402 % control sequences |\adl@colhtpd|, |\adl@vlineL| and |\adl@vlineR| be
2403 % |\relax| in order to keep them from being expanded by |\edef|\slash
2404 % |\xdef| for the preamble construction. The control sequences
2405 % |\adl@startmbox| and |\adl@endmbox| for \texttt{m}-columns of
2406 % \textsf{array} are also made |\let|-equal to |\relax|.
2407 %
2408 % Giving them their own definition is done by |\adl@preaminit| that is
2409 % called using |\afterassignment| after |\@preamble| is made by
2410 % |\adl@mkpream|, the original version of |\@mkpream|. If \textsf{array} is
2411 % not in use, |\@mkpream| is followed by an |\edef| of |\@preamble| to add
2412 % |\ialign| etc.\ and thus |\adl@preaminit| is properly called {\em after}
2413 % this final {\em assignment} to make |\@preamble|.
2414 %
2415 % With \textsf{array}, on the other hand, calling |\adl@preaminit| is safe
2416 % because |\@mkpream| is followed by |\xdef| for |\@preamble| too, but has
2417 % no effect because it is in the group for |\@mkpream|. This grouping,
2418 % however, gives us an easier way to give those control sequences their own
2419 % definition. That is, we simply initiate them with the definitions that
2420 % will be regained when the group is closed.
2421 %
2422 % \SpecialIndex{\adl@ncol}
2423 % \SpecialIndex{\adl@columns}
2424 % \SpecialIndex{\ifadl@leftrule}
2425 % The modified |\@mkpream| also initializes |\adl@ncol| and
2426 % |\ifadl@leftrule|, and set |\adl@columns| to the value of |\adl@ncol|
2427 % locally after the preamble is made. This has an effect in the case
2428 % without \textsf{array} because the body of |array|\slash|tabular| is in
2429 % the same grouping context of |\@mkpream|. With \textsf{array}, on the
2430 % other hand, this assignment has no effect but safe because it is included
2431 % in a group of |\@mkpream|'s own.
2432 % \end{macro}\end{macro}
2433 %
2434 % \begin{macrocode}
2435
2436 %% Making Preamble
2437
2438 \let\adl@mkpream\@mkpream
2439 \def\@mkpream#1{\let\adl@colhtdp\relax
2440 \let\adl@vlineL\relax \let\adl@vlineR\relax
2441 \let\adl@startmbox\relax \let\adl@endmbox\relax
2442 \global\adl@ncol\@ne \adl@leftruletrue
2443 \adl@mkpream{#1}\adl@columns\adl@ncol \afterassignment\adl@preaminit}
2444
2445 % \end{macrocode}
2446 %
2447 % \begin{macro}{\@addamp}
2448 % \changes{v1.4-2-1}{1999/06/25}
2449 % {Modified for new column counting in preamble construction.}
2450 % \SpecialIndex{\adl@currentcolumn}
2451 % \SpecialIndex{\adl@ncol}
2452 %
2453 % The macro |\@addamp| is also modified to add the code for incrementing the
2454 % counter |\adl@currentcolumn| to |\@preamble| with |&|. The counter
2455 % |\adl@ncol| is also incremented by |\@addamp| so that we can refer to its
2456 % value as \meta{c} of |\adl@vlineL|\slash|R|\@. This increment is done
2457 % |\global|-ly in order that we locally set |\adl@columns| to the counting
2458 % result outside of the group for |\@mkpream| of \textsf{array}. Therefore,
2459 % whether or not \textsf{array} is in use, |\adl@columns| will have a
2460 % correct value and will be correctly referred to by |\hdashline| to know how
2461 % many columns are specified in the preamble. Note that this |\global|
2462 % assignment is safe because the life time of |\adl@ncol| is same as that of
2463 % |\@preamble|.
2464 % \end{macro}
2465 %
2466 % \begin{macrocode}
2467 \def\@addamp{\if@firstamp\@firstampfalse \else
2468 \@addtopreamble{\global\advance\adl@currentcolumn\@ne &}%
2469 \global\advance\adl@ncol\@ne \fi}
2470
2471 % \end{macrocode}
2472 %
2473 % \SpecialIndex{\ifadl@usingarypkg}
2474 % Since the implementation of |\@testpach| and macros for class-0 characters
2475 % (i.e. |l|, |r| and |c|) is completely different between \LaTeX{} and
2476 % \textsf{array}, we have to have two versions switched by
2477 % |\adl@usingarypkg|.
2478 %
2479 % \subsubsection*{With \textsf{array}}
2480 %
2481 % \begin{macro}{\@testpach}
2482 % \changes{v1.4-2-1}{1999/06/25}
2483 % {The version for \textsf{array} is introduced.}
2484 % \SpecialIndex{\@arrayrule}
2485 % \SpecialIndex{\adl@arrayrule}
2486 % \SpecialIndex{\adl@arraydashrule}
2487 % \SpecialIndex{\adl@argarraydashrule}
2488 %
2489 % Although we introduced two preamble characters `|:|' and `|;|', we did not
2490 % introduce new character {\em class} because we want to minimize the
2491 % modification of original codes. Therefore, `|:|' and `|;|' is classified
2492 % into class-1 together with `\verb!|!'\@. Since these characters obviously
2493 % have their own appropriate operations, |\@testpach| is modified so that
2494 % |\@arrayrule|, which is invoked from |\@mkpream| in the case of class-1
2495 % character, is |\let|-equal to the macro corresponding to each character.
2496 % \end{macro}
2497 %
2498 % \begin{macrocode}
2499 \ifadl@usingarypkg
2500 \def\@testpach{\@chclass
2501 \ifnum \@lastchclass=6 \@ne \@chnum \@ne \else
2502 \ifnum \@lastchclass=7 5 \else
2503 \ifnum \@lastchclass=8 \tw@ \else
2504 \ifnum \@lastchclass=9 \thr@@
2505 \else \z@
2506 \ifnum \@lastchclass = 10 \else
2507 \edef\@nextchar{\expandafter\string\@nextchar}%
2508 \@chnum
2509 \if \@nextchar c\z@ \else
2510 \if \@nextchar l\@ne \else
2511 \if \@nextchar r\tw@ \else
2512 \z@ \@chclass
2513 \if\@nextchar |\@ne \let\@arrayrule\adl@arrayrule \else
2514 \if\@nextchar :\@ne \let\@arrayrule\adl@arraydashrule \else
2515 \if\@nextchar ;\@ne \let\@arrayrule\adl@argarraydashrule \else
2516 \if \@nextchar !6 \else
2517 \if \@nextchar @7 \else
2518 \if \@nextchar <8 \else
2519 \if \@nextchar >9 \else
2520 10
2521 \@chnum
2522 \if \@nextchar m\thr@@\else
2523 \if \@nextchar p4 \else
2524 \if \@nextchar b5 \else
2525 \z@ \@chclass \z@ \@preamerr \z@ \fi \fi \fi \fi \fi \fi
2526 \fi \fi \fi \fi \fi \fi \fi \fi \fi \fi \fi \fi}
2527
2528 % \end{macrocode}
2529 %
2530 % \begin{macro}{\@classz}
2531 % \changes{v1.4-2-1}{1999/06/25}
2532 % {Introduced because \textsf{array} uses it.}
2533 % \changes{v1.71}{2004/08/31}
2534 % {Modified to call \cs{adl@startmbox} and \cs{adl@endmbox} for
2535 % \textsf{array}'s \texttt{m}-columns.}
2536 % \SpecialIndex{\adl@putlrc}
2537 % \SpecialIndex{\adl@box}
2538 % \SpecialIndex{\adl@startmbox}
2539 % \SpecialIndex{\adl@endmbox}
2540 %
2541 % In \textsf{array}, |array| and |tabular| share common macro for class-0
2542 % named |\@classz|, which also generates the preamble for `|p|', `|m|' and
2543 % `|b|'. Thus we modify it to measure the height and depth of the
2544 % class-0 column by the macro |\adl@putlrc|, and to set |\adl@box| to the
2545 % box for `|p|' and its relatives. Note that a |m|-type preamble
2546 % (|@chnum|${}=3$) has to be generated to have |\adl@startmbox| and
2547 % |\adl@endmbox| in it because a |\vcenter| construct cannot be assigned
2548 % to |\adl@box| by |\setbox| directly.
2549 % \end{macro}
2550 %
2551 % \begin{macrocode}
2552 \def\@classz{\@classx
2553 \@tempcnta \count@
2554 \prepnext@tok
2555 \@addtopreamble{\ifcase \@chnum
2556 \hfil
2557 \adl@putlrc{\d@llarbegin \insert@column \d@llarend}\hfil \or
2558 \hskip1sp\adl@putlrc{\d@llarbegin \insert@column \d@llarend}\hfil \or
2559 \hfil\hskip1sp\adl@putlrc{\d@llarbegin \insert@column \d@llarend}\or
2560 \setbox\adl@box\hbox \adl@startmbox{\@nextchar}\insert@column
2561 \adl@endmbox\or
2562 \setbox\adl@box\vtop \@startpbox{\@nextchar}\insert@column \@endpbox \or
2563 \setbox\adl@box\vbox \@startpbox{\@nextchar}\insert@column \@endpbox
2564 \fi}\prepnext@tok}
2565 % \end{macrocode}
2566 %
2567 % \begin{macro}{\adl@class@start}
2568 % \changes{v1.4-2-1}{1999/06/25}
2569 % {Introduced for class number identification.}
2570 % \begin{macro}{\adl@class@iiiorvii}
2571 % \changes{v1.4-2-1}{1999/06/25}
2572 % {Introduced for class number identification.}
2573 %
2574 % Another stuff for compatibility is to refer to the class number for the
2575 % beginning of preamble which is different between \LaTeX{} and
2576 % \textsf{array}, and that for `|p|' or `|@|' to get the argument of `|;|' as
2577 % explained later. In the case with \textsf{array}, the former is class-4
2578 % and we use `|@|' (class-7) for the latter.
2579 % \end{macro}\end{macro}
2580 %
2581 % \begin{macrocode}
2582 \def\adl@class@start{4}
2583 \def\adl@class@iiiorvii{7}
2584
2585 % \end{macrocode}
2586 %
2587 % \subsubsection*{Without \textsf{array}}
2588 %
2589 % \begin{macro}{\@testpach}
2590 % \SpecialIndex{\@arrayrule}
2591 % \SpecialIndex{\adl@arrayrule}
2592 % \SpecialIndex{\adl@arraydashrule}
2593 % \SpecialIndex{\adl@argarraydashrule}
2594 %
2595 % The reason why and how we modify |\@testpach| of \LaTeX{} is same as those
2596 % of \textsf{array}.
2597 % \end{macro}
2598 %
2599 % \begin{macrocode}
2600 \else
2601 \def\@testpach#1{\@chclass \ifnum \@lastchclass=\tw@ 4\relax \else
2602 \ifnum \@lastchclass=\thr@@ 5\relax \else
2603 \z@ \if #1c\@chnum \z@ \else
2604 \if #1l\@chnum \@ne \else
2605 \if #1r\@chnum \tw@ \else
2606 \@chclass
2607 \if #1|\@ne \let\@arrayrule\adl@arrayrule \else
2608 \if #1:\@ne \let\@arrayrule\adl@arraydashrule \else
2609 \if #1;\@ne \let\@arrayrule\adl@argarraydashrule \else
2610 \if #1@\tw@ \else
2611 \if #1p\thr@@ \else \z@ \@preamerr 0\fi
2612 \fi \fi \fi \fi \fi \fi \fi \fi \fi}
2613
2614 % \end{macrocode}
2615 %
2616 % \begin{macro}{\@arrayclassz}
2617 % \begin{macro}{\@tabclassz}
2618 % \SpecialIndex{\adl@putlrc}
2619 %
2620 % Since \LaTeX{} has two macros for class-0, one for |array| and the other
2621 % for |tabular|, we have to modify both. Since the box for `|p|' is opened
2622 % by |\@@startpbox|, however, we may not worry about it.
2623 % \end{macro}\end{macro}
2624 %
2625 % \begin{macrocode}
2626 \def\@arrayclassz{\ifcase \@lastchclass \@acolampacol \or \@ampacol \or
2627 \or \or \@addamp \or
2628 \@acolampacol \or \@firstampfalse \@acol \fi
2629 \edef\@preamble{\@preamble
2630 \ifcase \@chnum
2631 \hfil\adl@putlrc{$\relax\@sharp$}\hfil
2632 \or \adl@putlrc{$\relax\@sharp$}\hfil
2633 \or \hfil\adl@putlrc{$\relax\@sharp$}\fi}}
2634 \def\@tabclassz{\ifcase \@lastchclass \@acolampacol \or \@ampacol \or
2635 \or \or \@addamp \or
2636 \@acolampacol \or \@firstampfalse \@acol \fi
2637 \edef\@preamble{\@preamble
2638 \ifcase \@chnum
2639 \hfil\adl@putlrc{\@sharp\unskip}\hfil
2640 \or \adl@putlrc{\@sharp\unskip}\hfil
2641 \or \hfil\hskip\z@ \adl@putlrc{\@sharp\unskip}\fi}}
2642 % \end{macrocode}
2643 %
2644 % \begin{macro}{\adl@class@start}
2645 % \changes{v1.4-2-1}{1999/06/25}
2646 % {Introduced for class number identification.}
2647 % \begin{macro}{\adl@class@iiiorvii}
2648 % \changes{v1.4-2-1}{1999/06/25}
2649 % {Introduced for class number identification.}
2650 % In \LaTeX{}, the beginning of preamble is class-6 and we use `|p|'
2651 % (class-3) to get the argument of `|;|'.
2652 % \end{macro}\end{macro}
2653 %
2654 % \begin{macrocode}
2655 \def\adl@class@start{6}
2656 \def\adl@class@iiiorvii{3}
2657 \fi
2658
2659 % \end{macrocode}
2660 %
2661 % Hereafter, codes for \LaTeX{} and \textsf{array} are common again.
2662 %
2663 % \begin{macro}{\adl@putlrc}
2664 % \changes{v1.5-2}{2002/06/26}
2665 % {\cs{adl@putlrc} must do \cs{unhbox}\cs{adl@box} to make glues
2666 % effective.}
2667 % \SpecialIndex{\adl@box}
2668 % \SpecialIndex{\adl@colhtdp}
2669 %
2670 % The macro |\adl@putlrc| is for class-0 preamble characters to set
2671 % |\adl@box| to the contents of a column, measure its height\slash depth by
2672 % |\adl@colhtdp| and put the box by |\unhbox| (not by |\box|) in order to
2673 % make the glues in the contents effective.
2674 % \end{macro}
2675 %
2676 % \begin{macrocode}
2677 \def\adl@putlrc#1{\setbox\adl@box\hbox{#1}\adl@colhtdp \unhbox\adl@box}
2678
2679 % \end{macrocode}
2680 %
2681 % \begin{macro}{\adl@arrayrule}
2682 % \changes{v1.4-2-1}{1999/06/25}
2683 % {Modified to replace \cs{adl@columns} with \cs{adl@ncol}.}
2684 % \begin{macro}{\adl@arraydashrule}
2685 % \changes{v1.4-2-1}{1999/06/25}
2686 % {Modified to replace \cs{adl@columns} with \cs{adl@ncol}.}
2687 % \begin{macro}{\adl@argarraydashrule}
2688 % \changes{v1.4-2-1}{1999/06/25}
2689 % {Modified to replace \cs{adl@columns} with \cs{adl@ncol}.}
2690 % \changes{v1.4-2-1}{1999/06/25}
2691 % {Modified to pretend \texttt{p} or \texttt{@} depending on if
2692 % \textsf{array} is in use.}
2693 % \begin{macro}{\adl@xarraydashrule}
2694 % \changes{v1.4-2-1}{1999/06/25}
2695 % {Modified to refer \cs{adl@class@start} rather than \LaTeX's 6.}
2696 % \changes{v1.4-2-2}{1999/06/25}
2697 % {Modified to add invisible rule of \cs{arrayrulewidth} wide if
2698 % \cs{ADLsome}\penalty\exhyphenpenalty\texttt{wide}.}
2699 % \changes{v1.7-3}{2004/5/21}
2700 % {Modified to add color arguments to \cs{adl@vline}\texttt{L}\slash
2701 % \texttt{R}}
2702 % \SpecialIndex{\@lastchclass}
2703 % \SpecialIndex{\adl@class@start}
2704 % \SpecialIndex{\ifadl@leftrule}
2705 % \SpecialIndex{\adl@vlineL}
2706 % \SpecialIndex{\adl@vlineR}
2707 % \SpecialIndex{\adl@ncol}
2708 % \SpecialIndex{\CT@arc@}
2709 % \SpecialIndex{\adl@dashgapcolor}
2710 % \SpecialIndex{\color}
2711 % \def\G{{\mit\Gamma}}
2712 %
2713 % The preamble parts for vertical solid- and dash-lines are constructed by
2714 % the macros |\adl@arrayrule| for `\verb!|!', |\adl@arraydashrule| for
2715 % `|:|', and |\adl@argarraydashrule| for `|;|'.\@ The macro;
2716 %
2717 % \begin{quote}
2718 % |\adl@xarraydashrule|\Meta{$c^L$}\Meta{$c^R$}|{|\meta{d}|/|\meta{g}|}|
2719 % \end{quote}
2720 %
2721 % is invoked by them to perform common operations. It at first checks the
2722 % preamble character is the first element of the preamble
2723 % (|\@lastchclass|${}=|\adl@class@start|$) or it follows another character
2724 % for vertical line (|\@lastchclass|${}=1$). If this is not satisfied, the
2725 % vertical line is put at the right edge of a column and thus
2726 % |\ifadl@leftrule| is set to false. Then it adds
2727 % |\adl@vlineL|\Meta{$\G_d$}\Meta{$\G_g$}
2728 % \Meta{$c^L$}|{|\meta{d}|/|\meta{g}|}| if |\ifadl@leftrule| is
2729 % true indicating the vertical line will appear at the left edge of the
2730 % column \meta{$c^L$}, or |\adl@vlineR|\Meta{$\G_d$}\Meta{$\G_g$}
2731 % \Meta{$c^R$}|{|\meta{d}|/|\meta{g}|}| otherwise. Note that \meta{$c^L$}
2732 % is always 1 for {\em main} preamble while \meta{$c^R$} is the column
2733 % number given by |\adl@ncol|, but \meta{$c^L$} may not be 1 for the
2734 % preamble of |\multicolumn| as described in \S\ref{sec:imp-mc}. Also note
2735 % that $\G_d$ and $\G_g$ are |\CT@arc@| and |\adl@dashgapcolor|
2736 % respectively whose bodies are |\color| for dashes and gaps specified by
2737 % |\arrayrulecolor| and |\dashgapcolor|, or |\relax| if they are not colored.
2738 %
2739 % In addition, an invisible |\vrule| of |\arrayrulewidth| wide is added if
2740 % both |\ADL#some#wide| and |\ADLactivate| are in effect, i.e.\ both
2741 % |\ifadl@zwrule| and |\ifadl@inactive| are false, to keep a space for
2742 % the vertical line having {\em real} width.
2743 % \end{macro}\end{macro}\end{macro}\end{macro}
2744 %
2745 % \begin{macro}{\adl@classv}
2746 % \begin{macro}{\adl@classvfordash}
2747 % \SpecialIndex{\@nextchar}
2748 % \SpecialIndex{\@chclass}
2749 % \SpecialIndex{\@classv}
2750 % \SpecialIndex{\adl@class@iiiorvii}
2751 %
2752 % The argument of `|;|' is not provided by |\adl@argarraydashrule| but is
2753 % directly passed from the preamble text through |\@nextchar|. This direct
2754 % passing is implemented by the following trick. The macro
2755 % |\adl@argarraydashrule| set |\@chclass| to |\adl@class@iiiorvii| to
2756 % pretend it is for `|p|' if \textsf{array} is not in use, or `|@|'
2757 % otherwise. Then it temporally changes the definition of |\@classv|, which
2758 % is incidentally for the argument of `|p|' and `|@|' in the case without/with
2759 % \textsf{array} respectively, to |\adl@classvfordash| to process the
2760 % argument of `|;|' rather than that of `|p|' or `|@|'\@. Then
2761 % |\adl@classvfordash| is invoked by |\@mkpream| and it adds the argument to
2762 % |\@preamble|. Finally, it restores the definition of |\@classv| and sets
2763 % |\@chclass| to 1 to indicate that the last item is a vertical line
2764 % specification.
2765 % \end{macro}\end{macro}
2766 %
2767 % \begin{macrocode}
2768 \def\adl@arrayrule{%
2769 \adl@xarraydashrule
2770 {\@ne}{\adl@ncol}{{\z@/\z@}}}
2771 \def\adl@arraydashrule{%
2772 \adl@xarraydashrule
2773 {\@ne}{\adl@ncol}%
2774 {{\dashlinedash/\dashlinegap}}}
2775 \def\adl@argarraydashrule{%
2776 \adl@xarraydashrule
2777 {\@ne}{\adl@ncol}{}%
2778 \@chclass\adl@class@iiiorvii \let\@classv\adl@classvfordash}
2779 \def\adl@xarraydashrule#1#2#3{%
2780 \ifnum\@lastchclass=\adl@class@start\else
2781 \ifnum\@lastchclass=\@ne\else
2782 \adl@leftrulefalse \fi\fi
2783 \ifadl@zwvrule\else \ifadl@inactive\else
2784 \@addtopreamble{\vrule\@width\arrayrulewidth
2785 \@height\z@ \@depth\z@}\fi \fi
2786 \ifadl@leftrule
2787 \@addtopreamble{\adl@vlineL{\CT@arc@}{\adl@dashgapcolor}%
2788 {\number#1}#3}%
2789 \else \@addtopreamble{\adl@vlineR{\CT@arc@}{\adl@dashgapcolor}%
2790 {\number#2}#3}\fi}
2791 \let\adl@classv\@classv
2792 \def\adl@classvfordash{\@addtopreamble{{\@nextchar}}\let\@classv\adl@classv
2793 \@chclass\@ne}
2794
2795 %%^L
2796 % \end{macrocode}
2797 %
2798 %
2799 % \subsection{Building Columns}
2800 % \label{sec:imp-col}
2801 %
2802 % \begin{macro}{\adl@preaminit}
2803 % \begin{macro}{\adl@colhtdp}
2804 % \begin{macro}{\adl@vlineL}
2805 % \begin{macro}{\adl@vlineR}
2806 %
2807 % If \textsf{array} is not in use, after the |\@preamble| is completed, the
2808 % control sequences for macros in it should regain their own definition.
2809 % The macro |\adl@preaminit| performs this operation for macros we
2810 % introduced, |\adl@colhtdp|, |\adl@vlineL| and |\adl@vlineR|. For the case
2811 % with \textsf{array}, we will call |\adl@preaminit| in \textsf{arydshln} to
2812 % initiate them with the definitions as described later.
2813 % \end{macro}\end{macro}\end{macro}\end{macro}
2814 %
2815 % \begin{macrocode}
2816
2817 %% Building Columns
2818
2819 \def\adl@preaminit{\let\adl@colhtdp\adl@@colhtdp
2820 \let\adl@vlineL\adl@@vlineL \let\adl@vlineR\adl@@vlineR}
2821
2822 % \end{macrocode}
2823 %
2824 % \begin{macro}{\adl@@colhtdp}
2825 % For the measurement of the height and depth of a row, |\adl@@colhtdp|
2826 % compares |\adl@height| and |\adl@depth| to the height and depth of
2827 % |\adl@box| which contains the main part of the column to be built, and
2828 % |\global|-ly updates the registers if they are less.
2829 % \end{macro}
2830 %
2831 % \begin{macrocode}
2832 \def\adl@@colhtdp{%
2833 \ifdim\adl@height<\ht\adl@box \global\adl@height\ht\adl@box \fi
2834 \ifdim\adl@depth<\dp\adl@box \global\adl@depth\dp\adl@box\fi}
2835
2836 % \end{macrocode}
2837 %
2838 % \begin{macro}{\adl@@vlineL}
2839 % \changes{v1.7-3}{2004/5/21}
2840 % {Color arguments are added.}
2841 % \begin{macro}{\adl@@vlineR}
2842 % \changes{v1.7-3}{2004/5/21}
2843 % {Color arguments are added.}
2844 % \begin{macro}{\adl@@ivline}
2845 % \changes{v1.7-3}{2004/5/21}
2846 % {Invocations of \cs{adl@setcolor} are added.}
2847 % \begin{macro}{\adl@setcolor}
2848 % \changes{v1.7-3}{2004/5/21}
2849 % {Introduced to color vertical lines.}
2850 % \begin{macro}{\adl@nocolor}
2851 % \changes{v1.7-3}{2004/5/21}
2852 % {Introduced to examine if coloring is specified.}
2853 % \begin{macro}{\adl@dashcolor}
2854 % \changes{v1.7-3}{2004/5/21}
2855 % {Introduced as the temporary variable of color specification of
2856 % dashes.}
2857 % \begin{macro}{\adl@gapcolor}
2858 % \changes{v1.7-3}{2004/5/21}
2859 % {Introduced as the temporary variable of color specification of
2860 % gaps.}
2861 % \SpecialIndex{\color}
2862 % \SpecialIndex{\current@color}
2863 % \def\G{{\mit\Gamma}} \let\g\gamma
2864 % The macro
2865 % |\adl@@vlineL|\meta{$\G_d$}\meta{$\G_g$}\meta{c}|{|\meta{d}|/|\meta{g}|}|
2866 % adds the element
2867 % $e=\langle c,d,g\rangle={}$
2868 % |\@elt|\Meta{c}\Meta{d}\Meta{g}\Meta{$\g_d$}\Meta{$\g_g$} to the tail of
2869 % the list |\adl@colsL| to construct $C^L_i$, where $\g_d$ and $\g_g$ are
2870 % the color specifications given by |\color| macros in $\G_d$ and $\G_g$.
2871 % The macro |\add@@vlineR| performs similar operation but the element is
2872 % added to the head of |\adl@colsR| for $C^R_i$ because it is processed
2873 % right-to-left manner. The argument \meta{d} and \meta{g} are extracted by
2874 % the macro |\adl@ivline| which converts given dimensional values of them to
2875 % integers. It also sets \meta{d} and \meta{g} to 0 (i.e. solid-line) if one
2876 % of given values are not positive, in order to make it sure that one dash
2877 % segment has positive length. Then it invokes |\adl@setcolor| to define
2878 % |\adl@dashcolor| and |\adl@gapcolor| with the color specification of
2879 % $\G_d$ and $\G_g$. Since |\adl@setcolor| locally expands |\color| macro
2880 % in $\G_d$ and $\G_g$ to define |\currnet@color| that becomes the body of
2881 % |\adl@dashcolor| ($\g_d$) and |\adl@gapcolor| ($\g_g$) with expansion,
2882 % different |\color| specifications of a color, such as |\color{red}|
2883 % and |\color[rgb]{1,0,0}|, will produce a unified result such as
2884 % |{rgb 1 0 0}|. If $\G_d$ or $\G_g$ is |\relax| which is the body of
2885 % |\adl@nocolor|, $\g_d$ or $\g_g$ is also |\relax| to indicate dashes are
2886 % colored (or not colored) as done in outer world and gaps are transparent.
2887 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
2888 %
2889 % \begin{macrocode}
2890 \def\adl@@vlineL#1#2#3#4{\adl@ivline#4\@nil{#1}{#2}%
2891 \xdef\adl@colsL{\adl@colsL
2892 \@elt{#3}{\number\@tempcnta}{\number\@tempcntb}%
2893 {\adl@dashcolor}{\adl@gapcolor}}}
2894 \def\adl@@vlineR#1#2#3#4{\adl@ivline#4\@nil{#1}{#2}%
2895 \xdef\adl@colsR{%
2896 \@elt{#3}{\number\@tempcnta}{\number\@tempcntb}%
2897 {\adl@dashcolor}{\adl@gapcolor}%
2898 \adl@colsR}}
2899 \def\adl@ivline#1/#2\@nil#3#4{%
2900 \@tempdima#1\relax \@tempcnta\@tempdima
2901 \@tempdima#2\relax \@tempcntb\@tempdima
2902 \ifnum\@tempcnta>\z@ \else \@tempcnta\z@ \@tempcntb\z@ \fi
2903 \ifnum\@tempcntb>\z@ \else \@tempcnta\z@ \@tempcntb\z@ \fi
2904 \adl@setcolor\adl@dashcolor{#3}\adl@setcolor\adl@gapcolor{#4}}
2905 \def\adl@setcolor#1#2{\def\@tempa{#2}\ifx\@tempa\adl@nocolor \def#1{\relax}%
2906 \else{#2\xdef#1{\current@color}}\fi}
2907 \def\adl@nocolor{\relax}
2908 % \end{macrocode}
2909 %
2910 % \begin{macro}{\adl@colhtdp}
2911 % \changes{v1.4-2-1}{1999/06/25}
2912 % {Initialized by calling \cs{adl@preaminit}.}
2913 % \begin{macro}{\adl@vlineL}
2914 % \changes{v1.4-2-1}{1999/06/25}
2915 % {Initialized by calling \cs{adl@preaminit}.}
2916 % \begin{macro}{\adl@vlineR}
2917 % \changes{v1.4-2-1}{1999/06/25}
2918 % {Initialized by calling \cs{adl@preaminit}.}
2919 % \SpecialIndex{\adl@preaminit}
2920 %
2921 % After |\adl@@colhtdp|, |\adl@@vlineL| and |\adl@@vlineR| are defined,
2922 % we call |\adl@preaminit| to |\let| their single |@| counterparts be equal
2923 % to them. Therefore, in the case with \textsf{array}, |\adl@colhtdp| etc.\
2924 % are temporarily |\relax| when |\@preamble| is being generated in the group
2925 % of |\@mkpream|, and regain their own definitions outside the group where
2926 % the completed |\@preamble| is referred to.
2927 % \end{macro}\end{macro}\end{macro}
2928 %
2929 % \begin{macrocode}
2930 \adl@preaminit
2931
2932 % \end{macrocode}
2933 %
2934 % \begin{macro}{\adl@inactivevl}
2935 % \changes{v1.4-2-3}{1999/06/25}
2936 % {Introduced to emulate `\texttt{:}' and \texttt{;} by
2937 % \protect\VERTBAR.}
2938 % \changes{v1.7-3}{2004/5/21}
2939 % {Modified to color the \cs{vline} by the first argument.}
2940 % \SpecialIndex{\ifadl@zwvrule}
2941 %
2942 % If |\ADLinactivate| is in effect, |\adl@vlineL|\slash|R| and
2943 % |\adl@@vlineL|\slash|R| are |\let|-equal
2944 % to |\adl@inactivevl|. This macro simply puts a |\vrule| by |\vline| with
2945 % |\color| (or |\relax|) in its first argument and
2946 % with\slash without negative |\hskip| of a half of |\arrayrulewidth| wide
2947 % depending on |\ifadl@zwvrule|, discarding other arguments.
2948 % \end{macro}
2949 %
2950 % \begin{macrocode}
2951 \def\adl@inactivevl#1#2#3#4{\ifadl@zwvrule \hskip-.5\arrayrulewidth \fi
2952 {#1\vline}\ifadl@zwvrule \hskip-.5\arrayrulewidth \fi}
2953
2954 % \end{macrocode}
2955 %
2956 %
2957 % \begin{macro}{\@@startpbox}
2958 % \begin{macro}{\@@endpbox}
2959 % \begin{macro}{\@startpbox}
2960 % \changes{v1.75}{2018/12/26}
2961 % {Introduced to cope with the \cs{color@begingroup}\slash
2962 % \cs{color@endgroup} problem.}
2963 % \begin{macro}{\@endpbox}
2964 % \changes{v1.4-2-1}{1999/06/25}
2965 % {Introduced because \textsf{array} uses it.}
2966 % \changes{v1.75}{2018/12/26}
2967 % {Modified to ensure that the macro has \cs{color@endgroup}
2968 % irrespective of \textsf{array}'s version.}
2969 % \begin{macro}{\adl@startmbox}
2970 % \changes{v1.71}{2004/08/31}
2971 % {Introduced to the bug fix of \textsf{array}'s \texttt{m}-columns.}
2972 % \changes{v1.75}{2018/12/26}
2973 % {Replace \cs{@startpbox} with \cs{adl@org@startpbox} to avoid the
2974 % color-grouping problem.}
2975 % \changes{v1.76}{2019/02/21}
2976 % {Replace \cs{adl@org@startpbox} with \cs{@startpbox} to do
2977 % color-grouping always and to invoke \cs{adl@LTstartpbox} through
2978 % \cs{@stratpbox} when \textsf{longtable} is in use.}
2979 % \begin{macro}{\adl@endmbox}
2980 % \changes{v1.71}{2004/08/31}
2981 % {Introduced to the bug fix of \textsf{array}'s \texttt{m}-columns.}
2982 % \changes{v1.76}{2019/02/21}
2983 % {Replace \cs{adl@org@endpbox} with what \cs{@endpbox} of
2984 % \textsf{array} has to perform color-grouping always.}
2985 % \SpecialIndex{\@finalstrut}
2986 % \SpecialIndex{\color@begingroup}
2987 % \SpecialIndex{\color@endgroup}
2988 % \SpecialIndex{\adl@box}
2989 % \SpecialIndex{\adl@colhtdp}
2990 % \SpecialIndex{\adl@LTstartpbox}
2991 %
2992 % The macros to make |\parbox| for `|p|', |\@@startpbox| and
2993 % \hbox{|\@@endpbox|}, are modified for height\slash depth measurement. The
2994 % code for |\@@endpbox| is based on that of \LaTeXe{} to fix the bug of
2995 % |\strut|-ing in \LaTeX-2.09, but |\@finalstrut| is manually expanded
2996 % because it is not available in \LaTeX-2.09.
2997 %
2998 % In \textsf{array}, these two macros are not used but |\@startpbox| and
2999 % |\@endpbox| are. Until v2.4h, the former may be untouched and the latter
3000 % can be |\let|-equal to |\@@endpbox|. However in v2.4i, |\color@begingroup|
3001 % and |\color@endgroup| are added to them to make the compatibility issue a
3002 % little bit complicated. That is, our version of |\@endpbox| would have to
3003 % have |\color@endgroup| if and only if \textsf{array} is v2.4i or later
3004 % because |\@startpbox| has |\color@begingroup| in these versions, if we
3005 % relied on the original |\@startpbox|. To avoid version dependent coding,
3006 % we copy the new definition of |\@starpbox| to ensure it has
3007 % |\color@begingroup| and let our own |\@endpbox| with height\slash depth
3008 % measurement have |\color@endgroup| irrespective of the version of
3009 % \textsf{array}. Note that the assigning the box having |'p'| or |'b'| to
3010 % |\adl@box| for the measurement is done in our own |\@classz| shown in
3011 % \S\ref{sec:imp-pream}.
3012 %
3013 % As for |m|-type columns, we need a special care because its body
3014 % |\vcenter| cannot be assigned directly to |\adl@box| by
3015 % |\setbox|\footnote{
3016 %
3017 % The author had forgotten this fact until Morten H\o gholm pointed out it.
3018 % Thanks Morten.}.
3019 %
3020 % Thus we enclose a |$\vcenter{|\ldots|}$| construct in a |\hbox| and assign
3021 % it to |\adl@box|. The construct is opened and closed by the macros
3022 % |\adl@startmbox| and |\adl@endmbox| with |\@startpbox| and what
3023 % |\@endpbox| of \textsf{array} 2.4j has, in order to perform color-grouping
3024 % regardless of the version of \textsf{array}. The latter macro also has
3025 % our own function to measure the height and depth of the |\hbox| by
3026 % |\adl@colhtdp|. Note that |\@startpbox| in |\adl@startmbox| can be
3027 % different from the definition made here and, more specifically, will be
3028 % |\adl@LTstartpbox| when \textsf{longtable} is in use. Also noote that the
3029 % mechanism with |\vcenter| was replaced with a vertical shift of a box for
3030 % '|m|' in v2.4f of \textsf{array}, but we stick the old mechanism to avoid
3031 % version dependent coding.
3032 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
3033 %
3034 % \begin{macrocode}
3035 \def\@@startpbox#1{\setbox\adl@box\vtop\bgroup \hsize#1\@arrayparboxrestore}
3036 \def\@@endpbox{\unskip \ifhmode \nobreak
3037 \vrule\@width\z@\@height\z@\@depth\dp\@arstrutbox \fi
3038 \par \egroup \adl@colhtdp \box\adl@box \hfil}
3039 \def\@startpbox#1{\bgroup
3040 \color@begingroup
3041 \setlength\hsize{#1}\@arrayparboxrestore
3042 \everypar{%
3043 \vrule\@height\ht\@arstrutbox\@width\z@
3044 \everypar{}}%
3045 }
3046 \def\@endpbox{\@finalstrut\@arstrutbox \color@endgroup \egroup
3047 \adl@colhtdp \box\adl@box \hfil}
3048 \def\adl@startmbox{\bgroup $\vcenter\@startpbox}
3049 \def\adl@endmbox{\@finalstrut\@arstrutbox \color@endgroup \egroup
3050 $\egroup \adl@colhtdp \box\adl@box \hfil}
3051
3052 %%^L
3053 % \end{macrocode}
3054 %
3055 %
3056 %
3057 % \subsection{Multi-columns}
3058 % \label{sec:imp-mc}
3059 %
3060 % \begin{macro}{\multicolumn}
3061 % \changes{v1.4-2-1}{1999/06/25}
3062 % {Modified for several reason.}
3063 % \begin{macro}{\adl@preamble}
3064 % \begin{macro}{\adl@mcaddamp}
3065 % \changes{v1.4-2-1}{1999/06/25}
3066 % {Introduced for the complaint on multiple columns if with
3067 % \textsf{array}.}
3068 % \begin{macro}{\adl@activatepbox}
3069 % \changes{v1.4-2-1}{1999/06/25}
3070 % {Introduced to do nothing if with \textsf{array}.}
3071 %
3072 % The macro |\multicolumn| is modified for the following.
3073 %
3074 % \begin{itemize}
3075 % \item
3076 % The macros to construct the parts of |\@preamble| for vertical lines,
3077 % |\adl@arrayrule|, |\adl@arraydashrule| and |\adl@argarraydashrule|, have
3078 % to perform operations slightly different from those for main preamble.
3079 % Thus they are |\def|-ined to multi-column version |\adl@mcarrayrule|, etc.
3080 % \changes{v1.3}{1998/10/08}
3081 % {\cs{def}-s for \cs{adl@mcarrayrule} etc. are enclosed in a group.}
3082 % These |\def|-initions are enclosed in a group so that they are not
3083 % affected to |array| or |tabular| which may occur in the third argument of
3084 % |\multicolumn|. In order to make |\@preamble| work well outside of the
3085 % group containing |\@makepream|, |\adl@preamble| is |\global|-ly |\let|-equal
3086 % to |\@preamble| just after |\@makepream| in the group and then
3087 % reverse |\let|-assignment is performed just after the group is closed.
3088 % These global assignment is unnecessary with \textsf{array} because
3089 % |\@preamlbe| is constructed |\global|-ly, but safe.
3090 %
3091 % Since this grouping nullifies the effect of |\adl@preaminit| called in
3092 % |\@mkpream|, we call |\adl@preaminit| again after the group closing.
3093 %
3094 % \item
3095 % In \textsf{array}, |\@addamp| to make |\@preamble| for |\multicolumn| has
3096 % a different definition from that for main one. Thus it is |\let|-equal
3097 % to |\adl@mcaddamp| whose definition is switched by |\ifadl@usingarypkg|.
3098 %
3099 % \item
3100 % If \textsf{array} is in use, |\@preamble| has to be |\xdef|-ed once again
3101 % by |\@addpreamble| with an |\@empty| argument after |\@mkpreamble| to
3102 % expand the contents of |\toks| registers. This is performed whether or
3103 % not with \textsf{array} because it is safe.
3104 %
3105 % \item
3106 % As done in |\@array|, |\set@typeset@protect| is replaced with direct
3107 % |\let|.
3108 %
3109 % \item
3110 % If without \textsf{array}, |\@startpbox| and |\@endpbox| should be
3111 % |\let|-equal to their |@@| counterparts, while should not with
3112 % \textsf{array}. Thus we define |\adl@activatepbox| to do or not to do so
3113 % depending on |\ifadl@usingarypkg|.
3114 %
3115 % \item
3116 % The counter |\adl@currentcolumn| is |\global|-ly incremented by the first
3117 % argument of |\multicolumn| (number of columns to be |\span|-ned).
3118 % \end{itemize}
3119 %
3120 % Note that |\adl@columns| is modified by |\@mkpream|, but it is not
3121 % referred to by |\adl@mcarrayrule| etc., and its value is restored before
3122 % referred to by |\hdashline|, etc.
3123 % \end{macro}\end{macro}\end{macro}\end{macro}
3124 %
3125 % \SpecialIndex{\adl@arrayrule}
3126 % \SpecialIndex{\adl@arraydashrule}
3127 % \SpecialIndex{\adl@argarraydashrule}
3128 % \SpecialIndex{\adl@mcarrayrule}
3129 % \SpecialIndex{\adl@mcarraydashrule}
3130 % \SpecialIndex{\adl@mcargarraydashrule}
3131 % \SpecialIndex{\ifadl@usingarypkg}
3132 % \SpecialIndex{\adl@currentcolumn}
3133 %
3134 % \begin{macrocode}
3135
3136 %% Multi-Columns
3137
3138 \def\multicolumn#1#2#3{\multispan{#1}\begingroup \begingroup
3139 \def\adl@arrayrule{\adl@mcarrayrule{#1}}%
3140 \def\adl@arraydashrule{\adl@mcarraydashrule{#1}}%
3141 \def\adl@argarraydashrule{\adl@mcargarraydashrule{#1}}%
3142 \let\@addamp\adl@mcaddamp
3143 \@mkpream{#2}\@addtopreamble\@empty
3144 \global\let\adl@preamble\@preamble \endgroup
3145 \let\@preamble\adl@preamble
3146 \def\@sharp{#3}\let\protect\relax
3147 \adl@activatepbox
3148 \adl@preaminit
3149 \@arstrut \@preamble\hbox{}\endgroup
3150 \global\advance\adl@currentcolumn#1\ignorespaces}
3151 \ifadl@usingarypkg
3152 \def\adl@mcaddamp{\if@firstamp\@firstampfalse \else\@preamerror5\fi}
3153 \let\adl@activatepbox\relax
3154 \else
3155 \let\adl@mcaddamp\@addamp
3156 \def\adl@activatepbox{\let\@startpbox\@@startpbox
3157 \let\@endpbox\@@endpbox}
3158 \fi
3159
3160 % \end{macrocode}
3161 %
3162 % \begin{macro}{\adl@mcarrayrule}
3163 % \begin{macro}{\adl@mcarraydashrule}
3164 % \begin{macro}{\adl@mcargarraydashrule}
3165 % \changes{v1.4-2-1}{1999/06/25}
3166 % {Modified to pretend \texttt{p} or \texttt{@} depending on if
3167 % \textsf{array} is in use.}
3168 % \SpecialIndex{\adl@arrayrule}
3169 % \SpecialIndex{\adl@arraydashrule}
3170 % \SpecialIndex{\adl@argarraydashrule}
3171 % \SpecialIndex{\adl@currentcolumn}
3172 %
3173 % The preamble parts for vertical lines are constructed by the macros
3174 % |\adl@mcarrayrule|, |\adl@mcarraydashrule| and |\adl@mcargarraydashrule|
3175 % to which the first argument \meta{$n$} of |\multicolumn| is passed to know
3176 % the number of columns to be |\span|-ned. They are similar to their
3177 % relatives for main preamble, |\adl@arrayrule|, etc., but the arguments
3178 % \meta{$c^L$} and \meta{$c^R$} passed to |\adl@xarraydashrule| are;
3179 % $$
3180 % c^L=c,\qquad c^R=c+n-1
3181 % $$
3182 % where $c={}$|\adl@currentcolumn|. This makes leading vertical lines drawn
3183 % at the left edge of the leftmost |\span|-ned column and trailing ones at
3184 % the right edge of the rightmost column.
3185 % \end{macro}\end{macro}\end{macro}
3186 %
3187 % \begin{macrocode}
3188 \def\adl@mcarrayrule#1{\@tempcnta#1\advance\@tempcnta\adl@currentcolumn
3189 \advance\@tempcnta\m@ne
3190 \adl@xarraydashrule
3191 {\adl@currentcolumn}{\@tempcnta}{{\z@/\z@}}}
3192 \def\adl@mcarraydashrule#1{\@tempcnta#1\advance\@tempcnta\adl@currentcolumn
3193 \advance\@tempcnta\m@ne
3194 \adl@xarraydashrule
3195 {\adl@currentcolumn}{\@tempcnta}%
3196 {{\dashlinedash/\dashlinegap}}}
3197 \def\adl@mcargarraydashrule#1{\@tempcnta#1\advance\@tempcnta\adl@currentcolumn
3198 \advance\@tempcnta\m@ne
3199 \adl@xarraydashrule
3200 {\adl@currentcolumn}{\@tempcnta}{}%
3201 \@chclass\adl@class@iiiorvii \let\@classv\adl@classvfordash}
3202
3203 %%^L
3204 % \end{macrocode}
3205 %
3206 %
3207 %
3208 % \subsection{End of Rows}
3209 % \label{sec:imp-eor}
3210 %
3211 % \begin{macro}{\@xarraycr}
3212 % \changes{v1.4-2-1}{1999/06/25}
3213 % {The version for \textsf{array} is introduced.}
3214 % \begin{macro}{\@xtabularcr}
3215 % \begin{macro}{\@xargarraycr}
3216 % \begin{macro}{\@yargarraycr}
3217 % \SpecialIndex{\ifadl@usingarypkg}
3218 %
3219 % At the end of the $i\th$ row, we have to calculate $h_i$ which is the height
3220 % plus depth of the row, and add elements \meta{$C^L_i,h_i$} and
3221 % \meta{$C^R_i,h_i$} to $R^L$ and $R^R$. To do this, |\cr|-s in the macros
3222 % \hbox{|\@xarraycr|}, |\@xtabularcr|, |\@xargarraycr| are replaced with our
3223 % own |\adl@cr|. The macro |\@yargarraycr|\meta{dimen} is also modified but
3224 % its |\cr| is replaced with |\adl@argcr|\meta{dimen} to add (negative)
3225 % |\dimen| to $h_i$. Note that |\@xargarraycr|\meta{dimen} uses ordinary
3226 % |\adl@cr| because the extra vertical space of \meta{dimen} is inserted to
3227 % the last column.
3228 %
3229 % Note that the implementation of |\@xarraycr| is slightly different between
3230 % \LaTeX{} and \textsf{array}, we have to have two versions and choose one.
3231 % \end{macro}\end{macro}\end{macro}\end{macro}
3232 %
3233 % \begin{macrocode}
3234
3235 %% End of row
3236
3237 \ifadl@usingarypkg
3238 \def\@xarraycr{\@ifnextchar[{\@argarraycr}{\ifnum0=`{}\fi\adl@cr}}
3239 \else
3240 \def\@xarraycr{\@ifnextchar[{\@argarraycr}{\ifnum0=`{\fi}${}\adl@cr}}
3241 \fi
3242 \def\@xtabularcr{\@ifnextchar[{\@argtabularcr}{\ifnum0=`{\fi}\adl@cr}}
3243 \def\@xargarraycr#1{\@tempdima#1\advance\@tempdima\dp\@arstrutbox
3244 \vrule\@height\z@\@depth\@tempdima\@width\z@
3245 \adl@cr}
3246 \def\@yargarraycr#1{\adl@argcr{#1}\noalign{\vskip #1}}
3247
3248 % \end{macrocode}
3249 %
3250 % \begin{macro}{\adl@cr}
3251 % \begin{macro}{\adl@argcr}
3252 % The macro |\adl@cr| and |\adl@argcr| perform |\cr| and then invoke the
3253 % common macro |\adl@@cr|\meta{x}. The argument \meta{x} is the extra
3254 % (negative) vertical space for |\adl@argcr|, while it is 0 for |\adl@cr|.
3255 % \end{macro}\end{macro}
3256 %
3257 % \begin{macro}{\adl@@cr}
3258 % \changes{v1.4-2-6}{1999/06/25}
3259 % {Modified to set \cs{adl@finaldepth}.}
3260 % \changes{v1.6}{2003/8/25}
3261 % {Modified to insert \cs{adl@discard}.}
3262 % \SpecialIndex{\@arstrutbox}
3263 % \SpecialIndex{\adl@totalheight}
3264 % \SpecialIndex{\adl@colsL}
3265 % \SpecialIndex{\adl@colsR}
3266 % \SpecialIndex{\adl@rowsL}
3267 % \SpecialIndex{\adl@rowsR}
3268 % \SpecialIndex{\adl@connect}
3269 % \SpecialIndex{\adl@discard}
3270 % \SpecialIndex{\adl@finaldepth}
3271 % \SpecialIndex{\adl@height}
3272 % \SpecialIndex{\adl@depth}
3273 % \SpecialIndex{\adl@currentcolumn}
3274 %
3275 % The macro |\adl@@cr|\meta{x} at first calculate $h_i$ as follows. The
3276 % registers |\adl@height|${}=\eta$ and |\adl@depth|${}=\delta$ have the
3277 % maximum height and depth of the columns in the row. However, they could
3278 % be smaller than the height and/or depth of |\@arstrutbox|, $\eta_s$ and
3279 % $\delta_s$. If so, the height and/or depth of the row are $\eta_s$ and
3280 % $\delta_s$. Therefore, $h_i$ is caluclated by;
3281 % $$
3282 % h_i = \max(\eta,\eta_s) + \max(\delta,\delta_s).
3283 % $$
3284 % Additionally, if the extra space \meta{x} is negative, a vertical space of
3285 % $x$ is inserted below the row\footnote{
3286 %
3287 % Before v1.54, negative \meta{x} shrinks the height of the row by $\vert
3288 % x\vert$. Although the former result may be more appropriate if the row
3289 % has vertical lines than the current because lines extrude to the next row
3290 % now, new feature is considered compatible with original
3291 % \texttt{array}\slash \texttt{tabular}.}.
3292 %
3293 % Thus the integer value of $h_i+x$ is
3294 % |\global|-ly added to |\adl@totalheight|, and the elements
3295 % \meta{$C^L_i{=}\cs{adl@colsL},h_i$} and
3296 % \meta{$C^R_i{=}\cs{adl@colsR},h_i$} are added to the tail of
3297 % $R^L={}$|\adl@rowsL| and $R^C={}$|\adl@rowsR|\@. If $x$ is not 0
3298 % (negative), {\it discard}$(x)$ or {\it connect}$(x)$ is also added after
3299 % \meta{$C^{L/R}_i,h_i$} according to the current environment (|longtable|
3300 % or not). In the real implementation, $R^L$ and $R^C$ has the following
3301 % format of \meta{rows}.
3302 %
3303 %
3304 % \begin{eqnarray*}
3305 % \meta{rows}&::=&
3306 % [\meta{row}\hbox{\tt;}]^*\\
3307 % \meta{row}&::=&
3308 % \hbox{{\tt(}\meta{cols}{\tt/}\meta{$h_i$}{\tt)}}\\
3309 % \meta{cols}&::=&
3310 % [\hbox{\cs{elt}\Meta{c}\Meta{d}\Meta{g}}]^*\ \vert\qquad\qquad&
3311 % \hbox to.3\textwidth{\ldots\ $C^L$ or $C^R$\hss}\cr
3312 % && \cs{adl@connect}\ \vert&
3313 % \hbox to.3\textwidth{\ldots\ for {\it connect}$(h_i)$\hss}\cr
3314 % && \cs{adl@discard}\ \vert&
3315 % \hbox to.3\textwidth{\ldots\ for {\it discard}$(h_i)$\hss}\cr
3316 % && \cs{relax}&
3317 % \hbox to.3\textwidth{\ldots\ for {\it disconnect}$(h_i)$\hss}\cr
3318 % \noalign{\vskip-\baselineskip}
3319 % \end{eqnarray*}
3320 %
3321 % Since |\adl@discard| is |\def|-ined as |\adl@connect| by |\adl@arrayinit|,
3322 % added |\adl@discard| transforms itself into |\adl@connect| if current
3323 % envrionment is not |longtable|. Otherwise, as we make |\adl@discard|
3324 % |\let|-equal to |\relax| when a |longtable| environment starts, it keeps
3325 % its own form.
3326 %
3327 % Then, |\adl@finaldepth| is set to |\adl@depth| if $x$ is zero, or to zero
3328 % otherwise (negative), in order to make the depth of |array|\slash|tabular|
3329 % equal to that of the last row. Finally, |\adl@colsL|, |\adl@colsR|,
3330 % |\adl@currentcolumn|, |\adl@height| and |\adl@depth| are reinitialized to
3331 % process the next row.
3332 % \end{macro}
3333 %
3334 % \begin{macrocode}
3335 \def\adl@cr{\cr\noalign{\adl@@cr\z@}}
3336 \def\adl@argcr#1{\cr\noalign{\adl@@cr{#1}}}
3337 \def\adl@@cr#1{
3338 \ifdim\adl@height<\ht\@arstrutbox \adl@height\ht\@arstrutbox\fi
3339 \ifdim\adl@depth<\dp\@arstrutbox \adl@depth\dp\@arstrutbox\fi
3340 \advance\adl@height\adl@depth
3341 \global\advance\adl@totalheight\adl@height
3342 \@tempdima#1\relax \global\advance\adl@totalheight\@tempdima
3343 \xdef\adl@rowsL{\adl@rowsL
3344 (\adl@colsL/\number\adl@height);%
3345 \ifdim#1=\z@\else (\adl@discard/\number\@tempdima);\fi}%
3346 \xdef\adl@rowsR{\adl@rowsR
3347 (\adl@colsR/\number\adl@height);%
3348 \ifdim#1=\z@\else (\adl@discard/\number\@tempdima);\fi}%
3349 \gdef\adl@colsL{}\gdef\adl@colsR{}
3350 \global\adl@currentcolumn\@ne
3351 \ifdim#1=\z@ \global\adl@finaldepth\adl@depth
3352 \else \global\adl@finaldepth\z@\fi
3353 \global\adl@height\z@ \global\adl@depth\z@}
3354
3355 %%^L
3356 % \end{macrocode}
3357 %
3358 %
3359 %
3360 % \subsection{Horizontal Lines}
3361 % \label{sec:imp-hline}
3362 %
3363 % \begin{macro}{\hline}
3364 % \changes{v1.4-2-6}{1999/06/25}
3365 % {Modified to set \cs{adl@finaldepth} to zero.}
3366 % \changes{v1.7-1}{2004/5/21}
3367 % {Modified to shift up if null-wide.}
3368 % \begin{macro}{\cline}
3369 % \changes{v1.4-2-6}{1999/06/25}
3370 % {Modified to set \cs{adl@finaldepth} to zero.}
3371 % \changes{v1.7-1}{2004/5/21}
3372 % {Modified to shift up if null-wide.}
3373 % \SpecialIndex{\adl@hline}
3374 % \SpecialIndex{\adl@xhline}
3375 % \SpecialIndex{\adl@connect}
3376 % \SpecialIndex{\adl@finaldepth}
3377 % \SpecialIndex{\adl@org@cline}
3378 %
3379 % The macro |\hline| is modified to insert |\vskip-\arrayrulewidth| before
3380 % drawing if |\ADLnullwidehline| is in effect, or to add the element ${\it
3381 % connect}(w)={}$|(\adl@connect/|\allowbreak|\number\arrayrulewidth)| to the
3382 % end of $R^L$ and $R^R$ by |\adl@hline| otherwise. The other modifications
3383 % are to set |\adl@finaldepth| to zero for the case that the last vertical
3384 % item is |\hline|, and to check if it is followed by not only |\hline| but
3385 % also |\hdashline| by |\adl@xhline|.
3386 %
3387 % The macro |\cline| is also modified to set |\adl@finaldepth| to zero. As
3388 % for the feature of |\ADLnullwidehline|, it inserts |\vskip-\arrayrulewidth|
3389 % to shift the line up befor drawing, and |\vskip\arrayrulewidth| after
3390 % drawing to cancel the negative skip inserted by |\adl@org@cline|.
3391 % \end{macro}\end{macro}
3392 %
3393 % \begin{macrocode}
3394
3395 %% Horizontal Lines
3396
3397 \def\hline{\noalign{\ifnum0=`}\fi
3398 \ifadl@zwhrule \vskip-\arrayrulewidth
3399 \else \adl@hline\adl@connect\arrayrulewidth \fi
3400 \hrule\@height\arrayrulewidth
3401 \global\adl@finaldepth\z@
3402 \futurelet\@tempa\adl@xhline}
3403 \def\cline#1{\noalign{\global\adl@finaldepth\z@
3404 \ifadl@zwhrule \vskip-\arrayrulewidth\fi}
3405 \adl@org@cline{#1}%
3406 \noalign{\ifadl@zwhrule \vskip\arrayrulewidth\fi}}
3407
3408 % \end{macrocode}
3409 %
3410 % \begin{macro}{\hdashline}
3411 % \changes{v1.4-2-5}{1999/06/25}
3412 % {Modified to make \cs{adl@hdashline} usable for
3413 % \cs{first}\slash\texttt{lasthdashline}.}
3414 % \begin{macro}{\adl@hdashline}
3415 % \changes{v1.4-2-5}{1999/06/25}
3416 % {Modified to be usable for
3417 % \cs{first}\slash\texttt{lasthdashline}.}
3418 % \changes{v1.7-1}{2004/5/21}
3419 % {Modified for null-wide horizontal lines.}
3420 % \begin{macro}{\adl@ihdashline}
3421 % \changes{v1.4-2-5}{1999/06/25}
3422 % {Introduced as the substitute of old \cs{adl@hdashline}.}
3423 % \changes{v1.7-1}{2004/5/21}
3424 % {\cs{adl@hline} is moved to \cs{adl@hdashline} for null-wide lines.}
3425 % \SpecialIndex{\adl@hcline}
3426 % \SpecialIndex{\dashlinedash}
3427 % \SpecialIndex{\dashlinegap}
3428 %
3429 % The macro |\hdashline| calls |\adl@hdashline| to open the |\noalign|
3430 % construct by the well-known trick |{\ifnum0=`}\fi| and then to invoke
3431 % |\adl@ihdashline| checking the existence of its optional argument
3432 % |[|\meta{dash}|/|\meta{gap}|]|. Before the invocation, it inserts
3433 % |\vskip-\arrayrulewidth| if |\ADLnullwidehline| is in effect, or adds
3434 % ${\it connect}(w)$ to the end of $R^L$ and $R^R$. Then |\adl@ihdashline|
3435 % closes the |\noalign| by |\ifnum0=`{\fi}| to start the pseudo row for the
3436 % horizontal dash-line. Before the dash-line is drawn by |\adl@hcline|
3437 % which is also used for |\cdashline|, all the columns are |\span|-ned by
3438 % giving |\adl@columns| to |\multispan|. Finally, the |\noalign| is opened
3439 % again and |\adl@xhline| is invoked to check whether |\h(dash)line| is
3440 % followed.
3441 % \end{macro}\end{macro}\end{macro}
3442 %
3443 % \begin{macro}{\adl@inactivehdl}
3444 % \changes{v1.4-2-3}{1999/06/25}
3445 % {Introduced to emulate \cs{hdashline} by \cs{hline}.}
3446 % \changes{v1.7-1}{2004/5/21}
3447 % {Modified to shift up if null-wide.}
3448 %
3449 % If |\ADLinactivate| is in effect, |\adl@ihdashline| is |\let|-equal
3450 % to |\adl@inactivehdl|. This macro simply puts a |\hrule| discarding its
3451 % arguments after inserting |\vskip|\allowbreak|-\arrayrulewidth| if
3452 % |\ADLnullwidehline| is in effect.
3453 % \end{macro}
3454 %
3455 % \begin{macrocode}
3456 \def\hdashline{\adl@hdashline\adl@ihdashline}
3457 \def\adl@hdashline#1{\noalign{\ifnum0=`}\fi
3458 \ifadl@zwhrule \vskip-\arrayrulewidth
3459 \else \adl@hline\adl@connect\arrayrulewidth \fi
3460 \@ifnextchar[%]
3461 {#1}%
3462 {#1[\dashlinedash/\dashlinegap]}}
3463 \def\adl@ihdashline[#1/#2]{\ifnum0=`{\fi}%
3464 \multispan{\adl@columns}\unskip \adl@hcline\z@[#1/#2]%
3465 \noalign{\ifnum0=`}\fi
3466 \futurelet\@tempa\adl@xhline}
3467 \def\adl@inactivehdl[#1/#2]{\ifadl@zwhrule \vskip-\arrayrulewidth \fi
3468 \hrule\@height\arrayrulewidth
3469 \futurelet\@tempa\adl@xhline}
3470
3471 % \end{macrocode}
3472 %
3473 % \begin{macro}{\adl@xhline}
3474 % \SpecialIndex{\@xhline}
3475 %
3476 % The macro |\adl@xhline| is the counterpart of the original |\@xhline|.
3477 % This is introduced to check the mixed sequence of |\hline| and
3478 % |\hdashline|, and to add the element ${\it
3479 % disconnect}(s)={}$|(\relax/\doublerulesep)| to the end of $R^L$ and $R^R$
3480 % by |\adl@hline| if a pair of |\h(dash)line| is found.
3481 % \end{macro}
3482 %
3483 % \begin{macrocode}
3484 \def\adl@xhline{\ifx\@tempa\hline \adl@ixhline\fi
3485 \ifx\@tempa\hdashline \adl@ixhline\fi
3486 \ifnum0=`{\fi}}
3487 \def\adl@ixhline{\vskip\doublerulesep \adl@hline\relax\doublerulesep}
3488
3489 % \end{macrocode}
3490 %
3491 % \begin{macro}{\adl@hline}
3492 % \SpecialIndex{\adl@totalheight}
3493 % \SpecialIndex{\adl@rowsL}
3494 % \SpecialIndex{\adl@rowsR}
3495 %
3496 % The macro |\adl@hline|\meta{cs}\meta{dimen} |\global|-ly adds the integer
3497 % value of \meta{dimen} to |\adl@totalheight| and adds the element
3498 % |(|\meta{cs}|/\number|\meta{dimen}|)| to the tail of $R^L$ and $R^R$. The
3499 % arguments \meta{cs}\meta{dimen} are |\adl@connect\arrayrulewidth| for
3500 % ${\it connect}(w)$ or |\relax\doublerulesep| for ${\it disconnect}(s)$.
3501 % \end{macro}
3502 %
3503 % \begin{macrocode}
3504 \def\adl@hline#1#2{\@tempcnta#2
3505 \global\advance\adl@totalheight\@tempcnta
3506 \xdef\adl@rowsL{\adl@rowsL
3507 (#1/\number\@tempcnta);}%
3508 \xdef\adl@rowsR{\adl@rowsR
3509 (#1/\number\@tempcnta);}}
3510
3511 % \end{macrocode}
3512 %
3513 % \begin{macro}{\cdashline}
3514 % \changes{v1.4-2-7}{1999/06/25}
3515 % {Modified to call renamed \cs{adl@cdline}.}
3516 % \begin{macro}{\adl@cdline}
3517 % \changes{v1.4-2-7}{1999/06/25}
3518 % {Renamed and modified to call renamed \cs{adl@cdlinea}\slash
3519 % \texttt{b}.}
3520 % \changes{v1.7-1}{2004/5/21}
3521 % {Modified to shift up if null-wide.}
3522 % \begin{macro}{\adl@cdlinea}
3523 % \changes{v1.4-2-7}{1999/06/25}
3524 % {Renamed.}
3525 % \begin{macro}{\adl@cdlineb}
3526 % \changes{v1.4-2-7}{1999/06/25}
3527 % {Renamed.}
3528 % \SpecialIndex{\@cline}
3529 % \SpecialIndex{\adl@cla}
3530 % \SpecialIndex{\adl@clb}
3531 % \SpecialIndex{\adl@org@cline}
3532 % \SpecialIndex{\adl@hcline}
3533 %
3534 % The macro |\cdashline| at first opens |\noalign| and then invokes
3535 % |\adl@cdline| checking the existence of its optional argument
3536 % |[|\meta{dash}|/|\meta{gap}|]|. The macro |\adl@cdline| first inserts
3537 % |\vskip-\arrayrulewidth| if |\ADLnullwidehline| is in effect. Then it
3538 % performs column |\span|-ing by the code based on that of |\@cline| in
3539 % \LaTeX-2.09 because \LaTeXe's version will not work with \LaTeX-2.09. The
3540 % main job is done by |\adl@hcline| after the target columns are |\span|-ned
3541 % by |\adl@cdlinea| or |\adl@cdlineb|.
3542 % \end{macro}\end{macro}\end{macro}\end{macro}
3543 %
3544 % \begin{macro}{\adl@inactivecdl}
3545 % \changes{v1.4-2-3}{1999/06/25}
3546 % {Introduced to emulate \cs{cdashline} by \cs{cline}.}
3547 % \changes{v1.7-1}{2004/5/21}
3548 % {Modified to invoke \cs{cline} rather than \cs{adl@orgcline} for
3549 % null-wide.}
3550 %
3551 % If |\ADLinactivate| is in effect, |\adl@cdline| is |\let|-equal
3552 % to |\adl@inactivecdl|. This macro simply calls our own |\cline|, after
3553 % closing the |\noalign| opened by |\cdashline|.
3554 % \end{macro}
3555 %
3556 % \begin{macrocode}
3557 \def\cdashline#1{\noalign{\ifnum0=`}\fi
3558 \@ifnextchar[%]
3559 {\adl@cdline[#1]}%
3560 {\adl@cdline[#1][\dashlinedash/\dashlinegap]}}
3561 \def\adl@cdline[#1-#2]{\ifadl@zwhrule \vskip-\arrayrulewidth \fi
3562 \global\adl@cla#1\relax
3563 \global\advance\adl@cla\m@ne
3564 \ifnum\adl@cla>\z@ \global\let\@gtempa\adl@cdlinea
3565 \else \global\let\@gtempa\adl@cdlineb\fi
3566 \global\adl@clb#2\relax
3567 \global\advance\adl@clb-\adl@cla \ifnum0=`{\fi}
3568 \@gtempa{-\arrayrulewidth}}
3569 \def\adl@cdlinea{\multispan\adl@cla &\multispan\adl@clb \unskip \adl@hcline}
3570 \def\adl@cdlineb{\multispan\adl@clb \unskip \adl@hcline}
3571
3572 \def\adl@inactivecdl[#1-#2][#3]{\ifnum0=`{\fi}\cline{#1-#2}}
3573
3574 % \end{macrocode}
3575 %
3576 % \begin{macro}{\adl@hcline}
3577 % \changes{v1.4-2-4}{1999/06/25}
3578 % {Modified to use \cs{adl@draw}.}
3579 % \changes{v1.54-3}{2003/08/25}
3580 % {Omit \cs{vskip} if the space is 0.}
3581 % \changes{v1.7-1}{2004/5/21}
3582 % {Modified not to shift null-wide \cs{cdashline} down.}
3583 % \SpecialIndex{\adl@draw}
3584 % \SpecialIndex{\adl@vrule}
3585 % \SpecialIndex{\adl@finaldepth}
3586 %
3587 % The macro |\adl@hcline|\meta{w}|[|\meta{d}|/|\meta{g}|]| draws a
3588 % horizontal dash-line of dash size $d$ and gap size $g$ for |\hdashline|
3589 % and |\cdashline| in the |\span|-ned columns by |\adl@draw|. As we will
3590 % discuss in \S\ref{sec:imp-draw}, the macro requires $d$ and $g$ are
3591 % passed through |\@tempdima| and \hbox{|\@tempdimb|}, and control sequences
3592 % \meta{rule}, \meta{skip} and \meta{box} are passed through its arguments
3593 % to make it usable for both horizontal and vertical lines. Then the
3594 % vertical space of $w$, $-$|\arrayrulewidth| for |\cdashline|, is
3595 % inserted if it is not 0 (for |\hdashline|) and |\ADLnullwidehline| is not
3596 % in effect.
3597 % \end{macro}
3598 %
3599 % \begin{macrocode}
3600 \def\adl@hcline#1[#2/#3]{\@tempdima#2\relax \@tempdimb#3\relax
3601 \adl@draw\adl@vrule\hskip\hbox \cr
3602 \noalign{\global\adl@finaldepth\z@ \ifdim#1=\z@\else
3603 \ifadl@zwhrule\else \vskip#1\fi\fi}}
3604
3605 % \end{macrocode}
3606 %
3607 % \begin{macro}{\firsthdashline}
3608 % \changes{v1.4-2-5}{1999/06/25}
3609 % {Introduced as the dashed version of \cs{firsthline}.}
3610 % \begin{macro}{\lasthdashline}
3611 % \changes{v1.4-2-5}{1999/06/25}
3612 % {Introduced as the dashed version of \cs{lasthline}.}
3613 % \SpecialIndex{\firsthline}
3614 % \SpecialIndex{\lasthline}
3615 % \SpecialIndex{\adl@hdashline}
3616 % \SpecialIndex{\adl@ihdashline}
3617 %
3618 % If \textsf{array} is in use, we wish to have dashed counterparts of
3619 % |\first|\slash|lasthline| named |\first|\slash|lasthdashline|, which
3620 % simply call |\adl@hdashline| with an argument to call
3621 % |\adl@first|\slash|lasthdashline| after closing |\noalign| opened by
3622 % |\adl@hdashline|.
3623 % \end{macro}\end{macro}
3624 %
3625 % \begin{macro}{\adl@defflhdl}
3626 % \changes{v1.4-2-5}{1999/06/25}
3627 % {Introduced for the tricky definition of \cs{adl@first}\slash
3628 % \texttt{lasthdashline}.}
3629 % \begin{macro}{\adl@idefflhdl}
3630 % \changes{v1.4-2-5}{1999/06/25}
3631 % {Introduced for the tricky definition of \cs{adl@first}\slash
3632 % \texttt{lasthdashline}.}
3633 % \begin{macro}{\adl@firsthdashline}
3634 % \changes{v1.4-2-5}{1999/06/25}
3635 % {Introduced as the body of \cs{firsthdashline}.}
3636 % \begin{macro}{\adl@lasthdashline}
3637 % \changes{v1.4-2-5}{1999/06/25}
3638 % {Introduced as the body of \cs{lasthdashline}.}
3639 % The macros |\adl@first|\slash|lasthdashline|, however, are defined in a
3640 % tricky manner to replace |\hline| in |\first|\slash|lasthline| with;
3641 %
3642 % \begin{quote}
3643 % |\adl@hdashline\adl@ihdashline[|\meta{dash}|/|\meta{gap}|]|
3644 % \end{quote}
3645 %
3646 % in order to avoid copy-and-replace. To do that, we define |\adl@defflhdl|
3647 % and |\adl@idefflhdl| in which the body of |\first|\slash|lasthline| is
3648 % expanded by |\exapndafter| and the parts preceding and following |\hline|
3649 % are extracted. Then the preceding part \meta{p}, the calling sequence of
3650 % |\adl@hdashline|, and the following part \meta{f} are connected to be the
3651 % body of |\adl@first|\slash|lasthdashline|. Thus we define
3652 % |\adl@firsthdashline| as follows.
3653 %
3654 % \begin{quote}
3655 % |\def\adl@firsthdashline|\texttt{[\#1/\#2]}|{%|\\
3656 % \mbox{\qquad}\meta{p}\\
3657 % \mbox{\qquad}|\adl@hdashline\adl@ihdashline|\texttt{[\#1/\#2]}\\
3658 % \mbox{\qquad}\meta{f}|}|
3659 % \end{quote}
3660 % \end{macro}\end{macro}\end{macro}\end{macro}
3661 %
3662 % \begin{macrocode}
3663 \ifadl@usingarypkg
3664 \def\firsthdashline{\adl@hdashline{\ifnum0=`{\fi}\adl@firsthdashline}}
3665 \def\lasthdashline{\adl@hdashline{\ifnum0=`{\fi}\adl@lasthdashline}}
3666
3667 \def\adl@defflhdl#1{\def\@tempa{#1}
3668 \expandafter\adl@idefflhdl}
3669 \def\adl@idefflhdl#1\hline#2\@nil{%
3670 \@namedef\@tempa[##1/##2]{#1\adl@hdashline\adl@ihdashline[##1/##2]#2}}
3671 \adl@defflhdl{adl@firsthdashline}\firsthline\@nil
3672 \adl@defflhdl{adl@lasthdashline}\lasthline\@nil
3673 \fi
3674
3675 %%^L
3676 % \end{macrocode}
3677 %
3678 %
3679 % \subsection{End of Environment}
3680 % \label{sec:imp-eoe}
3681 %
3682 % \begin{macro}{\endarray}
3683 % \changes{v1.74}{2018/09/26}
3684 % {Add conditional invokation of \cs{@arrayright}.}
3685 % \begin{macro}{\endtabular}
3686 % \changes{v1.7}{2004/5/21}
3687 % {Modified to refer proper \cs{endarray} depending on the
3688 % exsistance of \textsf{colortbl}.}
3689 % \begin{macro}{\endtabular*}
3690 % \SpecialIndex{\adl@arrayrestore}
3691 %
3692 % The macros to close the |array|/|tabular| environment, |\endarray| and
3693 % |\endtabular(*)|, are modified so that they invoke |\adl@endarray| to draw
3694 % vertical lines just before closing |\halign|, and |\adl@arrayrestore| to
3695 % restore registers and data structures |\global|-ly modified in the
3696 % environment. Note that \textsf{array} and related packages such as
3697 % \textsf{delarray} define a macro |\@arrayright| as the closing hook and
3698 % thus we invoke it if it is defined.
3699 % \end{macro}\end{macro}\end{macro}
3700 % 
3701 % \SpecialIndex{\@arrayright}
3702 %
3703 % \begin{macrocode}
3704
3705 %% End of Environment
3706
3707 \def\endarray{\adl@endarray \egroup \adl@arrayrestore \egroup
3708 \csname @arrayright\endcsname}
3709 \def\endtabular{\endarray $\egroup}
3710 \expandafter\let\csname endtabular*\endcsname\endtabular
3711
3712 % \end{macrocode}
3713 %
3714 % \begin{macro}{\adl@endarray}
3715 % \changes{v1.4-2-6}{1999/06/25}
3716 % {Modified to set the depth of \texttt{array}\slash\texttt{tabular}
3717 % to \cs{adl@finaldepth}.}
3718 % \begin{macro}{\adl@rows}
3719 % \begin{macro}{\adl@addvl}
3720 % \begin{macro}{\adl@vlrowL}
3721 % \begin{macro}{\adl@vlrowR}
3722 % \begin{macro}{\adl@vlrow}
3723 % \SpecialIndex{\adl@height}
3724 % \SpecialIndex{\adl@depth}
3725 % \SpecialIndex{\adl@@cr}
3726 % \SpecialIndex{\adl@currentcolumn}
3727 % \SpecialIndex{\adl@rowsL}
3728 % \SpecialIndex{\adl@rowsR}
3729 % \SpecialIndex{\adl@makevlrL}
3730 % \SpecialIndex{\adl@makevlrR}
3731 % \SpecialIndex{\adl@drawvl}
3732 % \SpecialIndex{\adl@finaldepth}
3733 %
3734 % The macro |\adl@endarray| at first closes the last row by |\crcr|. If this
3735 % |\crcr| has real effect, we have to invoke |\adl@@cr| to perfrom our own
3736 % end-of-row operations. We assume that the |\crcr| is effective if either
3737 % |\adl@height| or |\adl@depth| has a non-zero value\footnote{
3738 %
3739 % The author confesses that this rule is not strict and the introduction of a
3740 % switch could improve the strictness.}.
3741 %
3742 % Then the rows to draw vertical lines $L_1$, \ldots, $L_n$;
3743 % $$
3744 % \sigma_1 L_1 \sigma_2 L_2 \ldots L_{n-1}\sigma_n L_n\sigma_{n+1}
3745 % $$
3746 % are created in |\adl@vlrowL| and |\adl@vlrowR| by |\adl@makevlrL| and
3747 % |\adl@makevlrR|. In the real implementation,
3748 % $L_k=\meta{$\gamma_k,\pi_k,\delta_k,\xi_k,\tau_k,\beta_k$}$ is represented
3749 % as;
3750 %
3751 % \begin{quote}
3752 % |\adl@vl{|$\beta_k$|}{|$\tau_k-\beta_k$|}{|$\delta_k$|}{|$\xi_k$|}|.
3753 % \end{quote}
3754 %
3755 % Thus |\adl@vl| is made |\let|-equal to |\relax| when the rows are
3756 % constructed and to \hbox{|\adl@@vl|} when the rows are put.
3757 %
3758 % Since |\adl@makevlrL| and |\adl@makevlrR| shares common macros, they
3759 % conceptually have the following interface.
3760 %
3761 % \begin{quote}
3762 % \begin{tabbing}
3763 % \cs{adl@vlrow}${}={}$\cs{adl@makevlrL/R}(\=
3764 % \cs{adl@rows}\,:\,\meta{$R^L$ or $R^R$},\\
3765 % \> \cs{adl@currentcolumn}\,:\,\meta{start column},\\
3766 % \> \cs{adl@addvl}\,:\,\meta{macro to add an element})
3767 % \end{tabbing}
3768 % \end{quote}
3769 %
3770 % Thus they are invoked as;
3771 %
3772 % \begin{quote}\leavevmode
3773 % \cs{adl@vlrowL}${}={}$\cs{adl@makevrL}(
3774 % \cs{adl@rowsL},
3775 % 1,
3776 % \cs{adl@addvlL})\\
3777 % \cs{adl@vlrowR}${}={}$\cs{adl@makevrR}(
3778 % \cs{adl@rowsR},
3779 % \cs{adl@columns},
3780 % \cs{adl@addvlR})
3781 % \end{quote}
3782 %
3783 % \changes{v1.54-1}{2003/08/25}
3784 % {Rows for vertical lines are replaced by \cs{adl@drawvl}.}
3785 %
3786 % Finally, after constructed rows for vertical lines are put by
3787 % |\adl@drawvl|, a vertical skip of $-{}$|\adl@finaldepth| is inserted to
3788 % move back to the last baseline, and then an invisible |\vrule| of
3789 % |\adl@finaldepth| deep is put to make |array|\slash|tabular| has the depth
3790 % of the last real row or zero if it ends with a horizontal line.
3791 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
3792 %
3793 % \begin{macrocode}
3794 \def\adl@endarray{\crcr \noalign{
3795 \ifdim\adl@height=\z@
3796 \ifdim\adl@depth=\z@ \else \adl@@cr\z@ \fi
3797 \else \adl@@cr\z@ \fi
3798 \let\adl@vl\relax
3799 \def\adl@vlrow{}\adl@currentcolumn\@ne
3800 \let\adl@rows\adl@rowsL
3801 \let\adl@addvl\adl@addvlL
3802 \adl@makevlrL \global\let\adl@vlrowL\adl@vlrow
3803 \def\adl@vlrow{}\adl@currentcolumn\adl@columns
3804 \let\adl@rows\adl@rowsR
3805 \let\adl@addvl\adl@addvlR
3806 \adl@makevlrR \global\let\adl@vlrowR\adl@vlrow
3807 \global\let\adl@vl\adl@@vl}%
3808 \adl@drawvl
3809 \noalign{\vskip-\adl@finaldepth}%
3810 \omit\vrule\@width\z@\@height\z@\@depth\adl@finaldepth\cr}
3811
3812 % \end{macrocode}
3813 %
3814 % \begin{macro}{\adl@arrayrestore}
3815 % \SpecialIndex{\adl@height}
3816 % \SpecialIndex{\adl@depth}
3817 % \SpecialIndex{\adl@currentcolumn}
3818 % \SpecialIndex{\adl@totalheight}
3819 % \SpecialIndex{\adl@heightsave}
3820 % \SpecialIndex{\adl@depthsave}
3821 % \SpecialIndex{\adl@currentcolumnsave}
3822 % \SpecialIndex{\adl@totalheightsave}
3823 % \SpecialIndex{\adl@rowsL}
3824 % \SpecialIndex{\adl@rowsR}
3825 % \SpecialIndex{\adl@rowsLsave}
3826 % \SpecialIndex{\adl@rowsRsave}
3827 % \SpecialIndex{\adl@colsL}
3828 % \SpecialIndex{\adl@colsR}
3829 % \SpecialIndex{\adl@colsLsave}
3830 % \SpecialIndex{\adl@colsRsave}
3831 %
3832 % The macro |\adl@arrayrestore| restores the values of registers and data
3833 % structures, |\adl@height|, |\adl@depth|, |\adl@currentcolumn|,
3834 % |\adl@totalheight|, |\adl@rowsL|, |\adl@rowsR|, |\adl@colsL| and
3835 % |\adl@colsR|, saved by |\adl@arrayinit|.
3836 % \end{macro}
3837 %
3838 % \begin{macrocode}
3839 \def\adl@arrayrestore{%
3840 \global\adl@height\adl@heightsave
3841 \global\adl@depth\adl@depthsave
3842 \global\adl@currentcolumn\adl@currentcolumnsave
3843 \global\adl@totalheight\adl@totalheightsave
3844 \global\let\adl@rowsL\adl@rowsLsave
3845 \global\let\adl@rowsR\adl@rowsRsave
3846 \global\let\adl@colsL\adl@colsLsave
3847 \global\let\adl@colsR\adl@colsRsave}
3848
3849 %%^L
3850 % \end{macrocode}
3851 %
3852 %
3853 %
3854 % \def\defprog{
3855 % \def\WHILE{\hbox{\bf while}}
3856 % \def\DO{\hbox{\bf do}}
3857 % \def\BEGIN{\hbox{\bf begin}}
3858 % \def\END{\hbox{\bf end}}
3859 % \def\IF{\hbox{\bf if}}
3860 % \def\THEN{\hbox{\bf then}}
3861 % \def\ELSE{\hbox{\bf else}}
3862 % \def\TRUE{\hbox{\bf true}}
3863 % \def\FALSE{\hbox{\bf false}}
3864 % \let\_\gets
3865 % \def\GAMMA{{\mit\Gamma}}
3866 % \def\LAMBDA{{\mit\Lambda}}
3867 % \def\!{\stepcounter{enumi}\llap{\footnotesize(\arabic{enumi})\quad}}
3868 % }
3869 % \begin{figure}[p]\defprog
3870 % \begin{itemize}\item[]
3871 % \let\B\bot
3872 % \setcounter{enumi}{0}
3873 % \begin{tabbing}
3874 %\!$\LAMBDA\_\meta{}$; $R\_R^L$; $\gamma\_1$;\\
3875 %\!\WHILE\ $\gamma\leq\GAMMA$ \DO\ \BEGIN\\
3876 %\!\qquad\=
3877 % $\tau\_H$; $\beta\_H$; $\eta\_0$; $\delta\_ \meta{$-1,\B$}$;
3878 % $\xi\_ \meta{$-1,\B$}$;\\
3879 %\!\> ${\it conn}\_\FALSE$; ${\it double}\_\FALSE;$; $R'\_\meta{}$\\
3880 %\!\> \WHILE\ $R\neq\meta{}$ \DO\ \BEGIN\\
3881 %\!\>\qquad\= $\meta{$r,R$}\_R$;\\
3882 %\!\> \> $\meta{$C,h$}\_r$;\\
3883 %\!\> \> \IF\ $C=\meta{}$ \THEN\ \BEGIN\quad
3884 % ${\it add}(\tau,\beta,\delta,\xi)$; $\eta\_0$;\quad\END;\\
3885 %\!\> \> \ELSE\IF\ $C=\meta{connect}$ \THEN\quad $\eta\_h$;\\
3886 %\!\> \> \ELSE\ \BEGIN\\
3887 %\!\> \>\qquad\= $\meta{$e,C'$}=C$; $\meta{$c,d,g$}=e$;\\
3888 %\!\> \> \> \IF\ $c=\gamma$ \THEN\ \BEGIN\\
3889 %\!\> \> \>\qquad\= \IF\ $d=\delta\LAND g=\xi$ \THEN\ \BEGIN\\
3890 %\!\> \> \> \>\qquad\= \IF\ $\lnot{\it conn}$ \THEN\ \BEGIN\\
3891 %\!\> \> \> \> \>\qquad\= $\tau\_\beta+\eta$;
3892 % ${\it conn}\_\TRUE$;\\
3893 %\!\> \> \> \> \> \END;\\
3894 %\!\> \> \> \> \END;\\
3895 %\!\> \> \> \> \ELSE\ \BEGIN\\
3896 %\!\> \> \> \> \> ${\it add}(\tau,\beta,\delta,\xi)$;\\
3897 %\!\> \> \> \> \> $\delta\_d$; $\xi\_g$;
3898 % $\tau\_\beta+\eta$;
3899 % ${\it conn}\_\TRUE$;\\
3900 %\!\> \> \> \> \END;\\
3901 %\!\> \> \> \> \IF\ $C'=\meta{\meta{$\gamma,?,?$},$?$}$ \THEN
3902 % \quad ${\it double}\_\TRUE$;\\
3903 %\!\> \> \> \> $C\_C'$;\\
3904 %\!\> \> \> \END;\\
3905 %\!\> \> \> \ELSE\quad ${\it add}(\tau,\beta,\delta,\xi)$;\\
3906 %\!\> \> \> $\eta\_0$;\\
3907 %\!\> \> \END;\\
3908 %\!\> \> $\beta\_\beta-h$; $R'\_\meta{$R',\meta{$C,h$}$}$\\
3909 %\!\> \END;\\
3910 %\!\> ${\it add}(\tau,\beta,\delta,\xi)$; $R\_R'$;\\
3911 %\!\> \IF\ {\it double} \THEN\quad
3912 % $\LAMBDA\_\meta{$\LAMBDA,\cs{hskip}\cs{doublerulesep}$}$;\\
3913 %\!\> \ELSE\ \BEGIN\\
3914 %\!\> \> $\gamma\_\gamma+1$;\\
3915 %\!\> \> \IF\ $\gamma>\GAMMA$ \THEN\quad\=
3916 % $\LAMBDA\_\meta{$\LAMBDA,\cs{hfil}$}$;\\
3917 %\!\> \> \ELSE\>
3918 % $\LAMBDA\_
3919 % \meta{$\LAMBDA,\cs{hfil}\hbox{\tt\&}\cs{omit}$}$;\\
3920 %\!\> \END;\\
3921 %\!\END;\\
3922 %\!\\
3923 %\!{\bf procedure} ${\it add}(\tau,\beta,\delta,\xi)$ \BEGIN\\
3924 %\!\> \IF\ {\it conn} \THEN\ \BEGIN\\
3925 %\!\> \> $\LAMBDA\_
3926 % \meta{$\LAMBDA,\meta{$\beta,\tau-\beta,\delta,\xi$}$}$;
3927 % ${\it conn}\_\FALSE$;\\
3928 %\!\> \END;\\
3929 %\!\END;
3930 % \end{tabbing}
3931 % \end{itemize}
3932 % \caption{Conceptual Code of \cs{adl@makevlrL}}
3933 % \label{fig-makevlr}
3934 % \end{figure}
3935 %
3936 %
3937 %
3938 % \subsection{Drawing Vertical Lines}
3939 % \label{sec:imp-vl}
3940 %
3941 % \SpecialIndex{\adl@rowsL}
3942 % \SpecialIndex{\adl@rowsR}
3943 % \SpecialIndex{\adl@rows}
3944 % \SpecialIndex{\adl@vlrowL}
3945 % \SpecialIndex{\adl@vlrowR}
3946 % \SpecialIndex{\adl@vlrow}
3947 % \SpecialIndex{\adl@columns}
3948 % \SpecialIndex{\adl@currentcolumn}
3949 % \SpecialIndex{\adl@lastconn}
3950 % \SpecialIndex{\adl@dash}
3951 % \SpecialIndex{\adl@gap}
3952 % \SpecialIndex{\adl@vlrow}
3953 % \SpecialIndex{\ifadl@connected}
3954 % \SpecialIndex{\ifadl@doublerule}
3955 % \SpecialIndex{\adl@totalheight}
3956 %
3957 % \changes{v1.73-1}{2016/04/28}
3958 % {Add $\eta\EQ\cs{adl@lastconn}$, its initialization and updates, and
3959 % the addition to $\tau$.}
3960 %
3961 % Figure~\ref{fig-makevlr} shows the conceptual code of |\adl@makevlrL|.
3962 % The correspondance of variables in the code and control sequences in the
3963 % real implementation is as follows.
3964 % \begin{itemize}\item[]
3965 % $\begin{array}[t]{*4{r@{\;{:}\;}l}}
3966 % R^L& \cs{adl@rowsL}&
3967 % R& \cs{adl@rows}&
3968 % R'& \cs{@tempb}&
3969 % \mit\Lambda& \cs{adl@vlrowL}\\
3970 % \mit\Gamma& \cs{adl@columns}&
3971 % \gamma& \rlap{\cs{adl@currentcolumn}}\\
3972 % \tau& \cs{@tempcnta}&
3973 % \beta& \cs{@tempcntb}&
3974 % \eta& \rlap{\cs{adl@lastconn}}\\
3975 % \delta& \rlap{\cs{adl@dash}/\cs{adl@dashcolor}}&
3976 % \multicolumn2c{}&
3977 % \xi& \rlap{\cs{adl@gap}/\cs{adl@gapcolor}}&\multicolumn2c{}\\
3978 % H& \rlap{\cs{adl@totalheight}}\\
3979 % {\it conn}& \rlap{\cs{ifadl@connected}}&\multicolumn2c{}&
3980 % {\it double}& \rlap{\cs{ifadl@doublerule}}\\
3981 % \end{array}$
3982 % \end{itemize}
3983 %
3984 % \begin{macro}{\adl@makevlrL}
3985 % \changes{v1.4-2-8}{1999/06/25}
3986 % {Modified to replace \cs{hfil} with \cs{hss} to prevent drawing
3987 % vertical lines widen columns.}
3988 % \begin{macro}{\adl@makevlrR}
3989 % \changes{v1.4-2-8}{1999/06/25}
3990 % {Modified to replace \cs{hfil} with \cs{hss} to prevent drawing
3991 % vertical lines widen columns.}
3992 % The macro |\adl@makevlrL| corresponds to the line (2) and (31)--(36). Its
3993 % right-edge counterpart |\adl@makevlrR| has the same correspondance but the
3994 % lines (1)--(2) are;
3995 %
3996 % \begin{quote}\defprog\setcounter{enumi}{0}
3997 % \begin{tabbing}
3998 %\!$\LAMBDA\_\meta{}$; $R\_R^R$; $\gamma\_\GAMMA$;\\
3999 %\!\WHILE\ $\gamma>0$ \DO\ \BEGIN
4000 % \end{tabbing}
4001 % \end{quote}
4002 %
4003 % and (31)--(36) are;
4004 %
4005 % \begin{quote}\defprog\setcounter{enumi}{30}
4006 % \begin{tabbing}
4007 %\!\qquad\=
4008 % \IF\ {\it double} \THEN\quad
4009 % $\LAMBDA\_\meta{$\cs{hskip}\cs{doublerulesep},\LAMBDA$}$;\\
4010 %\!\> \ELSE\ \BEGIN\\
4011 %\!\>\qquad\= $\gamma\_\gamma-1$;\\
4012 %\!\> \> \IF\ $\gamma=0$ \THEN\quad\=
4013 % $\LAMBDA\_\meta{$\cs{hss},\LAMBDA$}$;\\
4014 %\!\> \> \ELSE\>
4015 % $\LAMBDA\_
4016 % \meta{$\hbox{\tt\&}\cs{omit}\cs{hss},\LAMBDA$}$;\\
4017 %\!\> \END;\\
4018 % \end{tabbing}
4019 % \end{quote}
4020 % \end{macro}\end{macro}
4021 % \begin{macrocode}
4022
4023 %% Drawing Vertical Lines
4024
4025 \def\adl@makevlrL{\adl@makevlr
4026 \ifadl@doublerule
4027 \edef\adl@vlrow{\adl@vlrow \hskip\doublerulesep}%
4028 \let\next\adl@makevlrL
4029 \else
4030 \advance\adl@currentcolumn\@ne
4031 \ifnum\adl@currentcolumn>\adl@columns \let\next\relax
4032 \edef\adl@vlrow{\adl@vlrow \hss}%
4033 \else \let\next\adl@makevlrL
4034 \edef\adl@vlrow{\adl@vlrow \hss &\omit}%
4035 \fi\fi\next}
4036 \def\adl@makevlrR{\adl@makevlr
4037 \ifadl@doublerule
4038 \edef\adl@vlrow{\hskip\doublerulesep \adl@vlrow}%
4039 \let\next\adl@makevlrR
4040 \else
4041 \advance\adl@currentcolumn\m@ne
4042 \ifnum\adl@currentcolumn=\z@ \let\next\relax
4043 \edef\adl@vlrow{\hss \adl@vlrow}%
4044 \else \let\next\adl@makevlrR
4045 \edef\adl@vlrow{&\omit \hss \adl@vlrow}%
4046 \fi\fi\next}
4047
4048 % \end{macrocode}
4049 %
4050 % \begin{macro}{\adl@makevlr}
4051 % \changes{v1.7-3}{2004/5/21}
4052 % {Modified to initialize \cs{adl@dashcolor} and \cs{adl@gapcolor}.}
4053 % \changes{v1.73-1}{2016/04/28}
4054 % {Add $\eta\EQ\cs{adl@lastconn}\gets0$.}
4055 % \SpecialIndex{\adl@addvl}
4056 % \SpecialIndex{\adl@lastconn}
4057 %
4058 % The macro |\adl@makevlr| corresponds to the lines (3)--(4) and (30).
4059 % \end{macro}
4060 %
4061 % \begin{macrocode}
4062 \def\adl@makevlr{\@tempcnta\adl@totalheight \@tempcntb\adl@totalheight
4063 \let\adl@lastconn\z@ \adl@dash\m@ne \adl@gap\m@ne
4064 \let\adl@dashcolor\relax \let\adl@gapcolor\relax
4065 \adl@connectedfalse \adl@doublerulefalse \def\@tempb{}%
4066 \expandafter\adl@imakevlr\adl@rows\@nil;%
4067 \adl@addvl
4068 \edef\adl@rows{\@tempb}}
4069
4070 % \end{macrocode}
4071 %
4072 % \begin{macro}{\adl@imakevlr}
4073 % \begin{macro}{\adl@iimakevlr}
4074 %
4075 % \changes{v1.73-1}{2016/04/28}
4076 % {Modify the definition of \cs{adl@connect} to pass $h$ to
4077 % \cs{adl@@connect}.}
4078 %
4079 % \begin{macro}{\adl@endmakevlr}
4080 % \SpecialIndex{\@elt}
4081 % \SpecialIndex{\adl@connect}
4082 %
4083 % The macro |\adl@imakevlr|\meta{$r$}|;| corresponds to the lines (5)--(6),
4084 % and the macro |\adl@iimakevlr(|\meta{$C$}|/|\meta{$h$}|)| to (7) and (28).
4085 % \end{macro}\end{macro}\end{macro}
4086 %
4087 % \begin{macrocode}
4088 \def\adl@imakevlr#1;{\def\@tempa{#1}\ifx\@tempa\@nnil \let\next\relax
4089 \else \adl@iimakevlr#1\let\next\adl@imakevlr \fi \next}
4090 \def\adl@iimakevlr(#1/#2){\let\@elt\adl@iiimakevlr
4091 \def\adl@connect{\adl@@connect#2}%
4092 \let\adl@endmakevlr\adl@endmakevlrcut
4093 #1\adl@endmakevlr
4094 \let\@elt\relax \let\adl@connect\relax
4095 \advance\@tempcntb-#2\edef\@tempb{\@tempb(\@tempc/#2);}}
4096
4097 % \end{macrocode}
4098 %
4099 % \begin{macro}{\adl@iiimakevlr}
4100 % \changes{v1.7-3}{2004/5/21}
4101 % {Modified to check color indentity.}
4102 % \changes{v1.73-1}{2016/04/28}
4103 % {Replace two occurences of $\tau\gets\beta$ with
4104 % $\tau\gets\beta+\eta$ and add $\eta\gets0$, where
4105 % $\eta\EQ\cs{adl@lastconn}$.}
4106 % \begin{macro}{\adl@ivmakevlr}
4107 % \changes{v1.7-3}{2004/5/21}
4108 % {Modified not to see $d$ and $g$ which now have colors.}
4109 % \begin{macro}{\adl@vmakevlr}
4110 % \begin{macro}{\adl@endmakevlrcut}
4111 % \changes{v1.73-1}{2016/04/28}
4112 % {Add $\eta\EQ\cs{adl@lastconn}\gets0$.}
4113 % \begin{macro}{\adl@endmakevlrconn}
4114 % \begin{macro}{\adl@@connect}
4115 % \changes{v1.73-1}{2016/04/28}
4116 % {Add $\eta\EQ\cs{adl@lastconn}\gets h$ with the added
4117 % argument $h$.}
4118 %
4119 % The correspondance of the lines (8)--(30) is a little bit complicated. As
4120 % shown above, |\adl@iimakevlr| expands $C$ attaching the sentinel
4121 % |\adl@endmakevlr|.
4122 %
4123 % \begin{enumerate}
4124 % \item
4125 % If $C\neq\meta{}$ and $C\neq\meta{connect}$, $C$ has at least one
4126 % |\@elt|\meta{$c$}\meta{$d$}\meta{$g$} which is made |\let|-equal to
4127 % |\adl@iiimakevlr| by |\adl@iimakevlr|. Thus the lines (10)--(21) and
4128 % (25)--(26) are performed by |\adl@iiimakevlr|.
4129 %
4130 % Then;
4131 % \begin{enumerate}
4132 % \item
4133 % if $c=\gamma$, |\@elt| becomes |\let|-equal to |\adl@ivmakevlr| which
4134 % corresponds to (22) in the case of $C'\neq\meta{}$. Then |\adl@vmakevlr|
4135 % is invoked for (23) and to eat the sentinel |\adl@endmakevlr|. If
4136 % $C'=\meta{}$, |\adl@endmakevlrconn| is invoked, because the sentinel
4137 % |\adl@endmakevlr| is made |\let|-equal to it by |\adl@iiimakevlr|, for
4138 % (23) (i.e.\ $C\gets\meta{}$).
4139 %
4140 % \item
4141 % if $c\neq\gamma$, |\adl@vmakevlr| is invoked to perform implicit $C\gets
4142 % C$ operation and to eat the sentinel.
4143 % \end{enumerate}
4144 %
4145 % \item
4146 % If $C=\meta{connect}$, i.e.\ it has only one element |\adl@connect|, the
4147 % macro |\adl@@connect| is invoked with $h$ because it is |\define|-dl to be
4148 % $|\adl@@connect|\meta{h}$. The macro performs (9) and implict $C\gets C\ {}
4149 % (=\meta{connect})$ eating the sentinel.
4150 %
4151 % \item
4152 % If $C=\meta{}$, |\adl@endmakevlrcut| that is |\let|-equal to the sentinel
4153 % |\adl@endmakevlr| is invoked to perform (8) and implicit $C\gets C
4154 % \ (=\meta{})$.
4155 % \end{enumerate}
4156 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
4157 % \SpecialIndex{\adl@addvl}
4158 %
4159 % \begin{macrocode}
4160 \def\adl@iiimakevlr#1#2#3#4#5{\let\@elt\adl@ivmakevlr \let\next\relax
4161 \ifnum#1=\adl@currentcolumn\relax
4162 \let\adl@endmakevlr\adl@endmakevlrconn
4163 \@tempswafalse
4164 \ifnum#2=\adl@dash\relax
4165 \ifnum#3=\adl@gap\relax
4166 \def\@tempa{#4}\ifx\@tempa\adl@dashcolor
4167 \def\@tempa{#5}\ifx\@tempa\adl@gapcolor
4168 \@tempswatrue
4169 \fi\fi\fi\fi
4170 \if@tempswa
4171 \ifadl@connected\else
4172 \@tempcnta\@tempcntb
4173 \advance\@tempcnta\adl@lastconn\relax
4174 \adl@connectedtrue
4175 \fi
4176 \else
4177 \adl@addvl
4178 \adl@dash#2\relax \adl@gap#3\relax
4179 \def\adl@dashcolor{#4}\def\adl@gapcolor{#5}%
4180 \@tempcnta\@tempcntb
4181 \advance\@tempcnta\adl@lastconn\relax
4182 \adl@connectedtrue
4183 \fi
4184 \else
4185 \adl@addvl
4186 \def\next{\adl@vmakevlr\@elt{#1}{#2}{#3}{#4}{#5}}%
4187 \fi
4188 \let\adl@lastconn\z@ \next}
4189 \def\adl@ivmakevlr#1{%
4190 \ifnum#1=\adl@currentcolumn \adl@doubleruletrue \fi
4191 \adl@vmakevlr\@elt{#1}}
4192 \def\adl@vmakevlr#1\adl@endmakevlr{\def\@tempc{#1}}
4193 \def\adl@endmakevlrcut{\adl@addvl \let\adl@lastconn\z@ \def\@tempc{}}
4194 \def\adl@endmakevlrconn{\def\@tempc{}}
4195 \def\adl@@connect#1\adl@endmakevlr{\def\adl@lastconn{#1}%
4196 \def\@tempc{\adl@connect}}
4197
4198 % \end{macrocode}
4199 %
4200 % \begin{macro}{\adl@addvlL}
4201 % \changes{v1.7-3}{2004/5/21}
4202 % {Modified to add colors to $\delta$ and $\xi$.}
4203 % \begin{macro}{\adl@addvlR}
4204 % \changes{v1.7-3}{2004/5/21}
4205 % {Modified to add colors to $\delta$ and $\xi$.}
4206 % The macro |\adl@addvlL| corresonds to the lines (38)--(42), i.e.\ the
4207 % procedure {\it add}. The macro |\adl@addvlR| performs simlar operations,
4208 % but its conceptual code is the following.
4209 %
4210 % \begin{quote}\defprog\setcounter{enumi}{37}
4211 % \begin{tabbing}
4212 %\!{\bf procedure} ${\it add}(\tau,\beta,\delta,\xi)$ \BEGIN\\
4213 %\!\qquad\=
4214 % \IF\ {\it conn} \THEN\ \BEGIN\\
4215 %\!\>\qquad\= $\LAMBDA\_
4216 % \meta{$\meta{$\beta,\tau-\beta,\delta,\xi$},\LAMBDA$}$;
4217 % ${\it conn}\_\FALSE$;\\
4218 %\!\> \END;\\
4219 %\!\END;
4220 % \end{tabbing}
4221 % \end{quote}
4222 % \end{macro}\end{macro}
4223 % \SpecialIndex{\ifadl@connected}
4224 % \SpecialIndex{\adl@vlrow}
4225 % \SpecialIndex{\adl@vl}
4226 % \SpecialIndex{\adl@dash}
4227 % \SpecialIndex{\adl@gap}
4228 %
4229 % \begin{macrocode}
4230 \def\adl@addvlL{\ifadl@connected
4231 \advance\@tempcnta-\@tempcntb
4232 \edef\adl@vlrow{\adl@vlrow
4233 \adl@vl{\number\@tempcntb}{\number\@tempcnta}%
4234 {\number\adl@dash}{\number\adl@gap}%
4235 {\adl@dashcolor}{\adl@gapcolor}}%
4236 \adl@connectedfalse \fi}
4237 \def\adl@addvlR{\ifadl@connected
4238 \advance\@tempcnta-\@tempcntb
4239 \edef\adl@vlrow{\adl@vl{\number\@tempcntb}{\number\@tempcnta}%
4240 {\number\adl@dash}{\number\adl@gap}%
4241 {\adl@dashcolor}{\adl@gapcolor}\adl@vlrow}%
4242 \adl@connectedfalse \fi}
4243
4244 % \end{macrocode}
4245 %
4246 % \begin{macro}{\adl@drawvl}
4247 % \changes{v1.54-1}{2003/08/25}
4248 % {Introduced to draw vertical lines correctly if \cs{ADLsomewide}.}
4249 % \begin{macro}{\adl@@vl}
4250 % \changes{v1.4-2-2}{1999/06/25}
4251 % {Modified to make vertical line null wide only if \cs{ADLnullwide}.}
4252 % \changes{v1.4-2-4}{1999/06/25}
4253 % {Modified to use \cs{adl@draw}.}
4254 % \changes{v1.54-1}{2003/08/25}
4255 % {Insert a negative skip to left/right of the line if \cs{ADLsomewide}.}
4256 % \changes{v1.7-3}{2004/5/21}
4257 % {Modified to color dashs and gaps.}
4258 % \begin{macro}{\adl@vl@leftskip}
4259 % \begin{macro}{\adl@vl@rightskip}
4260 % \SpecialIndex{\adl@vlrowL}
4261 % \SpecialIndex{\adl@vlrowL}
4262 % \SpecialIndex{\adl@vl}
4263 % \SpecialIndex{\adl@draw}
4264 % \SpecialIndex{\set@color}
4265 %
4266 % After the the macros |\adl@vlrowL| and |\adl@vlrowR| are constructed, they
4267 % are expanded to draw vertical lines by |\adl@drawvl|. Prior to the
4268 % expansion, the macro |\adl@drawvl| globally defines |\adl@vl@leftskip| and
4269 % |\adl@vl@rightskip|, which are the amount of negative spaces inserted to the
4270 % left\slash right of a vertical line, as follows.
4271 %
4272 % \begin{eqnarray*}
4273 % \cs{adl@vl@leftskip}&=&\cases{
4274 % \cs{arrayrulewidth}/2& if \cs{ifadl@zwrule}\cr
4275 % 0& else if leftside\cr
4276 % \cs{arrayrulewidth}& otherwise}\\
4277 % \cs{adl@vl@rightskip}&=&\cases{
4278 % \cs{arrayrulewidth}/2& if \cs{ifadl@zwrule}\cr
4279 % 0& else if rightside\cr
4280 % \cs{arrayrulewidth}& otherwise}
4281 % \end{eqnarray*}
4282 %
4283 % That is, if |\ADLnulwide| is in effect, a vertical line is surrounded by
4284 % horizontal spaces of $-|\arrayrulewidth|/2$ to adjust the center of the
4285 % line to the left or right edge of its column. Otherwise, a horizontal
4286 % space $-|\arrayrulewidth|$ is inserted after (before) the line is drawn to
4287 % adjust its left (right) edge to the left (right) edge of the column
4288 % \footnote{
4289 % Before v1.54, the horizontal spaces was not inserted if \cs{ADLsomewide}
4290 % and thus disconnected lines were not aligned vertically.}.
4291 %
4292 % Then the macros |\adl@vlrowL| and |\adl@vlrowR| are expanded. These
4293 % macros will have |\adl@vl|, which is made |\let|-equal to
4294 % \hbox{|\adl@@vl|} prior to the expansion, to draw a vertical line. The
4295 % macro
4296 % |\adl@@vl|\meta{$\beta$}\meta{$\lambda$}\meta{$\delta_l$}\meta{$\gamma_l$}
4297 % \meta{$\delta_c$}\meta{$\gamma_c$} ($x_l$ and $x_c$ are length and color)
4298 % draws a sloid line if $\gamma_l=0$ or a dash-line otherwise in a |\vbox| of
4299 % $\lambda=\tau-\beta$ high and |\raise|-s it by $\beta$. The method to
4300 % draw a dash line in the |\vbox| is analogous to that for horizontal line
4301 % shown in \S\ref{sec:imp-hline}, except that a line is surrounded by
4302 % horizontal spaces of |\adl@vl@leftskip| and |\adl@vl@rightskip|. Coloring
4303 % gaps is done by drawing a vertical rule setting $\gamma_c$ by |\set@color|
4304 % prior to dash line drawing if $\gamma_c$ is not |\relax|. To color
4305 % dashes or solid line, |\set@color| with $\delta_c$ is done if it is not
4306 % |\relax| before line drawing.
4307 % \end{macro}\end{macro}\end{macro}\end{macro}
4308 %
4309 % \begin{macrocode}
4310 \def\adl@drawvl{%
4311 \omit \relax \ifadl@zwvrule
4312 \gdef\adl@vl@leftskip{.5\arrayrulewidth}%
4313 \global\let\adl@vl@rightskip\adl@vl@leftskip
4314 \else \global\let\adl@vl@leftskip\z@
4315 \global\let\adl@vl@rightskip\arrayrulewidth
4316 \fi \adl@vlrowL \cr
4317 \omit \relax \ifadl@zwvrule
4318 \gdef\adl@vl@leftskip{.5\arrayrulewidth}%
4319 \global\let\adl@vl@rightskip\adl@vl@leftskip
4320 \else \global\let\adl@vl@leftskip\arrayrulewidth
4321 \global\let\adl@vl@rightskip\z@
4322 \fi \adl@vlrowR \cr}
4323
4324 \def\adl@@vl#1#2#3#4#5#6{\vbox to\z@{\vss\hbox{%
4325 \hskip-\adl@vl@leftskip
4326 \ifnum#3=\z@\else \def\@tempa{#6}\ifx\@tempa\adl@nocolor\else
4327 \raise#1sp\hbox{\let\current@color\@tempa \set@color
4328 \vrule height#2sp width\arrayrulewidth}%
4329 \hskip-\arrayrulewidth \fi \fi
4330 \raise#1sp\vbox to#2sp{
4331 \def\@tempa{#5}\ifx\@tempa\adl@nocolor\else
4332 \let\current@color\@tempa \set@color \fi
4333 \ifnum#3=\z@
4334 \hrule height#2sp depth\z@ width\arrayrulewidth
4335 \else \@tempdima#3sp \@tempdimb#4sp
4336 \adl@draw\adl@hrule\vskip\vbox
4337 \fi}%
4338 \hskip-\adl@vl@rightskip}}}
4339
4340 %%^L
4341 % \end{macrocode}
4342 %
4343 %
4344 %
4345 % \subsection{Drawing Dash-lines}
4346 % \label{sec:imp-draw}
4347 % \changes{v1.4-1}{1999/06/25}
4348 % {Section 4.12 is added.}
4349 %
4350 % \begin{macro}{\adl@vrule}
4351 % \changes{v1.4-2-4}{1999/06/25}
4352 % {Introduced to draw a dash for horizontal lines in \cs{adl@draw}.}
4353 % \begin{macro}{\adl@hrule}
4354 % \changes{v1.4-2-4}{1999/06/25}
4355 % {Introduced to draw a dash for vertical lines in \cs{adl@draw}.}
4356 % As explained later, horizontal and vertical lines are drawn by a common
4357 % macro |\adl@draw| to which the length of a dash segment, $d$, is passed
4358 % through |\@tempdima|. The macro also has an argument that is either
4359 % |\adl@vrule| to draw a dash for {\em horizontal} lines or |\adl@hrule| for
4360 % {\em vertical}. These two macros commonly have one argument \meta{f} to
4361 % draw a dash of $f\times d$ long and of |\arrayrulewidth| wide.
4362 % \end{macro}\end{macro}
4363 %
4364 % \begin{macrocode}
4365
4366 %% Draw Dash Lines (\adl@vrule/\adl@hrule, \hskip/\vskip, \hbox/\vbox)
4367
4368 \def\adl@vrule#1{\vrule\@width#1\@tempdima\@height\arrayrulewidth\relax}
4369 \def\adl@hrule#1{\hrule\@height#1\@tempdima\@width\arrayrulewidth\relax}
4370 % \end{macrocode}
4371 %
4372 % \begin{macro}{\adl@drawi}
4373 % \changes{v1.4-2-4}{1999/06/25}
4374 % {Introduced as \cs{adl@draw} in mode 1.}
4375 % \begin{macro}{\adl@drawii}
4376 % \changes{v1.4-2-4}{1999/06/25}
4377 % {Introduced as \cs{adl@draw} in mode 2.}
4378 % \begin{macro}{\adl@drawiii}
4379 % \changes{v1.4-2-4}{1999/06/25}
4380 % {Introduced as \cs{adl@draw} in mode 3.}
4381 % \begin{macro}{\adl@draw}
4382 % \changes{v1.4-2-4}{1999/06/25}
4383 % {Introduced as the mode and axis independent line drawing macro.}
4384 % \SpecialIndex{\xleaders}
4385 % \SpecialIndex{\cleaders}
4386 %
4387 % The macro |\adl@draw| is to draw a horizontal or vertical line. It is
4388 % |\let|-equal to one of |\adl@drawi|, |\adl@drawii| and |\adl@drawiii|
4389 % according to the drawing mode specified by |\ADLdrawingmode|. These three
4390 % macros have common interface, |\@tempdima| and \hbox{|\@tempdimb|} for the
4391 % length of dash and gap, $d$ and $g$, and three arguments \meta{rule},
4392 % \meta{skip} and \meta{box} with which |\adl@draw| is called in the
4393 % following manner.
4394 %
4395 % \begin{eqnarray*}
4396 % \hbox{|\adl@draw\adl@vrule\hskip\hbox|}&\ldots&\hbox{horizontal}\\
4397 % \hbox{|\adl@draw\adl@hrule\vskip\vbox|}&\ldots&\hbox{vertical}
4398 % \end{eqnarray*}
4399 %
4400 % The drawing methods in three modes have been explained in
4401 % \S\ref{sec:imp-ps2}. More specifically, |\adl@drawi| for mode 1, to which
4402 % |\adl@draw| is |\let|-equal by default, conceptually performs the
4403 % following operations.
4404 %
4405 % \begin{quote}
4406 % \def\arg#1{\hbox{{\tt\char`\{}#1{\tt\char`\}}}}
4407 % \meta{rule}\arg{1/2}\quad\meta{skip}$(g/2)$\\
4408 % |\xlearders|\meta{box}
4409 % \arg{\meta{skip}$(g/2)$ \meta{rule}\arg{1} \meta{skip}$(g/2)$}\\
4410 % \phantom{\texttt{.xleaders}}\meta{skip}(0 |plus| 1|fil minus| 1|fil|)\\
4411 % \meta{skip}$(g/2)$\quad\meta{rule}\arg{1/2}
4412 % \end{quote}
4413 %
4414 % The conceptual operations of |\adl@drawii| for mode 2 are as follows.
4415 %
4416 % \begin{quote}
4417 % \def\arg#1{\hbox{{\tt\char`\{}#1{\tt\char`\}}}}
4418 % \meta{rule}\arg{1/2}\quad\meta{skip}$(g/2)$\\
4419 % \meta{box}\arg{\meta{skip}$(g/2)$ \meta{rule}\arg{1} \meta{skip}$(g/2)$}\quad
4420 % \meta{skip}$(-d-g)$\\
4421 % |\xlearders|\meta{box}
4422 % \arg{\meta{skip}$(g/2)$ \meta{rule}\arg{1} \meta{skip}$(g/2)$}\\
4423 % \phantom{\texttt{.xleaders}}\meta{skip}(0 |plus| 1|fil minus| 1|fil|)\\
4424 % \meta{skip}$(-d-g)$\quad
4425 % \meta{box}\arg{\meta{skip}$(g/2)$ \meta{rule}\arg{1} \meta{skip}$(g/2)$}\\
4426 % \meta{skip}$(g/2)$\quad\meta{rule}\arg{1/2}
4427 % \end{quote}
4428 %
4429 % The macro |\adl@drawiii| for mode 3 is quite similar to |\adl@drawi|
4430 % except that |\xleaders| is replaced by |\cleaders|. This replacement is
4431 % done by temporarily |\let|-ing |\xleaders| be equal to |\cleaders|.
4432 % \end{macro}\end{macro}\end{macro}\end{macro}
4433 %
4434 % \SpecialIndex{\adl@box}
4435 %
4436 % \begin{macrocode}
4437 \def\adl@drawi#1#2#3{%
4438 #1{.5}#2.5\@tempdimb
4439 \xleaders#3{#2.5\@tempdimb #1{1}#2.5\@tempdimb}%
4440 #2\z@ plus1fil minus1fil\relax
4441 #2.5\@tempdimb #1{.5}}
4442 \def\adl@drawii#1#2#3{%
4443 \setbox\adl@box#3{#2.5\@tempdimb #1{1}#2.5\@tempdimb}%
4444 #1{.5}#2.5\@tempdimb
4445 \copy\adl@box #2-\@tempdima #2-\@tempdimb
4446 \xleaders\copy\adl@box#2\z@ plus1fil minus1fil\relax
4447 #2-\@tempdima #2-\@tempdimb \copy\adl@box
4448 #2.5\@tempdimb #1{.5}}
4449 \def\adl@drawiii#1#2#3{{\let\xleaders\cleaders \adl@drawi#1#2#3}}
4450 \let\adl@draw\adl@drawi
4451
4452 % \end{macrocode}
4453 %
4454 % \begin{macro}{\ADLdrawingmode}
4455 % \changes{v1.4-2-4}{1999/06/25}
4456 % {Introduced to specify drawing mode.}
4457 % The macro |\ADLdrawingmode|\Meta{m} defines the drawing mode by |\let|-ing
4458 % |\adl@draw| be equal to |\adl@drawi| if $m=1$, and so on. If \meta{m} is
4459 % neither 1, 2 nor 3, it is assumed as 1.
4460 % \end{macro}
4461 %
4462 % \begin{macrocode}
4463 \def\ADLdrawingmode#1{\ifcase #1%
4464 \let\adl@draw\adl@drawi \or
4465 \let\adl@draw\adl@drawi \or
4466 \let\adl@draw\adl@drawii \or
4467 \let\adl@draw\adl@drawiii \else
4468 \let\adl@draw\adl@drawi \fi}
4469
4470 %%^L
4471 % \end{macrocode}
4472 %
4473 %
4474 %
4475 % \subsection{Shorthand Activation}
4476 % \label{sec:imp-sh}
4477 % \changes{v1.4-1}{1999/06/25}
4478 % {Section 4.13 is added.}
4479 %
4480 % \begin{macro}{\adl@Array}
4481 % \changes{v1.4-2-3}{1999/06/25}
4482 % {Introduced as the body of \cs{Array}.}
4483 % \begin{macro}{\adl@Tabular}
4484 % \changes{v1.4-2-3}{1999/06/25}
4485 % {Introduced as the body of \cs{Tabular}.}
4486 % \begin{macro}{\adl@Tabularstar}
4487 % \changes{v1.4-2-3}{1999/06/25}
4488 % {Introduced as the body of \cs{Tabular*}.}
4489 % \begin{macro}{\adl@Longtable}
4490 % \changes{v1.6}{2003/8/25}
4491 % {Introduced as the body of \cs{Longtable}.}
4492 % \SpecialIndex{\array}
4493 % \SpecialIndex{\tabular}
4494 % \SpecialIndex{\tabular*}
4495 % \SpecialIndex{\longtable}
4496 %
4497 % The macros|\adl@Array|, |\adl@Tabular|, |\adl@Tabular*| and
4498 % |\adl@Longtable| start environments |array|, |tabular|, |tabular*| and
4499 % |longtable| respectively, turning |\ifadl@inactive| false to activate
4500 % dash-line functions. We will |\let| macros |\Array| etc.\ be equal to
4501 % them for shorthand activation.
4502 % \end{macro}\end{macro}\end{macro}\end{macro}
4503 %
4504 % \begin{macrocode}
4505
4506 %% Shorthand Activation
4507
4508 \def\adl@Array{\adl@inactivefalse \array}
4509 \def\adl@Tabular{\adl@inactivefalse \tabular}
4510 \def\adl@Tabularstar{\adl@inactivefalse \@nameuse{tabular*}}
4511 \def\adl@Longtable{\adl@inactivefalse \longtable}
4512
4513 % \end{macrocode}
4514 %
4515 % \begin{macro}{\@notdefinable}
4516 % \begin{macro}{\adl@notdefinable}
4517 % \changes{v1.4-2-3}{1999/06/25}
4518 % {Introduced to check if \cs{Array} etc. are definable.}
4519 % \SpecialIndex{\@ifdefinable}
4520 %
4521 % Before making |\Array| etc.\ |\let|-equal to |\adl@Array| etc., we have to
4522 % check if these macros having too natural names have already used. This
4523 % check is done by |\@ifdefinable| that will call |\@notdefinable| for the
4524 % complaint if undefinable. Since we want to complain with our own
4525 % warning message, |\@notdefinable| is temporarily |\def|-ined so that it
4526 % simply |\def|-ines a macro |\adl@notdefinable| as empty. Therefore,
4527 % |\adl@notdefinebale| will have some definition if one of |\Array|,
4528 % |\Tabular|, |\Tabular*| and |\Longtable| (if \textsf{longtable} is loaded)
4529 % cannot be defined, while it will stay undefined otherwise.
4530 % \end{macro}\end{macro}
4531 %
4532 % \begin{macrocode}
4533 \begingroup
4534 \def\@notdefinable{\gdef\adl@notdefinable{}}
4535 \@ifdefinable\Array\relax
4536 \@ifdefinable\Tabular\relax
4537 \expandafter\@ifdefinable\csname Tabular*\endcsname\relax
4538 \ifx\longtable\undefined\else \@ifdefinable\Longtable\relax \fi
4539 \endgroup
4540
4541 % \end{macrocode}
4542 %
4543 % \begin{macro}{\Array}
4544 % \changes{v1.4-2-3}{1999/06/25}
4545 % {Introduced as the always-active \cs{array}.}
4546 % \begin{macro}{\Tabular}
4547 % \changes{v1.4-2-3}{1999/06/25}
4548 % {Introduced as the always-active \cs{tabular}.}
4549 % \begin{macro}{\Tabular*}
4550 % \changes{v1.4-2-3}{1999/06/25}
4551 % {Introduced as the always-active \cs{tabular*}.}
4552 % \begin{macro}{\Longtable}
4553 % \changes{v1.6}{2003/8/25}
4554 % {Introduced as the always-active \cs{longtable}.}
4555 % \begin{macro}{\endArray}
4556 % \changes{v1.4-2-3}{1999/06/25}
4557 % {Introduced to \cs{end} the environment \texttt{Array}.}
4558 % \begin{macro}{\endTabular}
4559 % \changes{v1.4-2-3}{1999/06/25}
4560 % {Introduced to \cs{end} the environment \texttt{Tabular}.}
4561 % \begin{macro}{\endTabular*}
4562 % \changes{v1.4-2-3}{1999/06/25}
4563 % {Introduced to \cs{end} the environment \texttt{Tabular*}.}
4564 % \begin{macro}{\endLongtable}
4565 % \changes{v1.6}{2003/8/25}
4566 % {Introduced to \cs{end} the environment \texttt{Longtable}.}
4567 % \SpecialIndex{\PackageWarning}
4568 % \SpecialIndex{\@warning}
4569 % \SpecialIndex{\MessageBreak}%
4570 %
4571 % If |\adl@notdefinable| is |\undefined| indicating that all |\Array| etc.\
4572 % are definable, we |\let| them be equal to |\adl@Array| etc. We also
4573 % |\let| ending macros |\endArray| etc.\ be equal to |\endarray| etc.
4574 % Note that |\Longtable| and |\endLongtable| are defined only when
4575 % \textsf{longtable} is loaded, and |\endLongtable| is |\def|-ined as (not
4576 % being |\let|-equal to) |\endlongtable| because its definition of our own
4577 % is not given yet.
4578 %
4579 % Otherwise, we complain with a warning message put by |\PackageWarning| if
4580 % it is defined (i.e.\ \LaTeXe) or |\@warning| otherwise (i.e.\ \LaTeX-2.09).
4581 % \end{macro}\end{macro}\end{macro}\end{macro}
4582 % \end{macro}\end{macro}\end{macro}\end{macro}
4583 %
4584 % \begin{macrocode}
4585 \ifx\adl@notdefinable\undefined
4586 \let\Array\adl@Array
4587 \let\Tabular\adl@Tabular
4588 \expandafter\let\csname Tabular*\endcsname\adl@Tabularstar
4589 \let\endArray\endarray
4590 \let\endTabular\endtabular
4591 \expandafter\let\csname endTabular*\endcsname\endtabular
4592 \ifx\longtable\undefined\else
4593 \let\Longtable\adl@Longtable
4594 \def\endLongtable{\endlongtable}
4595 \fi
4596 \else
4597 \begingroup
4598 \ifx\longtable\undefined
4599 \def\@tempa{Array and Tabular are not defined because one of them\MessageBreak
4600 has been defined}
4601 \else
4602 \def\@tempa{Array/Tabular/Longtable are not defined because \MessageBreak
4603 one of them has been defined}
4604 \fi
4605 \ifx\PackageWarning\undefined
4606 \def\MessageBreak{^^J}
4607 \@warning\@tempa
4608 \else
4609 \let\on@line\empty
4610 \PackageWarning{arydshln}\@tempa
4611 \fi
4612 \endgroup
4613 \fi
4614
4615 % \end{macrocode}
4616 %
4617 % \begin{macro}{\ADLnoshorthanded}
4618 % \changes{v1.4-2-3}{1999/06/25}
4619 % {Introduced to nullify macros for shorthand activation.}
4620 % \changes{v1.6}{2003/8/25}
4621 % {\cs{Longtable} and \cs{endLongtable} are added.}
4622 % If a user wishes to define an environment named |Array| or |Tabular(*)|
4623 % (or |Longtable| if \textsf{longtable} is in use) by him/herself or by
4624 % loading other packages {\em after} \textsf{arydshln} is loaded,
4625 % |\newenvironment| for |Array| etc.\ will fail because they have already
4626 % been undefinable. The macro |\ADLnoshorthanded| makes them definable
4627 % again by |\let|-ing them and their ending counterparts be equal to
4628 % |\relax|.
4629 % \end{macro}
4630 %
4631 % \begin{macrocode}
4632 \def\ADLnoshorthanded{%
4633 \let\Array\relax
4634 \let\Tabular\relax
4635 \expandafter\let\csname Tabular*\endcsname\relax
4636 \let\endArray\relax
4637 \let\endTabular\relax
4638 \expandafter\let\csname endTabular*\endcsname\relax
4639 \ifx\longtable\undefined\else
4640 \let\Longtable\relax
4641 \let\endLongtable\relax \fi}
4642
4643 % \end{macrocode}
4644 %
4645 % \vskip15\baselineskip\endgraf\vskip-15\baselineskip\vskip0pt
4646 % \begin{macro}{\adl@act@arrayclassz}
4647 % \changes{v1.54-2}{2003/08/25}
4648 % {Introduced to activate \cs{@arrayclassz} etc. again.}
4649 % \begin{macro}{\adl@act@tabclassz}
4650 % \begin{macro}{\adl@act@classz}
4651 % \begin{macro}{\adl@act@@startpbox}
4652 % \begin{macro}{\adl@act@@endpbox}
4653 % \begin{macro}{\adl@act@startpbox}
4654 % \changes{v1.75}{2018/12/26}
4655 % {Introduced because \cs{@startpbox} may be different from the original.}
4656 % \begin{macro}{\adl@act@endpbox}
4657 % \begin{macro}{\adl@act@cr}
4658 % \begin{macro}{\adl@act@argcr}
4659 % \begin{macro}{\adl@act@cline}
4660 % \begin{macro}{\adl@act@endarray}
4661 % \begin{macro}{\adl@act@hline}
4662 % \begin{macro}{\adl@act@ihdashline}
4663 % \begin{macro}{\adl@act@cdline}
4664 % \begin{macro}{\adl@act@@vlineL}
4665 % \begin{macro}{\adl@act@@vlineR}
4666 % \SpecialIndex{\@arrayclassz}
4667 % \SpecialIndex{\@tabclassz}
4668 % \SpecialIndex{\@classz}
4669 % \SpecialIndex{\@@startpbox}
4670 % \SpecialIndex{\@@endpbox}
4671 % \SpecialIndex{\@endpbox}
4672 % \SpecialIndex{\adl@cr}
4673 % \SpecialIndex{\adl@argcr}
4674 % \SpecialIndex{\cline}
4675 % \SpecialIndex{\adl@endarray}
4676 % \SpecialIndex{\adl@hline}
4677 % \SpecialIndex{\adl@ihdashline}
4678 % \SpecialIndex{\adl@cdline}
4679 % \SpecialIndex{\adl@@vlineL}
4680 % \SpecialIndex{\adl@@vlineR}
4681 %
4682 % Finally here we define {\em active} version of |\@arrayclassz| named
4683 % |\adl@act@arrayclassz| etc. for |\adl@activate| (see \S\ref{sec:imp-init}).
4684 % The definitions are simply done by |\let|-ing |\adl@act@arrayclassz| equal
4685 % to |\@arrayclassz| etc\footnote{%
4686 %
4687 % Alternatively, we may define \cs{adl@act@arrayclassz} in place of
4688 % \cs{@arrayclassz} but the author chose this way to minimize the possiblity
4689 % of {\em enbug}.}.
4690 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
4691 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
4692 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
4693 % \end{macro}
4694 %
4695 % \begin{macrocode}
4696 \let\adl@act@arrayclassz\@arrayclassz
4697 \let\adl@act@tabclassz\@tabclassz
4698 \ifadl@usingarypkg \let\adl@act@classz\@classz \fi
4699 \let\adl@act@@startpbox\@@startpbox
4700 \let\adl@act@@endpbox\@@endpbox
4701 \let\adl@act@startpbox\@startpbox
4702 \let\adl@act@endpbox\@endpbox
4703 \let\adl@act@cr\adl@cr
4704 \let\adl@act@argcr\adl@argcr
4705 \let\adl@act@endarray\adl@endarray
4706 \let\adl@act@hline\adl@hline
4707 \let\adl@act@ihdashline\adl@ihdashline
4708 \let\adl@act@cdline\adl@cdline
4709 \let\adl@act@@vlineL\adl@@vlineL
4710 \let\adl@act@@vlineR\adl@@vlineR
4711
4712 %%^L
4713 % \end{macrocode}
4714 %
4715 %
4716 %
4717 % \subsection{Compatibility with \textsf{colortab}}
4718 % \label{sec:imp-colortab}
4719 %
4720 % \begin{macro}{\adl@CC@}
4721 % \begin{macro}{\CC@}
4722 % \changes{v1.5-1}{2000/07/12}
4723 % {Modified to save/restore globals before/after height measurement.}
4724 % \SpecialIndex{\LCC}
4725 % \SpecialIndex{\adl@arraysave}
4726 % \SpecialIndex{\adl@arrayrestore}
4727 %
4728 % The package \textsf{colortab} has a macro;
4729 %
4730 % \begin{quote}
4731 % |\LCC|\meta{colorspec}|\\|\meta{rows}|\ECC|
4732 % \end{quote}
4733 %
4734 % to color \meta{rows} referring \meta{colorspec}. The macro |\CC@|, the
4735 % heart of the coloring function, first makes a box with \meta{rows} using
4736 % |\@preamble| to measure the height of \meta{rows}, then makes a row putting
4737 % a heavy rule of the height in each column with a color command for the
4738 % column specified by \meta{colorspec}, and finally puts \meta{rows}
4739 % overlaying them on the colored rule. Therefore \meta{rows} is processed
4740 % twice by |\CC@| to update |\global| registers/structures incorrectly.
4741 %
4742 % Thus we modify |\CC@|, if the package \textsf{colortab} is provided, to
4743 % save |\global| stuff by |\adl@arraysave| before the height measurement and
4744 % restore them by |\adl@arrayrestore| after that.
4745 % \end{macro}\end{macro}
4746 %
4747 % \begin{macrocode}
4748
4749 %% Compatibility with colortab
4750
4751 \def\adl@CC@#1#2#3{%
4752 \ifcolortab
4753 \noalign{%
4754 \adl@arraysave
4755 \setbox\CT@box=\vbox{#1#3\crcr\egroup}%
4756 \adl@arrayrestore
4757 \CT@dim=\ht\CT@box
4758 \global\advance\CT@dim by \dp\CT@box
4759 \def\CT@next{}%
4760 \futurelet\next\CT@columncolor#2&\@nil}%
4761 \CT@next\cr
4762 \noalign{\vskip-\CT@dim}%
4763 \fi
4764 #3}
4765 \ifx\ColortabLoaded\undefined\else
4766 \let\CC@\adl@CC@
4767 \fi
4768
4769 %%^L
4770 % \end{macrocode}
4771 %
4772 %
4773 %
4774 % \subsection{Compatibility with \textsf{longtable}}
4775 % \label{sec:imp-longtable}
4776 %
4777 % \changes{v1.6}{2003/8/25}
4778 % {\S\ref{sec:imp-longtable} is added.}
4779 %
4780 % \SpecialOptIndex{longtable}{package}
4781 % \SpecialOptIndex{longtable}{environment}
4782 % \SpecialIndex{\output}
4783 % \SpecialIndex{\adl@rowsL}
4784 % \SpecialIndex{\adl@rowsR}
4785 %
4786 % Making \textsf{arydshln} compatible with \textsf{longtable} is a hard job
4787 % because a |longtable| consists of multiple {\em chunks} and each chunk is
4788 % a distinct |\halign|. We could draw vertical lines in each chunks as we
4789 % do with ordinary |array|\slash |table|. However this straightforward
4790 % solution should {\em break} dash-lines at invisible borders of chunks and
4791 % produce awful results.
4792 %
4793 % Therefore, this implementation draws dash-lines in |\output| routine in
4794 % which we have all the rows to be put in a page. The hard part is to know
4795 % which rows are being put in |\output|. This problem is solved by
4796 % extracting the leading part of $R^L$ (|\adl@rowsL|) and $R^R$
4797 % (|\adl@rowsR|) by the height\slash depth of the table fraction to be put
4798 % and removing the part from $R^{L/R}$.
4799 %
4800 %
4801 %
4802 % \subsubsection{Initialization}
4803 % \label{sec:imp-lt-init}
4804 % \changes{v1.7}{2003/05/21}
4805 % {Codes for \textsf{longtable} is surrounded by \cs{ifx}/\cs{fi}}
4806 %
4807 % First of all, the following switch and |\dimen| register are declared.
4808 % \begin{itemize}
4809 % \item
4810 % \begin{macro}{\ifadl@LTfirstpage}
4811 % |\ifadl@LTfirstpage| is tested in |\output| routine to examine if the page
4812 % being put has the first fraction of a |longtable|.
4813 % \end{macro}
4814 %
4815 % \item
4816 % \begin{macro}{\adl@LTpagetotal}
4817 % |\adl@LTpagetotal| is set to |\pagetotal| just before the first portion of a
4818 % |longtable| is added to the main vertical list. Since the |\box255| has
4819 % items preceding the |\longtable| and its first fraction, we can obtain the
4820 % height of the first fraction by subtracting |\adl@LTpagetotal| from the
4821 % height plus depth of |\box255|.
4822 % \end{macro}
4823 % \end{itemize}
4824 %
4825 %
4826 % \begin{macrocode}
4827
4828 %% Compatibility with longtable: initialization
4829
4830 \newif\ifadl@LTfirstpage
4831 \newdimen\adl@LTpagetotal
4832
4833 % \end{macrocode}
4834 %
4835 % Next, we skip everything if \textsf{longtable} is not in use, or
4836 % we have undefined-error when we refer to the definitions in it. Note that
4837 % since |\newif| cannot be in the |\ifx|\slash|\fi| construct, the
4838 % declarations above are excluded.
4839 %
4840 % \begin{macrocode}
4841 \ifx\longtable\undefined\else
4842
4843 % \end{macrocode}
4844 %
4845 %
4846 % \begin{macro}{\adl@LT@array}
4847 % \begin{macro}{\LT@array}
4848 % \changes{v1.73-2}{2016/04/28}
4849 % {Add \cs{let}-assignment of \cs{adl@LThdashline} to
4850 % \cs{adl@hdashline} so that the \textsf{longtable} version of
4851 % \cs{adl@hdashline} is effective only in \texttt{longtable}
4852 % environment rather than globally.}
4853 %
4854 % \begin{macro}{\adl@discard}
4855 % \SpecialIndex{\adl@arrayinit}
4856 % \SpecialIndex{\adl@hdashline}
4857 % \SpecialIndex{\adl@ihdashline}
4858 % \SpecialIndex{\adl@LTihdashline}
4859 % \SpecialIndex{\ifadl@inactive}
4860 %
4861 % Then we redefine the macro |\LT@array|, which is the heart of |\longtable|,
4862 % saving its original definition in |\adl@LT@array|. The modified
4863 % |\LT@array| first calls |\adl@arrayinit| to initialize the global data
4864 % structures, and sets |\ifadl@LTfirstpage| to true. Then |\adl@dashline|,
4865 % |\adl@idashline| and |\adl@discard| are made |\let|-equal to the
4866 % \textsf{longtable} versions |\adl@LTdashline| and |\adl@LTidashline|, and
4867 % |\relax| (to inhibit expansion) respectively. Then the macro calls
4868 % |\adl@LTinactivate| if |\adl@inactive| is true, and finally calls its
4869 % original version |\adl@LT@array|. Note that since |longtable| cannot be
4870 % nested;
4871 %
4872 % \begin{itemize}
4873 % \item
4874 % |\adl@arraysave| in |\adl@arrayinit| is unnecessary but safe, and thus
4875 % its invocation timing is not so sensitive; and
4876 %
4877 % \item
4878 % activator is not required.
4879 % \end{itemize}
4880 %
4881 % Also note that the assignment |\adl@ncol| to |\adl@columns| in
4882 % |\adl@arrayinit| is void and thus we will do it afterward.
4883 % \end{macro}\end{macro}\end{macro}
4884 %
4885 % \begin{macro}{\adl@LTinactivate}
4886 % \SpecialIndex{\adl@inactivate}
4887 % \SpecialIndex{\adl@ihdashline}
4888 % \SpecialIndex{\adl@LTinactivehdl}
4889 % \SpecialIndex{\endlongtable}
4890 % \SpecialIndex{\LT@make@row}
4891 % \SpecialIndex{\LT@echunk}
4892 % \SpecialIndex{\LT@end@hd@ft}
4893 % \SpecialIndex{\LT@kill}
4894 % \SpecialIndex{\LT@output}
4895 % \SpecialIndex{\adl@org@endlongtable}
4896 % \SpecialIndex{\adl@org@LT@make@row}
4897 % \SpecialIndex{\adl@org@LT@echunk}
4898 % \SpecialIndex{\adl@org@LT@end@hd@ft}
4899 % \SpecialIndex{\adl@org@LT@kill}
4900 % \SpecialIndex{\adl@org@LT@output}
4901 %
4902 % The macro |\adl@LTinactivate| first calls |\adl@inactivate| to do basic
4903 % inactivation and then |\let|-s the following control sequences be equal to
4904 % their counterparts in \textsf{longtable}.
4905 % \begin{itemize}\item[]
4906 % |\endlongtable|\quad
4907 % |\LT@make@row|\quad
4908 % |\LT@echunk|\quad
4909 % |\LT@end@hd@ft|\quad
4910 % |\LT@kill|\\
4911 % |\LT@output|
4912 % \end{itemize}
4913 % It also make |\adl@idashline| |\let|-equal to its inactive version because
4914 % we need the macro to find mixed |\hline| and |\hdasnline| sequence.
4915 % \end{macro}
4916 %
4917 % \begin{macrocode}
4918 \let\adl@LT@array\LT@array
4919 \def\LT@array{\adl@arrayinit \adl@LTfirstpagetrue
4920 \let\adl@discard\relax \let\adl@hdashline\adl@LThdashline
4921 \let\adl@ihdashline\adl@LTihdashline
4922 \ifadl@inactive \adl@LTinactivate \fi
4923 \adl@LT@array}
4924 \def\adl@LTinactivate{\adl@inactivate
4925 \let\endlongtable\adl@org@endlongtable
4926 \let\LT@make@row\adl@org@LT@make@row
4927 \let\LT@echunk\adl@org@LT@echunk
4928 \let\LT@end@hd@ft\adl@org@LT@end@hd@ft
4929 \let\LT@kill\adl@org@LT@kill
4930 \let\LT@output\adl@org@LT@output
4931 \let\adl@ihdashline\adl@LTinactivehdl}
4932
4933 % \end{macrocode}
4934 %
4935 % \begin{macro}{\adl@org@LT@make@row}
4936 % \begin{macro}{\LT@make@row}
4937 % \SpecialIndex{\adl@ncol}
4938 % \SpecialIndex{\adl@columns}
4939 % \SpecialIndex{\adl@LTstartpbox}
4940 % \SpecialIndex{\adl@@LTstartpbox}
4941 % \SpecialIndex{\@startpbox}
4942 % \SpecialIndex{\@@startpbox}
4943 % \SpecialIndex{\adl@LTendpbox}
4944 % \SpecialIndex{\@endpbox}
4945 % \SpecialIndex{\@@endpbox}
4946 % \SpecialIndex{\LT@startpbox}
4947 % \SpecialIndex{\adl@endmbox}
4948 % \SpecialIndex{\adl@@endmbox}
4949 % \SpecialIndex{\adl@LTendmbox}
4950 %
4951 % \changes{v1.72-1}{2016/03/19}
4952 % {Modified to add \cs{let}-assignments to \cs{adl@@endmbox} and
4953 % \cs{adl@endbmox} so that footnotes are correclty processed at the
4954 % closing of a \texttt{m}-type column.}
4955 % \changes{v1.75}{2018/12/26}
4956 % {Add description that \cs{adl@LTendpbox} is common for \cs{@endpbox}
4957 % and \cs{@@endpbox}.}
4958 % \changes{v1.76}{2019/02/21}
4959 % {Modified to add \cs{let}-assignment
4960 % $\cs{@startpbox}\EQ\cs{adl@LTstartpbox}$ if \textsf{array} is in
4961 % use, and to replace the RHS of \cs{let}-assignments of
4962 % \cs{@@startpbox} and \cs{@startpbox} for the case without
4963 % \textsf{array}, which was \cs{adl@LTstartpbox}, with newly
4964 % introduced \cs{adl@@LTstartpbox}.}
4965 % 
4966 % The macro |\LT@make@row| is redefined for additional initialization which
4967 % must be done after the original |\LT@array| performs its own
4968 % initialization. First, |\LT@make@row| itself is reset to its original
4969 % version |\adl@org@LT@make@row| to initialize stuff only once, since
4970 % |\LT@make@row| is called repeatedly at each chunk. Next |\adl@ncol| is
4971 % assigned to |\adl@columns| to give its value calculated in |\@mkpream|.
4972 %
4973 % Then macros to begin\slash end |p|-boxes are made |\let|-equal to our own
4974 % version because the original |\LT@array| has done it with
4975 % \textsf{longtable}'s own version. That is, if \textsf{array} is in use
4976 % |\@startpbox| is |\let|-equal to our own |\adl@LTstartpbox|, while
4977 % |\@@startpbox| and |\@startpbox| are |\let|-equal to another macro
4978 % |\adl@@LTstartpbox| of our own. On the other hand, |\@@endpbox| and
4979 % |\@endpbox| are commonly |\let|-equal to our own |\adl@LTendpbox|. Note
4980 % that these our own macros indirectly invoke |\color@begingroup| and
4981 % |\color@endgroup|, which are added to |\LT@startpbox| and |\LT@endpbox| of
4982 % \textsf{longtable} bundled in \textsf{latex-tools} 2019-01-05, to make the
4983 % color grouping effective regardless the version of \textsf{longtable}.
4984 % Also note that we need |\adl@LTendmbox| to close |m|-boxes through our own
4985 % closing macro |\adl@endmbox|, whose definition is kept in |\adl@@endmbox|,
4986 % for \textsf{longtable}-specific operations for footnotes.
4987 %
4988 % Finally, the original version |\adl@org@LT@make@row| is called.
4989 % \end{macro}\end{macro}
4990 %
4991 % \begin{macrocode}
4992 \let\adl@org@LT@make@row\LT@make@row
4993 \def\LT@make@row{\let\LT@make@row\adl@org@LT@make@row
4994 \adl@columns\adl@ncol
4995 \ifadl@usingarypkg
4996 \let\@startpbox\adl@LTstartpbox
4997 \else
4998 \let\@@startpbox\adl@@LTstartpbox
4999 \let\@startpbox\adl@@LTstartpbox
5000 \fi
5001 \let\@@endpbox\adl@LTendpbox
5002 \let\@endpbox\adl@LTendpbox
5003 \let\adl@@endmbox\adl@endmbox
5004 \let\adl@endmbox\adl@LTendmbox
5005 \adl@org@LT@make@row}
5006
5007 %%^L
5008 % \end{macrocode}
5009 %
5010 % The summary of the activation and inactivation specific to
5011 % \textsf{longtable} is shown in Table~\ref{tab:lt-act}.
5012 %
5013 %
5014 %
5015 % \begin{table}
5016 % \catcode`\|12
5017 % \caption{Active and Inactive \textsf{longtable} Operations}\label{tab:lt-act}
5018 % \def\!{\texttt{\char`\\\char`\\}}
5019 % \centerline{\begin{tabular}[t]{|l||l|l|}\hline
5020 % command&active&inactive\\\hline\hline
5021 % {\tt p b} (open)&&\\
5022 % \qquad with \textsf{array}& \cs{adl@act@classz}&
5023 % \cs{adl@org@classz}\\
5024 % & \quad$\to$\cs{adl@LTstartpbox}&
5025 % \quad$\to$\cs{LT@startpbox}\\
5026 % \qquad without \textsf{array}&\cs{adl@@LTstartpbox}&
5027 % \cs{LT@startpbox}\\\hline
5028 % {\tt m} (open)& \cs{adl@act@classz}&
5029 % \cs{adl@org@classz}\\
5030 % & \quad$\to$\cs{adl@startmbox}&
5031 % \quad$\to$\cs{LT@startpbox}\\
5032 % & \quad$\to$\cs{LT@startpbox}&\\\hline
5033 % {\tt p b} (close)& \cs{adl@LTendpbox}&
5034 % \cs{LT@endpbox}\\\hline
5035 % {\tt m} (close)& \cs{adl@LTendmbox}&
5036 % \cs{LT@endpbox}\\\hline
5037 % \cs{hline}& $\to$\cs{adl@act@hline}&
5038 % $\to$\cs{@gobbletwo}\\\hline
5039 % \cs{hdashline}& $\to$\cs{adl@LTihdashline}&
5040 % $\to$\cs{adl@LTinactivehdl}\\
5041 % & $\to$\cs{adl@act@hline}&
5042 % $\to$\cs{@gobbletwo}\\\hline
5043 % \cs{endlongtable}& modified version&
5044 % \cs{adl@org@endlongtable}\\
5045 % \cs{LT@make@row}&& \cs{adl@org@LT@make@row}\\
5046 % \cs{LT@echunk}&& \cs{adl@org@LT@echunk}\\
5047 % \cs{LT@end@hd@ft}&& \cs{adl@org@LT@end@hd@ft}\\
5048 % \cs{LT@kill}&& \cs{adl@org@LT@kill}\\
5049 % \cs{LT@output}&& \cs{adl@org@LT@output}\\\hline
5050 % \end{tabular}}
5051 % \end{table}
5052 %
5053 %
5054 %
5055 % \subsubsection{Ending Chunks}
5056 % \label{sec:imp-lt-echunk}
5057 %
5058 % \begin{macro}{\adl@org@endlongtable}
5059 % \begin{macro}{\endlongtable}
5060 % \begin{macro}{\adl@org@LT@echunk}
5061 % \begin{macro}{\LT@echunk}
5062 % \begin{macro}{\adl@LTlastrow}
5063 % \SpecialIndex{\adl@height}
5064 % \SpecialIndex{\adl@depth}
5065 % \SpecialIndex{\adl@@cr}
5066 %
5067 % When a chunk is closed with |\crcr|, we have to add the information of the
5068 % last row to $R^{L/R}=|\adl@rowsL|/|R|$ if the row is not finished by an
5069 % explicit |\\|. This is done by |\adl@LTlastrow| as we did at the first
5070 % job of |\adl@endarray|. Two chunk closing macros, |\endlongtable| and
5071 % |\LT@echunk|, are modified to call |\adl@LTlastrow| before its original
5072 % job done by |\adl@org@endlongtable| and |\adl@org@LT@echunk| respectively.
5073 % Note that |\adl@LTlastrow| only has |\crcr| and |\noalign| and thus
5074 % another |\crcr| in original |\endlongtable| and |\LT@echunk| is
5075 % no-operation as desired. Also note that |\adl@LTlastrow| is called twice
5076 % from |\endlongtable|, once from |\LT@echunk| in the original version, but
5077 % it is safe because the first call makes |\adl@height| and |\adl@depth|
5078 % zero and thus the second become no-operation.
5079 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
5080 %
5081 %
5082 % \begin{macrocode}
5083
5084 %% Compatibility with longtable: end chunk
5085
5086 \let\adl@org@endlongtable\endlongtable
5087 \def\endlongtable{\adl@LTlastrow \adl@org@endlongtable}
5088
5089 \let\adl@org@LT@echunk\LT@echunk
5090 \def\LT@echunk{\adl@LTlastrow \adl@org@LT@echunk}
5091
5092 \def\adl@LTlastrow{\crcr \noalign{
5093 \ifdim\adl@height=\z@
5094 \ifdim\adl@depth=\z@ \else \adl@@cr\z@ \fi
5095 \else \adl@@cr\z@ \fi}}
5096
5097 % \end{macrocode}
5098 %
5099 % \begin{macro}{\adl@org@LT@end@hd@ft}
5100 % \begin{macro}{\LT@end@hd@ft}
5101 % \begin{macro}{\adl@LThfsave}
5102 % \begin{macro}{\adl@LTth}
5103 % \begin{Macro}{\\adl@LTth\LT@firsthead}
5104 % \begin{Macro}{\\adl@LTth\LT@head}
5105 % \begin{Macro}{\\adl@LTth\LT@lastfoot}
5106 % \begin{Macro}{\\adl@LTth\LT@foot}
5107 % \begin{Macro}{\\adl@rowsL\LT@firsthead}
5108 % \begin{Macro}{\\adl@rowsL\LT@head}
5109 % \begin{Macro}{\\adl@rowsL\LT@lastfoot}
5110 % \begin{Macro}{\\adl@rowsL\LT@foot}
5111 % \begin{Macro}{\\adl@rowsR\LT@firsthead}
5112 % \begin{Macro}{\\adl@rowsR\LT@head}
5113 % \begin{Macro}{\\adl@rowsR\LT@lastfoot}
5114 % \begin{Macro}{\\adl@rowsR\LT@foot}
5115 % \SpecialIndex{\endfirsthead}
5116 % \SpecialIndex{\endhead}
5117 % \SpecialIndex{\endlastfoot}
5118 % \SpecialIndex{\endfoot}
5119 % \SpecialIndex{\LT@firsthead}
5120 % \SpecialIndex{\LT@head}
5121 % \SpecialIndex{\LT@lastfoot}
5122 % \SpecialIndex{\LT@foot}
5123 % \SpecialIndex{\adl@rowsL}
5124 % \SpecialIndex{\adl@rowsR}
5125 % \SpecialIndex{\adl@totalheight}
5126 %
5127 % Another chunk ending macro is |\LT@end@hd@ft|\meta{box} to close a
5128 % header\slash footer called by |\endfirsthead|, |\endhead|, |\endlastfoot|
5129 % and |\endfoot| with an argument \meta{box} being |\LT@firsthead|,
5130 % |\LT@head|, |\LT@lastfoot| and |\LT@foot| respectively.
5131 % In order to maintain the information of rows
5132 % $R^{L/R}=|\adl@rowsL|/|R|$ of headers/footers separately from the main one,
5133 % the modified |\LT@end@hd@ft| saves them together with
5134 % |\adl@totalheight| to weirdly named macros;
5135 %
5136 % \begin{quote}
5137 % |\\adl@LTth|\meta{box}\\
5138 % |\\adl@rowsL|\meta{box}\\
5139 % |\\adl@rowsR|\meta{box}
5140 % \end{quote}
5141 %
5142 % after closing the last row by |\adl@LTlastrow|. The |\string|
5143 % representation of the macros looks like;
5144 %
5145 % \begin{quote}
5146 % |\\adl@LTth\LT@firsthead|
5147 % \end{quote}
5148 % and so on. The saving operation is done by
5149 % the macro |\adl@LThfsave|\meta{box}\meta{info} and is equivalent to;
5150 %
5151 % \begin{quote}
5152 % |\global\let\|\meta{info}\meta{box}=\meta{info}
5153 % \end{quote}
5154 %
5155 % After the saving, three global variables are reinitialized. Calling
5156 % |\adl@LTlastrow| twice, once from the original version through
5157 % |\LT@echunk| is safe as described above.
5158 % \end{Macro}\end{Macro}\end{Macro}\end{Macro}
5159 % \end{Macro}\end{Macro}\end{Macro}\end{Macro}
5160 % \end{Macro}\end{Macro}\end{Macro}\end{Macro}
5161 % \end{macro}\end{macro}\end{macro}\end{macro}
5162 %
5163 % \begin{macrocode}
5164 \let\adl@org@LT@end@hd@ft\LT@end@hd@ft
5165 \def\LT@end@hd@ft#1{\adl@LTlastrow
5166 \noalign{\edef\adl@LTth{\number\adl@totalheight}%
5167 \adl@LThfsave#1\adl@LTth \global\adl@totalheight\z@
5168 \adl@LThfsave#1\adl@rowsL\gdef\adl@rowsL{}%
5169 \adl@LThfsave#1\adl@rowsR\gdef\adl@rowsR{}}
5170 \adl@org@LT@end@hd@ft#1}
5171 \def\adl@LThfsave#1#2{\expandafter\global\expandafter\let
5172 \csname\string#2\string#1\endcsname#2}
5173
5174 % \end{macrocode}
5175 %
5176 % \begin{macro}{\adl@org@LT@kill}
5177 % \begin{macro}{\LT@kill}
5178 % \begin{macro}{\adl@LTkill}
5179 % \begin{macro}{\adl@LTkillend}
5180 % The additional job for yet another chunk closer |\LT@kill| to kill a
5181 % template row is a little bit harder. Since the row information might have
5182 % been added by an explicit |\\| preceding |\kill|, we have to remove it
5183 % from the tail of |\adl@rowsL|\slash|R|, and subtract its $h_i$ from
5184 % |\adl@totalheight| because |\kill|-ed row may be in header\slash footer
5185 % definition. To do that, modified |\LT@kill| first ensures the information
5186 % addition by |\adl@LTlastrow|, then traverses |\adl@rowsL|\slash|R| adding
5187 % its non-last elements to |\@tempb| by the loop of |\adl@LTkill|, and
5188 % assigns |\@tempb| to |\adl@rowsL|\slash|R| globally by
5189 % |\adl@LTkillend| when |\adl@LTkill| finds the tail. The macro
5190 % |\adl@LTkillend| also sets the $h_i$ of the last element to |\@tempcnta|,
5191 % which is subtracted from |\adl@totalheight| globally. Finally, the
5192 % original version |\adl@org@LT@kill| is called.
5193 % \end{macro}\end{macro}\end{macro}\end{macro}
5194 %
5195 % \begin{macrocode}
5196 \let\adl@org@LT@kill\LT@kill
5197 \def\LT@kill{\adl@LTlastrow \noalign{
5198 \def\@tempb{}\expandafter\adl@LTkill\adl@rowsL\@nil\adl@rowsL
5199 \def\@tempb{}\expandafter\adl@LTkill\adl@rowsR\@nil\adl@rowsR
5200 \global\advance\adl@totalheight-\@tempcnta}%
5201 \adl@org@LT@kill}
5202 \def\adl@LTkill#1;#2{\def\@tempa{#2}%
5203 \ifx\@tempa\@nnil\def\next{\adl@LTkillend#1}%
5204 \else\edef\@tempb{\@tempb#1;}\def\next{\adl@LTkill#2}\fi
5205 \next}
5206 \def\adl@LTkillend(#1/#2)#3{\global\let#3\@tempb \@tempcnta#2\relax}
5207
5208 %%^L
5209 % \end{macrocode}
5210 %
5211 %
5212 %
5213 % \subsubsection{Horizontal Lines and \texttt{p}-Boxes}
5214 % \label{sec:imp-lt-hlpbox}
5215 %
5216 % \begin{macro}{\LT@hline}
5217 % \begin{macro}{\adl@LThdashline}
5218 % \changes{v1.7-1}{2004/5/21}
5219 % {Keep original without shift up because it is done by
5220 % \cs{adl@LThdline}.}
5221 % \changes{v1.73-2}{2016/04/28}
5222 % {Renamed from \cs{adl@hdashline} to make it effective only in
5223 % \texttt{longtable} environments.}
5224 % \begin{macro}{\adl@LTihdashline}
5225 % \begin{macro}{\adl@LTinactivehdl}
5226 % \begin{macro}{\adl@LThdlrow}
5227 % \SpecialIndex{\hline}
5228 % \SpecialIndex{\adl@hdashline}
5229 % \SpecialIndex{\adl@ihdashline}
5230 % \SpecialIndex{\adl@inactivehdl}
5231 % \SpecialIndex{\adl@hcline}
5232 %
5233 % The macro |\LT@hline|, \textsf{longtable} version of |\hline|, is
5234 % redefined to add pseudo row information to $R^{L/R}$ and to check mixed
5235 % sequence of |\hline| and |\hdashline|\footnote{
5236 %
5237 % In the original \textsf{longtable}, a sequence of three \cs{hline}-s are not
5238 % recognized. This buggy feature is fixed in this implementation.}.
5239 %
5240 % We also define the macro |\adl@LTihdashline[|\meta{dash}|/|\meta{gap}|]|
5241 % and its inactive counterpart |\adl@LTinactivehdl| as the
5242 % \textsf{longtable} version of |\adl@ihdashline| and |\adl@inactivehdl|.
5243 % These two macros, the main part of |\hdashline|, are redefined to make it
5244 % possible that |\hdashline| can be broken into two part by \TeX's page
5245 % breaker.
5246 %
5247 % These three macros call a common routine |\adl@LThdline| after defining
5248 % |\adl@LThdlrow| which makes a row of horizontal (dash) line drawn by
5249 % |\multispan| and |\leaders\hrule| or
5250 % |\adl@hcline[|\meta{dash}|/|\meta{gap}|]|.
5251 %
5252 % Note that we define |\adl@LThdashline| to make |\adl@hdashline|
5253 % \cs{let}-equal to it in \textsf{longtable} environments because its
5254 % version without \textsf{longtable} performs a part of the job done by
5255 % |\adl@LThdline| as shown soon.
5256 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
5257 %
5258 % \begin{macrocode}
5259
5260 %% Compatibility with longtable: horizontal lines and p-boxes
5261
5262 \def\LT@hline{\noalign{\ifnum0=`}\fi
5263 \gdef\adl@LThdlrow{\multispan{\LT@cols}\unskip
5264 \leaders\hrule\@height\arrayrulewidth\hfill\cr}%
5265 \adl@LThdline}
5266 \def\adl@LThdashline#1{\noalign{\ifnum0=`}\fi
5267 \@ifnextchar[%]
5268 {#1}%
5269 {#1[\dashlinedash/\dashlinegap]}}
5270 \def\adl@LTihdashline[#1/#2]{%
5271 \gdef\adl@LThdlrow{\multispan{\LT@cols}\unskip
5272 \adl@hcline\z@[#1/#2]}%
5273 \adl@LThdline}
5274 \def\adl@LTinactivehdl[#1/#2]{%
5275 \gdef\adl@LThdlrow{\multispan{\LT@cols}\unskip
5276 \leaders\hrule\@height\arrayrulewidth\hfill\cr}%
5277 \adl@LThdline}
5278
5279 % \end{macrocode}
5280 %
5281 %
5282 % \begin{macro}{\adl@LThdline}
5283 % \changes{v1.7-1}{2004/5/21}
5284 % {Modified to shift up if null-wide.}
5285 % \begin{macro}{\adl@LTxhline}
5286 % \begin{macro}{\adl@LTixhline}
5287 % \SpecialIndex{\@medpenalty}
5288 % \SpecialIndex{\@lowpenalty}
5289 % \SpecialIndex{\adl@hline}
5290 % \SpecialIndex{\adl@connect}
5291 % \SpecialIndex{\adl@discard}
5292 %
5293 % The macro |\adl@LThdline| called by above three macros first
5294 % inserts a vertical penalty 10000 to inhibit page break between the
5295 % horizontal line and preceding row. Then it inserts
5296 % |\vskip-\arrayrulewidth| with another break inhibitor if
5297 % |\ADLnullwidehline| is in effect, or adds the
5298 % pseudo row information {\it connect}(|\arrayrulewidth|) to $R^{L/R}$ by
5299 % |\adl@hline|\footnote{
5300 %
5301 % Or do noting if inactive and thus it is \cs{let}-equal to \cs{@gobbletwo}.}.
5302 %
5303 % Next, it draw a horizontal (dash) line by
5304 % |\adl@LThdlrow| and checks if the following control sequence is |\hline|
5305 % or |\hdashline| by |\futurelet| and |\adl@LTxhline|. If |\hline| or
5306 % |\hdashline| is the next token, |\adl@LTixhline| is called to insert a
5307 % vertical penalty of $-$|\@medpenalty| and a vertical space of
5308 % |\doublerulesep|. The macro |\adl@LTixhline| also adds {\it
5309 % disconnect}(|\doublerulesep|) to $R^{L/R}$ and makes |\adl@LThdlrow| void.
5310 % Otherwise, |\adl@LThdline| inserts a vertical penalty of $-$|\@lowpanalty|
5311 % and a vertical space of $-$|\arrayrulewidth| and draws the horizontal
5312 % (dash) line again by |\adl@LThdlrow|. Thus a page can be broken between
5313 % two overlaid horizontal (dash) lines\footnote{
5314 % 
5315 % If the page is broken, the horizontal line at the beginning of the
5316 % succeeding page has a width even if \cs{ADLnullwidehline} is in effect.}.
5317 % 
5318 % Two pseudo row information, {\it
5319 % discard}($-$|\arrayrulewidth|) for the negative vertical space which may
5320 % be discarded and {\it connect}(|\arrayrulewidth|) for the second
5321 % horizontal line, are also added to $R^{L/R}$.
5322 % \end{macro}\end{macro}\end{macro}
5323 %
5324 % \begin{macrocode}
5325 \def\adl@LThdline{\penalty\@M
5326 \ifadl@zwhrule \vskip-\arrayrulewidth \penalty\@M
5327 \else \adl@hline\adl@connect\arrayrulewidth \fi
5328 \ifnum0=`{\fi}%
5329 \adl@LThdlrow
5330 \noalign{\ifnum0=`}\fi
5331 \futurelet\@tempa\adl@LTxhline}
5332 \def\adl@LTxhline{\ifx\@tempa\hline \adl@LTixhline
5333 \else\ifx\@tempa\hdashline \adl@LTixhline
5334 \else \penalty-\@lowpenalty \vskip-\arrayrulewidth
5335 \adl@hline\adl@discard{-\arrayrulewidth}%
5336 \adl@hline\adl@connect\arrayrulewidth
5337 \fi\fi \ifnum0=`{\fi}%
5338 \adl@LThdlrow \noalign{\penalty\@M}}
5339 \def\adl@LTixhline{\penalty-\@medpenalty \vskip\doublerulesep
5340 \adl@hline\relax\doublerulesep \global\let\adl@LThdlrow\@empty}
5341
5342 % \end{macrocode}
5343 %
5344 % \begin{macro}{\adl@@LTstartpbox}
5345 % \changes{v1.76}{2019/02/21}
5346 % {Added for \texttt{p} columns without \textsf{array}.}
5347 % \begin{macro}{\adl@LTstartpbox}
5348 % \changes{v1.76}{2019/02/21}
5349 % {Modified to invoke \cs{color@begingroup} always.}
5350 % \begin{macro}{\adl@LTendpbox}
5351 % \changes{v1.75}{2018/12/26}
5352 % {Add description that the macro is used for both of \cs{@endpbox}
5353 % and \cs{@@endpbox}.}
5354 % \changes{v1.76}{2019/02/21}
5355 % {Modified to invoke \cs{color@endroup} always.}
5356 % \begin{macro}{\adl@LTendmbox}
5357 % \changes{v1.72-1}{2016/03/19}
5358 % {Added to process footnotes in \texttt{m}-type columns appropriately.}
5359 % 
5360 % \SpecialIndex{\@startpbox}
5361 % \SpecialIndex{\@@startpbox}
5362 % \SpecialIndex{\adl@box}
5363 % \SpecialIndex{\@@endpbox}
5364 % \SpecialIndex{\LT@startpbox}
5365 % \SpecialIndex{\adl@act@startpbox}
5366 % \SpecialIndex{\adl@act@endpbox}
5367 %
5368 % Macros for opening/closing |p|-boxes are fairly simple. The macro
5369 % |\adl@LTstartpbox{|\meta{w}|}| is |\let|-assigned to |\@startpbox| by
5370 % |\LT@make@row| to open a |p|-box of $w$ wide by
5371 % |\adl@act@startpbox| and performs a footnote related operation introduced
5372 % by \textsf{longtable}, when \textsf{array} is in use. If not, this macro
5373 % is invoked from |\adl@@LTstartpbox| which is |\let|-equal to
5374 % |\@@startpbox| and is to assign the |p|-box to |\adl@box|. Since
5375 % |\adl@act@startpbox| is for opening |p|-box with \textsf{array}, it has
5376 % |\color@begingroup| in it and thus the color grouping is effective
5377 % regardless the version of \textsf{longtable} or \textsf{array}.
5378 %
5379 % On the other hand, the closing macro |\adl@LTendpbox| for |p| (or
5380 % |d|)-boxes is |\let|-equal to |\@endpbox| and |\@@endpbox| for the cases
5381 % with\slash without \textsf{array}, and performs the footnote operations
5382 % after doing our own ones by |\adl@act@endpbox|, which also has
5383 % |\color@begingroup| for version-independent color grouping.
5384 %
5385 % As for |m|-boxes, the opening operation is done by |\adl@startmbox| in
5386 % which $|\@startpbox|\allowbreak=|\adl@LTstartpbox|$ is invoked for the
5387 % footnote-related operation. On the ohter hand, the closing operation is
5388 % done by |\adl@LTendmbox|, which is made |\let|-equal to |\adl@endmbox| by
5389 % our |\LT@make@row|, to perform our own operations by |\adl@@endmbox| in
5390 % which the originnal definition of |\adl@enmbox| is kept by |\LT@make@row|
5391 % too. Since |\adl@LTstartpbox| and |\adl@@endmbox| have
5392 % |\color@begingroup| and |\color@endgroup|, the color-grouping is done
5393 % regardless of the version of \textsf{longtable} and \textsf{array}.
5394 % \end{macro}\end{macro}\end{macro}\end{macro}
5395 %
5396 % \begin{macrocode}
5397 \def\adl@@LTstartpbox{\setbox\adl@box\vtop\adl@LTstartpbox}
5398 \def\adl@LTstartpbox#1{\adl@act@startpbox{#1}\let\@footnotetext\LT@p@ftntext}
5399 \def\adl@LTendpbox{\adl@act@endpbox \the\LT@p@ftn \global\LT@p@ftn{}}
5400 \def\adl@LTendmbox{\adl@@endmbox \the\LT@p@ftn \global\LT@p@ftn{}}
5401
5402 %%^L
5403 % \end{macrocode}
5404 %
5405 %
5406 %
5407 % \subsubsection{First Chunk}
5408 % \label{sec:imp-lt-start}
5409 %
5410 % \begin{macro}{\LT@start}
5411 % \SpecialIndex{\LT@firsthead}
5412 % \SpecialIndex{\adl@LTpagetotal}
5413 %
5414 % The macro |\LT@start| which puts (first) head and controls the page break
5415 % of the first page is modified for the following.
5416 %
5417 % \begin{itemize}
5418 % \item
5419 % After it inserts a vertical skip |\LTpre|, |\endgraf| is performed so that
5420 % the skip contributes to |\pagetotal|\footnote{
5421 %
5422 % This modification is necessary for the original \textsf{longtable}, or it
5423 % underestimates the room of the first page and leaves head and foot only.}.
5424 %
5425 % \item
5426 % When the |\box2| is |\vsplit| to get first item of the first chunk,
5427 % |\vbadness| is saved into |\@tempcnta|, set to 10000 to avoid unnecessary
5428 % |underfull| message\footnote{
5429 %
5430 % This is also necessary for the original version.},
5431 %
5432 % and restored from |\@tempcnta|.
5433 %
5434 % \item
5435 % The |\dimen| register |\adl@LTpagetotal| is set to |\pagetotal| to know
5436 % the total height of the items preceding |longtable|. Since the assignment
5437 % is performed after the inserted |\endgraf| and the intentional page break,
5438 % it should have real total height.
5439 %
5440 % \item
5441 % The box |\LT@firsthead| is put by |\copy| rather than |\box| because it is
5442 % referred to in the |\output| routine.
5443 % \end{itemize}
5444 %
5445 % This macro does not have inactive counterpart because the modification
5446 % shown above is desirable (first two) or not-harmful\footnote{
5447 %
5448 % Logically, at least.}
5449 %
5450 % (last two) to the original version.
5451 % \end{macro}
5452 %
5453 % \begin{macrocode}
5454
5455 %% Compatibility with longtable: first chunk
5456
5457 \def\LT@start{%
5458 \let\LT@start\endgraf
5459 \endgraf \penalty\z@ \vskip\LTpre \endgraf
5460 \dimen@\pagetotal
5461 \advance\dimen@ \ht\ifvoid\LT@firsthead\LT@head\else\LT@firsthead\fi
5462 \advance\dimen@ \dp\ifvoid\LT@firsthead\LT@head\else\LT@firsthead\fi
5463 \advance\dimen@ \ht\LT@foot
5464 \dimen@ii\vfuzz \@tempcnta\vbadness
5465 \vfuzz\maxdimen \vbadness\@M
5466 \setbox\tw@\copy\z@
5467 \setbox\tw@\vsplit\tw@ to \ht\@arstrutbox
5468 \setbox\tw@\vbox{\unvbox\tw@}%
5469 \vfuzz\dimen@ii \vbadness\@tempcnta
5470 \advance\dimen@\ht
5471 \ifdim\ht\@arstrutbox>\ht\tw@\@arstrutbox\else\tw@\fi
5472 \advance\dimen@\dp
5473 \ifdim\dp\@arstrutbox>\dp\tw@\@arstrutbox\else\tw@\fi
5474 \advance\dimen@ -\pagegoal
5475 \ifdim \dimen@>\z@\vfil\break \fi
5476 \global\adl@LTpagetotal\pagetotal
5477 \global\@colroom\@colht
5478 \ifvoid\LT@foot\else
5479 \advance\vsize-\ht\LT@foot
5480 \global\advance\@colroom-\ht\LT@foot
5481 \dimen@\pagegoal\advance\dimen@-\ht\LT@foot\pagegoal\dimen@
5482 \maxdepth\z@
5483 \fi
5484 \copy\ifvoid\LT@firsthead \LT@head \else \LT@firsthead \fi
5485 \output{\LT@output}}
5486
5487 %%^L
5488 % \end{macrocode}
5489 %
5490 %
5491 %
5492 % \subsubsection{Output Routine}
5493 % \label{sec:imp-lt-output}
5494 %
5495 % \begin{macro}{\adl@org@LT@output}
5496 % \begin{macro}{\LT@output}
5497 % \SpecialIndex{\LT@firsthead}
5498 % \SpecialIndex{\LT@head}
5499 % \SpecialIndex{\LT@lastfoot}
5500 % \SpecialIndex{\LT@foot}
5501 %
5502 % The output routine is the heart of the \textsf{longtable} compatible
5503 % implementation. The macro |\LT@output| which is set to |\output| by
5504 % |\LT@start| is modified from its original (and thus inactive) version
5505 % |\adl@org@LT@output| as follows.
5506 %
5507 % \begin{itemize}
5508 % \item
5509 % Three fractions of the original version to compile the final output image
5510 % of the table portion into |\box255| or the main vertical list are modified
5511 % to set the image into |\box255| unconditionally and to call
5512 % |\adl@LTdraw|\meta{foot}\meta{tail} which is the real heart of the
5513 % compatible implementation. The argument \meta{foot} is |\LT@foot| or
5514 % |\LT@lastfoot| according to the portion of the |longtable| to be output.
5515 % The argument \meta{tail} is |\vss| if the last item is it which is not
5516 % included in |\box255| yet, or |\@empty| otherwise. Since |\adl@LTdraw|
5517 % builds final output image drawing vertical (dash) lines in |\box255|, it is
5518 % put to the main vertical list if the |longtable| portion is the last one.
5519 %
5520 % \item
5521 % Since the boxes |\LT@head|, |\LT@foot| and |\LT@lastfoot| are referred to in
5522 % |\adl@LTdraw|, they are put by |\copy| rather than |\box|.
5523 % \end{itemize}
5524 % \end{macro}\end{macro}
5525 %
5526 % \begin{macrocode}
5527
5528 %% Compatibility with longtable: output routine
5529
5530 \let\adl@org@LT@output\LT@output
5531 \def\LT@output{%
5532 \ifnum\outputpenalty <-\@Mi
5533 \ifnum\outputpenalty > -\LT@end@pen
5534 \LT@err{floats and marginpars not allowed in a longtable}\@ehc
5535 \else
5536 \setbox\z@\vbox{\unvbox\@cclv}%
5537 \ifdim \ht\LT@lastfoot>\ht\LT@foot
5538 \dimen@\pagegoal
5539 \advance\dimen@-\ht\LT@lastfoot
5540 \ifdim\dimen@<\ht\z@
5541 \setbox\@cclv\vbox{\unvbox\z@\copy\LT@foot}%
5542 \adl@LTdraw\LT@foot\vss
5543 \@makecol
5544 \@outputpage
5545 \setbox\z@\vbox{\copy\LT@head}%
5546 \fi
5547 \fi
5548 \global\@colroom\@colht
5549 \global\vsize\@colht
5550 \setbox\@cclv\vbox{\unvbox\z@
5551 \copy\ifvoid\LT@lastfoot\LT@foot\else\LT@lastfoot\fi}%
5552 \adl@LTdraw\LT@lastfoot\@empty \box\@cclv
5553 \fi
5554 \else
5555 \setbox\@cclv\vbox{\unvbox\@cclv\copy\LT@foot}%
5556 \adl@LTdraw\LT@foot\vss
5557 \@makecol
5558 \@outputpage
5559 \global\vsize\@colroom
5560 \copy\LT@head
5561 \fi}
5562
5563 % \end{macrocode}
5564 %
5565 %
5566 % \begin{macro}{\adl@LTdraw}
5567 % \begin{macro}{\adl@LTinit}
5568 % \begin{macro}{\adl@LTheadL}
5569 % \begin{macro}{\adl@LTheadR}
5570 % \begin{macro}{\adl@LTfootL}
5571 % \begin{macro}{\adl@LTfootR}
5572 % \SpecialIndex{\adl@LTpagetotal}
5573 % \SpecialIndex{\ifadl@LTfirstpage}
5574 %
5575 % The macro |\adl@LTdraw|\meta{foot}\meta{tail} draws vertical
5576 % (dash) lines onto the image in |\box255|. First it measures the total
5577 % height $H$ (|\adl@totalheight|) of |longtable| rows in |\box255| and the
5578 % total height $H_b$ (|\@tempdima|) of its {\em body} which consists of the
5579 % rows without the header and footer, as follows where $H_{255}$, $H_h$ and
5580 % $H_t$ are the height plus depth of |\box255| and the effective header and
5581 % footer of the page respectively.
5582 %
5583 % \begin{eqnarray*}
5584 % T&=&\cases{\hbox{\cs{adl@LTpagetotal}}&if \cs{ifadl@LTfirstpage}\cr
5585 % 0&otherwise}\\
5586 % t&=&\cases{
5587 % \hbox{\cs{topskip} glue}&
5588 % if \texttt{longtable} is the first item of the page\cr
5589 % & ($\lnot(\hbox{\cs{ifadl@firstpage}}\land T{>}0)$)\cr
5590 % 0& otherwise}\\
5591 % H&=&H_{255}-t-T\\
5592 % H_b&=&H-H_h-H_t
5593 % \end{eqnarray*}
5594 %
5595 % The hard part is to measure $t$ because it is not |\topskip| but that
5596 % minus the first box of |\box255|. Thus we do not measure $t$ but remove
5597 % it from the box by the following tricky way. First we copy |\box255|
5598 % items into |\box0| adding a |\hrule| of 1\,sp high as its first item.
5599 % Then |\box0| is |\vsplit| to 1\,sp setting |\splittopskip| to 0.
5600 % Since the |\topskip| glue is the first item of |\box255| and the |\vsplit|
5601 % discards it at the breakpoint, |\box0| must have all the items in
5602 % |\box255| lead by 0 (|\splittopskip|) glue rather than |\topskip| glue.
5603 % Thus the height of |\box0| is $H_{255}-t$.
5604 %
5605 % \SpecialIndex{\LT@firsthead}
5606 % \SpecialIndex{\LT@head}
5607 % \SpecialIndex{\LT@lastfoot}
5608 % \SpecialIndex{\LT@foot}
5609 % \SpecialWeirdIndex{\\adl@LTth\LT@firsthead}
5610 % \SpecialWeirdIndex{\\adl@LTth\LT@head}
5611 % \SpecialWeirdIndex{\\adl@LTth\LT@lastfoot}
5612 % \SpecialWeirdIndex{\\adl@LTth\LT@foot}
5613 % \SpecialWeirdIndex{\\adl@rowsL\LT@firsthead}
5614 % \SpecialWeirdIndex{\\adl@rowsL\LT@head}
5615 % \SpecialWeirdIndex{\\adl@rowsL\LT@lastfoot}
5616 % \SpecialWeirdIndex{\\adl@rowsL\LT@foot}
5617 % \SpecialWeirdIndex{\\adl@rowsR\LT@firsthead}
5618 % \SpecialWeirdIndex{\\adl@rowsR\LT@head}
5619 % \SpecialWeirdIndex{\\adl@rowsR\LT@lastfoot}
5620 % \SpecialWeirdIndex{\\adl@rowsR\LT@foot}
5621 %
5622 % Subtraction of $H_h$ and $H_t$ is done by the macro
5623 % |\adl@LTinit{|\meta{hf}|}|\meta{box}, where \meta{hf} is |head| or |foot|
5624 % and \meta{box} is one of |\LT@firsthead|, |\LT@head| and \meta{foot}
5625 % (|\LT@lastfoot| or |\LT@foot|). This macro also copies the contents of
5626 % weirdly named structure such as |\\adl@rowsL\LT@head| into |\adl@LTheadL|
5627 % and so on\footnote{
5628 %
5629 % Copying by \cs{edef} can be replaced by \cs{let} with many
5630 % \cs{expandafter} but it is not comprehensible.}
5631 %
5632 % if \meta{box} is not void. Otherwise, |\adl@LTheadL| etc. is kept to
5633 % their initial value, |\@empty|.
5634 %
5635 % \SpecialIndex{\adl@makevlrL}
5636 % \SpecialIndex{\adl@makevlrL}
5637 % \SpecialIndex{\adl@rows}
5638 % \SpecialIndex{\adl@rowsL}
5639 % \SpecialIndex{\adl@rowsR}
5640 % \SpecialIndex{\adl@vlrow}
5641 % \SpecialIndex{\adl@vlrowL}
5642 % \SpecialIndex{\adl@vlrowR}
5643 % \SpecialIndex{\adl@addvl}
5644 % \SpecialIndex{\adl@addvlL}
5645 % \SpecialIndex{\adl@addvlR}
5646 % \SpecialIndex{\adl@currentcolumn}
5647 %
5648 % Next, we make rows for vertical lines by |\adl@makevlrL|\slash|R| after
5649 % extracting the leading part of $R^{L/R}$ corresponding to the {\em body}
5650 % by the macro
5651 % |\adl@LTsplit|\meta{$R^{L/R}$}\meta{$R^{L/R}_h$}\meta{$R^{L/R}_f$}, where
5652 % $R^{L/R}_h$ and $R^{L/R}_f$ are |\adl@LTheadL| and so on. Since the macro
5653 % defines |\adl@rows| given to |\adl@makevlL|\slash|R| to the sequence of
5654 % $R^{L/R}_h$, the extracted part of $R^{L/R}$ and $R^{L/R}_f$, the rows for
5655 % vertical lines for all the rows including header and footer are build in
5656 % |\adl@vlrowL| and |\adl@vlrowR| as in the ordinary case without
5657 % |longtable|.
5658 %
5659 % \SpecialIndex{\LT@bchunk}
5660 % \SpecialIndex{\LT@save@row}
5661 % \SpecialIndex{\c@LT@chunks}
5662 % \SpecialIndex{\LT@rows}
5663 % \SpecialIndex{\adl@vl}
5664 % \SpecialIndex{\adl@@vl}
5665 % \SpecialIndex{\adl@drawvl}
5666 %
5667 % Then the rows are put into |\box0| by calling |\LT@bchunk| with
5668 % |\adl@drawvl| (line drawing) and |\LT@save@row| (column widths
5669 % adjustment), saving\slash restoring counters |\LT@rows| and |\c@LT@chunks|
5670 % which |\LT@bchunk| globally updates. Since we refer to potentially
5671 % immature |\LT@save@row| here, some weird looking vertical lines could be
5672 % drawn but the result after convergence should be correct. Finally, the
5673 % contents of |\box255| followed by the vertical lines in |\box0| are put
5674 % back into |\box255| keeping its original depth and adding \meta{tail}
5675 % (|\vss| or nothing) to its end.
5676 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
5677 %
5678 % \begin{macrocode}
5679 \def\adl@LTdraw#1#2{%
5680 \@tempswatrue
5681 \ifadl@LTfirstpage\ifdim\adl@LTpagetotal>\z@\@tempswafalse \fi\fi
5682 \if@tempswa
5683 \setbox\z@\vbox{\hrule height1sp\unvcopy\@cclv}
5684 \splittopskip\z@
5685 \setbox\@ne\vsplit\z@ to1sp\relax
5686 \@tempdima\ht\z@
5687 \else \@tempdima\ht\@cclv \fi
5688 \advance\@tempdima\dp\@cclv
5689 \adl@totalheight\@tempdima
5690 \let\adl@LTheadL\@empty \let\adl@LTheadR\@empty
5691 \let\adl@LTfootL\@empty \let\adl@LTfootR\@empty
5692 \ifadl@LTfirstpage
5693 \global\adl@LTfirstpagefalse
5694 \advance\@tempdima-\adl@LTpagetotal
5695 \adl@totalheight\@tempdima
5696 \ifvoid\LT@firsthead
5697 \adl@LTinit{head}\LT@head
5698 \else \adl@LTinit{head}\LT@firsthead
5699 \fi
5700 \else \adl@LTinit{head}\LT@head \fi
5701 \ifvoid#1%
5702 \adl@LTinit{foot}\LT@foot
5703 \else \adl@LTinit{foot}#1\fi
5704 \let\adl@vl\relax \def\adl@discard{\adl@connect}%
5705 \def\adl@vlrow{}\adl@currentcolumn\@ne
5706 \adl@LTsplit\adl@rowsL\adl@LTheadL\adl@LTfootL
5707 \let\adl@addvl\adl@addvlL
5708 \adl@makevlrL \let\adl@vlrowL\adl@vlrow
5709 \def\adl@vlrow{}\adl@currentcolumn\adl@columns
5710 \adl@LTsplit\adl@rowsR\adl@LTheadR\adl@LTfootR
5711 \let\adl@addvl\adl@addvlR
5712 \adl@makevlrR \let\adl@vlrowR\adl@vlrow
5713 \let\adl@vl\adl@@vl
5714 \@tempcnta\LT@rows
5715 \LT@bchunk \adl@drawvl
5716 \LT@save@row\cr \egroup \setbox\@ne\lastbox \unskip \egroup
5717 \global\advance\c@LT@chunks\m@ne
5718 \global\LT@rows\@tempcnta
5719 \@tempdima\dp\@cclv
5720 \setbox\@cclv\vbox{\unvbox\@cclv \box\z@ \vskip-\@tempdima
5721 \hrule\@width\z@\@height\z@\@depth\@tempdima#2}}
5722 \def\adl@LTinit#1#2{\ifvoid#2\else
5723 \advance\@tempdima-\csname\string\adl@LTth\string#2\endcsname sp%
5724 \expandafter\edef\csname adl@LT#1L\endcsname{%
5725 \csname\string\adl@rowsL\string#2\endcsname}%
5726 \expandafter\edef\csname adl@LT#1R\endcsname{%
5727 \csname\string\adl@rowsR\string#2\endcsname}\fi}
5728
5729 % \end{macrocode}
5730 %
5731 % \begin{macro}{\adl@LTsplit}
5732 % \begin{macro}{\adl@LTxsplit}
5733 % \begin{macro}{\adl@LTrowrelax}
5734 % \begin{macro}{\adl@LTrowdiscard}
5735 % \begin{macro}{\adl@LTysplit}
5736 % \begin{macro}{\adl@LTisplit}
5737 % \begin{macro}{\adl@LTiisplit}
5738 % \begin{macro}{\adl@LTsplitend}
5739 % \SpecialIndex{\adl@rows}
5740 % \SpecialIndex{\adl@discard}
5741 %
5742 % The macro
5743 % |\adl@LTsplit|\meta{$R^{L/R}$}\meta{$R^{L/R}_h$}\meta{$R^{L/R}_f$} moves
5744 % leading elements in $R^{L/R}$ into $R'$ (|\adl@rows|) until total heights
5745 % of the elements summed in $h$ (|\@tempdimb|) reaches to $H_b$
5746 % (|\@tempdima|)\footnote{
5747 %
5748 % Although $h$ must become $H_b$ exactly in usual case, we stop the loop
5749 % when $h\geq H_b$ to avoid accidental overrun in unusual cases.}
5750 %
5751 % by a straightforward loop with the macros |\adl@LTisplit| to fetch the
5752 % $i$-th element and |\adl@LTiisplit| to get $h_i$. Before moving, however,
5753 % we have to remove discardable item(s)\footnote{
5754 %
5755 % Must be only one but the implementation allows two or more.}
5756 %
5757 % from the top of $R^{L/R}$. Since an element for a discardable item is
5758 % {\it disconnect} (|\relax|) or {\it discard} (|\adl@discard|), we check
5759 % the first part of the element by |\ifx|-comparison with |\adl@LTrowrelax|
5760 % and |\adl@LTrowdiscard| whose bodies are |\relax| and |\adl@discard| if the
5761 % |longtable| portion does not have a header ($R^{L/R}_h$ is |\@empty|).
5762 % Otherwise, the discardable item was not discarded because the first
5763 % item of the page is not it but the header.
5764 %
5765 % Note that since moving from $R^{L/R}$ to $R'$ is done by |\edef| and
5766 % |\adl@discard| is |\def|-ined as |\adl@connect| in |\adl@LTdraw|,
5767 % non-discarded {\it discard} transforms into {\it connect} in $R'$. Also
5768 % note that since the remaining part of $R^{L/R}$ is |\def|-ined as the body
5769 % of |\@tempb| which is globally |\let|-assigned to $R^{L/R}$ again,
5770 % |\adl@discard| survives in the new $R^{L/R}$.
5771 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
5772 % \end{macro}\end{macro}\end{macro}
5773 %
5774 % \begin{macrocode}
5775 \def\adl@LTsplit#1#2#3{\def\adl@rows{}\@tempdimb\z@
5776 \expandafter\adl@LTxsplit#1\@nil;%
5777 \edef\adl@rows{#2\adl@rows#3}%
5778 \global\let#1\@tempb}
5779 \def\adl@LTxsplit#1;{\def\@tempa{#1}%
5780 \ifx\@tempa\@nnil \def\@tempb{}\let\next\relax
5781 \else\ifx\adl@LTheadL\@empty \def\next{\adl@LTysplit#1}%
5782 \else \def\next{\adl@LTisplit#1;}\fi \fi
5783 \next}
5784 \def\adl@LTrowrelax{\relax}
5785 \def\adl@LTrowdiscard{\adl@discard}
5786 \def\adl@LTysplit(#1/#2){\def\@tempa{#1}%
5787 \ifx\@tempa\adl@LTrowrelax \let\next\adl@LTxsplit
5788 \else\ifx\@tempa\adl@LTrowdiscard \let\next\adl@LTxsplit
5789 \else \def\next{\adl@LTisplit(#1/#2);}\fi \fi
5790 \next}
5791 \def\adl@LTisplit#1;{\def\@tempa{#1}%
5792 \ifx\@tempa\@nnil \def\@tempb{}\let\next\relax
5793 \else\ifdim\@tempdimb<\@tempdima
5794 \adl@LTiisplit#1\let\next\adl@LTisplit
5795 \else \def\next{\adl@LTsplitend#1;}\fi \fi
5796 \next}
5797 \def\adl@LTiisplit(#1/#2){\edef\adl@rows{\adl@rows(#1/#2);}%
5798 \advance\@tempdimb#2sp}
5799 \def\adl@LTsplitend#1;\@nil;{\def\@tempb{#1;}}
5800 \fi
5801
5802 %%^L
5803 % \end{macrocode}
5804 %
5805 %
5806 %
5807 % \subsection{Compatibility with \textsf{colortbl}}
5808 % \label{sec:imp-colortbl}
5809 % \changes{v1.7}{2003/05/21}
5810 % {\S\ref{sec:imp-colortbl} is added.}
5811 %
5812 % The implementation to make \textsf{arydshln} compatible with
5813 % \textsf{colortbl} consists of the following three (almost
5814 % independent) issues.
5815 %
5816 % \begin{description}
5817 % \item[Cell coloring]
5818 % is the easiest part because it does not affect dash line drawing. Another
5819 % reason of the easiness is that \textsf{colortbl} packs each cell in a box
5820 % to measure its height for painting in the modified version of |\@classz|.
5821 % Thus we do not need to code |\@classz| for both of \textsf{colortbl} and
5822 % \textsf{arydshln}, but may sneak our own height\slash depth measurement
5823 % into |\@classz| of \textsf{colortbl}. Almost everything we have to pay
5824 % attention to is the compatibility of the initialization and finalization
5825 % of \textsf{colortbl} and \textsf{arydshln}.
5826 %
5827 % \item[Horizontal line coloring]
5828 % is relatively easy because it is almost enough to insert coloring macro
5829 % |\CT@arc@| before the line drawing. A little bit complicated part is the
5830 % gap coloring which is done by drawing a solid line of gap color before
5831 % dash line is drawn.
5832 %
5833 % \item[Vertical line coloring]
5834 % is the hardest part but almost everything is done in previous sections to
5835 % attach dash\slash gap color to each vertical line segment $e^i_j$ in the
5836 % list $C_i^L$ and $C_i^R$ of the $i$-th row information $r_i$. What we do
5837 % here is to fix the bugs of |\arrayrulecolor| and |\doublerulesepcolor| in
5838 % \textsf{colortbl} implementation and to add |\dashgapcolor|. If you put
5839 % |\arrayrulecolor| in |>{...}| construct to specify the color
5840 % of the vertical lines following the construct as the manual of
5841 % \textsf{colortbl} says, you will have an error message ``Misplaced
5842 % |\noalign|'' because the macro is expanded with |\noalign| in a column
5843 % body. Even if you somehow remove |\noalign| to avoid the error, you will
5844 % have a mysterious line coloring as follows:
5845 %
5846 % \begin{itemize}
5847 % \item
5848 % If you have |\arrayrulecolor| before the |\array|\slash|\tabular| starts,
5849 % |\arrayrule#color| in the preamble has no effect to vertical lines but
5850 % decides the color of horizontal lines except for those at the top of the
5851 % environment. Additional |\arrayrulecolor| at the beginning of a row has
5852 % no effect to vertical lines (as expected) but decides horizontal lines
5853 % following it (also as expected). The effect of |\doublerulesepcolor| is
5854 % same as |\arrayrulecolor|.
5855 %
5856 % \item
5857 % Otherwise, i.e. without |\arrayrulecolor| outside the environment,
5858 % |\arrayrule#color| in the preamble decides the color of vertical and
5859 % horizontal lines except for verticals preceding columns in the first row
5860 % and horizontals at the top of the environment. Additional
5861 % |\arrayrulecolor| at the beginning of a row decides all the vertical and
5862 % horizontal lines following it. On the other hand, |\doublerulesepcolor|
5863 % acts as if |\doublerulesepcolor{white}| is done outside the environment.
5864 % \end{itemize}
5865 %
5866 % The reason of the mysterious behavior is as follows. An
5867 % |\arrayrulecolor|, which globally |\def|-ines a macro |\CT@arc@| with a
5868 % body containing |\color|, in the preamble is not expanded nor evaluated in
5869 % the preamble construction phase but done when the first (and succeeding)
5870 % row is build. On the other hand, |\CT@arc@| attached to vertical line
5871 % drawing is expanded in the preamble construction phase. Thus if
5872 % |\CT@arc@| has been defined before the environment starts, vertical lines
5873 % are colored following the outside definition. Otherwise, since |\CT@arc@|
5874 % is |\let|-equal to |\relax|, it remains unchanged in the preamble
5875 % construction phase and expanded when each row is build referring to its
5876 % definition that |\arrayrulecolor| modifies in the row building phase.
5877 % Since the macro |\CT@drsc@| defined by |\doublerulesepcolor| is examined
5878 % if it is |\relax| or not in the preamble construction phase,
5879 % |\doublerulesepcolor| in the preamble has no effect regardless the
5880 % existence of the outside definition.
5881 %
5882 % Thus we have to expand and evaluate |\arrayrulecolor| and
5883 % |\doublerulecolor| in the preamble construction phase to define |\CT@arc@|
5884 % and |\CT@drsc@|. We also have to initialize |\CT@arc@| as an expandable
5885 % but non-operative token (e.g. a macro with a body of |\relax| as we do) to
5886 % make it is expanded in the preamble construction phase rather than the row
5887 % building.
5888 % \end{description}
5889 %
5890 %
5891 %
5892 % \subsubsection{Initialization, Cell Coloring and Finalization}
5893 % \label{sec:impl-colortbl-init}
5894 %
5895 % \begin{macro}{\CT@arc@}
5896 % \begin{macro}{\adl@dashgapcolor}
5897 % First of all, we initialize the macro |\CT@arc@|, which will be
5898 % |\def|-ined as |\color| to specify the color of solid lines and dash
5899 % segments by |\arrayrulecolor|, with a body of |\relax| because it will be
5900 % referred to by the vertical line drawing process even if \textsf{colortbl} is
5901 % not in use. We also initialize the macro |\adl@dashgapcolor| for the
5902 % color of gaps of dash lines similarly. Note that these macros are not
5903 % |\let|-equal to |\relax| but have bodies of |\relax| so that they are
5904 % replaced with |\relax| in the preamble construction phase rather than
5905 % surviving with their own name.
5906 % \end{macro}\end{macro}
5907 %
5908 % \begin{macrocode}
5909
5910 %% Compatibility with colortbl
5911
5912 \def\CT@arc@{\relax}
5913 \def\adl@dashgapcolor{\relax}
5914 % \end{macrocode}
5915 %
5916 % \SpecialIndex{\@ifpackageloaded}
5917 % Next we examine if \textsf{colortbl} is in use by |\@ifpackageloaded|, and
5918 % skip everything if not, or we have some errors especially when
5919 % \textsf{array} is not in use.
5920 %
5921 % \begin{macrocode}
5922 \@ifpackageloaded{colortbl}\@tempswatrue\@tempswafalse
5923 \if@tempswa
5924 % \end{macrocode}
5925 %
5926 % \begin{macro}{\adl@org@inactivate}
5927 % \begin{macro}{\adl@org@activate}
5928 % \begin{macro}{\adl@inactivate}
5929 % \begin{macro}{\adl@activate}
5930 % \changes{v1.75}{2018/12/26}
5931 % {Add inactivation of \cs{@startpbox}.}
5932 % \begin{macro}{\CT@setup}
5933 % \begin{macro}{\@endpbox}
5934 % Then we redefine |\adl@inactivate| and |\adl@activate| referring their
5935 % original version |\adl@org@inactivate| and |\adl@org@activate| so that
5936 % they make |\CT@setup| |\let|-equal to its original version |\adl@CT@setup|
5937 % if |\ADLinactivate| is in effect, or to our own version
5938 % |\adl@act@CT@setup| which will be defined soon. New |\adl@activate| also
5939 % {\em inactivates} \hbox{|\@startpbox|} and \hbox{|\@endpbox|} because our
5940 % own ones for column height\slash depth measuremnt is inappropriate with
5941 % \textsf{colortbl} as explained soon.
5942 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
5943 %
5944 % \begin{macrocode}
5945 \let\adl@org@inactivate\adl@inactivate
5946 \let\adl@org@activate\adl@activate
5947 \def\adl@inactivate{\adl@org@inactivate \let\CT@setup\adl@CT@setup}
5948 \def\adl@activate{\adl@org@activate \let\CT@setup\adl@act@CT@setup
5949 \let\@startpbox\adl@org@startpbox \let\@endpbox\adl@org@endpbox}
5950
5951 % \end{macrocode}
5952 %
5953 % \begin{macro}{\adl@CT@setup}
5954 % \begin{macro}{\CT@setup}
5955 % \begin{macro}{\adl@act@CT@setup}
5956 % \SpecialIndex{\adl@@colhtdp}
5957 % Cell coloring is done by |\@classz| preamble of \textsf{colortbl} in which
5958 % a column is packed in |\box0|. On the other hand, our own |\@classz| one
5959 % with \textsf{array} packs the column in |\adl@box| so that we measure its
5960 % height
5961 % and depth. Thus we have choices; to insert height\slash depth measurement
5962 % into \textsf{colrotbl}'s version; or to insert coloring into our own version.
5963 % Since the code of height\slash depth measurement is much simpler than the
5964 % coloring, we choose the first way. Thus the macro |\adl@act@CT@setup|,
5965 % which is |\let|-equal to |\CT@setup| and is invoked from |\@classz|
5966 % preamble after the column is packed into |\box0|, measures the height and
5967 % depth of |\box0| and sets |\adl@height| and/or |\adl@depth| to them if
5968 % they break the records as |\adl@@colhtdp| does with |\adl@box|, after it
5969 % invokes its original version |\adl@CT@setup|. Note that we compare
5970 % |\adl@height| with the height of |\box0| plus |\minrowclearance| because
5971 % it is the real height. Also note that we could insert the measurement
5972 % code into the modified version of \textsf{colortbls}'s |\@classz| placing
5973 % it just before the |\box0| is put where |\ht0| plus |\minrowclearance| is
5974 % caluculated, but did not because the author wished to make it clear that
5975 % |\@classz| is modified only for the bug fix of |\arrayrulecolor| and
5976 % |\doublerulesepcolor| (and to introduce |\dashgapcolor|).
5977 % \end{macro}\end{macro}\end{macro}
5978 %
5979 % \begin{macrocode}
5980 \let\adl@CT@setup\CT@setup
5981 \def\CT@setup{\adl@CT@setup
5982 \@tempdima\ht\z@ \advance\@tempdima\minrowclearance
5983 \ifdim\adl@height<\@tempdima \global\adl@height\@tempdima \fi
5984 \ifdim\adl@depth<\dp\z@ \global\adl@depth\dp\z@\fi}
5985 \let\adl@act@CT@setup\CT@setup
5986
5987 % \end{macrocode}
5988 %
5989 % \begin{macro}{\adl@activatepbox}
5990 % \SpecialIndex{\CT@cell@color}
5991 % \SpecialIndex{\CT@row@color}
5992 % \SpecialIndex{\CT@column@color}
5993 % \SpecialIndex{\CT@do@color}
5994 % \SpecialIndex{\multicolumn}
5995 % Another job for cell coloring is to make |\CT@|$x$|@color|
5996 % ($x\in\{\texttt{cell},\texttt{column},\texttt{do}\}$)
5997 % |\let|-equal to |\relax| before the body of |\multicolumn| is put so that
5998 % the |\columncolor| in the environment preamble does not affect the
5999 % |\span|-ned column. Note that resetting |\CT@cell@color| will be
6000 % unnecessary (but safe) because it is always reset after its invocation.
6001 % Also note that resetting |\CT@row@color| in \textsf{colortbl}'s
6002 % |\multicolumn| is a buggy feature because it should be effective, and thus
6003 % we remove it. Although we have our own |\multicolumn| for dash lines, we
6004 % keep it unchanged. Instead we redefine |\adl@activatepbox|, which is
6005 % usually |\relax| with \textsf{array}, to do the color resetting to
6006 % minimize recoding.
6007 % \end{macro}
6008 %
6009 % \begin{macrocode}
6010 \def\adl@activatepbox{\let\CT@cell@color\relax
6011 \let\CT@column@color\relax
6012 \let\CT@do@color\relax}
6013
6014 % \end{macrocode}
6015 %
6016 % \begin{macro}{\adl@CT@start}
6017 % \begin{macro}{\CT@start}
6018 % \begin{macro}{\adl@dashgapcolor@save}
6019 % \begin{macro}{\adl@CT@end}
6020 % \begin{macro}{\CT@end}
6021 % \begin{macro}{\endarray}
6022 % \changes{v1.74}{2018/09/26}
6023 % {Add conditional invokation of \cs{@arrayright}.}
6024 % \begin{macro}{\endArray}
6025 % Yet another job is the save\slash restore of color information at the
6026 % beginning and end of the environment. Since this is done by |\CT@start| and
6027 % |\CT@end|, we modify them to save\slash restore |\adl@dashgapcolor|
6028 % to/from |\adl@dashgapcolor@save| referring their original version
6029 % |\adl@CT@start| and |\adl@CT@end|. We also modify our own |\endarray| and
6030 % its shorthand active version |\endArray| so that |\CT@end| is invoked at
6031 % the end of environment together with |\@arrayright| if it is defined.
6032 % Note that we may not modify |\endtabular| because it refers |\endarray|.
6033 % Also note that |\CT@start| is invoked from |\@tabarray| which we keep
6034 % unchanged.
6035 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
6036 %
6037 % \begin{macrocode}
6038 \let\adl@CT@start\CT@start
6039 \def\CT@start{\adl@CT@start \let\adl@dashgapcolor@save\adl@dashgapcolor}
6040 \let\adl@CT@end\CT@end
6041 \def\CT@end{\adl@CT@end \global\let\adl@dashgapcolor\adl@dashgapcolor@save}
6042 \def\endarray{\adl@endarray \egroup \adl@arrayrestore \CT@end \egroup
6043 \csname @arrayright\endcsname}
6044 \ifx\adl@notdefinable\undefined \let\endArray\endarray \fi
6045
6046 % \end{macrocode}
6047 %
6048 % \subsubsection{Horizontal Line Coloring}
6049 % \label{sec:impl-colortbl-hline}
6050 %
6051 % \begin{macro}{\hline}
6052 % \begin{macro}{\adl@inactivehdl}
6053 % \begin{macro}{\adl@ixhline}
6054 % \SpecialIndex{\CT@arc@}
6055 % \SpecialIndex{\CT@drsc@}
6056 % \SpecialIndex{\cline}
6057 % To color |\hline| and inactivated |\hdashline|, we modify our own |\hline|
6058 % and |\adl@inactivehdl| inserting the line coloring macro |\CT@arc@| before
6059 % drawing by |\hrule| and pushing the coloring\slash drawing into a group.
6060 % We also modify |\adl@ixhline| to draw a colored horizontal rule of
6061 % |\doublerulesep| wide with the color defined in |\CT@drsc@| if it is not
6062 % |\relax|, rather than to insert a vertical skip. Note that the |\cline|
6063 % coloring is done by \textsf{colortbl}'s |\cline| renamed as
6064 % |\adl@org@cline| and invoked from our own one.
6065 % \end{macro}\end{macro}\end{macro}
6066 %
6067 % \begin{macrocode}
6068 \def\hline{\noalign{\ifnum0=`}\fi
6069 \ifadl@zwhrule \vskip-\arrayrulewidth
6070 \else \adl@hline\adl@connect\arrayrulewidth \fi
6071 {\CT@arc@ \hrule\@height\arrayrulewidth}%
6072 \global\adl@finaldepth\z@
6073 \futurelet\@tempa\adl@xhline}
6074 \def\adl@inactivehdl[#1/#2]{\ifadl@zwhrule \vskip-\arrayrulewidth \fi
6075 {\CT@arc@ \hrule\@height\arrayrulewidth}%
6076 \futurelet\@tempa\adl@xhline}
6077 \def\adl@ixhline{{\ifx\CT@drsc@\relax \vskip \else
6078 \CT@drsc@\hrule\@height \fi \doublerulesep}%
6079 \adl@hline\relax\doublerulesep}
6080 % \end{macrocode}
6081 %
6082 % \begin{macro}{\adl@ihdashline}
6083 % \begin{macro}{\adl@act@ihdashline}
6084 % \begin{macro}{\adl@cdline}
6085 % \begin{macro}{\adl@act@cdline}
6086 % To draw a horizontal dash line with colored dashes and also colored gaps,
6087 % we drastically modified |\adl@ihdashline| for |\hdashline| and
6088 % |\adl@cdline| for |\cdashline|. First, they invoke |\adl@hclinesetup|
6089 % that makes the prefix of a |\multispan|-ned row from the first to last
6090 % columns for |\hdashline| or given columns for |\cdashline|. Then the line
6091 % is drawn by the modified version of |\adl@hcline|. We have to declare
6092 % these macros are active ones again.
6093 % \end{macro}\end{macro}\end{macro}\end{macro}
6094 %
6095 % \begin{macrocode}
6096 \def\adl@ihdashline[#1/#2]{\adl@hclinesetup\@ne\adl@columns
6097 \adl@hcline\z@[#1/#2]%
6098 \noalign{\ifnum0=`}\fi
6099 \futurelet\@tempa\adl@xhline}
6100 \let\adl@act@ihdashline\adl@ihdashline
6101 \def\adl@cdline[#1-#2]{\ifadl@zwhrule \vskip-\arrayrulewidth \fi
6102 \adl@hclinesetup{#1}{#2}%
6103 \adl@hcline{-\arrayrulewidth}}
6104 \let\adl@act@cdline\adl@cdline
6105 % \end{macrocode}
6106 %
6107 % \begin{macro}{\adl@hclinesetup}
6108 % \begin{macro}{\adl@cdlinea}
6109 % \begin{macro}{\adl@cdlineb}
6110 % The macro |\adl@hclinesetup|\meta{f}\meta{t} makes the prefix of a
6111 % |\multispan|-ned row from the column $f$ to $t$ and |\global|-ly defines
6112 % it as |\@gtempa|. This is done by a code very similar to original
6113 % |\adl@cdline| (and thus \LaTeX-2.09's |\cline|) but the invocation of
6114 % |\adl@hcline| is removed from |\adl@cdliena| and |\adl@cdlineb|, one of
6115 % which is |\@gtempa|.
6116 % \end{macro}\end{macro}\end{macro}
6117 %
6118 % \begin{macrocode}
6119 \def\adl@hclinesetup#1#2{\global\adl@cla#1\relax
6120 \global\advance\adl@cla\m@ne
6121 \ifnum\adl@cla>\z@ \global\let\@gtempa\adl@cdlinea
6122 \else \global\let\@gtempa\adl@cdlineb\fi
6123 \global\adl@clb#2\relax
6124 \global\advance\adl@clb-\adl@cla \ifnum0=`{\fi}}
6125 \def\adl@cdlinea{\multispan\adl@cla &\multispan\adl@clb \unskip}
6126 \def\adl@cdlineb{\multispan\adl@clb \unskip}
6127 % \end{macrocode}
6128 %
6129 % \begin{macro}{\adl@hcline}
6130 % \begin{macro}{\adl@paintdashgap}
6131 % \SpecialIndex{\adl@draw}
6132 % The modified version of |\adl@hcline|\meta{w}|[|\meta{d}|/|\meta{g}|]|
6133 % draws a colored horizontal dash line of dash size $d$ and gap size $g$ and
6134 % insert vertical skip of $w$. First it |\span|-s columns by |\@gtempa| and
6135 % checks if the body of |\adl@dashgapcolor| is something other than
6136 % |\relax|. If so, i.e.\ it has |\color|,
6137 % |\adl@paintdashgap| is invoked to
6138 % draw a horizontal rule of |\color| by |\leaders| as the background
6139 % of the dash line, to insert |\nobreak| (for \textsf{longtable}) and a
6140 % negative space for canceling the width of the
6141 % rule, and to |\span| the columns again. Then |\adl@hcline| draws the
6142 % colored dash line, over the background if the gaps are colored, by
6143 % inserting |\CT@arc@| before the invocation of |\adl@draw|.
6144 % \end{macro}\end{macro}
6145 %
6146 % \begin{macrocode}
6147 \def\adl@hcline#1[#2/#3]{\@gtempa
6148 \ifx\adl@dashgapcolor\adl@nocolor \else \adl@paintdashgap \fi
6149 {\@tempdima#2\relax \@tempdimb#3\relax
6150 \CT@arc@ \adl@draw\adl@vrule\hskip\hbox}\cr
6151 \noalign{\global\adl@finaldepth\z@ \ifdim#1=\z@\else
6152 \ifadl@zwhrule\else \vskip#1\fi\fi}}
6153 \def\adl@paintdashgap{{\adl@dashgapcolor
6154 \leaders\hrule\@height\arrayrulewidth\hfill}\cr
6155 \noalign{\penalty\@M \vskip-\arrayrulewidth}\@gtempa}
6156
6157 % \end{macrocode}
6158 %
6159 % \subsubsection{Vertical Line Coloring}
6160 % \label{sec:impl-colortbl-vline}
6161 %
6162 % \begin{macro}{\arrayrulecolor}
6163 % \begin{macro}{\CT@arc@}
6164 % \begin{macro}{\doublerulesepcolor}
6165 % \begin{macro}{\CT@drsc@}
6166 % \begin{macro}{\dashgapcolor}
6167 % \begin{macro}{\adl@dashgapcolor}
6168 % \begin{macro}{\adl@defcolor}
6169 % \begin{macro}{\adl@idefcolor}
6170 % \begin{macro}{\adl@noalign}
6171 % \begin{macro}{\nodashgapcolor}
6172 % \SpecialIndex{\@array}
6173 % A bug of \textsf{colortbl}'s |\arrayrulecolor| and |\doublerulesepcolor|
6174 % is that they are defined like;
6175 %
6176 % \begin{quote}
6177 % |\ifdim\baselineskip=\z@ \noalign \fi{\gdef\CT@arc@{\color...}}|
6178 % \end{quote}
6179 %
6180 % This aims to do |\noalign{\gdef...}| in \textsf{array}\slash
6181 % \textsf{tabular} and do |{\gdef...}| outside but has two problems: First,
6182 % if they are in |>{...}| construct, they are expanded with |\noalign|
6183 % inappropriately when the argument of |>| is expanded. Second, they may
6184 % appear at a place where |\baselineskip| is 0 but is outside of
6185 % \textsf{array}\slash\textsf{tabular} and will cause the misplaced
6186 % |\noalign| error. To solve the second problem, we introduced
6187 % |\adl@noalign| which is set to |\noalign| in the environment by our own
6188 % |\@array|, and |\relax| outside. We also introduced
6189 % |\adl@defcolor|\meta{cs}\meta{opt} for the common job to define \meta{cs}
6190 % as |\color| with \meta{opt}, in |\noalign| if necessary, by
6191 % |\adl@idefcolor|. Thus |\arrayrulecolor| and |\doublerulesepcolor| are
6192 % modified to define |\CT@arc@| and |\CT@drsc@| using |\adl@defcolor|, and
6193 % our own |\dashgapcolor| is defined
6194 % similarly to define |\adl@dashgapcolor|. Another macro |\nodashgapcolor|
6195 % to nullify |\dashgapcolor| is also defined with |\adl@noalign| to reset
6196 % |\adl@dashgapcolor| to |\relax|.
6197 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
6198 % \end{macro}\end{macro}\end{macro}\end{macro}\end{macro}
6199 %
6200 % \begin{macrocode}
6201 \def\arrayrulecolor{\adl@defcolor\CT@arc@}
6202 \def\doublerulesepcolor{\adl@defcolor\CT@drsc@}
6203 \def\dashgapcolor{\adl@defcolor\adl@dashgapcolor}
6204 \def\adl@defcolor#1#2#{\adl@idefcolor{#1}{#2}}
6205 \def\adl@idefcolor#1#2#3{\adl@noalign{\gdef#1{\color#2{#3}}}}
6206 \let\adl@noalign\relax
6207 \def\nodashgapcolor{\adl@noalign{\gdef\adl@dashgapcolor{\relax}}}
6208
6209 % \end{macrocode}
6210 %
6211 % \begin{macro}{\@classz}
6212 % \begin{macro}{\adl@act@classz}
6213 % \begin{macro}{\adl@org@classz}
6214 % The tougher bug of \textsf{colortbl} is the expansion timing of
6215 % |\arrayrulecolor| and |\dobule#rule#sep#color| in a |>|-argument. We
6216 % have to modify |\@classz| to extract them from
6217 % |\toks|\break|\@tempcnta| as its
6218 % original version does for |\columncolor|. Thus we inserted the invocation
6219 % of |\adl@extract@arc| for |\arrayrulecolor|, |\adl@extract@drsc| for
6220 % |\doublerulesep#color|, and |\adl@extract@dgc| for |\dashgapcolor| just
6221 % after the invocation of |\CT@extract|. Note that the other part of
6222 % |\@classz| is not modified logically, but done for author's preference of
6223 % indentation. Also note that both |\adl@act@classz| and |\adl@org@classz|
6224 % are |\let|-equal to the modified |\@classz| because we have to be bug free
6225 % even if |\ADLinactive| is in effect.
6226 % \end{macro}\end{macro}\end{macro}
6227 %
6228 % \begin{macrocode}
6229 \def\@classz{\@classx
6230 \@tempcnta\count@ \prepnext@tok
6231 \expandafter\CT@extract\the\toks\@tempcnta\columncolor!\@nil
6232 \expandafter\adl@extract@arc\the\toks\@tempcnta\arrayrulecolor!\@nil
6233 \expandafter\adl@extract@drsc
6234 \the\toks\@tempcnta\doublerulesepcolor!\@nil
6235 \expandafter\adl@extract@dgc\the\toks\@tempcnta\dashgapcolor!\@nil
6236 \@addtopreamble{%
6237 \setbox\z@\hbox\bgroup\bgroup
6238 \ifcase \@chnum
6239 \hskip\stretch{.5}\kern\z@
6240 \d@llarbegin
6241 \insert@column
6242 \d@llarend\hskip\stretch{.5}%
6243 \or \d@llarbegin \insert@column \d@llarend \hfill
6244 \or \hfill \kern\z@ \d@llarbegin \insert@column \d@llarend
6245 \or $\vcenter
6246 \@startpbox{\@nextchar}\insert@column \@endpbox $%
6247 \or \vtop \@startpbox{\@nextchar}\insert@column \@endpbox
6248 \or \vbox \@startpbox{\@nextchar}\insert@column \@endpbox
6249 \fi
6250 \egroup\egroup
6251 \begingroup
6252 \CT@setup
6253 \CT@column@color
6254 \CT@row@color
6255 \CT@cell@color
6256 \CT@do@color
6257 \endgroup
6258 \@tempdima\ht\z@
6259 \advance\@tempdima\minrowclearance
6260 \vrule\@height\@tempdima\@width\z@
6261 \unhbox\z@}%
6262 \prepnext@tok}
6263 \let\adl@act@classz\@classz
6264 \let\adl@org@classz\@classz
6265
6266 % \end{macrocode}
6267 %
6268 % \begin{macro}{\adl@def@extract}
6269 % \begin{macro}{\adl@extract@arc}
6270 % \begin{macro}{\adl@extract@arc@b}
6271 % \begin{macro}{\CT@arc@}
6272 % \begin{macro}{\adl@extract@drsc}
6273 % \begin{macro}{\adl@extract@drsc@b}
6274 % \begin{macro}{\CT@drsc@}
6275 % \begin{macro}{\adl@extract@dgc}
6276 % \begin{macro}{\adl@extract@dgc@b}
6277 % \begin{macro}{\adl@dashgapcolor}
6278 % The definitions of |\adl@extract@|$x$ ($x\in\{\texttt{arc}, \texttt{drsc},
6279 % \texttt{dgc}\}$) are quite similar to each other. For example
6280 % |\adl@extract@arc| is defined as follows.
6281 %
6282 %\begin{verbatim}
6283 %\def\adl@extract@arc#1\arrayrulecolor#2#3\@nil{%
6284 % \if!#2\toks\@tempcnta{#1}\let\@tempa\relax%
6285 % \else\if[#2%]
6286 % \def\@tempa{\adl@extract@arc@b{#1}#3\@nil}%
6287 % \else \def\CT@arc@{\color{#2}}%
6288 % \def\@tempa{\adl@extract@arc#1#3\@nil}%
6289 % \fi\fi \@tempa}
6290 %\def\adl@extract@arc@b#1#2]#3{%
6291 % \def\CT@arc@{\color[#2]{#3}}%
6292 % \adl@extract@arc#1}
6293 %\end{verbatim}
6294 %
6295 % This code extracts {\em all the} occurrences of
6296 % |\arrayrulecolor[|\meta{m}|]|\Meta{c} from the token register and
6297 % |\def|-ines |\CT@arc@| as |\color[|\meta{m}|]|\Meta{c}. Note that
6298 % |\CT@extract| does a similar job for |\columncolor| but it mistakingly
6299 % ignores the possibility that the token register has two or more
6300 % |\columncolor|\footnote{
6301 %
6302 % Fixing this bug is not our business.}.
6303 %
6304 % Anyway, if we copy the code above and replace
6305 % `|@arc|' with `|@drsc|', |\arrayrulecolor| with |\doublerulesepcolor|, and
6306 % |\CT@arc@| with \hbox{|\CT@drsc@|}, we will have |\adl@extract@drsc|(|@b|)
6307 % for
6308 % |\doublerulesepcolor|. The code for |\adl@extract@dgc|(|@b|) will be also
6309 % obtained similarly. However, having three relatives for a almost common
6310 % job is too awful. Thus we introduce;
6311 %
6312 % \begin{quote}
6313 % |\adl@def@extract|\meta{key}\meta{umac}\meta{cmac}
6314 % \end{quote}
6315 %
6316 % to define the macros |\adl@extract@|\textit{key} and
6317 % |\adl@extract@|\textit{key}|@b| for the user interface macro \meta{umac}
6318 % in which a color macro \meta{cmac} is defined with |\color|. For example,
6319 % we will obtain |\adl@extract@arc|(|@b|) shown above by;
6320 %
6321 % \begin{quote}
6322 % |\adl@def@extract{arc}\arrayrulecolor\CT@arc@|
6323 % \end{quote}
6324 %
6325 % Note that |\color| is made |\relax| in the preamble construction phase
6326 % by \textsf{colortbl}'s \break |\@mkpream| and regain its proper meaning
6327 % after the phase.
6328 % \end{macro}\end{macro}\end{macro}\end{macro}
6329 % \end{macro}\end{macro}\end{macro}
6330 % \end{macro}\end{macro}\end{macro}
6331 %
6332 % \begin{macrocode}
6333 \def\adl@def@extract#1#2#3{%
6334 \expandafter\def\csname adl@extract@#1\endcsname##1#2##2##3\@nil{%
6335 \if!##2\toks\@tempcnta{##1}\let\@tempa\relax
6336 \else\if[##2%]
6337 \def\@tempa{\@nameuse{adl@extract@#1@b}{##1}##3\@nil}%
6338 \else \def#3{\color{##2}}%
6339 \def\@tempa{\@nameuse{adl@extract@#1}##1##3\@nil}%
6340 \fi\fi \@tempa}
6341 \expandafter\def\csname adl@extract@#1@b\endcsname##1##2]##3{%
6342 \def#3{\color[##2]{##3}}%
6343 \@nameuse{adl@extract@#1}##1}}
6344 \adl@def@extract{arc}\arrayrulecolor\CT@arc@
6345 \adl@def@extract{drsc}\doublerulesepcolor\CT@drsc@
6346 \adl@def@extract{dgc}\dashgapcolor\adl@dashgapcolor
6347
6348 % \end{macrocode}
6349 %
6350 %
6351 %
6352 % \subsubsection{Compatibility with \textsf{longtable}}
6353 % \label{sec:impl-colortbl-longtable}
6354 %
6355 % \begin{macro}{\LT@hline}
6356 % \begin{macro}{\adl@LTihdashline}
6357 % \begin{macro}{\adl@LTinactivehdl}
6358 % \begin{macro}{\adl@LTixhline}
6359 % Yet another compatiblity issue is to cope with both \textsf{longtable} and
6360 % \textsf{colortbl}. We redefine |\LT@hline| and |\LT@inactivehdl| in order
6361 % to put |\CT@arc@| before line drawing and to push them in a group.
6362 % Modified |\adl@LTidashline| first
6363 % invokes |\adl@hclinesetup| and open |\noalign| because it is closed by
6364 % |\adl@hclinesetup|. The contents of |\adl@LThdlrow| for
6365 % |\adl@LTidashline| is simply |\adl@hcline| because it does |\multispan|
6366 % now. The macro |\adl@LTixhline| is modified to paint the |\doublerulesep|
6367 % gap by |\leaders\hrule| with color of |\CT@drsc@| if it is not |\relax|.
6368 % \end{macro}\end{macro}\end{macro}\end{macro}
6369 %
6370 %
6371 % \begin{macrocode}
6372 \ifx\longtable\undefined\else
6373 \def\LT@hline{\noalign{\ifnum0=`}\fi
6374 \gdef\adl@LThdlrow{\multispan{\LT@cols}\unskip{\CT@arc@
6375 \leaders\hrule\@height\arrayrulewidth\hfill}\cr}%
6376 \adl@LThdline}
6377 \def\adl@LTihdashline[#1/#2]{\adl@hclinesetup\@ne\adl@columns
6378 \noalign{\ifnum0=`}\fi
6379 \gdef\adl@LThdlrow{\adl@hcline\z@[#1/#2]}%
6380 \adl@LThdline}
6381 \def\adl@LTinactivehdl[#1/#2]{%
6382 \gdef\adl@LThdlrow{\multispan{\LT@cols}\unskip{\CT@arc@
6383 \leaders\hrule\@height\arrayrulewidth\hfill}\cr}%
6384 \adl@LThdline}
6385 \def\adl@LTixhline{%
6386 \ifx\CT@drsc@\relax \gdef\adl@LThdlrow{\noalign{
6387 \penalty-\@medpenalty \vskip\doublerulesep}}
6388 \else \gdef\adl@LThdlrow{\noalign{\penalty\@M}%
6389 \multispan{\LT@cols}\unskip{\CT@drsc@
6390 \leaders\hrule\@height\doublerulesep\hfill}\cr}\fi
6391 \ifnum0=`{\fi}\adl@LThdlrow \noalign{\ifnum0=`}\fi
6392 \adl@hline\relax\doublerulesep \global\let\adl@LThdlrow\@empty}
6393 \fi
6394 \fi
6395 % \end{macrocode}
6396 %
6397 % \iffalse
6398 %</package>
6399 % \fi
6400 %
6401 %
6402 %
6403 % \IndexPrologue{\newpage\section*{Index}
6404 % Italicized number refers to the page where the specification and usage of
6405 % corresponding entry are described, while underlined is for the
6406 % implementation of the entry.
6407 %
6408 % To find a control sequence, remove prefixes \cs{@}, \cs{adl@} and
6409 % \cs{ifadl@} from its name if it has one of them.}
6410 % \Finale
6411 % \newpage
6412 % \def\EQ{=} \def\GT{>} \def\BAR{|} \def\NEQ{\neq}
6413 % \PrintChanges
6414 \endinput