Debian Codegraph @

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

raw
1 % \iffalse meta-comment
2 % Copyright (C) 2015--2026 by Edgar Olthof
3 %
4 % This file may be distributed and/or modified under the
5 % conditions of the LaTeX Project Public License, either
6 % version 1.3 of this license or (at your option) any later
7 % version. The latest version of this license is in
8 %
9 % http://www.latex-project.org/lppl.txt
10 %
11 % and version 1.3 or later is part of all distributions of
12 % LaTeX version 1999/12/01 or later.
13 %
14 % \fi
15 % \iffalse
16 %<package>\ProvidesPackage{fcolumn}
17 %<package> [2026/05/20 v1.4.5 Automatic Currency Package (EHTO)]
18 %<package>\NeedsTeXFormat{LaTeX2e}
19 %<package>\RequirePackage{array}[v2.6d]
20 %<*driver>
21 %\DocumentMetadata{lang=en,tagging=on} % Uncommenting this line is OK, except for
22 % the verbatim tables (they contain the percentage sign) and for the captions in
23 % normal tabulars (in longtable everything is fine).
24 \documentclass{ltxdoc}
25 \usepackage{dcolumn}
26 \usepackage{booktabs}
27 \usepackage[tableposition=top,aboveskip=0pt]{caption}
28 % The next lines are a fix to the caption package, as that package is incompatible with tagging.
29 %\RemoveFromHook{begindocument}[latex-lab-testphase-float]
30 %\makeatletter \ExplSyntaxOn
31 %\socket_new_plug:nnn{tagsupport/parbox/before}{caption}
32 % {\tagpdfparaOn %restart para tagging
33 % \tl_if_empty:NTF\@current@float@struct{\tag_struct_begin:n{tag=Caption,firstkid}}
34 % {\tag_struct_begin:n{tag=Caption,parent=\@current@float@struct,firstkid}}}
35 %\socket_new_plug:nnn{tagsupport/parbox/after}{caption}{\tag_struct_end:}
36 %\l@addto@macro\caption@beginex@hook{%
37 % \tagpdfparaOff %leavevmode in parbox should not start paragraph structure
38 % \AssignSocketPlug{tagsupport/parbox/before}{caption}
39 % \AssignSocketPlug{tagsupport/parbox/after}{caption}}
40 %\ExplSyntaxOff \makeatother % End fix
41 \DeclareCaptionStyle{normal}{labelfont={rm,bf,small},textfont={rm,small},
42 indention=0pt,labelsep=colon,justification=raggedright}
43 \captionsetup[table]{style=normal}
44 \captionsetup{singlelinecheck=false}
45 \usepackage{url}
46 \usepackage{eurosym}
47 \usepackage{color}
48 \usepackage[hang,symbol*,perpage]{footmisc}
49 \setlength\footnotemargin{10pt}
50 \usepackage{fcolumn}
51 \usepackage{longtable}
52 \EnableCrossrefs
53 \CodelineIndex
54 % The next definition of \LT@output is to fix a flaw in longtable.sty.
55 % It can be removed once the fix is implemented; expected completion
56 % date around end Q1, 2026, according to David Carlisle.
57 \ExplSyntaxOn \makeatletter \def\LT@output{%
58 \ifnum\outputpenalty <-\@Mi
59 \ifnum\outputpenalty > -\LT@end@pen
60 \LT@err{floats~ and~ marginpars~ not~ allowed~ in~ a~ longtable}\@ehc
61 \else
62 \setbox\z@\vbox{\unvbox\@cclv}%
63 \ifdim \ht\LT@lastfoot>\ht\LT@foot
64 \dimen@\pagegoal
65 \advance\dimen@\ht\LT@foot
66 \advance\dimen@-\ht\LT@lastfoot
67 \ifdim\dimen@<\ht\z@
68 \setbox\@cclv\vbox{\unvbox\z@\copy\LT@foot
69 % \vss % old line
70 \vskip 0pt plus \maxdimen minus \normalbaselineskip
71 }\@makecol
72 \@expl@@@mark@update@singlecol@structures@@
73 \@outputpage
74 \global\vsize\@colroom
75 \setbox\z@\vbox{\box\LT@head}%
76 \fi
77 \fi
78 \unvbox\z@\box\ifvoid\LT@lastfoot\LT@foot\else\LT@lastfoot\fi
79 \UseTaggingSocket{tbl/longtable/foot}
80 \fi
81 \else
82 \setbox\@cclv\vbox{\unvbox\@cclv\copy\LT@foot
83 % \vss % old line
84 \vskip 0pt plus \maxdimen minus \normalbaselineskip
85 }\UseTaggingSocket{tbl/longtable/foot}\@makecol
86 \@expl@@@mark@update@singlecol@structures@@
87 \@outputpage
88 \global\vsize\@colroom
89 \copy\LT@head\nobreak
90 \fi}
91 \makeatother \ExplSyntaxOff
92 \begin{document}
93 \DocInput{\jobname.dtx}
94 \end{document}
95 %</driver>
96 %\fi
97 %\RecordChanges
98 %\CheckSum{1278}
99 %\CharacterTable
100 % {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
101 % 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
102 % Digits \0\1\2\3\4\5\6\7\8\9
103 % Exclamation \! Double quote \" Hash (number) \#
104 % Dollar \$ Percent \% Ampersand \&
105 % Acute accent \' Left paren \( Right paren \)
106 % Asterisk \* Plus \+ Comma \,
107 % Minus \- Point \. Solidus \/
108 % Colon \: Semicolon \; Less than \<
109 % Equals \= Greater than \> Question mark \?
110 % Commercial at \@ Left bracket \[ Backslash \\
111 % Right bracket \] Circumflex \^ Underscore \_
112 % Grave accent \` Left brace \{ Vertical bar \|
113 % Right brace \} Tilde \~}
114 %
115 %\changes{v0.1}{2015/02/06}{First working version.}
116 %\changes{v1.0}{2015/03/07}{Three-argument version is working properly.}
117 %\changes{v1.1}{2015/09/14}{Automatic checking of column balance performed when number of
118 % F-columns is even (behaviour can be overridden). Empty entries are now recognised and
119 % correctly treated as such, except for the one ended by the double backslash. Not serious;
120 % workaround possible. Furthermore optimisation of code: minimised the number of private
121 % counts and resetting of column counter done in a nicer way.}
122 %\changes{v1.1.1}{2015/09/27}{Installation procedure changed from .ins-in-.dtx to separate
123 % .ins and .dtx after discussion with Karl Berry, as well as some minor code improvements.}
124 %\changes{v1.1.2}{2015/09/29}{Some inconsistencies between explanatory text and actual
125 % code removed.}
126 %\changes{v1.2}{2019/03/06}{Input parsing changed after comment from Frank Mittelbach.
127 % He (Frank) also gave various suggestions for improving robustness or user friendliness
128 % of this package. This version is only backwards compatible when zero decimal digits were
129 % and are specified as modifier.}
130 %\changes{v1.3}{2021/01/03}{Christian Hoff requested the possibility to provide extra formatting
131 % information to a column, e.g., colouring. That was implemented by having an extra parameter
132 % to the generic F-column. The current solution is not very robust, as font and/or size change
133 % in math environment are very tricky, but providing colour information works, see the example
134 % in the main text. This version is now compatible with package longtable and backwards
135 % compatible to fcolumn v1.2: it only adds functionality.}
136 %\changes{v1.4}{2021/11/21}{More robust version on the extra formatting information: alternative
137 % formatting after the comma (if any), instead of additional information. Corrected a few typos.}
138 %\changes{v1.4.1}{2022/02/01}{Definition needed for longtable only if that package is loaded.
139 % General update, streamlining code. Excess digits silently ignored when all zero.}
140 %\changes{v1.4.2}{2023/07/25}{Better documentation. Removed redundant endgroup/begingroup
141 % pairs and other small improvements. Unbalanced columns are reported by stating their
142 % difference. Implementation of option ``red''. No changes in user experience.}
143 %\changes{v1.4.3}{2026/01/10}{Attempt to make fcolumn tagging-aware. Included my name and
144 % e-mail in README (CTAN obligation). Changed makefile to produce a fcolumn.tgz that contains
145 % files in fcolumn/. Removed some typos.}
146 %\changes{v1.4.4}{2026/01/16}{Repair of v1.4.3, which was a complete disaster in terms of tagging,
147 % resulting in a broken package. Hooks will be used to introduce code in array.sty and longtable.sty
148 % instead of overwriting definitions once those Hooks are implemented there. Until that time
149 % fcolumn falls back to the original solution of redefining three macros in those packages.
150 % Changed makefile to produce fcolumn.zip instead of fcolumn.tgz. Changed README file: it
151 % had extension .md in v1.4.3 but was (and is) raw ASCII.}
152 %\changes{v1.4.5}{2026/05/20}{Intermediate version ensuring a smooth transition when going
153 % to the 2026-06-01 LaTeX kernel.}
154 %\GetFileInfo{fcolumn.sty}
155 %\DoNotIndex{\@@,\@@endpbox,\@@startpbox,\@@@@endpbox,\@@@@startpbox,\@acol,\@arstrut,\@arstrutbox}
156 %\DoNotIndex{\@classiv,\@classz,\@currentHref,\@depth,\cs,\@endpbox,\@firstampfalse,\@firstamptrue}
157 %\DoNotIndex{\@gobble,\@halignto,\@height,\@mkpream,\@ne,\@nextchar,\@preamble,\@sharp,\@startpbox}
158 %\DoNotIndex{\@tabacol,\@tabarray,\@tabclassiv,\@tabclassz,\@tempdima,\@tempswafalse,\@tempswatrue}
159 %\DoNotIndex{\@temptokena,\@tfor,\@whilesw,\@width,\advance,\arraystretch,\baselineskip,\bgroup}
160 %\DoNotIndex{\break,\count,\count@,\cr,\def,\divide,\do,\dp,\edef,\else,\egroup,\fi,\fill,\g,\global}
161 %\DoNotIndex{\hfil,\hline,\ht,\hsize,\if,\if@firstamp,\if@tempswa,\ifcase,\ifdim,\ifnum,\ifodd,\ifx}
162 %\DoNotIndex{\int,\kill,\lineskip,\LT@@hl,\LT@@@@hl,\LT@@tabarray,\LT@@@@tabarray,\LT@bchunk}
163 %\DoNotIndex{\LT@caption,\LTcaptype,\LTchunksize,\LT@endpbox,\LT@hline,\LT@kill,\LT@LL@FM@cr}
164 %\DoNotIndex{\LT@make@row,\LT@mcol,\LT@no@pgbk,\LT@nofcols,\LT@rows,\LT@setprevdepth,\LT@startpbox}
165 %\DoNotIndex{\LT@tabularcr,\LTleft,\LTright,\let,\loop,\mathsurround,\m@ne,\message,\multiply}
166 %\DoNotIndex{\newpage,\nopagebreak,\or,\pagebreak,\par,\prepnext@tok,\protect,\refstepcounter}
167 %\DoNotIndex{\relax,\repeat,\rlap,\setbox,\space,\stepcounter,\string,\strutbox,\tabcolsep}
168 %\DoNotIndex{\tabskip,\tabularnewline,\the,\theHLT@tables,\the@toks,\thr@@,\tbl,\tl,\tw@,\unhbox}
169 %\DoNotIndex{\vbox,\vcenter,\vrule,\vskip,\vtop,\wd,\xdef,\z@,\z@skip}
170 %\title{The \textsf{fcolumn} package\thanks{This file has version number~\fileversion,
171 %last revised~\filedate.}}
172 %\author{Edgar Olthof\\
173 %\texttt{edgar <dot> olthof <at> inter <dot> nl <dot> net}}
174 %\date{Printed \today}
175 %\maketitle
176 %\begin{abstract}
177 %\noindent In financial reports, text and currency amounts are regularly put in
178 %one table, e.g., a year balance or a profit-and-loss overview. This
179 %package provides the settings for automatically typesetting and checking
180 %such columns, including the sum line (preceded by a rule of the correct
181 %width), using the specifier |f|.
182 %\end{abstract}
183 %
184 %\section{Introduction}\label{intro}
185 %The package |fcolumn| provides the macros for an extra tabular specifier
186 %that makes creating financial tables easy. The column specifier~|f|
187 %itself is rather simple; it is the predefined version of a generic
188 %column~|F|. The generic version expects four arguments: 1)~grouping
189 %character of the integer part on output, 2)~decimal mark used on output,
190 %3)~compact additional information on input/output characteristics, and
191 %4)~anything, but primarily used for providing formatting information, see below.
192 %
193 %The f-column in the current version of the package is defined for the
194 %continental European standard: |\newcolumntype{f}{F.,{3,2}{}}|. This
195 %means that a number like 12345,67 will be typeset as $12{.}345{,}67$.
196 %People in the Anglo-Saxon world would rather code
197 %|\newcolumntype{f}{F,.{3,2}{}}| for the same input, yielding
198 %$12{,}345{.}67$ as output for the number given above. The default
199 %value for |#3| is |3,2|, indicating that grouping of the integer
200 %part is by three digits, that a comma is used in the \TeX-source to
201 %indicate the decimal mark, and that the decimal part consists of
202 %two digits. However, if in your country or company grouping is done
203 %with a thinspace every four digits, that the decimal mark in the
204 %source should be the character |p|, and there are three digits after
205 %the decimal mark---that happens to be a |\cdot|---, then simply specify
206 %|\newcolumntype{f}{F\,\cdot{4p3}{}}| in that case. The input could
207 %be 123456p78 then, yielding $12{\,}3456{\cdot}780$ as output.
208 %
209 %By default two digits are used for the decimal part, so if you really
210 %want no decimal digits (in that case of course also skipping the decimal
211 %mark) you have to explicitly specify~|x,0| for some~|x|. If you want no
212 %grouping character, specify~|0,x|.
213 %
214 %As the fourth parameter you can insert anything just before the typesetting
215 %of an amount in a column takes place. Its purpose is to add additional
216 %formatting information, e.g., |\color{blue}| to have the contents of a column
217 %coloured blue, but it can be misused, so use with care. And it can't do all!
218 %
219 %\floatsep=5pt plus 2pt minus 3pt
220 %\textfloatsep=5pt plus 2pt minus 3pt
221 %\renewcommand{\bottomfraction}{0.9}
222 %\renewcommand{\textfraction}{0.1}
223 %\begin{table}[htb]
224 %\caption{Example Table.} \label{tab:fcraw}
225 %\begin{tabular}{@{}lflf@{}}
226 %\multicolumn4{@{}c@{}}{\bfseries Balance sheet}\\
227 %\toprule
228 %properties & \leeg{31 dec 2014} & debts & \leeg{31 dec 2014}\\
229 %\midrule
230 %house & 200000 & equity capital & 50000 \\
231 %bank account & -603,23 & mortgage & 150000 \\
232 %savings & 28000 \\
233 %cash & 145,85 & profit & 27542,62 \\
234 %\sumline
235 %\bottomrule
236 %\end{tabular}
237 %\end{table}
238 %This package requires and loads the |array| package~\cite{array}. To
239 %show where and how the |F|-column is used, let's look at some typical
240 %financial information as shown in Table~\ref{tab:fcraw}
241 %and how this is entered in \LaTeX\ (Table~\ref{tab:fcrawverb}).
242 %\begin{table}[htb]
243 %\caption{Verbatim version of Example Table~\ref{tab:fcraw}.}
244 %\label{tab:fcrawverb} \topsep0pt
245 %\begin{verbatim}
246 %\begin{table}[htb]
247 %\caption{Example Table.} \label{tab:fcraw}
248 %\begin{tabular}{@{}lflf@{}}
249 %\multicolumn4{@{}c@{}}{\bfseries Balance sheet}\\
250 %\toprule
251 %properties & \leeg{31 dec 2014} & debts & \leeg{31 dec 2014}\\
252 %\midrule
253 %house & 200000 & equity capital & 50000 \\
254 %bank account & -603,23 & mortgage & 150000 \\
255 %savings & 28000 \\
256 %cash & 145,85 & profit & 27542,62 \\
257 %\sumline
258 %\bottomrule
259 %\end{tabular}
260 %\end{table}
261 %\end{verbatim}
262 %\end{table}
263 %All the work was done by the column specifier~|f| (for ``finance'').
264 %In this case it constructs the |\sumline|, typesets the numbers,
265 %calculates the totals, determines the widths of the sumrules, and checks
266 %whether the two columns are in balance; if not, the user is warned via a
267 %|\PackageWarning|. Of course for nice settings the |booktabs|
268 %package~\cite{booktabs} was used, but that is not the point here.
269 %
270 %This package is heavily inspired by the |dcolumn| package by
271 %David Carlisle~\cite{dcolumn}; some constructions are more or less
272 %copied from that package. From version~1.3 onwards it incorporates the idea
273 %of Christian Hoff of providing additional (formatting) information per column.
274 %A rather contrived example is given in Table~\ref{tab:fcformat}, combining colour and fonts.
275 %\newcolumntype{q}[1]{F.,{3,2}{#1}}
276 %\begin{table}[htb]
277 %\caption{Example Table with column formatting.} \label{tab:fcformat}
278 %\begin{tabular}{@{}lq{\color{blue}\mathsf,\mathbf}lq{\color{magenta}}@{}}
279 %\multicolumn4{@{}c@{}}{\bfseries Balance sheet}\\
280 %\toprule
281 %properties & \leeg{31 dec 2014} & debts & \leeg{31 dec 2014}\\
282 %\midrule
283 %house & 200000 & equity capital & 50000 \\
284 %bank account & -603,23 & mortgage & 150000 \\
285 %savings & 28000 \\
286 %cash & 145,85 & profit & 27542,62 \\
287 %\sumline
288 %\bottomrule
289 %\end{tabular}
290 %\end{table}
291 %How this is entered in \LaTeX\ is shown in Table~\ref{tab:fcformatverb}. The font changing commands
292 %like |\mathsf| and |\mathbf| act on an argument, hence require braces,
293 %but these are already provided internally for this purpose. For that reason this
294 %type of commands must be given last, without braces (and if you don't specify a font
295 %changing command, these extra internal braces are just redundant). The fourth argument to the
296 %\begin{table}[htb]
297 %\caption{Almost verbatim version of Example Table~\ref{tab:fcformat}.}
298 %\label{tab:fcformatverb} \topsep0pt
299 %\begin{verbatim}
300 %\newcolumntype{q}[1]{F.,{3,2}{#1}}
301 %\begin{table}[htb]
302 %\caption{Example Table with column formatting.} \label{tab:fcformat}
303 %\begin{tabular}{@{}lq{\color{blue}\mathsf,\mathbf}lq{\color{magenta}}@{}}
304 %\multicolumn4{@{}c@{}}{\bfseries Balance sheet}\\
305 %...
306 %... (Same financial contents as in Table 1.)
307 %...
308 %\end{tabular}
309 %\end{table}
310 %\end{verbatim}
311 %\end{table}
312 %new columntype may consist of two parts, separated by a comma. In that case, the part to
313 %the left of the comma is applied to the data entered by the user and the right part, if
314 %non-empty, is the replacement formatting for the result. The example in Table~\ref{tab:fcformat}
315 %shows this: the bold font is only used in the |\sumline| and there is no colour specification,
316 %so that's back to the default (black). If you want formatting for the whole column, like magenta
317 %in the case of the last column of Table~\ref{tab:fcformat}, leave out the comma.
318 %Font size changes, e.g., ``|\huge|'' in ``|\huge\mathbf|''
319 %as parameter to column type |q| are ignored by \LaTeX, since the formatting
320 %information is used in math environment, which has its own way of handling this.
321 %This isn't bad, as size changes in one column, without overall changes to the
322 %table look terrible. If you want something huge, make a |\huge| table.
323 %
324 %As will be demonstrated by Table~\ref{tab:fclong} in the next page break, the package
325 %|fcolumn| now also works with |longtable|~\cite{longtable}, irrespective of the order
326 %in which these packages are loaded. The raw formatting of the multipage table
327 %is shown in Table~\ref{tab:fclongverb}. While constructing this example with tagging,
328 %see section~\ref{tagging}, it turned out that |booktabs| caused problems that were
329 %solved by including option |firstaid|. For more information on how to handle |\endhead|
330 %and its ilk, see the documentation of |longtable|~\cite{longtable}. Not shown here are
331 %the new |fcolumn| formatting possibilities (like new fonts and/or colours), but it has
332 %been checked they do work in combination with |longtable|. The packages
333 %|supertabular|~\cite{supertabular} and |fcolumn| also work happily together, so you can
334 %choose which package you want if you need this. There is no race condition with
335 %|supertabular| either.
336 %\begin{longtable}[l]{@{}lflf@{}}
337 %\caption{Table showing compatibility of fcolumn and
338 % longtable.\label{tab:fclong}}\\
339 %\multicolumn4{@{}c@{}}{\bfseries Balance sheet}\\
340 %\toprule
341 %properties & \leeg{31 dec 2014} & debts & \leeg{31 dec 2014}\\
342 %\midrule
343 %\endfirsthead
344 %\caption[]{\textit{(continued from previous page)\/}}\\
345 %\toprule
346 %properties & \leeg{31 dec 2014} & debts & \leeg{31 dec 2014}\\
347 %\midrule
348 %\endhead
349 %\bottomrule
350 %\multicolumn4{@{}r@{}}{\small\textit{(Table continues on
351 % next page)\/}}\\
352 %\endfoot
353 %\bottomrule
354 %\endlastfoot
355 %house & 200000 & equity capital & 50000 \\
356 %bank account & -603,23 & mortgage & 150000 \\
357 %savings 1 & 5600 \\
358 %savings 2 & 5600 \\
359 %savings 3 & 5600 \\
360 %savings 4 & 5600 \\
361 %savings 5 & 5600 \\
362 %cash & 145,85 & profit & 27542,62 \\
363 %\sumline
364 %\end{longtable}
365 %
366 %\begin{table}[htb]
367 %\caption{Almost verbatim version of Example Table~\ref{tab:fclong}.}
368 %\label{tab:fclongverb} \topsep0pt
369 %\begin{verbatim}
370 %\begin{longtable}[l]{@{}lflf@{}}
371 %\caption{Table showing compatibility of fcolumn and
372 % longtable.\label{tab:fclong}}\\
373 %\multicolumn4{@{}c@{}}{\bfseries Balance sheet}\\
374 %\toprule
375 %properties & \leeg{31 dec 2014} & debts & \leeg{31 dec 2014}\\
376 %\midrule
377 %\endfirsthead
378 %\caption[]{\textit{(continued from previous page)\/}}\\
379 %\toprule
380 %properties & \leeg{31 dec 2014} & debts & \leeg{31 dec 2014}\\
381 %\midrule
382 %\endhead
383 %\bottomrule
384 %\multicolumn4{@{}r@{}}{\small\textit{(Table continues on
385 % next page)\/}}\\
386 %\endfoot
387 %\bottomrule
388 %\endlastfoot
389 %house & 200000 & equity capital & 50000 \\
390 %...
391 %... (Somewhat altered financial contents compared to
392 %... Table 1 to demonstrate the page break.)
393 %...
394 %\sumline
395 %\end{longtable}
396 %\end{verbatim}
397 %\end{table}
398 %Unlike in the package |dcolumn|~\cite{dcolumn}, the alignment on the decimal mark in
399 %|fcolumn| is achieved by right aligning numbers that all have the same number of
400 %digits to the right of this mark, in combination with the fact that all decimal glyphs
401 %have the same width\footnote{The so-called tabular numerals. If the font you use has
402 %proportional numerals, the alignment won't work in general.}. This width is font specific,
403 %so changing fonts within a column may ruin the alignment, see, e.g., the first |F|-column
404 %in Table~\ref{tab:fcformat}.
405 %\section{Commands}
406 %The user only needs to know five commands and two options; these are given here.
407 %\begin{macro}{F}
408 %In the tabular the column specifier |F| can be given with arguments, or
409 %the predefined version~|f|, where the four arguments of |F| are |.|,
410 %|,|, |{3,2}|, and |{}|. If you want |g| to be your own definition like the
411 %curious one given in Section~\ref{intro}, then specify
412 %|\newcolumntype{g}{F\,\cdot{4p3}{}}| prior to using |g| in a tabular.
413 %
414 %Entries in an |F|-column are, from that moment on, treated as numbers unless
415 %explicitly escaped by |\leeg|, see below. The numbers are typeset
416 %according to the template the user gives with his/her |F|-column.
417 %The ``middle'' character of |#3| is an important switch: it does more
418 %than just setting the input decimal mark. By default the input grouping
419 %character is the dot, except when the dot is specified as input decimal
420 %mark; in that case the comma is acting as input grouping character. With
421 %this convention continental Europe and the Anglo-Saxon part of the world are
422 %served. This is hard coded, but using input grouping characters is optional anyway.
423 %\end{macro}
424 %\begin{macro}{\sumline}
425 %The numbers in an |F|-column are typeset as a financial amount, but the real
426 %benefit comes with the |\sumline|. It does three things:
427 %\iffalse This used to be done with compactenum from paralist but that package
428 %is not compatible with tagging \fi
429 %
430 %\begingroup \hangindent=12pt \hangafter=1
431 %\noindent 1)~It calculates and shows the total of the column so far and the maximum
432 % width encountered so far, including the width of the total;
433 %
434 %\hangindent=12pt \hangafter=1
435 %\noindent 2)~It generates a rule with width calculated in the first item;
436 %
437 %\hangindent=12pt \hangafter=1
438 %\noindent 3)~It checks the columns that are supposed to balance whether or not
439 % they actually do. If so, nothing happens. Otherwise a |\PackageWarning| is
440 % given that column~$i$ and~$j$ do not balance, where $i$ and $j$ are the relevant
441 % columns. This is only done automatically if the total number of |F|-columns
442 % is even, e.g., if there are six |F|-columns, then~1 is checked against~4, 2
443 % against~5, and 3 against~6. If the number of |F|-columns is odd then anything
444 % could be possible in that table and nothing is assumed about structure within
445 % the table. This behaviour can be overridden, see below.
446 %
447 %\endgroup
448 %By default the vertical separation between the rule and the total is
449 %2\thinspace pt, but this can be changed by the optional argument to
450 %|\sumline|. Give, e.g., |\sumline[10pt]|, in case you want this spacing
451 %to be 10\thinspace pt. And you may even give two options, like in
452 %|\sumline[10pt][5pt]|, in which the second option is the extra space below
453 %the summary row. In fact that second option is parsed to |\\|, that is
454 %implicit in |\sumline|.
455 %\end{macro}
456 %\begin{macro}{\resetsumline}
457 %Suppose you want to typeset one tabular with the profit-and-loss of many
458 %projects individually. In case the layout of those tabulars is similar it
459 %were nice if all columns were aligned. This can be done by making it one
460 %big tabular with a fresh start for each project. The macro |\resetsumline|
461 %is used for that: it resets all totals and all column widths, see for example
462 %Table~\ref{tab:fcmult}.
463 %\begin{table}[htb]
464 %\caption{Example: multiple projects.} \label{tab:fcmult}
465 %\begin{tabular}{@{}lfflff@{}}
466 %\multicolumn6{@{}c@{}}{\bfseries Project~1}\\
467 %\toprule
468 %expense & \leeg{actual} & \leeg{budget} &
469 %income & \leeg{actual} & \leeg{budget} \\
470 %\midrule
471 %food & 450,20 & 500 & tickets & 1200 & 1000 \\
472 %drinks & 547,5 & 400 \\
473 %music & 180 & 100 \\
474 %profit & 22,3 \\
475 %\sumline[2pt][10pt]
476 %\resetsumline
477 %\multicolumn6{@{}c@{}}{\bfseries Project~2}\\
478 %\toprule
479 %expense & \leeg{actual} & \leeg{budget} &
480 %income & \leeg{actual} & \leeg{budget} \\
481 %\midrule
482 %food & 250 & 300 & tickets & 400 & 450 \\
483 %drinks & 100 & 80 \\
484 %music & 80 & 70 & loss & 30 \\
485 %\sumline
486 %\bottomrule
487 %\end{tabular}
488 %\end{table}
489 %Note that the rules in the first and third |F|-columns of Project~1 cover
490 %$1{.}200{,}00$ whereas in Project~2 those rules are narrower since they
491 %only cover $430{,}00$; still the columns are aligned (similar story for
492 %|F|-columns two and four). The verbatim way of setting up Table~\ref{tab:fcmult}
493 %is given in Table~\ref{tab:fcmultverb}.
494 %\begin{table}[htb]
495 %\caption{Verbatim version of Table~\ref{tab:fcmult}.}
496 %\label{tab:fcmultverb} \topsep0pt
497 %\begin{verbatim}
498 %\begin{table}[htb]
499 %\caption{Example: multiple projects.} \label{tab:fcmult}
500 %\begin{tabular}{@{}lfflff@{}}
501 %\multicolumn6{@{}c@{}}{\bfseries Project~1}\\
502 %\toprule
503 %expense & \leeg{actual} & \leeg{budget} &
504 %income & \leeg{actual} & \leeg{budget} \\
505 %\midrule
506 %food & 450,20 & 500 & tickets & 1200 & 1000 \\
507 %drinks & 547,5 & 400 \\
508 %music & 180 & 100 \\
509 %profit & 22,3 \\
510 %\sumline[2pt][10pt]
511 %\resetsumline
512 %\multicolumn6{@{}c@{}}{\bfseries Project~2}\\
513 %\toprule
514 %expense & \leeg{actual} & \leeg{budget} &
515 %income & \leeg{actual} & \leeg{budget} \\
516 %\midrule
517 %food & 250 & 300 & tickets & 400 & 450 \\
518 %drinks & 100 & 80 \\
519 %music & 80 & 70 & loss & 30 \\
520 %\sumline
521 %\bottomrule
522 %\end{tabular}
523 %\end{table}
524 %\end{verbatim}
525 %\end{table}
526 %\end{macro}
527 %\begin{macro}{\leeg}
528 %If an |F|-column should be empty then simply leave it empty. If, however, it is
529 %not empty but the entry should be treated as text---even if it is a number---,
530 %this can be done with |\leeg|. It expects an argument and this argument is
531 %typeset in the column. The common case is where |p.m.| (\textit{pro memoria})
532 %is entered or in column headers.
533 %\end{macro}
534 %\begin{macro}{\checkfcolumns}
535 %The automatic column balance check can also be done manually. If
536 %|F|-columns~1 and~4 should balance and you want them to be checked,
537 %then simply say |\checkfcolumns14|. With more than nine |F|-columns
538 %you may be forced to say something like |\checkfcolumns{10}{12}|. If
539 %|\checkfcolumns| is used, the automatic check is disabled. Multiple
540 %|\checkfcolumns|s are supported; if |F|-columns~1, 2, and~3 should
541 %balance, you specify |\checkfcolumns12| and |\checkfcolumns23|.
542 %There is no explicit command to disable all checking, but
543 %|\checkfcolumns11| obviously does the trick, at the expense of
544 %some trivial calculations.
545 %\end{macro}
546 %\begin{macro}{\ifstrict@ccounting}
547 %\begin{macro}{\ifminusr@d}
548 %In the rare occasion that a negative number occurs in a financial table,
549 %the sign of that number can be an explicit minus sign, i.e.,~`$-$' or the
550 %number is coloured red, or it is typeset between parentheses, and there
551 %may be even other ways. By default (for aesthetic reasons) |fcolumn|
552 %typesets it with a minus~sign, but strict accounting prescribes that the
553 %number should be put between parentheses. The latter can be accomplished
554 %by setting |\strict@ccountingtrue|, but since this contains a non-letter,
555 %it is more \LaTeX-like to invoke |fcolumn| with the option |strict|, i.e.,
556 %|\usepackage[strict]{fcolumn}|, which sets this flag. If you want the
557 %negative numbers to be coloured red, use the option |red|, i.e.,
558 %|\usepackage[red]{fcolumn}|, but note that |strict| and |red|
559 %are mutually exclusive: if both are used, |red| is reset because
560 %|strict| is strict.
561 %\end{macro}
562 %\end{macro}
563 %\section{Tagging}\label{tagging}
564 %Additional code for making tagged PDF~\cite{tagging} is available by
565 %generating the style file with option ``tagging''. If you don't want
566 %this or if your \TeX-distribution doesn't support tagging (yet), it
567 %is possible to exclude these lines. This is still highly experimental,
568 %and to properly process the document you are reading now you can't
569 %use tagging by having |\DocumentMetadata| as first line, because |ltxdoc|
570 %doesn't work well with tagging (a next version will be done with |l3doc|;
571 %it has been checked that that works fine). For ``normal'' documents,
572 %e.g., using |\documentclass{article}| you can start with
573 %|\DocumentMetadata{tagging=on}| as first line, see the example file
574 %fcexample.tex, that is also included in the distribution.
575 %In this version the math-tagging is switched off again in
576 %the |F|-columns of a |tabular| environment, since the math there is fake math.
577 %\StopEventually{\PrintChanges\PrintIndex}
578 %\section{The macros}
579 %Here follows the actual code; most users will stop reading here, but
580 %some examples that may be interesting are presented below. Previous
581 %versions can be loaded for not-up-to-date systems, but v1.4.3 is lacking
582 %here, because it had a severe bug.
583 % \begin{macrocode}
584 \DeclareRelease{v1.4.2}{2023-07-25}{fcolumn-2023-07-25.sty}
585 \DeclareRelease{v1.4.4}{2026-01-26}{fcolumn-2026-01-16.sty}
586 \DeclareCurrentRelease{v1.4.5}{2026-05-20}
587 \NeedsTeXFormat{LaTeX2e}[2025/06/01]
588 % \end{macrocode}
589 %\subsection{Options}
590 %\begin{macro}{option red}
591 %\begin{macro}{option strict}
592 %There are two options. If |red| is set, negative numbers are displayed in red.
593 %If |strict| is set, strict accounting rules are used in display and this overrules
594 %a possibly set |red|. Depending on the two booleans associated with these options
595 %the |color| package~\cite{color} will be loaded or not.
596 % \begin{macrocode}
597 \newif\ifminusr@d \minusr@dfalse
598 \newif\ifstrict@ccounting \strict@ccountingfalse
599 \DeclareOption{red}{\minusr@dtrue}
600 \DeclareOption{strict}{\strict@ccountingtrue}
601 \ProcessOptions \ifminusr@d \ifstrict@ccounting \minusr@dfalse
602 \PackageWarningNoLine{fcolumn}{Option `red' is reset due to use
603 of `strict'}\else\usepackage{color}\fi\fi
604 % \end{macrocode}
605 %\end{macro}
606 %\end{macro}
607 %\subsection{Definitions}
608 %\begin{macro}{column F}
609 %\begin{macro}{column f}
610 %The column specifier |F| is the generic one, and |f| is the default
611 %(continental European) one for easy use. Note that the definition of the
612 %column type |f| does not use private macros (no~|@|), so overriding its
613 %definition is easy for a user.
614 % \begin{macrocode}
615 \newcolumntype{F}[4]{>{\b@fi{#1}{#2}{#3}{#4}}r<{\e@fi}}
616 \newcolumntype{f}{F.,{3,2}{}}
617 % \end{macrocode}
618 %\end{macro}
619 %\end{macro}
620 %\begin{macro}{\FCsc@l}
621 %\begin{macro}{\FCtc@l}
622 %Two \meta{count}s are defined, that both start at zero: the \meta{count}
623 %|\FCsc@l|, that keeps track at which |F|-column the tabular is working on
624 %and the \meta{count} |\FCtc@l|, that records the number of |F|-columns
625 %that were encountered so far in the document. Later in the package the
626 %code can be found for generating a new \meta{count} and a new \meta{dimen}
627 %if the number of requested |F|-columns is larger than currently available.
628 %This is of course the case when an |F|-column is used for the first time.
629 % \begin{macrocode}
630 \newcount\FCsc@l \FCsc@l=0 \newcount\FCtc@l \FCtc@l=0
631 % \end{macrocode}
632 %\end{macro}
633 %\end{macro}
634 %\begin{macro}{\geldm@cro}
635 %The macro |\geldm@cro| takes a number and by default interprets this as
636 %an amount expressed in cents (dollar cents, euro cents, centen, Pfennige,
637 %centimes, kopecks, groszy) and typesets it as the amount in entire
638 %currency units (dollars, euros, guldens, Marke, francs, rubles, z\l oty)
639 %with comma as decimal mark and the dot as grouping character
640 %(thousand separator if the first part of |#1| is~3).
641 %As explained, this can be changed. It uses three private booleans:
642 %|\ifwiths@p|, |\ifstrict@ccounting|, and |\ifminusr@d|. The latter two
643 %are used to typeset negative numbers in a special way. By default it
644 %doesn't do this: a minus sign is used.
645 % \begin{macrocode}
646 \newif\ifwiths@p
647 % \end{macrocode}
648 %Actually |\geldm@cro| is only a wrapper around |\g@ldm@cro|.
649 % \begin{macrocode}
650 \def\geldm@cro#1#2{\withs@pfalse
651 \afterassignment\g@ldm@cro\count@#1\relax{#2}}
652 % \end{macrocode}
653 %\end{macro}
654 %\begin{macro}{\g@ldm@cro}
655 %After setting the environment for formatting, this macro starts by looking
656 %at the sign of |#2|: if it is negative, it prints the correct indicator (a parenthesis,
657 %a minus sign, or nothing at all in case of |red|), assigns the absolute value
658 %of~|#2| to |\count2| and goes on. Note that |\geldm@cro| and therefore |\g@ldm@cro|
659 %are always used within |$|s, so it is really a minus sign that is printed, not a hyphen.
660 %All calculations are done with |\count0|, |\count1|, etc., i.e., without |F|-column-specific
661 %\meta{count}s because it is all done locally. Leaving the tabular environment will
662 %restore their values. This is also true for the effect of |\FCform@t|, so that
663 %formatting information is local to this column. The reason for inserting the |{|
664 %between |\FCform@t| and |\ifnum| (and the accompanying |}| just before finishing
665 %this definition) is to facilitate the possible use of |\mathbf| or any other font
666 %changing command as the last item in |\FCform@t|.
667 %
668 %When |\FCform@t| includes a colour change, this is overruled for negative numbers
669 %when option |red| is set. For that reason it is unwise to use a red colour for the
670 %whole column in combination with the option |red|: sign information is obscured then.
671 % \begin{macrocode}
672 \def\g@ldm@cro#1\relax#2{\ifstrict@ccounting\def\bm@nus{(}
673 \def\em@nus{)}\else\def\em@nus{}\ifminusr@d\def\bm@nus{\color{red}}
674 \else\def\bm@nus{-}\fi\fi\FCform@t{\ifnum#2<0 \bm@nus\count2=-#2
675 \else\count2=#2 \fi
676 % \end{macrocode}
677 %Calculate the entire currency units: this is the result of $x/a$ as integer
678 %division, with $a=10^n$ and $n$ the part of |#1| after the separator (if any).
679 %Here the first character of |#1| is discarded, so the separator in |#1| is not strict:
680 %you could also specify |3.2| instead of |3,2| (or even |3p2|).
681 % \begin{macrocode}
682 \count4=\ifx\relax#1\relax 2 \else\@gobble#1\relax\fi
683 \count3=0
684 \loop\ifnum\count3<\count4
685 \divide\count2 by 10 \advance\count3 by \@ne
686 \repeat
687 % \end{macrocode}
688 %Note that |\count3| now equals |\count4|: this going up-and-down will be used more often,
689 %it saves several assignments. The value in |\count2| is then output by |\g@ldens| using
690 %the separation given (and stored in |\count@|).
691 % \begin{macrocode}
692 \g@ldens{\the\count@}%
693 % \end{macrocode}
694 %If there is a decimal part\dots
695 % \begin{macrocode}
696 \ifnum\count3>0 {\decim@lmark}
697 % \end{macrocode}
698 %Next the decimal part is dealt with. Now $x\bmod a$ is calculated in the
699 %usual way: $x-(x/a)*a$ with integer division. The minus sign necessary
700 %for this calculation is introduced in the next line by changing the
701 %comparison from |<| to |>|.
702 % \begin{macrocode}
703 \ifnum#2>0 \count2=-#2\else\count2=#2 \fi
704 \loop\ifnum\count3>0
705 \divide\count2 by 10 \advance\count3 by \m@ne
706 \repeat
707 % \end{macrocode}
708 %The value of |\count3| is now~0, so counting up again.
709 % \begin{macrocode}
710 \loop\ifnum\count3<\count4
711 \multiply\count2 by 10 \advance\count3 by \@ne
712 \repeat
713 \ifnum#2>0 \advance\count2 by #2
714 \else \advance\count2 by -#2
715 \fi
716 \zerop@d{\number\count3}{\number\count2}%
717 \fi
718 % \end{macrocode}
719 %If the negative number is indicated by putting it between parentheses,
720 %then the closing parenthesis should stick out of the column, otherwise
721 %the alignment of this entry in the column is wrong. This is done by an
722 %|\rlap| and therefore does not influence the column width. For the rightmost
723 %column this means that this parenthesis may even stick out of the table.
724 %I don't like this, therefore I chose to put |\strict@ccountingfalse|.
725 %Change if you like, by setting the option |strict|.
726 %
727 %If overflow was detected, an exclamation mark is output to the right of
728 %the value that caused this. This of course ruins the appearance of the
729 %table, but in this case that serves a clear goal: there's something wrong
730 %and you should know.
731 % \begin{macrocode}
732 \ifx\FCs@gn\m@ne\ifnum#2<0 \rlap{\em@nus~!}
733 \else\rlap{\phantom{\em@nus}~!}\fi
734 \else\ifnum#2<0 \rlap{\em@nus}\fi\fi}}
735 % \end{macrocode}
736 % \end{macro}
737 %\begin{macro}{\g@ldens}
738 %Here the whole currency units are dealt with. The macro |\g@ldens| is
739 %used recursively, therefore the double braces; this allows to use
740 %|\count0| locally. This also implies that tail recursion is not
741 %possible here, but that is not very important, as the largest number
742 %(which is $2^{31}-1$) will only cause a threefold recursion using the
743 %default |3,2| (ninefold when using~|1,0|, but who does that?). The
744 %largest amount this package can deal with is therefore 2.147.483.647
745 %(using |3,0|). For most people this is probably more than enough if
746 %the currency is euros or dollars. And otherwise clearly state that you
747 %use a currency unit of~k\EUR{} (or even M\EUR{} for the very rich).
748 %
749 %There is no obvious interpretation of |#1| being zero or negative,
750 %therefore this is used as an indicator that no grouping character should be used.
751 % \begin{macrocode}
752 \def\g@ldens#1{{\count3=\count2 \count0=#1
753 % \end{macrocode}
754 %First divide by $10^n$, where $n$ is |#1|.
755 % \begin{macrocode}
756 \ifnum\count0<1 \count0=3 \fi
757 \loop \ifnum\count0>0 \divide\count2 by 10 \advance\count0 by \m@ne
758 \repeat
759 % \end{macrocode}
760 %Here is the recursive part,
761 % \begin{macrocode}
762 \ifnum\count2>0 \g@ldens{#1}\fi
763 % \end{macrocode}
764 %and then reconstruct the rest of the number.
765 % \begin{macrocode}
766 \count0=#1
767 \ifnum\count0<1 \count0=3 \fi
768 \loop \ifnum\count0>0 \multiply\count2 by 10 \advance\count0 by \m@ne
769 \repeat
770 \count2=-\count2
771 \advance\count2 by \count3 \du@zendprint{#1}}}
772 % \end{macrocode}
773 %\end{macro}
774 %\begin{macro}{\du@zendprint}
775 %The macro |\du@zendprint| takes care of correctly printing the separator
776 %and possible trailing zeros. The former, however, is only done if |#1|
777 %is larger than zero.
778 % \begin{macrocode}
779 \def\du@zendprint#1{\ifwiths@p\ifnum#1>0 {\sep@rator}\fi
780 \zerop@d{#1}{\number\count2}%
781 \else\zerop@d1{\number\count2}\fi\global\withs@ptrue}
782 % \end{macrocode}
783 %\end{macro}
784 %\begin{macro}{\zerop@d}
785 %The macro |\zerop@d| uses at least |#1| digits for printing the
786 %number |#2|, padding with zeros when necessary. Note: |#1| being
787 %zero or negative is a flag that it should be interpreted as 3.
788 %A bit ugly, but it works, since the related code knows about this.
789 %
790 %It is done within an extra pair of braces, so that |\count0|
791 %and |\count1| can be used without disturbing their values in other macros.
792 % \begin{macrocode}
793 \def\zerop@d#1#2{{\count0=1 \count1=#2
794 % \end{macrocode}
795 %First determine the number of digits of |#2| (expressed in the decimal
796 %system). This number is in |\count0| and is at least~1.
797 % \begin{macrocode}
798 \loop \divide \count1 by 10 \ifnum\count1>0 \advance\count0 by \@ne
799 \repeat
800 % \end{macrocode}
801 %If |#1| is positive, the number of zeros to be padded is the maximum
802 %of~0 and |#1-\count0| (the latter can be negative), so a
803 %simple loop suffices. If it is zero or negative, this is a signal
804 %that it should be interpreted as 3 (and no separator will be output).
805 % \begin{macrocode}
806 \ifnum#1>0 \loop\ifnum\count0<#1\relax 0\advance\count0 by \@ne
807 \repeat\else\advance\count0 by -3
808 \loop\ifnum\count0<0 0\advance\count0 by \@ne \repeat
809 \fi\number#2}}
810 % \end{macrocode}
811 %\end{macro}
812 %\begin{macro}{\zetg@ld}
813 %This macro takes care of several things: it increases the subtotal for a
814 %given |F|-column, it checks whether or not that subtotal has overflown, it
815 %records the largest width of the entries in that column and it typesets
816 %|#1| via |\geldm@cro|.
817 % \begin{macrocode}
818 \def\zetg@ld#1#2{\count0=#2\relax \let\FCs@gn=\@ne
819 % \end{macrocode}
820 %First it checks whether there is a risk of overflow in this step. If
821 %$A$ and $B$ are two \TeX-registers and $B$ is to be added to $A$,
822 %overflow cannot occur if 1)~one is (or both are) zero or 2)~if $A$ and~$B$
823 %have different signs; otherwise be careful. Note that \TeX\ does not
824 %check for overflow when performing an |\advance| (done in section~1238
825 %of Ref.~\cite{CandTB}), in contrast to |\multiply|, see section~105.
826 % \begin{macrocode}
827 \ifnum\count0<0 \ifnum\csname FCtot@\romannumeral\FCsc@l\endcsname<0
828 \let\FCs@gn=\m@ne \fi\fi
829 \ifnum\count0>0 \ifnum\csname FCtot@\romannumeral\FCsc@l\endcsname>0
830 \let\FCs@gn=\m@ne \fi\fi
831 \global\advance\csname FCtot@\romannumeral\FCsc@l\endcsname by \count0
832 \ifx\FCs@gn\m@ne
833 % \end{macrocode}
834 %They have the same sign, hence risk of overflow. Record the sign
835 %of |\count0| (and of the original total of this column; we just established
836 %they were the same) in |\FCs@gn|. Table~\ref{tab:fcoverflow}
837 %\begin{table}[htb]
838 %\def\PackageError#1#2#3{\PackageWarning{#1}{#2: #3}}
839 %\caption{Examples on overflow.}
840 %\label{tab:fcoverflow}
841 %\begin{tabular}{@{}lfff@{}}
842 %\multicolumn4{@{}c@{}}{\bfseries Projects}\\
843 %\toprule
844 %income & \leeg{31 dec 2014} & \leeg{31 dec 2015} & \leeg{31 dec 2016} \\
845 %\midrule
846 %item~1 & 20000000 & 20000000 & 20000000 \\
847 %item~2 & 10000000 & 2000000 & -1500000 \\
848 %item~3 & 5000000 & -1500000 & 2000000 \\
849 %\checkfcolumns23
850 %\sumline
851 %\bottomrule
852 %\end{tabular}
853 %\end{table}
854 %shows what can go wrong if the numbers are too large: in the left
855 %|F|-column the sumline is incorrect and the number that caused
856 %the overflow is indicated by an exclamation mark. In the middle
857 %|F|-column, overflow occurs twice and because this is once positive,
858 %once negative here, cancellation of errors occurs and the sumline
859 %is correct in the end. Nevertheless, it is advised to swap the two
860 %items that caused the overflow, as shown in the right |F|-column.
861 %
862 %Check that the sign of the updated column total is still correct.
863 %If so, |\FCs@gn| is reset (to |\@ne|) at the end of this chunk.
864 % \begin{macrocode}
865 \ifnum\count0>0 \let\FCs@gn\@ne \fi
866 \count0=\csname FCtot@\romannumeral\FCsc@l\endcsname
867 \ifnum\FCs@gn<0 \count0=-\count0 \fi
868 \ifnum\count0<0
869 \let\FCs@gn=\m@ne
870 \PackageError{fcolumn}{Register overflow}{Overflow occurred in
871 fcolumn \number\FCsc@l\space near or at line \the\inputlineno.
872 You can press\MessageBreak <enter> now and I'll proceed, but
873 check your table. The offending entry\MessageBreak is indicated
874 with an exclamation mark in the output.}%
875 \else\let\FCs@gn=\@ne
876 \fi
877 \fi
878 % \end{macrocode}
879 %The value of |\FCs@gn| is used in |\geldm@cro| below.
880 % \begin{macrocode}
881 \setbox0=\hbox{$\geldm@cro{#1}{#2}$}%
882 \ifdim\wd0>\csname FCwd@\romannumeral\FCsc@l\endcsname
883 \global\csname FCwd@\romannumeral\FCsc@l\endcsname=\wd0
884 \fi\unhbox0}
885 % \end{macrocode}
886 %\end{macro}
887 %
888 %The \meta{count}s |\FC@l| and |\FC@r| capture the parts to the left and to the
889 %right of the decimal mark, respectively.
890 % \begin{macrocode}
891 \newcount\FC@l \newcount\FC@r
892 % \end{macrocode}
893 %Some auxiliary definitions for capturing compacted information.
894 % \begin{macrocode}
895 \def\setucc@de#1#2\relax{\uccode`\~=`#1 }
896 \def\assignform@t#1,#2,#3\assignform@t{\def\FCform@t{#1}%
897 \def\FCform@tt{#2}\ifx\FCform@tt\@empty \def\FCform@tt{#1}\fi}
898 % \end{macrocode}
899 %\begin{macro}{\m@thcodeswitch}
900 %As will be shown below, once the first digit, or sign, or decimal mark, or
901 %grouping character is scanned, the decimal digits should loose their activeness.
902 %That is done here for the digits in a rather blunt way, by putting their
903 %|\mathcode|s to zero if |#1| equals~0, since the actual |\mathcode| is not
904 %important---as long as it is not |"8000|---because the digits are not used for
905 %typesetting. (And even if they were; it's inside |\box0|, whose contents will be
906 %discarded.) When the |$| in |\e@fi| is encountered, the digits get back their
907 %original |\mathcode|s so that the actual typesetting in |\zetg@ld| is correct again.
908 %With a non-zero |#1| the activeness is switched on.
909 % \begin{macrocode}
910 \def\m@thcodeswitch#1{\count0=10 \loop\ifnum\count0>0
911 \advance\count0 by \m@ne\mathcode\expandafter`\the\count0=
912 \ifnum#1=0 0 \else "8000 \fi\repeat}
913 % \end{macrocode}
914 %\end{macro}
915 %\def\<#1>{\leavevmode\hbox{$\langle$#1\/$\rangle$}}
916 %\begin{macro}{\b@fi}
917 %The macro |\b@fi| provides the beginning of the financial column. It
918 %will be inserted in the column to capture the number entered by the user.
919 %The separator and decimal mark are within a math environment, so you
920 %can indeed specify |\,| instead of |\thinspace|, but there are extra
921 %braces around them in the code where they are used, so it doesn't
922 %affect the spacing between the digits (trick copied from |dcolumn|,
923 %Ref.~\cite{dcolumn}).
924 % \begin{macrocode}
925 \def\b@fi#1#2#3#4{\sep@xt#1\sep@xt\def\decim@lmark{#2}\def\sp@l{#3}
926 \assignform@t#4,,\assignform@t\global\advance\FCsc@l by \@ne
927 \global\FC@l=0 \global\FC@r=1
928 % \end{macrocode}
929 %The next statement is missing in |\@array| (see section~\ref{adaptations})
930 %for good reasons; it indicates to the tagging engine that what follows is
931 %fake math, not to be tagged. But for |fcolumn| that's exactly what's going
932 %on, so include it here, but only in columns with specifier~|F|, i.e.,
933 %in the current macro.
934 % \begin{macrocode}
935 %<*tagging>
936 \m@th
937 %</tagging>
938 % \end{macrocode}
939 %The value specified by the user is then captured by |\FC@l| and
940 %this is done in a special way: |\FC@l| is assigned globally within
941 %|\box0|. Why? To use it as scribbling paper to examine what the
942 %user entered, without dumping it into the horizontal list.
943 %
944 %There are four parts to an |F|-column entry, all parts optional,
945 %making 16 combinations. The sequence is
946 %(in the Backus--Naur notation of Ref.~\cite{CandTA}):
947 %\<sign> \<integer constant> \<decimal mark>
948 %\<integer constant>. Here \<sign> is a plus or minus character with
949 %category code~12, \<integer constant> is a sequence of zero
950 %or more (decimal) \<digit>s, and \<decimal mark> is the middel
951 %part of |#3|, i.e., the comma in |3,2| or the period in |3.2|. If
952 %the \<decimal mark> is absent with no space characters between the
953 %two \<integer constant> terms, these merge, making four redundant
954 %entries. One of the combinations is \<empty>, a sequence of characters
955 %outside the ones that are made active, e.g., plain text or nothing at all:
956 %this is the only combination that doesn't put anything in an |F|-column---and
957 %was the most difficult part to handle.
958 %
959 %The minus sign must be captured separately, because in an entry like
960 %|-0,07| the 7~cents are negative, but this cannot be seen from the
961 %part to the left of the decimal mark, since $-0$ is~$0$ in \TeX\ (in
962 %fact in most computer languages, but not in~MIX~\cite{TAOCP1}), so
963 %|\ifnum-0<0| yields |false|. |\FCs@gn| is a general purpose flag. Its
964 %first use is to capture the sign.
965 % \begin{macrocode}
966 \let\FCs@gn=\@ne\relax \setbox0\hbox\bgroup$
967 % \end{macrocode}
968 %Do the scan inside a box and inside math mode. Start with defining all
969 %characters that may appear as the first one in an |F|-column as active;
970 %digits first. This needs a complicated |\edef| with accompanying |\noexpand|
971 %because the starting digit for |\FC@l| must be packed inside each definition.
972 % \begin{macrocode}
973 \count@=10 \loop\ifnum\count@>0 \advance\count@ by \m@ne
974 \uccode`\~=\expandafter`\the\count@ \uppercase{\edef~}{\noexpand
975 \m@thcodeswitch0 \global\FC@l=\the\count@}\repeat
976 % \end{macrocode}
977 %For the input decimal mark something extra is needed: if it is the
978 %first character in an |F|-column (like in |,07|), it should also
979 %restore the |\mathcode|s of the digits. Checking whether or not it
980 %is the first is easy, since in that case the |\mathcode|s of the
981 %decimal digits are still |"8000|. The assignment to |\FC@r| starts
982 %with~1 (with no space following), so that appended digits get captured
983 %correctly, even if they start with~0; postprocessing of |\FC@r| is done
984 %in |\e@fi|. The input decimal mark switches itself off as active
985 %character, so at most one input decimal mark is allowed (N.B.:
986 %this makes sense).
987 % \begin{macrocode}
988 \afterassignment\setucc@de\count@#3\relax
989 \uppercase{\def~}{\ifnum\mathcode`\0="8000 \m@thcodeswitch0 \fi
990 \afterassignment\deactdecm@rk\count@#3\relax \global\FC@r=1}
991 % \end{macrocode}
992 %The input grouping character effectively expands to ``nothing, i.e.,
993 %ignore'' in a complicated way: it ignores the character and resumes
994 %scanning the number. The test prior to that action is needed if
995 %the grouping character is the first character encountered in the
996 %|F|-column. Which part to continue with depends on whether or not an
997 %input decimal mark was encountered; that can be checked by looking
998 %at its |\mathcode|.
999 %
1000 %The input grouping character is the dot ``|.|'', except when that character
1001 %was already chosen as input decimal mark. In that case, the grouping
1002 %character will be the comma. This is easy to check because the |\uccode|
1003 %of |`\~| is still preserved.
1004 % \begin{macrocode}
1005 \ifnum\uccode`\~=`. \uccode`\~=`,\relax\else \uccode`\~=`.\relax\fi
1006 \uppercase{\def~}{\ifnum\mathcode`\0="8000 \m@thcodeswitch0 \fi
1007 \afterassignment\d@cm\count@#3\relax
1008 % \end{macrocode}
1009 %The |\expandafter| below is necessary because the global assignment
1010 %should act after the |\fi|.
1011 % \begin{macrocode}
1012 \ifnum\count@=\mathcode`- \expandafter\global\FC@l=\the\FC@l
1013 \else \expandafter\global\FC@r=\the\FC@r\fi}%
1014 % \end{macrocode}
1015 %The signs are relatively simple: record the sign, restore |\mathcode|s if
1016 %needed (it should be: a minus sign between digits screws up everything),
1017 %and start scanning the number.
1018 % \begin{macrocode}
1019 \uccode`\~=`+\relax \uppercase{\def~}{\ifnum\mathcode`\0="8000
1020 \m@thcodeswitch0 \fi\global\FC@l=0}
1021 \uccode`\~=`-\relax \uppercase{\def~}{\ifnum\mathcode`\0="8000
1022 \m@thcodeswitch0 \fi\global\let\FCs@gn\m@ne \global\FC@l=0}
1023 % \end{macrocode}
1024 %Now actually activate all these codes.
1025 % \begin{macrocode}
1026 \mathcode`-="8000 \mathcode`+="8000 \mathcode`.="8000
1027 % \end{macrocode}
1028 %These three remain active until the |$| in |\e@fi| is encountered.
1029 %The following ones will, except in the \<empty> case, have their activeness
1030 %turned off at some time.
1031 % \begin{macrocode}
1032 \m@thcodeswitch1 \afterassignment\actdecm@rk\count@#3\relax}
1033 % \end{macrocode}
1034 %\end{macro}
1035 %\begin{macro}{\sep@xt}
1036 %\begin{macro}{\actdecm@rk}
1037 %\begin{macro}{\deactdecm@rk}
1038 %\begin{macro}{\d@cm}
1039 %Again a few small macros to do important work. At first |\sep@xt| to extract
1040 %the first character of |#1|, which in most cases will be the only character.
1041 %Then |\actdecm@rk| and |\deactdecm@rk| to activate and deactive respectively the
1042 %decimal mark. Finally, |\d@cm| captures a |\mathcode| for further investigation.
1043 % \begin{macrocode}
1044 \def\sep@xt#1#2\sep@xt{\def\sep@rator{#1}}
1045 \def\actdecm@rk#1#2\relax{\ifx#1.\relax \mathcode`,="8000
1046 \else \mathcode`#1="8000 \fi}
1047 \def\deactdecm@rk#1#2\relax{\mathcode`#1=0 }%
1048 \def\d@cm#1#2{\count@=\mathcode`#1 }
1049 % \end{macrocode}
1050 %\end{macro}
1051 %\end{macro}
1052 %\end{macro}
1053 %\end{macro}
1054 %\begin{macro}{\e@fi}
1055 %If the digits are still active then either nothing, or only
1056 %characters that did not deactivate the digits were entered. In
1057 %both cases the output should be \<empty>. To flag this situation outside
1058 %the group that started with the opening |$| of |\b@fi|, |\FC@r| is set
1059 %globally to a negative value. This doesn't harm, because it didn't
1060 %contain relevant information anyway. Outside the group, the sign
1061 %of |\FC@r| can be tested then. This is a slight misuse of this
1062 %|\count|, but now it's documented. In effect, |\FC@r| can only be~$-1$,
1063 %1, or at least~10, so the comparison |\ifnum\FC@r>0| does not miss~0.
1064 % \begin{macrocode}
1065 \def\e@fi{\ifnum\mathcode`\0="8000 \global\FC@r=\m@ne\fi$\egroup
1066 \ifnum\FC@r>0
1067 % \end{macrocode}
1068 %If there was no decimal mark or if there was a decimal mark but no
1069 %decimal part, |\FC@r| will still be~1, which doesn't parse well with
1070 %|\secd@xt|, so a zero is appended, i.e., yielding~|10|.
1071 % \begin{macrocode}
1072 \ifnum\FC@r=1 \FC@r=10 \fi
1073 % \end{macrocode}
1074 %Next is a loop for bringing the decimal part in the correct way to
1075 %the integer part. This loop is performed the number of decimal
1076 %digits to be printed (the~2 in |3,2| of the default setting). The higher
1077 %this number is, e.g., 6~when |3,6| is chosen as |#3|---but who does that?---,
1078 %the lower the maximum initial value for |\FC@l| can be, before a low-level
1079 %\TeX\ arithmetic overflow will occur, so don't overdo it.
1080 % \begin{macrocode}
1081 \afterassignment\i@ts\count@\sp@l
1082 \loop\ifnum\count0>0 \multiply\FC@l by 10 \expandafter\secd@xt
1083 \number\FC@r\secd@xt \advance\count0 by \m@ne \repeat
1084 % \end{macrocode}
1085 %This also means that if more decimal digits than this are provided, the excess
1086 %digit(s) will not be handled. This is truncation, not rounding! If all
1087 %truncated digits are zero, this truncation is exact and they are silently ignored,
1088 %see the example Table~\ref{tab:fctrunc},
1089 %\begin{table}[htb]
1090 %\newcolumntype{d}{D,,{-1}}
1091 %\caption{Truncating excess digits.}
1092 %\label{tab:fctrunc}
1093 %\begin{tabular}{@{}ldfl@{}}
1094 %\toprule
1095 %composer & \multicolumn1c{raw entry} & \leeg{debt} & remark \\
1096 %\midrule
1097 %Berg & 123,450 & 123,450 & silently ignoring digit ``0'' \\
1098 %Eisler & 234,563 & 234,563 & warning: digit ``3'' ignored \\
1099 %Sch\"onberg & 345,6704 & 345,6704 & warning: digits ``04'' ignored \\
1100 %Webern & 2,3456 & 2,3456 & warning: digits ``56'' ignored, i.e.,\\
1101 % & & & \quad without rounding this entry to 2,35 \\
1102 %\sumline
1103 %\bottomrule
1104 %\end{tabular}
1105 %\end{table}
1106 %that was created with |@{}ldfl@{}| (|d| for centering on the decimal
1107 %mark~\cite{dcolumn}) as tabular key. If, however, at least one of them is
1108 %not zero, a |\PackageWarning| will be given, showing the discarded digit(s).
1109 %There is nothing magical about the constant~19 below: it is simply the
1110 %concatenation of~|1| and the largest decimal digit. The smallest next value
1111 %that |\FC@r| can have is~100, so all values from~19 up to and including~99
1112 %would have worked here.
1113 % \begin{macrocode}
1114 \ifnum\expandafter\@gobble\number\FC@r>0
1115 \PackageWarning{fcolumn}{Excess digit\ifnum\FC@r>19 s\fi\space
1116 ``\expandafter\tw@l\number\FC@r\relax'' in decimal part
1117 \MessageBreak ignored near or}
1118 \fi
1119 % \end{macrocode}
1120 %Don't forget to correct for the sign (once this is done, |\FCs@gn|
1121 %is free again and can and will be used for other purposes).
1122 %Then output the result.
1123 % \begin{macrocode}
1124 \ifx\FCs@gn\m@ne\relax\FC@l=-\FC@l\fi\zetg@ld{\sp@l}{\FC@l}\fi}
1125 % \end{macrocode}
1126 %\end{macro}
1127 %\begin{macro}{\i@ts}
1128 %\begin{macro}{\tw@l}
1129 %Macro |\e@fi| uses two very tiny macros, given here. In previous versions
1130 %these were defined inside |\e@fi|, so they also got undefined when |\e@fi|
1131 %ended. The current solution makes execution a bit faster.
1132 % \begin{macrocode}
1133 \def\i@ts#1#2{\count0=#2} \def\tw@l#1#2\relax{#2}
1134 % \end{macrocode}
1135 %\end{macro}
1136 %\end{macro}
1137 %\begin{macro}{\secd@xt}
1138 %The second digit from the left is needed from a string of characters
1139 %representing a decimal number (that should be at least 10 and start with
1140 %a~|1|, but that is guaranteed by |\e@fi|). The macro |\secd@xt| extracts
1141 %that digit, which is then added to |\FC@l|. A new number is assigned to
1142 %|\FC@r|, that consists of the digits of |1#2|, unless |#2| was empty; in
1143 %that case~10 is assigned. In this way |\FC@r| is prepared for insertion
1144 %in the next invocation of |\secd@xt|. In iterating: 1234 yields~134,
1145 %yields~14, yields~10, stays~10, etc.
1146 % \begin{macrocode}
1147 \def\secd@xt1#1#2\secd@xt{\advance\FC@l by #1
1148 \FC@r=1#2 \ifnum\FC@r=1 \FC@r=10 \fi}
1149 % \end{macrocode}
1150 %\end{macro}
1151 %\subsection{The sumline, close to a postamble}
1152 %\begin{macro}{\@mksumline}
1153 %The construction of the sumline is much easier than that of the preamble (see
1154 %Section~\ref{adaptations}) for several reasons. It may be safely assumed that
1155 %the preamble specifier is grammatically correct because it has already been
1156 %screened by |\@mkpream|. Furthermore, most entries will simply add nothing
1157 %to |\s@ml@ne|, e.g., |@|, |!|, and \verb+|+ can be fully ignored. Ampersands
1158 %are only inserted by |c|, |l|, |r|, |p|, |m|, and |b|. So, if |\x| is a macro
1159 %that prints the desired result of the column (see below), then a specifier
1160 %like |lflf| will yield the sumline |&\x&&\x\\|. Had the specifier been
1161 %\verb+l|f||@{ }l|f+, then the same sumline must be constructed: all
1162 %difficulties are already picked up and solved in the creation of the preamble.
1163 %
1164 %In reality the sumline must be constructed from the expanded form of the
1165 %specifier, so |@{}lf@{}| will expand as
1166 %|@{}l>{\b@fi.,{3,2}{}}r<{\e@fi}@{}|. The rules for constructing the
1167 %\begingroup \parskip0pt
1168 %\iffalse Clumsy way to have bulleted items because package paralist
1169 % is not compatible with tagging \fi
1170 %sumline are now very simple:
1171 %
1172 %\setbox0\hbox{$\bullet$~}
1173 %\noindent\copy0 add an ampersand when |c|, |l|, |r|, |p|, |m|, or |b| is found,
1174 %unless it is the first one (this \hbox to \wd0{\hfil}is the same as in the preamble);
1175 %
1176 %\noindent\copy0 add a |\x| when |<{\e@fi}| is found;
1177 %
1178 %\noindent\copy0 ignore everything else;
1179 %
1180 %\noindent\box0 close with a |\\|.
1181 %
1182 %\noindent (For completeness' sake it should be mentioned that prior to the |\\| also the column
1183 %\endgroup
1184 %check is inserted, see |\aut@check|.)
1185 %To discriminate, a special version of |\@testpach|~\cite{array}
1186 %could be written, but that is not necessary: |\@testpach| can do all the work,
1187 %although much of it will be discarded. Here speed is sacrificed for space
1188 %and this can be afforded because the creation of the sumline is done only
1189 %once per |tabular| or |longtable|.
1190 %
1191 %The start is copied from |\@mkpream|.
1192 % \begin{macrocode}
1193 \def\@mksumline#1{\gdef\s@ml@ne{}\@lastchclass 4 \@firstamptrue
1194 % \end{macrocode}
1195 %At first the column number is reset and the actual code for what was
1196 %called |\x| above is made inactive.
1197 % \begin{macrocode}
1198 \global\FCsc@l=0 \let\prr@sult=\relax
1199 % \end{macrocode}
1200 %Then |\@mkpream| is picked up again.
1201 % \begin{macrocode}
1202 \@temptokena{#1}\@tempswatrue\@whilesw\if@tempswa\fi{\@tempswafalse
1203 \the\NC@list}\count0\m@ne\let\the@toks\relax\prepnext@tok
1204 % \end{macrocode}
1205 %Next is the loop over all tokens in the expanded form of the specifier. The change
1206 %with respect to |\@mkpream| is that the body of the loop is now only dealing
1207 %with F-classes~0, 2, and 10. What to do in those cases is of course different from
1208 %what to do when constructing the preamble, so special definitions are created, see
1209 %below.
1210 % \begin{macrocode}
1211 \expandafter\@tfor\expandafter\@nextchar\expandafter:\expandafter=\the
1212 \@temptokena\do{\@testpach\ifcase\@chclass\@classfz\or\or\@classfii\or
1213 \or\or\or\or\or\or\or\@classfx\fi\@lastchclass\@chclass}%
1214 % \end{macrocode}
1215 %And the macro is finished by applying the |\aut@check| and appending the
1216 %|\\| to the sumline. Note that the |\aut@check| is performed {\itshape
1217 %in\/} the last column, but since it does not put anything in the
1218 %horizontal list---it only writes to screen and transcript file---, this
1219 %is harmless.
1220 % \begin{macrocode}
1221 \xdef\s@ml@ne{\s@ml@ne\noexpand\aut@check\noexpand\\}}
1222 % \end{macrocode}
1223 %\end{macro}
1224 %\begin{macro}{\@addtosumline}
1225 %Macro |\@addtosumline|, as its name already suggests, adds something to the
1226 %sumline, like its counterpart |\@addtopreamble| did to the preamble.
1227 % \begin{macrocode}
1228 \def\@addtosumline#1{\xdef\s@ml@ne{\s@ml@ne #1}}
1229 % \end{macrocode}
1230 %\end{macro}
1231 %\begin{macro}{\@classfx}
1232 %Class~f10 for the sumline creation is a stripped down version of |\@classx|: add an
1233 %ampersand unless it is the first. It deals with the specifiers |b|, |m|, |p|, |c|,
1234 %|l|, and |r|.
1235 % \begin{macrocode}
1236 \def\@classfx{\if@firstamp \@firstampfalse \else \@addtosumline &\fi}
1237 % \end{macrocode}
1238 %\end{macro}
1239 %\begin{macro}{\@classfz}
1240 %Class~f0 is applicable for specifiers |c|, |l|, and |r|, and if the arguments
1241 %of |p|, |m|, or |b| are given. The latter three cases, with |\@chnum| is 0,
1242 %1, or~2 should be ignored and the first three cases are now similar to class~f10.
1243 % \begin{macrocode}
1244 \def\@classfz{\ifnum\@chnum<\thr@@ \@classfx\fi}
1245 % \end{macrocode}
1246 % \end{macro}
1247 %\begin{macro}{\@classfii}
1248 %Here comes the nice and nasty part. Class~f2 is applicable if a~|<| is
1249 %specified. This is tested by checking |\@lastchclass|, which should be
1250 %equal to~8. Then it is checked that the argument to |<| is indeed
1251 %|\e@fi|. This check is rather clumsy but this was the first way, after
1252 %many attempts, that worked. It is necessary because the usage of |<|
1253 %is not restricted to |\e@fi|: the user may have specified other
1254 %\LaTeX-code using~|<|.
1255 % \begin{macrocode}
1256 \def\@classfii{\ifnum\@lastchclass=8
1257 \edef\t@stm{\expandafter\string\@nextchar}
1258 \edef\t@stn{\string\e@fi} \ifx\t@stm\t@stn
1259 % \end{macrocode}
1260 %If both tests yield |true|, i.e., we encountered a |<{\e@fi}| where
1261 %we expect one to find, then add the macro to typeset everything (this
1262 %is what was mentioned |\x| above for brevity).
1263 % \begin{macrocode}
1264 \@addtosumline{\prr@sult}
1265 % \end{macrocode}
1266 %But we're not done yet: in the following lines of code the appropriate
1267 %\meta{count}s and \meta{dimen}s are created, if necessary. Note that
1268 %|\FCsc@l| was set to~0 in the beginning of |\@mksumline|, so it is
1269 %well-defined when |\@classfii| is used.
1270 % \begin{macrocode}
1271 \global\advance\FCsc@l by \@ne \ifnum\FCsc@l>\FCtc@l
1272 % \end{macrocode}
1273 %Apparently the number of requested columns is larger than the currently
1274 %available number of relevant \meta{count}s and \meta{dimen}s, so new ones
1275 %should be created. What is checked here is merely the existence of
1276 %|\FCtot@<some romannumeral>|. If it already exists---although it may not
1277 %even be a \meta{count}; that cannot be checked---it is not created by
1278 %|fcolumn| and an error is given. In case it is a \meta{count} you're just
1279 %lucky, and you could ignore that error, although any change to this
1280 %\meta{count} is global anyway, so things will be overwritten. In the
1281 %case it is not a \meta{count}, things will go haywire and you'll soon
1282 %find out. The remedy then is to rename your \meta{count} prior to
1283 %|fcolumn| to avoid this name clash.
1284 % \begin{macrocode}
1285 \expandafter\ifx\csname FCtot@\romannumeral\FCsc@l\endcsname\relax
1286 \expandafter\newcount\csname FCtot@\romannumeral\FCsc@l\endcsname
1287 \else
1288 \PackageError{fcolumn}{Name clash for <count>}{\expandafter\csname
1289 FCtot@\romannumeral\FCsc@l\endcsname is already defined and it may
1290 not even be a <count>. If you're\MessageBreak sure it is a <count>,
1291 you can press <enter> now and I'll proceed, but things\MessageBreak
1292 will get overwritten.}%
1293 \fi
1294 % \end{macrocode}
1295 %And the same is applicable for the \meta{dimen}: in case of a name
1296 %clash you have to rename your \meta{dimen} prior to |fcolumn|.
1297 % \begin{macrocode}
1298 \expandafter\ifx\csname FCwd@\romannumeral\FCsc@l\endcsname\relax
1299 \expandafter\newdimen\csname FCwd@\romannumeral\FCsc@l\endcsname
1300 % \end{macrocode}
1301 %If the creation was successful, the \meta{count} |\FCtc@l| should be
1302 %increased.
1303 % \begin{macrocode}
1304 \global\FCtc@l=\FCsc@l
1305 \else
1306 \PackageError{fcolumn}{Name clash for <dimen>}{\expandafter\csname
1307 FCwd@\romannumeral\FCsc@l\endcsname is already defined and it may
1308 not even be a <dimen>. If you're\MessageBreak sure it is a <dimen>,
1309 you can press <enter> now and I'll proceed, but things\MessageBreak
1310 will get overwritten.}%
1311 \fi
1312 \fi
1313 \fi
1314 \fi}
1315 % \end{macrocode}
1316 %Once created, it is not necessary to initialise them here because that
1317 %is done later in one go.
1318 %\end{macro}
1319 %\begin{macro}{\sumline}
1320 %The command for the sumline has one optional argument: the separation
1321 %between the rule and the total. By default this is 2\thinspace pt, but
1322 %the user may, e.g., specify |\sumline[10pt]| if that separation needs to be
1323 %10\thinspace pt. The assignment needs to be global, because it is done
1324 %in the first column of the tabular, but is valid for the whole line.
1325 % \begin{macrocode}
1326 \newdimen\s@mlinesep
1327 \def\sumline{\@ifnextchar[\s@mline{\s@mline[2pt]}}
1328 \def\s@mline[#1]{\global\s@mlinesep=#1 \s@ml@ne}
1329 % \end{macrocode}
1330 %In the introduction it was stated that |\sumline| has two options, but in
1331 %reality that second option is the option to |\\|, that is issued by |\s@ml@ne|.
1332 %\end{macro}
1333 %\begin{macro}{\prr@sult}
1334 %The macro |\prr@sult| actually puts the information together. It starts
1335 %like |\leeg|.
1336 % \begin{macrocode}
1337 \def\prr@sult{$\egroup \let\e@fi=\relax \let\FCform@t=\FCform@tt
1338 % \end{macrocode}
1339 %Then the information for the last line is computed. It is not sufficient
1340 %to calculate the width of the result (in points) to use that as the width
1341 %of the rule separating the individual entries and the result. It may happen
1342 %that the sum is wider (in points) than any of the entries, e.g., when the
1343 %result of $6+6$ (using specifier~|3,2|) is typeset. The width of the rule
1344 %should be equal to the width of |\hbox{$12{,}00$}| then. On the other
1345 %hand the width of the rule when summing $24$ and $-24$ should be that of
1346 %|\hbox{$-24{,}00$}| (or |\hbox{$(24{,}00$}|, see above), not the width of
1347 %the result |\hbox{$0{,}00$}|. Therefore the maximum of all entry widths,
1348 %including the result, was calculated. This excludes the extension to the
1349 %right in case parentheses are used, again for aesthetic reasons.
1350 % \begin{macrocode}
1351 \setbox0=\hbox{$\geldm@cro{\sp@l}{\number\csname
1352 FCtot@\romannumeral\FCsc@l\endcsname}$}%
1353 \ifdim\wd0>\csname FCwd@\romannumeral\FCsc@l\endcsname
1354 \global\csname FCwd@\romannumeral\FCsc@l\endcsname=\wd0
1355 \fi
1356 \vbox{\hrule width \csname FCwd@\romannumeral\FCsc@l\endcsname
1357 \vskip\s@mlinesep
1358 \hbox to \csname FCwd@\romannumeral\FCsc@l\endcsname{\hfil\unhbox0}}}
1359 % \end{macrocode}
1360 %\end{macro}
1361 %\subsection{Other checks}
1362 %\begin{macro}{\leeg}
1363 %This macro is used to overrule the default behaviour of the pair |\b@fi|
1364 %and |\e@fi|. It starts with ending the groups in the
1365 %same way that |\e@fi| would normally do. Then the effect of |\e@fi|
1366 %(that is still in the preamble) is neutralised by |\let|ting it to
1367 %be |\relax|. This |\let| is only local to the current column. Then
1368 %the argument to |\leeg| is processed in the normal way for a right
1369 %aligned column.
1370 %
1371 %Since the user may from time to time also need a column entry other than
1372 %a number in the table, e.g., |\leeg{p.m.}|, this definition is without
1373 %at-sign. By defining |\leeg| in this way, instead of |\multicolumn1r{}|
1374 %(which contains |\omit|), the default spacing in the column is retained.
1375 %It doesn't alter the width of the sumrule, but has its normal effect on
1376 %the column width, so be careful: don't insert the unabridged version of
1377 %Romeo and Juliet~\cite{RJ} here. It is not typeset in math mode, nor
1378 %does it use the extra (formatting) information of |#4| of the fcolumn,
1379 %so you're completely free here.
1380 % \begin{macrocode}
1381 \def\leeg#1{$\egroup \let\e@fi=\relax #1}
1382 % \end{macrocode}
1383 %Note that anything may be given as argument to |\leeg|, so in principle
1384 %it can also be used to cheat: |\leeg{0,03}| will insert the text~|0,03|
1385 %in the table but it doesn't increase the totals of that column by~3
1386 %(assuming |3,2| coding for the separations). But you won't cheat, won't you?
1387 %\end{macro}
1388 %\begin{macro}{\res@tsumline}
1389 %Since all changes to the totals and widths of the columns are global,
1390 %they have to be reset actively at the start of a tabular or array.
1391 %That is an action by itself, but it may occur more often, on request
1392 %of the user, therefore a special macro is defined. A side effect of
1393 %this macro is that |\FCsc@l| is reset to~0. This is an advantage:
1394 %it should be zero at the beginning of a line in the table (for other
1395 %lines this is done by the |\\|).
1396 % \begin{macrocode}
1397 \def\res@tsumline{\FCsc@l=\FCtc@l\loop\ifnum\FCsc@l>0
1398 \global\csname FCtot@\romannumeral\FCsc@l\endcsname=0
1399 \global\csname FCwd@\romannumeral\FCsc@l\endcsname=\z@
1400 \advance\FCsc@l by \m@ne\repeat}
1401 % \end{macrocode}
1402 %\end{macro}
1403 %\begin{macro}{\resetsumline}
1404 %To reset a sumline within a table, it should be done within a |\noalign|.
1405 % \begin{macrocode}
1406 \def\resetsumline{\noalign{\res@tsumline}}
1407 % \end{macrocode}
1408 %\end{macro}
1409 %\begin{macro}{\aut@check}
1410 %If the number of |F|-columns is even, it is assumed that they are part of
1411 %two sets of columns of which each column of the first set should balance
1412 %the appropriate column of the second set. If on the other hand the number
1413 %of columns is odd, then at least one column has nothing to balance against
1414 %and no automatic checking occurs. It is correct to check for oddness of
1415 %|\FCsc@l| since this |\aut@check| is only performed in the last column of
1416 %the tabular: the value of |\FCsc@l| now equals the number of columns used
1417 %in the current tabular (and may be less than |\FCtc@l|).
1418 %
1419 %The output is only to screen and the transcript file; it doesn't change
1420 %the appearance of your document, so in case the assumption is wrong you
1421 %can safely ignore the result and go on. The \meta{count}s~0 and~1 are
1422 %used here and this can be done because any content of those \meta{count}s
1423 %from previous calculations has become irrelevant at this moment.
1424 %
1425 %If the list |\FC@chklist| is empty, the list for the automatic check is
1426 %generated (which will remain empty if |\FCsc@l| is odd).
1427 % \begin{macrocode}
1428 \def\FC@chklist{}
1429 \def\aut@check{\ifx\@empty\FC@chklist\relax
1430 \ifodd\FCsc@l\else
1431 \count0=\@ne \count1=\FCsc@l \divide\count1 by \tw@
1432 \loop\ifnum\count1<\FCsc@l \advance\count1 by \@ne
1433 \xdef\FC@chklist{\FC@chklist\number\count0,\number\count1;}%
1434 \advance\count0 by\@ne \repeat
1435 \fi
1436 \fi
1437 % \end{macrocode}
1438 %Then this list is peeled off and processed.
1439 %The comparison is done between the internal representation of the totals,
1440 %so this only makes sense when the decimal part, i.e., the |2| of |3,2| in
1441 %the default value for |#3| of~|f| is the same for both columns. That is
1442 %almost always the case due to the nature of this type of tabulars. The
1443 %\meta{count} |\FC@l| is free to be used for the calculations needed here.
1444 % \begin{macrocode}
1445 \loop\ifx\FC@chklist\@empty\else
1446 \expandafter\fre@t\FC@chklist\fre@t
1447 \FC@l=\csname FCtot@\romannumeral\count0\endcsname
1448 \advance\FC@l by -\csname FCtot@\romannumeral\count1\endcsname
1449 \ifnum\FC@l=0 \else \ifnum\FC@l<0 \FC@l=-\FC@l \fi
1450 \PackageWarning{fcolumn}{Representations of F-columns \number
1451 \count0 \space and \number\count1 \space differ by\MessageBreak
1452 \number\FC@l\space due to \string\sumline\space near or}%
1453 \fi
1454 \repeat}
1455 % \end{macrocode}
1456 %When |\aut@check| is finished, |\FC@chklist| is empty again, i.e., well
1457 %prepared for the next time it is used. This also means that the default
1458 %behaviour kicks in again: if that's not what you want, you should
1459 %specify the appropriate |\checkfcolumns| lines again.
1460 %\end{macro}
1461 %\begin{macro}{\fre@t}
1462 %This function eats the first two numbers off |\FC@chklist|.
1463 % \begin{macrocode}
1464 \def\fre@t#1,#2;#3\fre@t{\count0=#1 \count1=#2 \xdef\FC@chklist{#3}}
1465 % \end{macrocode}
1466 %\end{macro}
1467 %\begin{macro}{\checkfcolumns}
1468 %But the assumptions for |\aut@check| may be wrong, therefore manual
1469 %control on this checking is also made possible here. The macro
1470 %|\checkfcolumns| provides a way to the user to check that the appropriate
1471 %columns are balanced (as it should in a balance). Arguments |#1| and
1472 %|#2| are the |F|-column numbers to compare. It is the responsibility of
1473 %the user to provide the correct numbers here, otherwise bogus output
1474 %is generated. If this manual check is inserted, the automatic check
1475 %will not be performed.
1476 % \begin{macrocode}
1477 \def\checkfcolumns#1#2{\noalign{\xdef\FC@chklist{\FC@chklist #1,#2;}}}
1478 % \end{macrocode}
1479 %\end{macro}
1480 %
1481 %\subsection{Adaptations to existing macros}\label{adaptations}
1482 %\begin{macro}{\@array}
1483 %The definition of |\@array| had to be extended slightly because it should
1484 %also include |\@mksumline| (acting on the same |#2| that |\@mkpream| gets).
1485 %This change is transparant: it only adds functionality and if you don't
1486 %use that, you won't notice the difference. This addition to |\@array|
1487 %by redefining it is distracting and error-prone and is better implemented
1488 %via the Hook-mechanism~\cite{hook}; from the June~1, 2026 version onwards,
1489 %package |array| will contain two dedicated Hooks~\cite{LTN43} for adding
1490 %various chunks, among which the code on |\@mksumline| in |\@array| and for
1491 %resetting |\FCsc@l| in |\ar@ialign|, see below. But till that date it is
1492 %|fcolumn| that should take care of those Hooks and that is done here.
1493 % \begin{macrocode}
1494 \IfFormatAtLeastF{2026-06-01}{\NewHookWithArguments{tbl/init}{2}
1495 \NewHook{tbl/row/end}}
1496 % \end{macrocode}
1497 %Now fill those Hooks. The |#2| is the actual |#2| from |\@array|.
1498 % \begin{macrocode}
1499 \AddToHookWithArguments{tbl/init}{\begingroup\@mksumline{#2}%
1500 \res@tsumline\endgroup}
1501 % \end{macrocode}
1502 %No arguments for this one.
1503 % \begin{macrocode}
1504 \AddToHook{tbl/row/end}{\global\FCsc@l=0 }
1505 % \end{macrocode}
1506 %Redefining starts by just copying the original definition from v2.6d (or later)
1507 %of the |array| package~\cite{array}, compacted.
1508 % \begin{macrocode}
1509 %<*tagging>
1510 \ExplSyntaxOn
1511 %</tagging>
1512 \IfFormatAtLeastF{2026-06-01}{\def\@array[#1]#2{\@tempdima\ht
1513 \strutbox\advance\@tempdima by \extrarowheight\setbox\@arstrutbox
1514 \hbox{\vrule\@height\arraystretch\@tempdima\@depth\arraystretch
1515 \dp\strutbox\@width\z@}%
1516 %<*tagging>
1517 \tbl_init_cell_data_for_table:
1518 %</tagging>
1519 \begingroup\@mkpream{#2}%
1520 %<*tagging>
1521 \tbl_count_table_cols:
1522 %</tagging>
1523 \xdef\@preamble{\noexpand\ar@ialign\@halignto
1524 \bgroup\@arstrut
1525 %<*tagging>
1526 \UseTaggingSocket{tbl/row/begin}
1527 \tbl_init_cell_data_for_row:
1528 %</tagging>
1529 \@preamble\tabskip\z@\cr}\endgroup
1530 % \end{macrocode}
1531 %After this |\endgroup| a |\begingroup| is issued; this ensures that
1532 %|\@mksumline| that is created there experiences the same environment that
1533 %|\@mkpream| saw. As a side product of |\@mksumline| also the \meta{count}s
1534 %for the totals and \meta{dimen}s for the widths of the columns are created.
1535 %All columns should start fresh, i.e., totals are~0 and widths are 0\thinspace pt.
1536 %This can be done within the group, as these resettings are global.
1537 % \begin{macrocode}
1538 \UseHookWithArguments{tbl/init}{2}{}{#2}
1539 % \end{macrocode}
1540 %From here on it is just the old definition of |array.sty|.
1541 % \begin{macrocode}
1542 \@arrayleft\if #1t\vtop\else\if#1b\vbox\else\vcenter\fi\fi
1543 \bgroup\let\@sharp ##\let\protect\relax\lineskip\z@\baselineskip\z@
1544 \mathsurround \z@ \let\\\@arraycr \let\tabularnewline\\\let\par\@empty
1545 %<*tagging>
1546 \UseTaggingSocket{tbl/init}
1547 %</tagging>
1548 \@preamble}
1549 % \end{macrocode}
1550 %Because |\@array| was changed here and it is this version that should
1551 %be used, |\@@array| should be |\let| equal to |\@array| again.
1552 % \begin{macrocode}
1553 \let\@@array=\@array\def\ar@ialign{\everycr{\noalign{%
1554 %<*tagging>
1555 \tbl_if_row_was_started:T{\UseTaggingSocket{tbl/row/end}}
1556 \tbl_update_cell_data_for_next_row:
1557 %</tagging>
1558 % \end{macrocode}
1559 %\end{macro}
1560 %Here comes an accompanying change: after each |\\| (or |\cr| for that matter)
1561 %the \meta{count} |\FCsc@l| should be reset. This is easiest done with
1562 %|\everycr|, but |\everycr| is essentially put to |{}| by |\ar@ialign|, so
1563 %that definition should change. The resetting should be done globally.
1564 % \begin{macrocode}
1565 \UseHook{tbl/row/end}
1566 % \end{macrocode}
1567 %Then the definition is picked up again.
1568 % \begin{macrocode}
1569 }}\tabskip\z@skip\halign}}
1570 % \end{macrocode}
1571 %Much of the techniques here are repeated in |\LT@array|, see the next section.
1572 %
1573 %\subsection{Support for multipage tables}
1574 %Packages |longtable|~\cite{longtable} and |supertabular|~\cite{supertabular}
1575 %can be used for tables that span multiple pages. Package |supertabular| works
1576 %with |fcolumn| out of the box (no changes needed), but may lead to different
1577 %column widths on individual pages. If you don't want that, use |longtable|.
1578 %The packages |fcolumn| and |longtable| also work together because only one
1579 %definition of |longtable| needs to be changed if it is present. That change
1580 %is done via a Hook~\cite{hook}, so it no longer matters in which order
1581 %|fcolumn| and |longtable| are loaded. More happens under the hood if also
1582 %|caption|~\cite{caption} is loaded, and even that is running fine. But with
1583 %the near infinite number of packages that exist nowadays I did not check
1584 %all combinations, so I may have overlooked some.
1585 % \begin{macrocode}
1586 %<*tagging>
1587 %<@@=tbl>
1588 %</tagging>
1589 % \end{macrocode}
1590 %\begin{macro}{\LT@array}
1591 %And here is the only definition of |longtable| that needs to be extended
1592 %to make |fcolumn| work with that package. The lines are compacted a bit
1593 %w.r.t.\ the original |longtable| code; if you want to study the
1594 %code, have a look at the documentation~\cite{longtable}. It is defined, even
1595 %if |longtable| is not loaded. Similar Hooks as in |array| will be placed in
1596 %package |longtable|, after which this section can be removed in
1597 %the next version.
1598 % \begin{macrocode}
1599 \IfFormatAtLeastF{2026-06-01}{\AtBeginDocument{\def\LT@array[#1]#2{%
1600 %<*tagging>
1601 \UseTaggingSocket{tbl/init}\tl_if_empty:eTF{\LTcaptype}%
1602 {\tl_gset:Ne\@currentHref{LT@tables.\theHLT@tables}}{%
1603 %</tagging>
1604 \@kernel@refstepcounter{\LTcaptype}\stepcounter{LT@tables}
1605 %<*tagging>
1606 \tl_gset:Ne\@currentHref{\LTcaptype.\cs_if_exist_use:c
1607 {theH\LTcaptype}}}\tbl_gzero_row_count:
1608 \UseTaggingSocket{tbl/longtable/init}
1609 %</tagging>
1610 \if l#1\LTleft\z@\LTright\fill\else\if r#1\LTleft\fill\LTright\z@
1611 \else\if c#1\LTleft\fill\LTright\fill\fi\fi\fi\let\LT@mcol
1612 \multicolumn\let\LT@@@@tabarray\@tabarray\let\LT@@@@hl\hline
1613 \def\@tabarray{\let\hline\LT@@@@hl\let\multicolumn\LT@mcol
1614 \LT@@@@tabarray}\let\\\LT@tabularcr\let\tabularnewline\\
1615 \def\newpage{\noalign{\break}}\def\pagebreak{\noalign{\ifnum`}=0
1616 \fi\@testopt{\LT@no@pgbk-}4}\def\nopagebreak{\noalign{\ifnum`}=0
1617 \fi\@testopt\LT@no@pgbk4}\let\hline\LT@hline\let\kill\LT@kill
1618 \let\caption\LT@caption\@tempdima\ht\strutbox\let\@endpbox
1619 \LT@endpbox\ifx\extrarowheight\@undefined\let\@acol\@tabacol\let
1620 \@classz\@tabclassz\let\@classiv\@tabclassiv\def\@startpbox{%
1621 \vtop\LT@startpbox}\let\@@@@startpbox\@startpbox\let\@@@@endpbox
1622 \@endpbox\let\LT@LL@FM@cr\@tabularcr\else\advance\@tempdima
1623 \extrarowheight\col@sep\tabcolsep\let\@startpbox\LT@startpbox
1624 \let\LT@LL@FM@cr\@arraycr\fi\setbox\@arstrutbox\hbox{\vrule
1625 \@height\arraystretch\@tempdima\@depth\arraystretch\dp\strutbox
1626 \@width\z@}\let\@sharp##\let\protect\relax
1627 \begingroup\@mkpream{#2}%
1628 %<*tagging>
1629 \tbl_count_table_cols:
1630 %</tagging>
1631 \xdef\LT@bchunk{%
1632 %<*tagging>
1633 \tbl_inbetween_rows:
1634 %</tagging>
1635 \global\advance\c@LT@chunks\@ne\global\LT@rows\z@\setbox\z@\vbox
1636 \bgroup\LT@setprevdepth\tabskip\LTleft\noexpand\halign to\hsize
1637 \bgroup\tabskip\z@\@arstrut
1638 %<*tagging>
1639 \UseTaggingSocket{tbl/row/begin}\tbl_init_cell_data_for_row:
1640 %</tagging>
1641 \@preamble
1642 \tabskip\LTright\cr}\endgroup
1643 % \end{macrocode}
1644 %Until this line it was just the code for |\LT@array| from package~|longtable|.
1645 %The next line is new to |\LT@array|. Its purpose is the same as in |\@array| above.
1646 % \begin{macrocode}
1647 \UseHookWithArguments{tbl/init}{2}{}{#2}
1648 % \end{macrocode}
1649 %A few lines of the original definition again.
1650 % \begin{macrocode}
1651 \expandafter\LT@nofcols\LT@bchunk&\LT@nofcols\LT@make@row\m@th
1652 \let\par\@empty\res@tsumline\everycr{\noalign{%
1653 %<*tagging>
1654 \@@_trace:n {--longtable-->~chunk~row:~ \the\LT@rows\space
1655 row:~ \the\g_@@_row_int \space column:~ \the\g_@@_col_int}
1656 \tbl_if_row_was_started:T {\UseTaggingSocket{tbl/row/end}
1657 \tbl_inbetween_rows:}\int_compare:nNnT\LT@rows<\LTchunksize
1658 {\tbl_gincr_row_count:} % Next is row about to start
1659 %</tagging>
1660 % \end{macrocode}
1661 %And here is the last addition.
1662 % \begin{macrocode}
1663 \UseHook{tbl/row/end}
1664 % \end{macrocode}
1665 %From here on |\LT@array| is picked up again.
1666 % \begin{macrocode}
1667 }}\lineskip\z@\baselineskip\z@\LT@bchunk}}}
1668 %<*tagging>
1669 \ExplSyntaxOff
1670 %<@@=>
1671 %</tagging>
1672 % \end{macrocode}
1673 %\end{macro}
1674 %
1675 %When |longtable| splits a table, i.e., is doing what you loaded it for, modern
1676 %versions of pdf\/\LaTeX\ complain about infinite glue shrinkage. This is due to
1677 %a |\vss| in the definition of |\LT@output|. It's not an error and can be ignored,
1678 %but it is annoying since it interrupts generating the output file. The author
1679 %of |longtable| announced a fix in |\LT@output| in the next version of that
1680 %package~\cite{infinite}.
1681 %
1682 %That's it!
1683 %
1684 %\section*{Acknowledgement}
1685 %Thanks to Karl Berry for valuable comments regarding the consistency of the
1686 %installation procedure of this version. Frank Mittelbach gave various useful
1687 %suggestions for improving the input parsing as well as hints to make the package
1688 %more \LaTeX-like. He also challenged me to make |fcolumn| compatible with
1689 %|longtable|. Christian Hoff's request on column formatting triggered many
1690 %happy hours of coding.
1691 %\bibliographystyle{plain}
1692 %\begin{thebibliography}{10}
1693 %\itemsep0pt \parsep0pt \parskip0pt
1694 %\bibitem{array} Frank Mittelbach and David Carlisle. A new implementation of
1695 % \LaTeX's |tabular| and |array| environment: \url{www.ctan.org/pkg/array}.
1696 %\bibitem{booktabs} Simon Fear and Danie Els, the |booktabs| package. Publication
1697 % quality tables in \LaTeX: \url{www.ctan.org/pkg/booktabs}.
1698 %\bibitem{dcolumn} David Carlisle, the |dcolumn| package: \url{www.ctan.org/pkg/dcolumn}.
1699 %\bibitem{longtable} David Carlisle, the |longtable| package: \url{www.ctan.org/pkg/longtable}.
1700 %\bibitem{supertabular} Theo Jurriens ^^A I checked with Theo that `Jurriens' is correct; not `Jurri\"ens' (EHTO).
1701 % and Johannes Braams, the |supertabular| package, to be found at: \url{www.ctan.org/pkg/supertabular}.
1702 %\bibitem{tagging} Ulrike Fischer, the |tagpdf| package: \url{www.ctan.org/pkg/tagpdf}.
1703 %\bibitem{color} David Carlisle, the |color| package: \url{www.ctan.org/pkg/color}.
1704 %\bibitem{CandTB} Donald Knuth, {\it Computers \& Typesetting/B, ``\TeX: the program,''\/}
1705 % Addison-Wesley, Reading (1991).
1706 %\bibitem{CandTA} Donald Knuth, {\it Computers \& Typesetting/A, ``The \TeX book,''\/}
1707 % Addison-Wesley, Reading (1991).
1708 %\bibitem{TAOCP1} Donald Knuth, {\it The Art of Computers Programming, volume~1,
1709 % ``Fundamental Algorithms,''\/} Addison-Wesley, Reading (1997).
1710 %\bibitem{RJ} William Shakespeare, Romeo and Juliet, a tragedy (1597).
1711 %\bibitem{hook} Frank Mittelbach, ``\LaTeX's hook management'',
1712 % \url{www.latex-project.org/help/documentation/lthooks-doc.pdf}.
1713 %\bibitem{LTN43} \LaTeX-News 43: \url{www.latex-project.org/news/latex2e-news/ltnews43.pdf}.
1714 %\bibitem{caption} Axel Sommerfeldt, the |caption| package: \url{www.ctan.org/pkg/caption}.
1715 %\bibitem{infinite} See discussion at \url{https://github.com/latex3/latex2e/issues/1907}.
1716 %\end{thebibliography}
1717 %\Finale
1718 \endinput