summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/fcolumn/fcolumn.dtx
blob: 5dcef07b612b41394999499cbc5956e67f922d67 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
% \iffalse meta-comment
% Copyright (C) 2015--2019 by Edgar Olthof
%
% This file may be distributed and/or modified under the
% conditions of the LaTeX Project Public License, either
% version 1.3 of this license or (at your option) any later
% version.  The latest version of this license is in
%
%    http://www.latex-project.org/lppl.txt
%
% and version 1.3 or later is part of all distributions of
% LaTeX version 1999/12/01 or later.
%
% \fi
% \iffalse
%<package>\ProvidesPackage{fcolumn}
%<package>          [2019/04/24 v1.2 Automatic Currency Package (EHTO)]
%<package>\NeedsTeXFormat{LaTeX2e}
%<package>\RequirePackage{array}[v2.4k]
%<*driver>
\documentclass{ltxdoc}
\usepackage{fcolumn}          
\usepackage{booktabs}
\usepackage[tableposition=top,aboveskip=0pt]{caption}
\DeclareCaptionStyle{normal}{labelfont={rm,bf,small},textfont={rm,small},%
indention=0pt,labelsep=colon,justification=raggedright}
\captionsetup[table]{style=normal}
\captionsetup{singlelinecheck=false}
\usepackage{paralist}
\usepackage{url}
\usepackage[hyperfootnotes=false,hidelinks]{hyperref}
\usepackage{eurosym}
\EnableCrossrefs
\CodelineIndex
\RecordChanges
\begin{document}
   \DocInput{fcolumn.dtx}
\end{document}
%</driver>
% \fi
% \CheckSum{1078}
% \CharacterTable
%  {Upper-case    \A\B\C\D\E\F\G\H\I\J\K\L\M\N\O\P\Q\R\S\T\U\V\W\X\Y\Z
%   Lower-case    \a\b\c\d\e\f\g\h\i\j\k\l\m\n\o\p\q\r\s\t\u\v\w\x\y\z
%   Digits        \0\1\2\3\4\5\6\7\8\9
%   Exclamation   \!     Double quote  \"     Hash (number) \#
%   Dollar        \$     Percent       \%     Ampersand     \&
%   Acute accent  \'     Left paren    \(     Right paren   \)
%   Asterisk      \*     Plus          \+     Comma         \,
%   Minus         \-     Point         \.     Solidus       \/
%   Colon         \:     Semicolon     \;     Less than     \<
%   Equals        \=     Greater than  \>     Question mark \?
%   Commercial at \@     Left bracket  \[     Backslash     \\
%   Right bracket \]     Circumflex    \^     Underscore    \_
%   Grave accent  \`     Left brace    \{     Vertical bar  \|
%   Right brace   \}     Tilde         \~}
%
% \changes{v0.1}{2015/02/06}{First working version.}
% \changes{v1.0}{2015/03/07}{Three-argument version is working properly.}
% \changes{v1.1}{2015/09/14}{Automatic checking of column balance performed when number of
%   F-columns is even (behaviour can be overridden).  Empty entries are now recognised and 
%   correctly treated as such, except for the one ended by the double backslash. Not serious;
%   workaround possible. Furthermore optimisation of code: minimised the number of private 
%   counts and resetting of column counter done in a nicer way.}
% \changes{v1.1.1}{2015/09/27}{Installation procedure changed from .ins-in-.dtx to separate
%   .ins and .dtx after discussion with Karl Berry as well as some minor code improvements.}
% \changes{v1.1.2}{2015/09/29}{Some inconsistencies between explanatory text 
%   and actual code removed.}
% \changes{v1.2}{2019/03/06}{Input parsing changed after comment from Frank
%   Mittelbach.  He (Frank) also gave various suggestions for improving robustness
%   or user friendlyness of this package.  This version is only backwards compatible
%   when zero decimal digits were and are specified as modifier.}
% \GetFileInfo{fcolumn.sty}
% \DoNotIndex{\@arstrut,\@arstrutbox,\@depth,\@firstampfalse,\@firstamptrue,\@gobble,\@halignto}
% \DoNotIndex{\@height,\@mkpream,\@ne,\@nextchar,\@preamble,\@sharp,\@tempdima,\@tempswafalse}
% \DoNotIndex{\@tempswatrue,\@\temptokena,\@tfor,\@whilesw,\@width,\advance,\arraystretch,\baselineskip}
% \DoNotIndex{\begingroup,\bgroup,\count,\count@,\cr,\csname,\def,\divide,\do,\dp,\edef,\egroup}
% \DoNotIndex{\else,\endcsname,\endgroup,\expandafter,\extrarowheight,\fi,\gdef,\halignto,\hbox}
% \DoNotIndex{\hfil,\ht,\if,\if@firstamp,\if@tempswa,\ifcase,\lineskip,\loop,\m@ne,\m@th,\message}
% \DoNotIndex{\multiply,\or,\par,\prepnext@tok,\protect,\repeat,\setbox,\space,\string,\strutbox}
% \DoNotIndex{\tabskip,\tabularnewline,\the,\the@toks,\thr@@,\tw@,\unhbox,\vbox,\vcenter,\vrule}
% \DoNotIndex{\vskip,\vtop,\wd,\xdef,\z@,\z@skip}
% \title{The \textsf{fcolumn} package\thanks{This file has version
% number \fileversion, last revised \filedate.}}
% \author{Edgar Olthof\\
% \texttt{edgar <dot> olthof <at> inter <dot> nl <dot> net}}
% \date{Printed \today}
% \maketitle
% \begin{abstract}
% In financial reports, text and currency amounts are regularly put in
% one table, e.g., a year balance or a profit-and-loss overview.  This
% package provides the settings for automatically typesetting and checking
% such columns, including the sum line (preceded by a rule of the correct
% width), using the specifier |f|.
% \end{abstract}
%
% \section{Introduction}\label{intro}
% The package |fcolumn| provides the macros for an extra tabular specifier
% that makes creating financial tables easy.  The column specifier |f| 
% itself is rather simple; it is the predefined version of a generic 
% column |F|.  The generic version expects three arguments: 1)~grouping 
% character of the integer part on output, 2)~decimal mark used on output,
% and 3)~compact additional information on input/output characteristics,
% see below.
%
% The f-column in the current version of the package is defined for the 
% continental European standard: |\newcolumntype{f}{F{.}{,}{3,2}}|.  This
% means that a number like 12345,67 will be typeset as $12{.}345{,}67$.  
% People in the Anglo-saxon world would rather code 
% |\newcolumntype{f}{F{,}{.}{3,2}}| for the same input, yielding 
% $12{,}345{.}67$ as output for the number given above.  The default
% value for |#3| is |3,2|, indicating
% that grouping of the integer part is by three digits, that a comma is
% used in the \TeX-source to indicate the decimal separator, and that
% the decimal part consists of two digits. 
% If however, in your country or company grouping is done with a thinspace every 
% four digits, that the separator in the source should be the character |p|,
% and there are three digits after the decimal mark---that happens to be a 
% |\cdot|---, then simply specify |\newcolumntype{f}{F{\,}{\cdot}{4p3}}| in 
% that case.  The input could be 12345p678 then, yielding $1{\,}2345{\cdot}678$
% as output.
%
% By default two digits are used for the decimal part, so if you really
% want no decimal digits (in that case of course also skipping the decimal
% mark) you have to explicitly specify |x,0|. If you want no grouping
% character, specify |0,x|.
%
% This package requires and loads the |array| package~\cite{array}.  To
% show where and how the |F|-column is used, let's look at some typical 
% financial information as shown in Table~\ref{tab:ex1}
% \begin{table}[htb]
% \caption{Example Table.}
% \label{tab:ex1}
% \begin{tabular}{@{}lflf@{}}
% \multicolumn4c{\bfseries Balance sheet}\\
% \toprule
% properties & \leeg{31 dec 2014} & debts & \leeg{31 dec 2014}\\
% \midrule
% house &      200000    & equity capital&  50000    \\
% bank account & -603,23 & mortgage &      150000    \\
% savings &     28000    & &                         \\
% cash &          145,85 & profit  &        27542,62 \\
% \sumline
% \bottomrule
% \end{tabular}
% \end{table}
% and how this is entered in \LaTeX\ (Table~\ref{tab:ex2}).
% \begin{table}[htb]
% \caption{Verbatim version of Example Table.}
% \label{tab:ex2}
% \begin{verbatim}
%\begin{table}[htb]
%\caption{Example Table.}
%\label{tab:ex1}
%\begin{tabular}{@{}lflf@{}}
%\multicolumn4c{\bfseries Balance sheet}\\
%\toprule
%properties & \leeg{31 dec 2014} & debts & \leeg{31 dec 2014}\\
%\midrule
%house &      200000    & equity capital&  50000    \\
%bank account & -603,23 & mortgage &      150000    \\
%savings &     28000    & &                         \\
%cash &          145,85 & profit  &        27542,62 \\
%\sumline
%\bottomrule
%\end{tabular}
%\end{table}
% \end{verbatim}
% \end{table}
% All the work was done by the column specifier ``f'' (for ``finance'').  
% In this case it constructs the |\sumline|, typesets the numbers, 
% calculates the totals, determines the widths of the sumrules, and checks
% whether the two columns are in balance; if not, the user is warned via a
% |\PackageWarning|.  Of course for nice settings the |booktabs| 
% package~\cite{booktabs} was used, but that is not the point here.
%
% This package is heavily inspired by the |dcolumn| package by
% David Carlisle~\cite{dcolumn}, some constructions are more or less 
% copied from that package.  Version~1.2 (this version) uses a few ideas 
% from, and incorporates quite some suggestions by, Frank Mittelbach.
%
% \section{Commands}
% The user only needs to know six commands or constructions.  These six are given here.
% \begin{macro}{F}
% In the tabular the column specifier |F| can be given with arguments, or 
% the predefined version~|f|, where the three arguments of |F| are |{.}|, 
% |{,}|, and |{3,2}|.  If you want |g| to be your own definition like the 
% curious one given in Section~\ref{intro}, then specify 
% |\newcolumntype{g}{F{\,}{\cdot}{4p3}}| prior to using |g| in a tabular.
% 
% Entries in an F-column are, from that moment on, treated as numbers unless
% explicitly escaped by |\leeg|, see below.  The numbers are typeset 
% according to the template the user gives with his/her |F|-column.
% The ``middle'' character of |#3| is an important switch: it does more
% than just setting the input decimal mark.  By default the input grouping 
% character is the dot, except when the dot is specified as input decimal
% mark; in that case the comma is acting as input grouping character.  With 
% this convention the continental Europe and Anglo-saxon part of the world 
% is served. 
% \end{macro}
% \begin{macro}{\sumline}
% The numbers in an |F|-column are typeset as a financial amount, but the real
% benefit comes with the |\sumline|. It does three things:
% \begin{compactenum}[1)]
% \item It calculates the total of the column so far and the maximum width 
% encountered so far, including the width of the total;
% \item It generates a rule with width calculated in the first item;
% \item It checks the columns that are supposed to balance whether or not they
% actually do.  If so, nothing happens.  If not, a |\PackageWarning| is given 
% that column $i$ and $j$ do not balance, where $i$ and $j$ are the relevant 
% columns.  This is only done if the total number of |F|-columns is even, 
% e.g., if there are six |F|-columns, then~1 is checked against~4, 2 against~5, 
% and 3 against~6.  If the number of |F|-columns is odd then anything could be
% possible in that table and nothing is assumed about structure within the 
% table.  This behaviour can be overridden, see below.
% \end{compactenum}
% By default the vertical separation between the rule and the total is 
% 2\thinspace pt, but this can be changed by the optional argument to 
% |\sumline|. Give, e.g., |\sumline[10pt]|, in case you want this spacing
% to be 10\thinspace pt.
% \end{macro}
% \begin{macro}{\resetsumline}
% Suppose you want to typeset one tabular with the profit-and-loss of many
% projects individually.  The layout of those tabulars is the same and it 
% were nice if all columns were aligned.  This can be done by making it one
% big tabular with a fresh start for each project.  The macro 
% |\resetsumlines| is used for that: it resets all totals and all column 
% widths, see for example Table~\ref{tab:ex3}.
% \begin{table}[htb]
% \caption{Example: multiple projects.}
% \label{tab:ex3}
% \begin{tabular}{@{}lfflff@{}}
% \multicolumn6c{\bfseries Project~1}\\
% \toprule
% expense & \leeg{actual} & \leeg{budget} & 
% income & \leeg{actual} & \leeg{budget} \\
% \midrule
% food &         450,20 & 500    & tickets &  1200    &  1000    \\
% drinks &       547,5  & 400    \\
% music &        180    & 100    \\
% profit &        22,3  \\
% \sumline
% \resetsumline
% \multicolumn6c{\bfseries Project~2}\\
% \toprule
% expense & \leeg{actual} & \leeg{budget} & 
% income & \leeg{actual} & \leeg{budget} \\
% \midrule
% food &         250    & 300    & tickets &  400    &  450    \\
% drinks &       100    &  80    \\
% music &         80    &  70    & loss &      30    \\
% \sumline
% \bottomrule
% \end{tabular}
% \end{table}
% Note that the rules in the first and third |F|-columns of project~1 cover
% $1{.}200{,}00$ whereas in project~2 those rules are narrower since they 
% only cover $430{,}00$; still the columns are aligned.  The verbatim way
% of setting up Table~\ref{tab:ex3} is given in Table~\ref{tab:ex4}.
% \begin{table}[htb]
% \caption{Verbatim version of Table~\ref{tab:ex3}.}
% \label{tab:ex4}
% \begin{verbatim}
%\begin{table}[htb]
%\caption{Example: multiple projects.}
%\label{tab:ex3}
%\begin{tabular}{@{}lfflff@{}}
%\multicolumn6c{\bfseries Project~1}\\
%\toprule
%expense & \leeg{actual} & \leeg{budget} & 
%income & \leeg{actual} & \leeg{budget} \\
%\midrule
%food &         450,2  & 500    & tickets &  1200    &  1000    \\
%drinks &       547,5  & 400    \\
%music &        180    & 100    \\
%profit &        22,3  \\
%\sumline
%\resetsumline
%\multicolumn6c{\bfseries Project~2}\\
%\toprule
%expense & \leeg{actual} & \leeg{budget} & 
%income & \leeg{actual} & \leeg{budget} \\
%\midrule
%food &         250    & 300    & tickets &  400    &  450    \\
%drinks &       100    &  80    \\
%music &         80    &  70    & loss &      30    \\
%\sumline
%\bottomrule
%\end{tabular}
%\end{table}
% \end{verbatim}
% \end{table}
% \end{macro}
% \begin{macro}{\leeg}
% If an |F|-column should be empty then simply leave it empty.  If however
% it should  not be empty but the entry should be treated as text---even 
% if it is a number---, this can be done with |\leeg|.  It expects an 
% argument and this argument is typeset in the column.  The common case is
% where |p.m.| (\textit{pro memoria}) is entered.  In contrast to v1.1.2 
% of this package, now even an empty |F|-column followed by |\\| is allowed.
% \end{macro}
% \begin{macro}{\checkfcolumns}
% The automatic column balance check can also be done manually.  If 
% |F|-columns~1 and~4 should balance and you want them to be checked,
% then simply say |\checkfcolumns14|.  With more than nine |F|-columns
% you may be forced to say something like |\checkfcolumns{10}{12}|. If
% |\checkfcolumns| is used, the automatic check is disabled.  Multiple
% |\checkfcolumns|s are supported; if |F|-columns~1, 2, and~3 should 
% balance, you specify |\checkfcolumns12| and |\checkfcolumns23|.
% There is no explicit command to disable all checking, but
% |\checkfcolumns11| obviously also serves that purpose.
% \end{macro}
% \begin{macro}{\ifstrict@ccounting}
% In the rare occasion that a negative number occurs in a financial table,
% the sign of that number can be an explicit minus sign ($-$) or the number
% is coloured red, or it is typeset between parentheses, and there may be
% even other ways.  By default (for aesthetic reasons) |fcolumn| typesets
% it with a minus~sign, but strict accounting prescibes that the number 
% should be put between parentheses.  The latter can be accomplished by 
% setting |\strict@ccountingtrue|. 
% \end{macro}
% \StopEventually{\PrintChanges\PrintIndex}
% \section{The macros}
% Here follows the actual code.
% \subsection{Option}
% \begin{macro}{option strict}
% There is one option. If set, strict accounting rules are used in display.
%    \begin{macrocode}
\newif\ifstrict@ccounting \strict@ccountingfalse
\DeclareOption{strict}{\strict@ccountingtrue}
\ProcessOptions
%    \end{macrocode}
% \end{macro}
%\subsection{Definitions}
% \begin{macro}{column F}
% \begin{macro}{column f}
% The column specifier |F| is the generic one, and |f| is the default 
% (continental European) one for easy use. Note that the definition of the
% column type |f| does not use private macros (no~|@|), so overriding its 
% definition is easy for a user.
%    \begin{macrocode}
\newcolumntype{F}[3]{>{\b@fi{#1}{#2}{#3}}r<{\e@fi}}
\newcolumntype{f}{F{.}{,}{3,2}}
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \begin{macro}{\FCsc@l}
% \begin{macro}{\FCtc@l}
% Two \meta{count}s are defined, that both start at zero: the \meta{count} 
% |\FCsc@l|, that keeps track at which F-column the tabular is working on 
% and the \meta{count} |\FCtc@l|, that records the number of F-columns 
% that were encountered so far.  Later in the package the code can be 
% found for generating a new \meta{count} and a new \meta{dimen} if the
% number of requested |F|-columns is larger than currently available.  
% This is of course the case when an |F|-column is used for the first time.
%    \begin{macrocode}
\newcount\FCsc@l \FCsc@l=0 \newcount\FCtc@l \FCtc@l=0
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \begin{macro}{\geldm@cro}
% The macro |\geldm@cro| takes a number and by default interprets this as
% an amount expressed in cents (dollar cents, euro cents, centen, Pfennige,
% centimes, kopecks, groszy) and typesets it as the amount in entire 
% currency units (dollars, euros, guldens, Marke, francs, rubles, z\l oty)
% with comma as decimal separator and the dot as grouping character
% (thousand separator if the first part of |#1| is~3).
% As explained, this can be changed.  It uses two private booleans: 
% |\withs@p| and |\strict@ccounting|.  The latter is used to typeset
% negative numbers between parentheses.  By default it doesn't do this:
% a minus sign is used.
%    \begin{macrocode}
\newif\ifwiths@p
%    \end{macrocode}
% Actually |\geldm@cro| is only a wrapper around |\g@ldm@cro|.
%    \begin{macrocode}
\def\geldm@cro#1#2{\withs@pfalse 
\afterassignment\g@ldm@cro\count@#1\relax{#2}}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@ldm@cro}
% This macro starts by looking at the sign of |#2|: if it is negative, it prints 
% the correct indicator (a parenthesis or a minus sign), assigns the absolute value 
% of~|#2| to |\count2| and goes on.  Note that |\geldm@cro| and therefore |\g@ldm@cro|
% are always used within |$|s, so it is really a minus sign that is printed, not a hyphen.
% All calculations are done with |\count0|, |\count1|, etc.\ i.e., without |F|-column-specific
% \meta{count}s because it is all done locally.  Leaving the tabular environment will 
% restore their values.
%    \begin{macrocode}
\def\g@ldm@cro#1\relax#2{\ifnum#2<0 \ifstrict@ccounting (\else -\fi
  \count2=-#2 \else\count2=#2 
\fi
%    \end{macrocode}
% Calculate the entire currency units: this is the result of $x/a$ as integer
% division, with $a=10^n$ and $n$ the part of |#1| after the separator (if any).
% Here the first character of |#1| is discarded, so the separator in |#1| is not strict:
% you could also specify |3.2| instead of |3,2| (or even |3p2|).
%   \begin{macrocode}
\count4=\ifx\relax#1\relax 2 \else \@gobble#1\relax\fi
\count3=0
\loop\ifnum\count3<\count4 
  \divide\count2 by 10 \advance\count3 by \@ne
\repeat
%    \end{macrocode}
% Note that |\count3| now equals |\count4|: this going up-and-down will be used more often,
% it saves several assignments.  The value in |\count2| is then output by |\g@ldens| using 
% the separation given (and stored in |\count@|).
%    \begin{macrocode}
\g@ldens{\the\count@}%
%    \end{macrocode}
% If there is a decimal part\dots
%    \begin{macrocode}
\ifnum\count3>0\decim@lmark
%    \end{macrocode}
% Next the decimal part is dealt with.  Now $x\bmod a$ is calculated in the 
% usual way: $x-(x/a)*a$ with integer division.  The minus sign necessary 
% for this calculation is introduced in the next line by changing the 
% comparison from |<| to |>|.
%    \begin{macrocode}
  \ifnum#2>0 \count2=-#2\else\count2=#2 \fi
  \loop\ifnum\count3>0
    \divide\count2 by 10 \advance\count3 by \m@ne
  \repeat
%    \end{macrocode}
% The value of |\count3| is now~0, so counting up again.
%    \begin{macrocode}
  \loop\ifnum\count3<\count4
    \multiply\count2 by 10 \advance\count3 by \@ne
  \repeat
  \ifnum#2>0 \advance\count2 by #2
  \else \advance\count2 by -#2
  \fi
  \zerop@d{\number\count3}{\number\count2}%
\fi
%    \end{macrocode}
% If the negative number is indicated by putting it between parentheses,
% then the closing parenthesis should stick out of the column, otherwise
% the alignment of this entry in the column is wrong.  This is done by an
% |\rlap| and therefore does not influence the column width.  For the last
% column this means that this parenthesis may even stick out of the table.
% I don't like this, therefore I chose to put |\strict@ccountingfalse|.
% Change if you like, by setting the option |strict|.
% 
% If overflow was detected, an exclamation mark is output to the right of
% the value that caused this.  This of course ruins the appearance of the
% table, but in this case that serves a clear goal: there's something wrong
% and you should know.
%    \begin{macrocode}
\ifx\FCs@gn\m@ne \ifnum#2<0 \ifstrict@ccounting
  \rlap{)~!}\else\rlap{~!}\fi\else\ifstrict@ccounting
  \rlap{\phantom{)}~!}\else\rlap{~!}\fi\fi
\else \ifnum#2<0 \ifstrict@ccounting\rlap{)}\fi\fi
\fi}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@ldens}
% Here the whole currency units are dealt with.  The macro |\g@ldens| is 
% used recursively, therefore the double braces; this allows to use 
% |\count0| locally.  This also implies that tail recursion is not 
% possible here, but that is not very important, as the largest number
% (which is $2^{31}-1$) will only cause a threefold recursion using the
% default |3,2| (ninefold when using |1,0|, but who does that?).  The 
% largest amount this package can deal with is therefore 2.147.483.647 
% (using |3,0|).  For most people this is probably more than enough if
% the currency is euros or dollars.  And otherwise make clear that you
% use a currency unit of~k\EUR{} (or even M\EUR{} for the very rich).
% The author is thinking of ways to use two counters for each number.
% The maximum then becomes $2^{63}-1$.  Even expressed in cents this 
% would lead to a maximum of slightly more than 92.2\thinspace P\EUR{};
% about 100 times the current world economy~\cite{world_economy}. Yet
% another method is to used Heiko Oberdiek's package |bigintcalc|:
% then only memory restrictions apply.  This, however, requires a major
% rewrite of |fcolumn|. For now, version~1.2 sticks to the moderate amounts.
%
% There is no straightforward interpretation of |#1| being zero or 
% negative, therefore this is used as an indicator that no grouping
% character should be used.
%    \begin{macrocode}
\def\g@ldens#1{{\count3=\count2 \count0=#1
%    \end{macrocode}
% First divide by $10^n$, where $n$ is |#1|.
%    \begin{macrocode}
\ifnum\count0<1 \count0=3 \fi
\loop \ifnum\count0>0 \divide\count2 by 10 \advance\count0 by \m@ne
\repeat
%    \end{macrocode}
% Here is the recursive part,
%    \begin{macrocode}
\ifnum\count2>0 \g@ldens{#1}\fi
%    \end{macrocode}
% and then reconstruct the rest of the number.
%    \begin{macrocode}
\count0=#1
\ifnum\count0<1 \count0=3 \fi
\loop \ifnum\count0>0 \multiply\count2 by 10 \advance\count0 by \m@ne
\repeat
\count2=-\count2
\advance\count2 by \count3 \du@zendprint{#1}}}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\du@zendprint}
% The macro |\du@zendprint| takes care for correctly printing the separator
% and possible trailing zeros. The former, however, is only done if |#1|
% is larger than zero.
%    \begin{macrocode}
\def\du@zendprint#1{\ifwiths@p\ifnum#1>0 \sep@rator\fi
 \zerop@d{#1}{\number\count2}%
\else\zerop@d1{\number\count2}\fi\global\withs@ptrue}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\zerop@d}
% The macro |\zerop@d| uses at least |#1| digits for printing the 
% number |#2|, padding with zeros when necessary. Note: |#1| being
% zero or negative is a flag that it should be interpreted as 3.
% A bit ugly, but it works, since the related code knows about this.
%
% It is done within an extra pair of braces, so that |\count0| 
% and |\count1| can be used without disturbing their values in other macros.
%    \begin{macrocode}
\def\zerop@d#1#2{{\count0=1 \count1=#2
%    \end{macrocode}
% First determine the number of digits of |#2| (expressed in the decimal
% system).  This number is in |\count0| and is at least~1.
%    \begin{macrocode}
\loop \divide \count1 by 10 \ifnum\count1>0 \advance\count0 by \@ne
\repeat
%    \end{macrocode}
% If |#1| is positive, the number of zeros to be padded is 
% $\max(0,|#1-\count0|)$ (the second argument can be negative), so a
% simple loop suffices.  If it is zero or negative, this is a signal
% that it should be interpreted as 3 (and no separator will be output).
%    \begin{macrocode}
\ifnum#1>0
 \loop \ifnum\count0<#1\relax 0\advance\count0 by \@ne
 \repeat 
\else
 \advance\count0 by -3
 \loop \ifnum\count0<0 0\advance\count0 by \@ne
 \repeat 
\fi\number#2}}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\zetg@ld}
% This macro takes care for several things: it increases the subtotal for a
% given F-column, it checks whether or not that subtotal has overflown, it
% records the largest width of the entries in that column and it typesets 
% |#1| via |\geldm@cro|.
%    \begin{macrocode}
\def\zetg@ld#1#2{\count0=#2\relax \let\FCs@gn=\@ne
%    \end{macrocode}
% First it checks whether there is a risk of overflow in this step.  If
% $A$ and $B$ are two \TeX-registers and $B$ is to be added to $A$, 
% overflow will not occur if one is (or both are) zero or if $A$ and~$B$
% have different signs.  Otherwise, be careful. Note that \TeX\ does not
% check for overflow when performing an |\advance| (done in section~1238
% of Ref.~\cite{CandTB}), in contrast to |\multiply|, see section~105.
%    \begin{macrocode}
\ifnum\count0<0 
  \ifnum\csname FCtot@\romannumeral\FCsc@l\endcsname<0
    \let\FCs@gn=\m@ne 
  \fi
\fi
\ifnum\count0>0
  \ifnum\csname FCtot@\romannumeral\FCsc@l\endcsname>0
    \let\FCs@gn=\m@ne 
  \fi
\fi
\global\advance\csname FCtot@\romannumeral\FCsc@l\endcsname by \count0
\ifx\FCs@gn\m@ne 
%    \end{macrocode}
% They had the same sign: risk of overflow. Record the sign
% of |\count0| (and of the original total of this column; they were 
% the same) in |\FCs@gn|. Table~\ref{tab:ex5} 
% \begin{table}[htb]
% \def\PackageError#1#2#3{\PackageWarning{#1}{#2}}
% \caption{Examples on overflow.}
% \label{tab:ex5}
% \begin{tabular}{@{}lfff@{}}
% \multicolumn4c{\bfseries Projects}\\
% \toprule
% income & \leeg{31 dec 2014} & \leeg{31 dec 2015} & \leeg{31 dec 2016} \\
% \midrule
% item~1 &  20000000 & 20000000 &20000000 \\
% item~2 &  10000000 &  2000000 &-2000000 \\
% item~3 &   5000000 & -2000000 & 2000000 \\
% \checkfcolumns23
% \sumline
% \bottomrule
% \end{tabular}
% \end{table}
% shows what can go wrong if the numbers are too large: in the left
% |F|-column the sumline is incorrect and the number that caused 
% the overflow is indicated by an exclamation mark. In the middle 
% |F|-column, overflow occurs twice and because this is once positive,
% once negative here, cancellation of errors occurs and the sumline
% is correct in the end.  Nevertheless, it is advised to swap the two
% items that caused the overflow, as shown in the right |F|-column.
%
% Since the absolute value of |\FCs@gn| is unity, no overflow will
% occur in the multiplication step below.
%    \begin{macrocode}
  \ifnum\count0>0 \let\FCs@gn\@ne \fi
  \count0=\csname FCtot@\romannumeral\FCsc@l\endcsname
  \multiply\count0 by \FCs@gn
  \ifnum\count0<0 
    \let\FCs@gn=\m@ne
    \PackageError{fcolumn}{Register overflow}{Overflow occurred
    in fcolumn \number\FCsc@l. Check your table.}%
  \else\let\FCs@gn=\@ne
  \fi
\fi
%    \end{macrocode}
% The value of |\FCs@gn| is used in |\geldm@cro| below.
%    \begin{macrocode}
\setbox0=\hbox{$\geldm@cro{#1}{#2}$}%
\ifdim\wd0>\csname FCwd@\romannumeral\FCsc@l\endcsname
 \global\csname FCwd@\romannumeral\FCsc@l\endcsname=\wd0
\fi\unhbox0}
%    \end{macrocode}
% \end{macro}
% \def\<#1>{\leavevmode\hbox{$\langle$#1\/$\rangle$}}
% \begin{macro}{\b@fi}
% The macro |\b@fi| provides the beginning of the financial column.  It
% will be inserted in the column to capture the number entered by the user
% The separator and decimal mark are within a math environment, so you
% can indeed specify |\,| instead of |\thinspace|, but there is an extra
% brace around, so it doesn't affect the spacing between the digits 
% (trick copied from |dcolumn|, Ref.~\cite{dcolumn}).
%
% The \meta{count}s |\FC@l| captures the part to the left of the decimal 
% marker, |\FC@r| that to the right.
%    \begin{macrocode}
\newcount\FC@l \newcount\FC@r
\def\FC@chklist{}
\def\setucc@de#1#2\relax{\uccode`\~=`#1 }
\def\b@fi#1#2#3{%
%    \end{macrocode}
% An intermediate macro |\sep@xt| to extract the first character
% of |#1|, which in most cases will be the only character.
%    \begin{macrocode}
\def\sep@xt##1##2\end{\def\sep@rator{{##1}}}%
\sep@xt#1\end\def\decim@lmark{{#2}}%
\def\sp@l{#3}\global\advance\FCsc@l by \@ne
\global\FC@l=0 \global\FC@r=1
%    \end{macrocode}
% The value specified by the user is then captured by |\FC@l| and 
% this is done in a special way: |\FC@l| is assigned globally within
% |\box0|.  Why?  To use it as scribbling paper to examine what the 
% user entered, without dumping it into the horizontal list.
%
% There are four parts to an |F|-column entry, all parts optional,
% making 16 combinations.  The sequence is
% (in the Backus--Naur notation of Ref.~\cite{CandTA}):
% \<sign> \<integer constant> \<decimal marker>
% \<integer constant>. Here \<sign> is a plus or minus character with 
% category code~12, \<integer constant> is a sequence of zero
% or more (decimal) \<digit>s, and \<decimal marker> is the middel 
% part of |#3|, i.e., the comma in |3,2| or the period in |3.2|.
% If the \<decimal marker> is absent with no space characters between 
% the two \<integer constant> terms, these merge, making four redundant
% entries.  One of the 
% combinations is \<empty>, a sequence of exactly zero non-space tokens:
% this is the only combination that doesn't put anything in an |F|-column.
%
% The minus sign must be captured separately, because in an entry like
% |-0,07| the 7~cents are negative, but this cannot be seen from the
% integer part, since $-0$ is $0$ in \TeX\ (in fact in most computer
% languages, but not in~MIX~\cite{TAOCP1}), so |\ifnum-0<0| yields false.
% |\FCs@gn| is a general purpose flag.  Its first use is to capture
% the sign.
%    \begin{macrocode}
\let\FCs@gn=\@ne\relax \setbox0\hbox\bgroup$%   
%    \end{macrocode}
% Do the scan inside a box and inside math mode.  Start with
% defining all characters that may appear as the first one in an |F|-column
% as active.
%    \begin{macrocode}
\uccode`\~=`0\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=0}
\uccode`\~=`1\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=1}
\uccode`\~=`2\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=2}
\uccode`\~=`3\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=3}
\uccode`\~=`4\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=4}
\uccode`\~=`5\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=5}
\uccode`\~=`6\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=6}
\uccode`\~=`7\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=7}
\uccode`\~=`8\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=8}
\uccode`\~=`9\relax \uppercase{\def~}{\restorem@thcodes \global\FC@l=9}
%    \end{macrocode}
% For the input decimal mark something extra is needed: if it is the
% first character in an |F|-column (like in |,07|), it should also
% restore the |\mathcode|s of the digits.  Checking whether or not it
% is the first is easy, since in that case the |\mathcode|s of the 
% decimal digits is still |"8000|.  The assignment to |\FC@r| starts 
% with~1, so that appended digits get captured correctly, even if they
% start with~0.  Postprocessing of |\FC@r| is done in |\e@fi|.
% The input decimal mark switches itself off as active character, 
% so at most one input decimal separator is allowed (N.B.:
% this makes sense).
%    \begin{macrocode}
\def\deactdecm@rk##1##2\relax{\mathcode`##1=0 }%
\afterassignment\setucc@de\count@#3\relax
\uppercase{\def~}{\ifnum\mathcode`\0=\mathcode`- \restorem@thcodes\fi
\afterassignment\deactdecm@rk\count@#3\relax \global\FC@r=1}%
%    \end{macrocode}
% The input grouping character effectively expands to ``nothing, i.e.,
% ignore'' in a complicated way: it ignores the character and resumes 
% scanning the number.  The test prior to that action is needed if
% the grouping character is the first character encountered in the
% |F|-column.  Which part to continue with depends on whether or not an
% input decimal mark was encountered; that can be checked by looking
% at its |\mathcode|.
% 
% The input grouping character is the dot |.|, except when that character
% was already chosen as input decimal mark.  In that case, the grouping 
% character will be the comma. This is easy to check because the |\uccode|
% of |`\~| is still preserved.  
%    \begin{macrocode}
\def\d@cm##1##2{\count@=\mathcode`##1 }
\ifnum\uccode`\~=`. \uccode`\~=`,\relax\else \uccode`\~=`.\relax\fi
\uppercase{\def~}{\ifnum\mathcode`\0=\mathcode`- \restorem@thcodes\fi
\afterassignment\d@cm\count@#3\relax
%    \end{macrocode}
% The |\expandafter| below is necessary because the global assignment
% should act after the |\fi|.
%    \begin{macrocode}
\ifnum\count@=\mathcode`- \expandafter\global\FC@l=\the\FC@l
\else \expandafter\global\FC@r=\the\FC@r\fi}%
%    \end{macrocode}
% The signs are relatively simply: record the sign, restore |\mathcode|s if
% needed (it should be: a minus sign between digits screws up everything),
% and start scanning the number.
%    \begin{macrocode}
\uccode`\~=`+\relax
\uppercase{\def~}{\ifnum\mathcode`\0=\mathcode`-
\restorem@thcodes\fi\global\FC@l=0}%
\uccode`\~=`-\relax 
\uppercase{\def~}{\ifnum\mathcode`\0=\mathcode`-
\restorem@thcodes\fi\global\let\FCs@gn\m@ne \global\FC@l=0}%
%    \end{macrocode}
% Now actually activate all these codes.  The first is simple, but
% after that, one can't say |"8000| anymore because |0| acts as active.
% But copying |\mathcode|s still works.
%    \begin{macrocode}
\mathcode`-="8000 \mathcode`+=\mathcode`- \mathcode`.=\mathcode`- 
%    \end{macrocode}
% These three remain active until the |$| in |\e@fi| is encountered.
% The following ones will, in the general case, have their activeness 
% turned off at some time.
%    \begin{macrocode}
\def\actdecm@rk##1##2\relax{\ifx##1.\relax \mathcode`,=\mathcode`-
\else \mathcode`##1=\mathcode`- \fi}%
\afterassignment\actdecm@rk\count@#3\relax
\mathcode`\0=\mathcode`- \mathcode`\1=\mathcode`- 
\mathcode`\2=\mathcode`- \mathcode`\3=\mathcode`- 
\mathcode`\4=\mathcode`- \mathcode`\5=\mathcode`- 
\mathcode`\6=\mathcode`- \mathcode`\7=\mathcode`- 
\mathcode`\8=\mathcode`- \mathcode`\9=\mathcode`- }
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\e@fi}
% If the digits are still active then either nothing was entered or only
% characters that did not deactivate the digits were entered.  In 
% either case the output should be \<empty>.  To flag this outside
% the group that started with the opening |$| of |\b@fi|, |\FC@r| is set
% globally to a negative value.  This doesn't harm, because it didn't 
% contain relevant information anyway.  Outside the group, the sign
% of |\FC@r| can then be tested.  This is a slight misuse of this
% |\count|, but now it's documented.
%
% If there was no decimal sign or if there was a decimal sign but no 
% decimal part, |\FC@r| will still be~1, which doesn't parse well with
% |\secd@xt|, so a zero is appended.
%    \begin{macrocode}
\def\e@fi{\ifnum\mathcode`\0=\mathcode`- \global\FC@r=\m@ne\fi$\egroup
\ifnum\FC@r>0
 \ifnum\FC@r<10 \multiply\FC@r by 10 \fi
%    \end{macrocode}
% Next is a loop for bringing the decimal part in the correct way to the 
% the integer part.  The loop is performed the number of decimal
% digits to be printed (the~2 in |3,2| of the default setting).  This also
% means that if you specified more decimal digits than this, the excess 
% digit(s) will not be handled and a |\PackageWarning| will be given.
%    \begin{macrocode}
 \def\i@ts##1##2{\count0=##2}
 \afterassignment\i@ts\count@\sp@l
 \loop\ifnum\count0>0 \multiply\FC@l by 10
  \expandafter\secd@xt\number\FC@r\end \advance\count0 by \m@ne
 \repeat
 \ifnum\FC@r>10
   \def\tw@l##1##2\relax{##2}
   \PackageWarning{fcolumn}{Excess digit\ifnum\FC@r>100 s\fi\space
   ``\expandafter\tw@l\number\FC@r\relax'' in decimal part
   \MessageBreak ignored}
 \fi
%    \end{macrocode}
% Don't forget to correct for the sign (once this is done, |\FCs@gn|
% is free again and can and will be used for other purposes).
% Then output the result.
%    \begin{macrocode}
 \ifx\FCs@gn\m@ne\relax \FC@l=-\FC@l \fi
 \zetg@ld{\sp@l}{\FC@l}%
\fi}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\secd@xt}
% The second digit from the left is needed from a decimal number. The
% macro |\secd@xt| extracts that digit, provided that the number has at
% least two digits, but that is guaranteed by |\e@fi|.  That second digit
% is then added to |\FC@l|.  A new number is assigned to |\FC@r|, that 
% consists of the digits of |#1#3|.  If |#3| was empty, a zero is appended.
% In this way |\FC@r| is prepared for insertion in the next invocation of 
% |\secd@xt|.  In iterating: 1234 yields~134, yields~14, yields~10,
% stays~10, etc.
%    \begin{macrocode}
\def\secd@xt#1#2#3\end{\advance\FC@l by #2
\FC@r=#1#3 \ifnum\FC@r<10 \multiply\FC@r by 10 \fi}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\restorem@thcodes}
% As shown above, once the first digit, or sign, or decimal separator, or
% grouping character is scanned, the decimal digits should loose their
% activeness.  That is done here rather blunt, since the actual 
% |\mathcode| is not important---as long as it is not |"8000|---because
% the digits are not used for typesetting (and even if they were; it's
% inside |\box0|, whose contents will be discarded).  When the |$| 
% in |\e@fi| is encountered, the digits get back their original
% |\mathcode|s so that the actual typesetting in |\zetg@ld| is
% correct again.
%    \begin{macrocode}
\def\restorem@thcodes{\mathcode`\0=0 \mathcode`\1=0 
\mathcode`\2=0 \mathcode`\3=0 \mathcode`\4=0 \mathcode`\5=0
\mathcode`\6=0 \mathcode`\7=0 \mathcode`\8=0 \mathcode`\9=0 }
%    \end{macrocode}
% \end{macro}
% \subsection{Adaptations to existing macros}
% \begin{macro}{\@array}
% The definition of |\@array| had to be extended slightly because it should
% also include |\@mksumline| (acting on the same |#2| as |\@mkpream| gets).
% This change is transparant: it only adds functionality and if you don't 
% use that, you won't notice the difference.  It starts by just copying the
% original definition from v2.4k of the |array| package~\cite{array}.
%    \begin{macrocode}
\def\@array[#1]#2{%
\@tempdima \ht \strutbox
\advance \@tempdima by\extrarowheight
\setbox \@arstrutbox \hbox{\vrule
           \@height \arraystretch \@tempdima
           \@depth \arraystretch \dp \strutbox
           \@width \z@}%
%    \end{macrocode}
% Here comes the first change: after each |\\| (or |\cr| for that matter)
% the \meta{count} |\FCsc@l| should be reset.  This is easiest done with
% |\everycr|, but |\everycr| is put to |{}| by |\ialign|, so that definition
% should change.  The resetting should be done globally.
%    \begin{macrocode}
\def\ialign{\everycr{\noalign{\global\FCsc@l=0 }}%
 \tabskip\z@skip\halign}
%    \end{macrocode}
% Then the definition is picked up again.
%    \begin{macrocode}
\begingroup
\@mkpream{#2}%
\xdef\@preamble{\noexpand \ialign \@halignto
\bgroup\@arstrut\@preamble\tabskip\z@\cr}%
\endgroup 
%    \end{macrocode}
% The combination |\endgroup| followed by |\begingroup| seems redundant, 
% but that is not the case: the |\endgroup| restores everything that was 
% not |\global|.  With the following |\begingroup| it is ensured that 
% |\@mksumline| experiences the same settings as |\@mkpream| did.
%    \begin{macrocode}
\begingroup
\@mksumline{#2}%
\endgroup
%    \end{macrocode}
% As a side product of |\@mksumline| also the \meta{count}s for the totals 
% and \meta{dimen}s for the widths of the colums are created.  The columns 
% should start fresh, i.e., totals are~0 and widths are 0\thinspace pt.
%    \begin{macrocode}
\res@tsumline
%    \end{macrocode}
% From here on it is just the old definition of |array.sty|.
%    \begin{macrocode}
\@arrayleft
\if #1t\vtop \else \if#1b\vbox \else \vcenter \fi \fi
\bgroup
\let \@sharp ##\let \protect \relax
\lineskip \z@
\baselineskip \z@
\m@th
\let\\\@arraycr \let\tabularnewline\\\let\par\@empty \@preamble} 
%    \end{macrocode}
% Because |\@array| was changed here and it is this version that should
% be used, |\@@array| should be |\let| equal to |\@array| again.
%    \begin{macrocode}
\let\@@array=\@array
%    \end{macrocode}
% \end{macro}
% \subsection{The sumline, close to a postamble}
% \begin{macro}{\@mksumline}
% The construction of the sumline is much easier than that of the preamble
% for several reasons.  It may be safely assumed that the preamble specifier
% is grammatically correct because it has already been screened by |\@mkpream|.
% Furthermore most entries will simply add nothing to |\s@ml@ne|, e.g., |@|, |!|,
% and \verb+|+ can be fully ignored. Ampersands are only inserted by |c|, |l|, |r|,  
% |p|, |m|, and |b|.  So a specifier like |@{}lflf@{}| will yield the sumline
% |&\a&&\a\\|, (where |\a| is a macro that prints the desired result of the column, 
% see later).  Had the specifier been \verb+l|f||@{   }l|f+, then the same sumline must
% be constructed: all difficulties are already picked up and solved in the creation
% of the preamble.
%
% In reality the sumline must be constructed from the expanded form of the
% specifier, so |@{}lf@{}| will expand as
% |@{}l>{\b@fi{.}{,}{3,2}}r<{\e@fi}@{}|. The rules for constructing the 
% sumline are now very simple:
% \begin{compactitem}[$\bullet$]
% \item add an ampersand when |c|, |l|, |r|, |p|, |m|, or |b| is found, 
% unless it is the first one (this is the same as in the preamble);
% \item add a |\a| when |<{\e@fi}| is found;
% \item ignore everything else;
% \item close with a |\\|.
% \end{compactitem}
% (In reality also the column check is inserted just before the |\\|, see
% |\aut@check|.)  To discriminate, a special version of |\@testpach| could
% be written, but that is not necessary: |\@testpach| can do all the work,
% although much of it will be discarded.  Here speed is sacrificed for space
% and this can be afforded because the creation of the sumline is done only 
% once per |\tabular|.
% 
% The start is copied from |\@mkpream|.
%    \begin{macrocode}
\def\@mksumline#1{\gdef\s@ml@ne{}\@lastchclass 4 \@firstamptrue
%    \end{macrocode}
% At first the column number is reset and the actual code for what was
% called |\a| above is made inactive.
%    \begin{macrocode}
\global\FCsc@l=0
\let\prr@sult=\relax
%    \end{macrocode}
% Then |\@mkpream| is picked up again.
%    \begin{macrocode}
\@temptokena{#1}
\@tempswatrue
\@whilesw\if@tempswa\fi{\@tempswafalse\the\NC@list}%
\count0\m@ne
\let\the@toks\relax
\prepnext@tok
%    \end{macrocode}
% Next is the loop over all tokens in the expanded form of the specifier.  The change
% with respect to |\@mkpream| is that the body of the loop is now only dealing
% with F-classes~0, 2, and 10.  What to do in those cases is of course different from
% what to do when constructing the preamble, so special definitions are created, see
% below.
%    \begin{macrocode}
\expandafter \@tfor \expandafter \@nextchar
 \expandafter :\expandafter =\the\@temptokena \do
 {\@testpach
 \ifcase \@chclass \@classfz
  \or \or \@classfii \or
  \or \or \or \or \or \or \or \@classfx \fi
 \@lastchclass\@chclass}%
%    \end{macrocode}
% And the macro is finished by applying the |\aut@check| and appending the 
% |\\| to the sumline.  Note that the |\aut@check| is performed {\itshape 
% in\/} the last column, but since it does not put anything in the 
% horizontal list---it only writes to screen and transcript file---, this
% is harmless.
%    \begin{macrocode}
\xdef\s@ml@ne{\s@ml@ne\noexpand\aut@check\noexpand\\}}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\@addtosumline}
% Macro |\@addtosumline|, as its name already suggests, adds something to the 
% sumline, like its counterpart |\@addtopreamble| did to the preamble.
%    \begin{macrocode}
\def\@addtosumline#1{\xdef\s@ml@ne{\s@ml@ne #1}}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\@classfx}
% Class~f10 for the sumline creation is a stripped down version of |\@classx|: add an
% ampersand unless it is the first.  It deals with the specifiers |b|, |m|, |p|, |c|,
% |l|, and |r|.
%    \begin{macrocode}
\def\@classfx{\if@firstamp \@firstampfalse \else \@addtosumline &\fi} 
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\@classfz}
% Class~f0 is applicable for specifiers |c|, |l|, and |r|, and if the arguments of |p|,
% |m|, or |b| are given.  The latter three cases, with |\@chnum| is 0, 1, or~2 should be
% ignored and the first three cases are now similar to class~f10.
%    \begin{macrocode}
\def\@classfz{\ifnum\@chnum<\thr@@ \@classfx\fi}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\@classfii}
% Here comes the nice and nasty part.  Class~f2 is applicable if a~|<| is 
% specified.  This is tested by checking |\@lastchclass|, which should be 
% equal to~8.  Then it is checked that the argument to |<| is indeed 
% |\e@fi|.  This check is rather clumsy but this was the first way, after
% many attempts, that worked.  It is necessary because the usage of |<|
% is not restricted to |\e@fi|: the user may have specified other 
% \LaTeX-code using~|<|.
%    \begin{macrocode}
\def\@classfii{\ifnum\@lastchclass=8 
 \edef\t@stm{\expandafter\string\@nextchar}
 \edef\t@stn{\string\e@fi}
 \ifx\t@stm\t@stn
%    \end{macrocode}
% If both tests yield |true| then add the macro to typeset everything.
%    \begin{macrocode}
  \@addtosumline{\prr@sult}
%    \end{macrocode}
% But we're not done yet: in the following lines of code the appropriate \meta{count}s and
% \meta{dimen}s are created, if necessary.  Note that |\FCsc@l| was set to~0 in the 
% beginning of |\@mksumline|, so it is well-defined when |\@classfii| is used.
%    \begin{macrocode}
  \global\advance\FCsc@l by \@ne 
  \ifnum\FCsc@l>\FCtc@l
%    \end{macrocode}
% Apparently the number of requested columns is larger than the currently
% available number of relevant \meta{count}s and \meta{dimen}s, so new ones
% should be created.  What is checked here is merely the existence of 
% |\FCtot@<some romannumeral>|.  If it already exists---although it may not
% even be a \meta{count}; that cannot be checked---it is not created by 
% |fcolumn| and an error is given.  In case it is a \meta{count} you're just
% lucky, and you could ignore that error, although any change to this 
% \meta{count} is global anyway, so things will be overwritten.  In the
% case it is not a \meta{count}, things will go haywire and you'll soon
% find out.  The remedy then is to rename your \meta{count} prior to 
% |fcolumn| to avoid this name clash.
%    \begin{macrocode}
   \expandafter\ifx\csname FCtot@\romannumeral\FCsc@l\endcsname\relax
    \expandafter\newcount\csname FCtot@\romannumeral\FCsc@l\endcsname
   \else
    \PackageError{fcolumn}{Name clash for <count>}{\expandafter\csname
    FCtot@\romannumeral\FCsc@l\endcsname is already defined and it may
    not even be a <count>. If you're\MessageBreak sure it is a <count>,
    you can press <enter> now and I'll proceed, but things\MessageBreak
    will get overwritten.}%
   \fi
%    \end{macrocode}
% And the same is applicable for the \meta{dimen}: in case of a name 
% clash you have to rename your \meta{dimen} prior to |fcolumn|.
%    \begin{macrocode}
   \expandafter\ifx\csname FCwd@\romannumeral\FCsc@l\endcsname\relax
    \expandafter\newdimen\csname FCwd@\romannumeral\FCsc@l\endcsname
%    \end{macrocode}
% If the creation was successful, the \meta{count} |\FCtc@l| should be 
% increased.
%    \begin{macrocode}
    \global\FCtc@l=\FCsc@l
   \else
    \PackageError{fcolumn}{Name clash for <dimen>}{\expandafter\csname
    FCwd@\romannumeral\FCsc@l\endcsname is already defined and it may 
    not even be a <dimen>.  If you're\MessageBreak sure it is a <dimen>,
    you can press <enter> now and I'll proceed, but things\MessageBreak
    will get overwritten.}%
   \fi
  \fi
 \fi
\fi}
%    \end{macrocode}
% Once created, it is not necessary to initialise them here because that
% is done later in one go.
% \end{macro}
% \begin{macro}{\sumline}
% The command for the sumline has one optional argument: the separation
% between the rule and the total. By default this is 2\thinspace pt, but
% the user may specify |\sumline[10pt]| if that separation needs to be
% 10\thinspace pt.  The assignment needs to be global, because it is done
% in the first column of the tabular, but is valid for the whole line.
%    \begin{macrocode}
\newdimen\s@mlinesep
\def\sumline{\@ifnextchar[\s@mline{\s@mline[2pt]}}
\def\s@mline[#1]{\global\s@mlinesep=#1 \s@ml@ne}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\prr@sult}
% The macro |\prr@sult| actually puts the information together.  It starts
% like |\leeg|.
%    \begin{macrocode}
\def\prr@sult{$\egroup \let\e@fi=\relax
%    \end{macrocode}
% Then the information for the last line is computed.  It is not sufficient
% to calculate the width of the result (in points) to use that as the width
% of the rule separating the individual entries and the result.  It may be 
% that the sum is larger (in points) than any of the entries, e.g., when the
% result of $6+6$ (using specifier~|3,2|) is typeset. The width of the rule 
% should be equal to the width of |\hbox{$12{,}00$}| then.  On the other
% hand the width of the rule when summing $24$ and $-24$ should be that of 
% |\hbox{$-24{,}00$}| (or |\hbox{$(24{,}00$}|, see above), not the width of
% the result |\hbox{$0{,}00$}|.  Therefore the maximum of all entry widths,
% including the result, was calculated.
%    \begin{macrocode}
\setbox0=\hbox{$\geldm@cro{\sp@l}{\number\csname
FCtot@\romannumeral\FCsc@l\endcsname}$}%
\ifdim\wd0>\csname FCwd@\romannumeral\FCsc@l\endcsname
 \global\csname FCwd@\romannumeral\FCsc@l\endcsname=\wd0
\fi
\vbox{\hrule width \csname FCwd@\romannumeral\FCsc@l\endcsname
\vskip\s@mlinesep
\hbox to \csname FCwd@\romannumeral\FCsc@l\endcsname{\hfil\unhbox0}}}
%    \end{macrocode}
% \end{macro}
% \subsection{Other checks}
% \begin{macro}{\leeg}
% This macro is used to overrule the default behaviour of the pair |\b@fi| 
% and |\e@fi|.  It starts with ending the groups in the 
% same way that |\e@fi| would normally do.  Then the effect of |\e@fi|
% (that is still in the preamble) is annihilated by |\let|ting it to 
% be |\relax|.  This |\let| is only local to the current column.  Then
% the argument to |\leeg| is treated in a similar way as |\e@fi| would
% do with a typeset number.
%
% Since the user may from time to time also need a column entry other than 
% a number in the table, e.g., |\leeg{p.m.}|, this definition is without 
% at-sign. By defining |\leeg| in this way, instead of |\multicolumn1r{}|
% (which contains |\omit|), the default spacing in the column is retained.
% It has its normal effect on the column width, but doesn't alter the 
% width of the sumrule.
%    \begin{macrocode}
\def\leeg#1{$\egroup \let\e@fi=\relax #1}
%    \end{macrocode}
% Note that anything may be given as argument to |\leeg|, so in principle
% it can also be used to cheat: |\leeg{0,03}| will insert |0,03| in the 
% table but it doesn't increase the totals of that column by~3 (assuming 
% |3,2| coding for the separations).  But you won't cheat, won't you?  
% It may affect the width, so be careful: don't insert the unabridged 
% version of Romeo and Julia here.
% \end{macro}
% \begin{macro}{\res@tsumline}
% Since all changes to the totals and widths of the columns are global,
% they have to be reset actively at the start of a tabular or array.
% That is an action by itself, but it may occur more often, on request
% of the user, therefore a special macro is defined.  A side effect of
% this macro is that |\FCsc@l| is reset to~0. This is an advantage:
% it should be zero at the beginning of a line in the table (for other
% lines this is done by the |\\|).
%    \begin{macrocode}
\def\res@tsumline{\FCsc@l=\FCtc@l\loop\ifnum\FCsc@l>0
 \global\csname FCtot@\romannumeral\FCsc@l\endcsname=0
 \global\csname FCwd@\romannumeral\FCsc@l\endcsname=\z@
 \advance\FCsc@l by \m@ne\repeat}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\resetsumline}
% To reset a sumline within a table, it should be done within a |\noalign|.
%    \begin{macrocode}
\def\resetsumline{\noalign{\res@tsumline}}
%    \end{macrocode}   
% \end{macro}
% \begin{macro}{\aut@check}
% If the number of F-columns is even, it is assumed that they are part of 
% two sets of columns of which each column of the first set should balance
% the appropriate column of the second set.  If on the other hand the number
% of columns is odd, then at least one column has nothing to balance against
% and no checking occurs.  It is correct to check for oddness of |\FCsc@l| 
% since this |\aut@check| is only performed in the last column of the 
% tabular: the value of |\FCsc@l| now equals the number of columns used 
% in the current tabular (and may differ from |\FCtc@l|).
%
% The output is only to screen and the transcript file; it doesn't change
% the appearance of your document, so in case the assumption is wrong you
% can safely ignore the result and go on.  The \meta{count}s 0 and~1 are 
% used here and this can be done because any content of those \meta{count}s
% from previous calculations has become irrelevant at this moment.
%
% If the list |\FC@chklist| is empty, the list for the automatic check is
% generated (which will remain empty if |\FCsc@l| is odd).
%    \begin{macrocode}
\def\aut@check{\ifx\@empty\FC@chklist\relax
 \ifodd\FCsc@l\else 
  \count0=\@ne \count1=\FCsc@l 
  \divide\count1 by \tw@
  \loop\ifnum\count1<\FCsc@l
    \advance\count1 by \@ne
    \xdef\FC@chklist{\FC@chklist\number\count0,\number\count1;}%
    \advance\count0 by\@ne
  \repeat
  \fi
\fi
%    \end{macrocode}
% Then this list is peeled off and processed.
%    \begin{macrocode}
\loop 
  \ifx\FC@chklist\@empty\let\FCs@gn=\@ne\else\let\FCs@gn=\m@ne\fi
  \ifx\FCs@gn\m@ne
    \expandafter\fre@t\FC@chklist\end
    \ifnum\csname FCtot@\romannumeral\count0\endcsname=
      \csname FCtot@\romannumeral\count1\endcsname\else
      \PackageWarningNoLine{fcolumn}{F-columns \number\count0 \space
      and \number\count1 \space do not balance}%
    \fi
\repeat}
%    \end{macrocode}
% When |\aut@check| is finished, |\FC@chklist| is empty again, i.e., well
% prepared for the next time it is used.  This also means that the default
% behaviour kicks in again: if that's not what you want, you should
% specify the appropriate |\checkfcolumns| lines again.
% \end{macro}
% \begin{macro}{\fre@t}
% This function eats the first two numbers off |\FC@chklist|.
%    \begin{macrocode}
\def\fre@t#1,#2;#3\end{\count0=#1 \count1=#2 \xdef\FC@chklist{#3}}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\checkfcolumns}
% But the assumptions for |\aut@check| may be wrong, therefore manual
% control on this checking is also made possible here. The macro 
% |\checkfcolumns| provides a way to the user to check that the appropriate 
% columns are balanced (as it should in a balance).  Arguments |#1| and 
% |#2| are the |F|-column numbers to compare.  It is the responsibility of 
% the user to provide the correct numbers here, otherwise bogus output 
% is generated.  If this manual check is inserted, the automatic check
% will not be performed.
%    \begin{macrocode}
\def\checkfcolumns#1#2{\noalign{\xdef\FC@chklist{\FC@chklist #1,#2;}}}
%    \end{macrocode}
% \end{macro}
% That's it!
% \section*{Acknowledgement}
% Thanks to Karl Berry for valuable comments regarding the consistency of
% the installation procedure of this version.  Frank Mittelbach gave 
% various useful suggestions for improving the input parsing as well as
% hints to make the package more \LaTeX-like. 
% \bibliographystyle{plain}
% \begin{thebibliography}{9}
% \itemsep0pt
% \parsep0pt
% \parskip0pt
% \bibitem{array} Frank Mittelbach and David Carlisle. 
%    A new implementation of \LaTeX's |tabular| and |array| environment.
% \bibitem{booktabs} Simon Fear.  Publication quality tables in \LaTeX.
% \bibitem{dcolumn} David Carlisle.  The |dcolumn| package.
% \bibitem{world_economy} According to the IMF \url{www.imf.org}.
% \bibitem{CandTB} Donald Knuth, Computers \& Typesetting/B, \TeX: the program.
% \bibitem{CandTA} Donald Knuth, Computers \& Typesetting/A, The \TeX book.
% \bibitem{TAOCP1} Donald Knuth, The Art of Computers Programming, volume~1.
% \end{thebibliography}
% \Finale
\endinput