summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/gauss/gauss.sty
blob: a6da7faa1477c3bb758e12332ef832cc53b8ce53 (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
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
% TODO: 
%  *  ALLOW FINE TUNING OF ARROW ETC ARRANGEMENT BY PARAMETERS
%     (Suggestion by Antonis Tsolomitis)
%  *  BUG: ROWOPMINSIZE AND COLOPMINSIZE DONT WORK IN \swap01 etc.
%     (MAXIMUM COMPUTATION MISSIG)
%  *  BUG: ARROWHT HAS TWO CONTRADICTING MEANINGS
%
% \title{\texttt{gauss.sty} -- A Package for Typesetting Matrix Operations}
% \author{Manuel Kauers}
% \maketitle
% 
% \MakeShortVerb{\|}
%
% \newenvironment{example}
% {\par\goodbreak\medskip
%  \begin{minipage}[c]{.45\textwidth}
%   \def\switch{\end{minipage}\begin{minipage}[c]{.45\textwidth}\hfil$}
%   \obeylines
% }{$\hfil\end{minipage}\medskip\goodbreak\par}
% 
% \begin{abstract}
%  This package provides \LaTeX-macros for typesetting operations on a matrix.
%  By an ``operation on a matrix'' we understand a \textit{row operation}
%  or a \textit{column operation}. 
%
%  The user interface of the package is very straightforward and easy to understand
%  while the results look quite pretty.
% \end{abstract}
%
% \tableofcontents
%
% \section{Usage}
%
% If you find yourself in search of a package that enables you to easily typeset 
% constructions like
% \[
%  \begin{gmatrix}[v]
%   1 & 0 & 5 & 7 & 2 \\
%   3 & 1 & 1 & 5 & 1 \\
%   1 & 0 & -7 & 1 & 4 \\
%   4 & 3 & 6 & 5 & 4\\
%   1 & 7 & 9 & 4 & 3 \\
%   0 & 0 & 8 & 0 & -1
%  \rowops
%   \add[-3]01
%   \add[-1]02
%   \swap34
%   \mult5{\cdot   0}
%   \add[x^2-1]53
%  \colops
%   \swap01
%   \mult3{\cdot   1}
%   \add[0]24
%  \end{gmatrix} = \begin{gmatrix}[v]
%   0 & 1 & 5 & 7 & 2 \\
%   1 & 0 & -14 & -16 & -5 \\
%   0 & 0 & -12 & -6 & 2 \\
%   7 & 1 & 9 & 4 & 3 \\
%   3 & 4 & 6 & 5 & 4\\
%   0 & 0 & 0 & 0 & 0
%  \end{gmatrix},
% \]
% then this package is what you need. 
% It defines a new matrix environment which is extended by comprehensive macros for
% typesetting so-called ``operations'' on the matrix.
% An operation is either a row operation or a column operation, and may involve one or 
% two lines. 
% Examples of such operations arise in the context of Gaussian elimination for solving
% systems of linear equations in linear algebra: swaping rows, adding the multiple of one
% row to another, and multiply a row by a constant factor.
%
% \subsection{How to typeset matrix operations}
%
% \begin{environment}{gmatrix}
% The package defines a new matrix environment |gmatrix| which
% behaves just like \LaTeX's and \AmS\LaTeX's |matrix|. It takes an optional 
% parameter \meta{delimtype} to select the matrix delimiters. So, |gmatrix[p]|
% corresponds to |pmatrix|, |gmatrix[v]| to |vmatrix|, and so on.
%
% The |gmatrix| environment can be used exaclty like its brothers and sisters
% defined by \LaTeX\ and \AmS\LaTeX, for instance:
% \begin{example}
%  |\begin{gmatrix}[p]|
%  | a & b \\|
%  | c & d|
%  |\end{gmatrix}|
%  \switch
%  \begin{gmatrix}[p] a&b\\ c&d\end{gmatrix}
% \end{example}
% The content of the |gmatrix| environment consists of three parts: matrix, row operations,
% and column operations. The latter two are optional parts, and the ordering of them is
% arbitrary (i.e.\ row operations may be stated before column operations and vice versa). 
% The matrix part is required, and it must be the first one.
% \end{environment}
% 
% \begin{macro}{\rowops}
% \begin{macro}{\colops}
% To skip to the next section, there are two comands |\rowops| which swiches to the row 
% operation section, and |\colops| which switches to the column operation section.
% \end{macro}
% \end{macro}
%
% \begin{macro}{\mult}
% \begin{macro}{\add}
% \begin{macro}{\swap}
% Within the operation sections, you have to state the sequence of operations which are to
% be typeset. There are the three commands |\mult|, |\add|, and |\swap| to do this. These
% commands are specified as follows:
%
% \begin{enumerate}
% \item |\mult{i}{\cdot   b}| typesets the operation ``multiply the $i$th row (or column)
%   by~$b$'',
% \item |\swap[a][b]{i}{j}| typesets the operation ``swap the $i$th and the $j$th row 
%   (or column)''.
%   $a$~and~$b$ are labels to typeset at the end of the arrows, similar to the $\cdot   b$ of
%   the |\mult| command. The command does nothing if $i=j$.
% \item |\add[a][b]{i}{j}| typesets the operation ``add the $a$-fold of row (or column)~$i$ to
%   row (or column)~$j$. $b$~is a label for the $j$th line. The command does nothing if $i=j$.
% \end{enumerate}
% 
% In the standard implementation, optional arguments of |\swap| and the second optional
% argument of |\add| are ignored. See Section~\ref{ssec:atp} for how to enable them. 
%
% Rows are counted top-down, and columns are counted from left to right. The uppermost row
% and the leftmost column have the index~0. There is also the posibility to use |*| as index
% which causes the typesetting of several operations where |*| runs over all indices. For
% example, |\mult{*}{5}| in the |\rowops| section of a $n\times   n$ matrix is equivalent to
% state |\mult{0}{5}|,\dots,|\mult{|$n-1$|}{5}|.
% \end{macro}
% \end{macro}
% \end{macro}
%
% \subsection{Examples}
%
% \begin{itemize}
% \item A matrix with row operations
% \begin{example}
%  |\begin{gmatrix}[p]|
%  | 1 & 2 & 3 \\|
%  | 4 & 5 & 6 \\|
%  | 7 & 8 & 9|
%  |\rowops|
%  | \swap{0}{1}|
%  | \mult{0}{\cdot   7}|
%  | \add[5]{1}{2}|
%  |\end{gmatrix}|
% \switch
%  \begin{gmatrix}[p] 1&2&3\\4&5&6\\7&8&9 \rowops
%  \swap01\mult0{\cdot   7}\add[5]12
%  \end{gmatrix}
% \end{example}
% \item The same operations in an other ordering
% \begin{example}
%  |\begin{gmatrix}[p]|
%  | 1 & 2 & 3 \\|
%  | 4 & 5 & 6 \\|
%  | 7 & 8 & 9|
%  |\rowops|
%  | \add[5]{1}{2}|
%  | \swap{0}{1}|
%  | \mult{0}{\cdot   7}|
%  |\end{gmatrix}|
% \switch
%  \begin{gmatrix}[p] 1&2&3\\4&5&6\\7&8&9 \rowops
%  \add[5]12\swap01\mult0{\cdot   7}
%  \end{gmatrix}
% \end{example}
% \item A matrix with column operations
% \begin{example}
%  |\begin{gmatrix}[p]|
%  | 1 & 2 & 3 \\|
%  | 4 & 5 & 6 \\|
%  | 7 & 8 & 9|
%  |\colops|
%  | \swap{0}{1}|
%  | \mult{0}{\cdot   7}|
%  | \add[5]{1}{2}|
%  |\end{gmatrix}|
% \switch
%  \begin{gmatrix}[p] 1&2&3\\4&5&6\\7&8&9 \colops
%  \swap01\mult0{\cdot   7}\add[5]12
%  \end{gmatrix}
% \end{example}
% \item A matrix with both row and column operations
% \begin{example}
%  |\begin{gmatrix}[v]|
%  | 1 & 2 & 3 \\|
%  | 4 & 5 & 6 \\|
%  | 7 & 8 & 9|
%  |\rowops|
%  | \swap{1}{2}|
%  | \mult{2}{\cdot   3}|
%  | \add[-5]{1}{0}|
%  | \add[-3]{1}{2}|
%  |\colops|
%  | \swap{0}{1}|
%  | \mult{0}{\cdot   7}|
%  | \add[5]{1}{2}|
%  |\end{gmatrix}|
% \switch
%  \begin{gmatrix}[v] 1&2&3\\4&5&6\\7&8&9 \rowops
%  \swap12\mult2{\cdot   3}\add[-5]10\add[-3]12 \colops
%  \swap01\mult0{\cdot   7}\add[5]12
%  \end{gmatrix}
% \end{example}
% \item Multiple operations using the |*| index
% \begin{example}
%  |\begin{gmatrix}[p]|
%  | 1&2&3&4\\|
%  | 5&6&7&8\\|
%  | 9&10&11&12\\|
%  | 13&14&15&16|
%  |\rowops|
%  | \add[x]{0}{*}|
%  |\end{gmatrix}|
% \switch
%  \begin{gmatrix}[p]
%   1&2&3&4\\
%   5&6&7&8\\
%   9&10&11&12\\
%   13&14&15&16
%  \rowops \add[x]0*
%  \end{gmatrix}
% \end{example}
%  Note that the first row is not added to itself, because |\add[x]{0}{0}| has no effect.
%  You can also use two stars:
% \begin{example}
%  |\begin{gmatrix}[p]|
%  | 1&2&3\\|
%  | 4&5&6\\|
%  | 7&8&9|
%  |\rowops|
%  | \add{*}{*}|
%  |\end{gmatrix}|
% \switch
%  \kern-1.5em\begin{gmatrix}[p]
%   1&2&3\\ 4&5&6\\ 7&8&9
%  \rowops \add**
%  \end{gmatrix}\kern-2em
% \end{example}
% \item The package clearly also handels a matrix with larger entries correctly:
% \[
%  \begin{gmatrix}[p]
%   a & b & c & d & e \\
%   0 & 0 & \displaystyle\int\limits_a^b f(x)\,dx & 0 & 0 \\
%   a & b & c & d & e 
%  \rowops
%   \mult{1}{:\displaystyle\int\limits_a^b f(x)\,dx}%
%   \add[-c]10 \add[-1]02 
%  \end{gmatrix}
% \]
% Even nested |gmatrix|es are possible:
% \[
% \def\littleA#1#2#3#4{\begin{gmatrix}[p]#1&#2\\#3&#4\rowops \add[-#3/#1]01\end{gmatrix}}
% \def\littleB#1#2#3#4#5{\begin{gmatrix}[p]#1&#2\\#3&#4\rowops \mult0{\cdot   #5}\end{gmatrix}}
% \def\littleC#1#2#3#4{\begin{gmatrix}[p]#1&#2\\#3&#4\rowops \swap01\end{gmatrix}}
% \kern-1.5em
% \begin{gmatrix}[v]
%  \littleA 2233 & \littleC 1234 & \littleA abcd \\
%  \rule[-20pt]{0pt}{45pt}\littleB 0110\pi & \littleC vwxy & \littleC 1xx{x^2} \\
%  \littleB 12345 & \littleA \cdot   \cdot   \cdot   \cdot   & \littleB 54321
% \rowops
%  \add[\pi^2/6]01
%  \mult1{\cdot   42}
%  \swap02
% \end{gmatrix}
% \kern-1.5em
% \]
% \end{itemize}
%
% \subsection{Adapting the package}\label{ssec:atp}
%
% \subsubsection{Distances and dimensions}
%
% The appearance of the operation lines and arrows depends strongly on the values of the
% dimension parameters described in this section.
%
% \def\test{\begin{gmatrix}[p]a&b&c\\d&e&f\\g&h&i\rowops
%  \add[x]01\add[y]12\mult1{\cdot   y}\swap02\end{gmatrix}}%
% \def\ttest#1=#2.{\[#1=#2\relax\test\]}%
%
% \begin{macro}{\rowarrowsep}
% \begin{macro}{\colarrowsep}
%  |\rowarrowsep| denotes the distance from the matrix to the operations. 
%  For example, |\rowarrowsep=10pt| yields
%  \ttest\rowarrowsep=5pt.
%  and |\rowarrowsep=50pt| yiels
%  \ttest\rowarrowsep=50pt.
%  The corresponding dimension for column operations is |\colarrowsep|.
% \end{macro}
% \end{macro}
%  \begin{macro}{\opskip}
%   |\opskip| is the distance between two consecutive operations.
%  For example, |\opskip=6pt| yields
%  \ttest\opskip=6pt.
%  and |\opskip=30pt| yields
%  \ttest\opskip=25pt.
%  The |\opskip| length is responsible for both row and column operations.
% \end{macro}
% \begin{macro}{\labelskip}
%  |\labelskip| is the distance between an operation arrow and its labels.
%  For example, |\labelskip=3pt| yields
%  \ttest\labelskip=3pt.
%  and |\labelskip=15pt| yields
%  \ttest\labelskip=15pt.
%  The |\labelskip| length is responsible for both row and column operations.
% \end{macro}
% \begin{macro}{\rowopminsize}
% \begin{macro}{\colopminsize}
%  The length |\rowopminsize| is the minimum amount of a horizontal operation
%  segment to go to the right. 
%  For example, |\rowopminsize=3pt| yields
%  \ttest\rowopminsize=3pt.
%  If the horizontal segment ends with an arrow tip and |\rowopminsize| is less than 
%  the width of |\leftarrow|, then the width of |\leftarrow| is taken. In the above 
%  example, this is the case in the |\add[x]{0}{1}| operation. An example for an 
%  exact use of a small value of |\rowopminsize| is the upper horizontal line of 
%  |\add[y]{1}{2}|.
%  For comparation, |\rowopminsize=30pt| yields
%  \ttest\rowopminsize=30pt.
%  The corresponding value for column operations is |\colopminsize|.
% \end{macro}
% \end{macro}
%
% \subsubsection{Labels}
%
% The typesetting of a label can be changed by redefining the macros which are responsible
% for label typesetting. Each label parameter of |\mult|, |\add|, and |\swap| is passed to
% special ``fontifier'' macros which take one argument and fontify it according to the
% semantical requirements. Here is a list of those fontifier macros and their default 
% definitions:
%
% \begin{macro}{\rowmultlabel}
%  |\rowmultlabel| is the label of a |\mult| operation in the |\rowops| section.
%  Its default definition is \verb?{|\,#1}?.
% \end{macro}
% \begin{macro}{\colmultlabel}
%  |\colmultlabel| is the respective macro for the |\colops| section. It is defined
%  to 
% \begin{example}
%  |\underline{\hbox to 1.2em{$\hss\mathstrut{}#1\hss$}}|\kern-20em
% \switch
% \end{example}
% \noindent by default.
% \end{macro}
% \begin{macro}{\rowswapfromlabel}
%  |\rowswapfromlabel| is the label of a |\swap| operation in the |\rowops| section
%  which is to place at the first of the two rows. It is defaultly defined to |{}|, i.e.\
%  the label parameter is ignored.
% \end{macro}
% \begin{macro}{\colswapfromlabel}
%  |\colswapfromlabel| is the respective macro for the |\colops| section which is
%  also empty by default.
% \end{macro}
% \begin{macro}{\rowswaptolabel}
%  |\rowswaptolabel| is like |\rowswapfromlabel|, but for the other row. It is empy
%  by default.
% \end{macro}
% \begin{macro}{\colswaptolabel}
%  |\colswaptolabel| is |\rowswaptolabel|'s brother for the |\colops| section.
% \end{macro}
% \begin{macro}{\rowaddfromlabel}
%  |\rowaddfromlabel| is the macro for the label of the from-line of an |\add| command.
%  It is defined to |{\scriptstyle#1}| by default.
% \end{macro}
% \begin{macro}{\coladdfromlabel}
%  |\coladdfromlabel| is respective macro for the column operations.
% \end{macro}
% \begin{macro}{\rowaddtolabel}
%  |\rowaddtolabel| fontifies the label of the to-line of an |\add| command. This macro
%  is defined to |{\scriptscriptstyle +}| by default, i.e.\ it ignores the parameter.
% \end{macro}
% \begin{macro}{\coladdtolabel}
%  |\coladdtolabel| is the respective command for the column operation. It behaves
%  likewise.
% \end{macro}
%
% For the following example, all of the above labels were defined to |{#1}|, i.e.\ to identity.
%
% \begin{example}
%  |\begin{gmatrix}[p]|
%  | a & b & c \\|
%  | d & e & f \\|
%  | g & h & i|
%  |\colops|
%  | \mult0{m}|
%  | \add[af][at]01|
%  | \swap[sf][st]02|
%  |\rowops|
%  | \mult0{m}|
%  | \add[af][at]01|
%  | \swap[sf][st]02|
%  |\end{gmatrix}|
% \switch
% \def\rowmultlabel#1{#1}
% \def\colmultlabel#1{#1}
% \def\rowswapfromlabel#1{#1}
% \def\colswapfromlabel#1{#1}
% \def\rowswaptolabel#1{#1}
% \def\colswaptolabel#1{#1}
% \def\rowaddfromlabel#1{#1}
% \def\coladdfromlabel#1{#1}
% \def\rowaddtolabel#1{#1}
% \def\coladdtolabel#1{#1}
%  \begin{gmatrix}[p]
%   a & b & c \\
%   d & e & f \\
%   g & h & i 
%  \colops
%   \mult0{m}
%   \add[af][at]01
%   \swap[sf][st]02
%  \rowops
%   \mult0{m}
%   \add[af][at]01
%   \swap[sf][st]02
%  \end{gmatrix}
% \end{example}
% 
% \subsubsection{Matrix delimiters}
%
% \begin{macro}{\newmatrix}
%  It is possible to define new delimiter specifiers to |gmatrix|, say |gmatrix[X]|,
%  by defining a matrix environment |Xmatrix|.
%  A definition of |Xmatrix| which fulfills the requirements needed for compatibility
%  with |gmatrix| is provided automatically by the call of
%
%  \begin{example}
%  |\newmatrix{|\meta{left-delim}|}{|\meta{right-delim}|}{X}|,\kern-20em
%  \switch
%  \end{example}
%
%  which defines the environment |Xmatrix|. The arguments \meta{left-delim} and 
%  \meta{right-delim} need to be compatible to the |\left|-|\right| mechanism of \TeX.
%  As soon as |Xmatrix| exists, it is also possible to use |X| as optional argument
%  to |gmatrix|.
%
%  By convention, the suffix is one single character. If you try to enter |g@| or
%  the empty string as suffix, nothing is done, otherwise, the definition works
%  as well.
% \end{macro}
%
% \subsection{Features}
%
% \begin{itemize}
%  \item You need not care about the width or height of some macro cells, 
%   operations are always aligned well, i.e. centered to the column or row.
%  \item Operation elements will not intersect each other, unless you give
%   some very huge labels.
%  \item There is no restriction to the order of operation commands, so you can
%   choose an arbitrary order to achive the best typographic result.
%  \item If no operations are given, the result is exactly the result of
%   the \AmS-\TeX\ |matrix| environment. 
%  \item Unlike \AmS's |matrix| environment, there is no limit to the matrix' size
%   in our reimplementation |gmatrix|.
%  \item Nested |gmatrix|'s are possible.
% \end{itemize}
%
% \subsection{Trap doors and hints}
%
% \begin{itemize}
%  \item The last row \emph{must not} end with an |\\|, but each other line 
%   should end with |\\|.
%  \item The last row is used internally to measure the column's widths.
%   Therefore, if you want to point to a column~$i$, then the last row must have
%   at least $i+1$ entries.
%  \item In row operations, the package considers the width of labels
%   (that is, the width of factors in |\mult| and |\add|). But you have to
%   take care that your labels are not higher than the corresponding line,
%   otherwise they may intersect with other arrows or labels.
%  \item analogously for column operations.
%  \item The package should also run without the |amsmath| package, but if you
%   use that package (which is assumed to be the usual situation), you have to
%   load |gauss| after |amsmath|.
% \end{itemize}
%
% \subsection{Bug parade}
%
% A list of submitted bugs and suggested work-arounds or fixes.
% If you face any bug that is not in the list below, feel free to contact me
% at |manuel@kauers.de|.
%
% \begin{itemize}
% \item Hans Frederik Nordhaug faced problems with versions of \AmS-\LaTeX\ 
%  that don't define |*matrix| environments as expected (e.g.\ version 2.13).
%  The current version of |gauss| therefore redefines all those environments 
%  using our |\newmatrix| tool, and requires |amsmath| to be loaded prior to 
%  the |gauss| package.
% \item Morten H{\o}gholm suggested the introduction of fontifying macros and
%  the use of changeable lengths as discussed in Section~\ref{ssec:atp}. 
%  He also suggested some very fine typographic tunings. 
% \item Herbert Voss found that a |\unitlength=1pt| was missing to make the
%  behaviour of the package independent of redefinitions of |\unitlength|
%  outside |gmatrix|. (Fixed.)
% \item Michael Hagedorn noticed that signs in entries a treated like
%  binary relations, i.e., wrong spacing is used. (Fixed.)
% \end{itemize}
%
% \StopEventually
% 
% \section{Implementation}
%
%    \begin{macrocode}
\ProvidesPackage{gauss}[2003/01/14]
\RequirePackage{amsmath}
\makeatletter
%    \end{macrocode}
%
% To avoid naming conflicts with other packages, our private control 
% sequences all start with |\g@|. 
% Permanently public are only the |gmatrix| environment, the fontifying macros (like
% |\rowaddfromlabel|), and the dimensions (like |\opskip|).
%
% The |amsmath| package is not necessarily needed, but if used, it has to be
% loaded prior to the |gauss| package. This is forced by the |\RequirePackage|
% command.
% 
% \subsection{Allocation of registers and definition of common macros}
%
% Boxes,\dots
%    \begin{macrocode}
\newbox\g@trash
\newbox\g@matrixbox
\newbox\g@eastbox
\newbox\g@northbox
\newbox\g@label
\newbox\g@b@tmp
\newbox\g@b@tmpa
\newbox\g@b@tmpb
%    \end{macrocode}
% \dots counters,\dots
%    \begin{macrocode}
\newcount\g@maxrow 
\newcount\g@maxcol
\newcount\g@maxrow@old
\newcount\g@maxcol@old
\newcount\g@c@tmp
\newcount\g@c@tmpa
%    \end{macrocode}
% \dots and dimensions \dots
%    \begin{macrocode}
\newdimen\g@axisHeight
\newdimen\g@linethickness
\newdimen\g@tab
\newdimen\g@arrowht
\newdimen\g@arrowwd
\newdimen\g@d@tmp
\newdimen\g@d@tmpa
\newdimen\g@d@tmpb
\newdimen\g@d@tmpc
\newdimen\g@d@tmpd
\newdimen\g@d@tmpe
%    \end{macrocode}
% are allocated.
%
% \begin{macro}{\g@for}
% For frequent use, we define a special loop mechanism, which allowes to
% iterate over a given interval from a lower bound to a higher one, or
% reversely. The code to execute is given as the third argument of |\g@for|,
% using |#1| for the iteration variable that is substituted by |\the\loopCount|
% for each value in the given bounds.
%
% Each of the bounds is also visited. Example: The following code prints out
% the numbers from 1 to 37, inclusively:
%
% \begin{example}
% |\g@for1\to37\do{#1 }|
% \switch
% \end{example}
%
% We first need some more control sequences: |\g@loopContent| is defined to the loop's
% body when the loop is entered.
% |\g@loopCount| is the variable to increment or decrement with each 
% iteration. |\g@loopEnd| marks the value at which to stop the loop,
% and |\g@loopStep| contains the direction, i.e. $|\g@loopStep|=-1$ iff
% $|\g@loopEnd| < \meta{start value}$.
%    \begin{macrocode}
\def\g@loopContent#1{}
\newcount\g@loopCount\g@loopCount=0
\newcount\g@loopEnd\g@loopEnd=1
\newcount\g@loopStep\g@loopStep=1
%    \end{macrocode}
% The |\g@loop| command executes the loop initialized by |\g@for|.
% Each iteration is executed in its own group to avoid side effects and
% expecially to provide nested loops.
%    \begin{macrocode}
\def\g@loop{%
 % base case?
 \ifnum\g@loopCount=\g@loopEnd\else
  % no: execute loop body
  {\expandafter\g@loopContent\expandafter{\the\g@loopCount}}%
  % increment or decrement the loop variable
  \advance\g@loopCount\g@loopStep
  % call \g@loop recursivly.
  \g@loop
 \fi
}
%    \end{macrocode}
% Finally, here is the definition of |\g@for|. Each value in the interval
% from |#1| to |#2|, including |#1| and |#2| is visited exactly one time.
%    \begin{macrocode}
\def\g@for#1\to#2\do#3{%
 \def\g@loopContent##1{#3}%
 \g@loopCount=#1
 \g@loopEnd=#2
 \ifnum\g@loopEnd>\g@loopCount%
  \g@loopStep=1
  \else\g@loopStep=-1
 \fi
 \advance\g@loopEnd\g@loopStep % inclusive upper bound
 \g@loop
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\g@checkBounds}
% The next tool is used by the generic operation commands to check whether or not
% a given index is valid. If $|#2|\leq|#3|\leq|#4|$ does not hold, a package
% error is thrown that tells the user what happened.
%
% Parameter |#1| contains `r' or `c' to denote ``rows'' or ``columns'', 
% respectively. This piece of information is only used within the construction of
% the error message.
%
% \begin{macro}{\ifg@indexCorrect}
% The result of |\g@checkBounds| is returned via |\ifg@indexCorrect|.
%    \begin{macrocode}
\newif\ifg@indexCorrect
\def\g@checkBounds#1#2#3#4{%
 \g@indexCorrectfalse
 \ifnum#2>#3%
  \PackageError{gauss}{\g@shorterror{#1} #3<#2}{\g@longerror{#1}}
 \else
  \ifnum#3>#4%
   \PackageError{gauss}{\g@shorterror{#1} #3>#4}{\g@longerror{#1}}
  \else
   \g@indexCorrecttrue
  \fi
 \fi
}
%    \end{macrocode}
% We skip the definitions of |\g@shorterror| and |\g@longerror| which serve to 
% produce error messages. 
\def\g@shorterror#1{\ifx r#1 Row \else Column \fi index out of bounds: }
\def\g@longerror#1{%
 An index of an operation points to a \ifx r#1 row \else column \fi %
 that does not exist.\MessageBreak
 Note that the index of the %
 \ifx r#1 bottom row \else leftmost column \fi is 0 while the index of the %
 \ifx r#1 top row \else rightmost column \fi is <number of %
 \ifx r#1 rows\else columns\fi - 1> .%
}
% \end{macro}\end{macro}
%
% \begin{macro}{\g@downarrow}
% For drawing horizontal arrows of arbitrary length, we use the construction
%
% \begin{example} 
% |\hbox to|\meta{width}|{$\leftarrowfill$}|\kern-20em
% \switch\end{example}
%
% \noindent which uses Plain-\TeX's |\leftarrowfill|. Unfortunately, there is no
% vertical correspondence to that mechanism and thus, we have construct something
% like this by ourselves. We will do so by reimplementing a mechanism that is used
% by |\left| and |\right| to construct delimiters of arbitrary height.
%
%    \begin{macrocode}
\DeclareMathSymbol{\g@downarrowSymb}{\mathord}{largesymbols}{`\y}
\DeclareMathSymbol{\g@vertlineSymb}{\mathord}{largesymbols}{`\?}
\def\g@vertline{\hbox{$\g@vertlineSymb$}\kern-\lineskip}%
%    \end{macrocode}
%
% After allocating the basic symbols, we define |\g@downarrow| by a recursion
% which fills up a vbox with the necessary number of |\g@vertline|'s and a
% final |\g@downarrowSymb|. 
%
% The resulting vbox has exactly the height given in |#1| (as \TeX-length), and
% no depth. If |#1| is less than a minimum height, then it is set to that minimum
% height.
%
%    \begin{macrocode}
\def\g@downarrow#1{\vbox{%
 \vfill
 \baselineskip\z@\relax
 \g@d@tmpc=#1\relax
 \ifdim \g@d@tmpc<\g@arrowht
  \g@d@tmpc\g@arrowht\relax
 \fi
 \g@vlineRec
 \kern\g@d@tmpc 
 \setbox\g@trash=\hbox{$\g@downarrowSymb$}%
 \hbox{\raise\dp\g@trash\box\g@trash}%
}}
\def\g@vlineRec{%
 \advance\g@d@tmpc-\g@arrowht
 \ifdim \g@d@tmpc<\z@ \else
  \g@vertline
  \g@vlineRec
 \fi
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsection{Converting floasts and lengths to each other}
%
% \begin{macro}{\g@defdim}%
% \begin{macro}{\g@defdouble}%
% \begin{macro}{\g@dim}%
% \begin{macro}{\g@double}%
% The typesetting of matrix operations is done by use of the |picture| 
% environment of \LaTeX. The macros of that environment require plain
% numbers, possibly containing a decimal point. Though it is not clearly
% correct, we will call that data format \emph{float} or \emph{double}.
% 
% |picture|'s macros do not work if you give them dimensions as input.
% And since the results of measuring a matrix are necessarily dimensions,
% we need a mechanism to convert dimensions to floats and vice versa.
%
% This mechanism is the topic of the current section.
%
% In fact, we almost provide our own data structure whose values can be shown
% as \TeX\ dimensions or as floats. You can ``construct a new instance'' of
% that structure either by a dimension (using |\g@defdim|) or by a double
% (using |\g@defdouble|). In both cases, a macro is defined to be the 
% corresponding double value. 
%
% Given an instance of our data structure, i.e.\ given a double, you can get
% its double representation using |\g@double| (this just typesets the double
% representation), and you can store its value into a \TeX\ dimension using
% |\g@dim|.
%
% Macros for manipulation on floats are defined in the following section.
% 
% \medskip
% We first need a macro that cuts away the ``pt''. This is rather tricky because
% the ``pt'' that arises in the result of some |\the|\meta{counter} has not the
% catcodes as expected. We can redefine them temporarily but we have to note that
% constructions like |\g@defdim{|\meta{identifier}|}{12pt}| (i.e.\ giving the length
% directly) are no longer possible, since the ``pt'' of a directly given length
% has the ``normal'' catcodes.
%    \begin{macrocode}
\edef\redo#1{\catcode`p=#1\catcode`t=#1\relax}
\redo{12}
\def\g@del#1pt{#1}
\redo{11}
%    \end{macrocode}
% Defining a float by a dimension. The first argument expects an idetifier
% (identifiers are arbitrary strings), and the second argument expects a
% \TeX\ dimension \emph{register}, i.e. some control sequence |\cs| that 
% evaluates to ``\dots pt'' if you say |\the\cs|. 
%
% It is not possible to specify a double by directly give a length. Use
% |\g@defdouble| below in that case.
%    \begin{macrocode}
\def\g@defdim#1#2{%
 \edef\g@defdim@arg{\the #2}%
 \edef\g@defdim@arg{\expandafter\g@del\g@defdim@arg}%
 \g@defdouble{#1}{\g@defdim@arg}%
}
%    \end{macrocode}
% And here is |\g@defdouble|. |#1| should be an identifier and |#2| should
% be the value to store in float |#1|. To avoid naming conflics with other
% macros, |#2| is stored into a macro based on |g@@| and the content of |#1|.
%    \begin{macrocode}
\def\g@defdouble#1#2{%
 \expandafter\expandafter\expandafter\global
 \expandafter\edef\csname g@@#1 \endcsname{#2}%
}
%    \end{macrocode}
% We now come to the macros for ``reading'' a float. These are |\g@dim| (to
% read the dimensional representation) and |\g@double| (for the double 
% representation).
%
% An error will occur if you try to read the value of a float that was not
% previously defined. (``Missing number, treated as zero.'')
%
% First |\g@dim|: Let |#1| be the identifier and |#2| the \TeX\ dimension
% registern to store the value of |#1| in.
%    \begin{macrocode}
\def\g@dim#1#2{%
 \edef\g@dim@arg{\g@double{#1}}%
 #2=\g@dim@arg\p@\relax
}
%    \end{macrocode}
% And |\g@double| is even simpler:
%    \begin{macrocode}
\def\g@double#1{%
 \csname g@@#1 \endcsname
}
%    \end{macrocode}
% \end{macro}\end{macro}\end{macro}\end{macro}
%
% \subsection{Macros for calculus on floats}
%
% We need some macros that provide simple arithmetic calculation on 
% floats. Those are defined now.
%
% \begin{macro}{\g@advance}
% Given a float $f_1$, the following macro performs $f_1 := f_1 + f_2$
% where $f_2$ may be either a \TeX\ dimension or a float:
% If |\csname|$f_2$|\encsname| does not evaluate to some control sequence, 
% it is assumed to denote a \TeX\ dimension (e.g. |5pt|, or |\the\something|)
%    \begin{macrocode}
\def\g@advance#1#2{%
 \g@dim{#1}{\g@d@tmpa}% f_1 := #1
 \expandafter\ifx\csname g@@#2 \endcsname\relax
  \g@d@tmpb=#2% f_2 := #2 (TeX dimension)
 \else
  \g@dim{#2}{\g@d@tmpb}% f_2 := #2 (float)
 \fi
 \advance\g@d@tmpa\g@d@tmpb\relax% f_1 += f_2
 \g@defdim{#1}{\g@d@tmpa}% #1 := f_1
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\g@min}\begin{macro}{\g@minD}
% Given two floats $f_1, f_2$ and a \TeX\ dimension $d_3$, 
% the following macro performs $d_3 := \min\{f_1, f_2\}$.
%    \begin{macrocode}
\def\g@min#1#2#3{%
 \g@dim{#1}{\g@d@tmpa}% f_1 := #1
 \g@dim{#2}{\g@d@tmpb}% f_2 := #2
 \ifdim \g@d@tmpa<\g@d@tmpb
   #3=\g@d@tmpa
 \else
   #3=\g@d@tmpb
 \fi
 \relax
}
%    \end{macrocode}
% There is a so called $D$-version of the latter macro. By use of |\g@min|, 
% this macro also calculates $\min\{f_1,f_2\}$, but its result is translated
% into a double representation which is then stored in control sequence |#3|.
%    \begin{macrocode}
\def\g@minD#1#2#3{%
 \g@min{#1}{#2}{\g@d@tmpc}%
 \edef\g@minD@arg{\the\g@d@tmpc}%
 \edef\g@minD@arg{\expandafter\g@del\g@minD@arg}%
 \edef#3{\g@minD@arg}%
}
%    \end{macrocode}
% \end{macro}\end{macro}
% \begin{macro}{\g@max}\begin{macro}{\g@maxD}
% And here is are the two opposite macros of the preceeding two.
%    \begin{macrocode}
\def\g@max#1#2#3{%
 \g@dim{#1}{\g@d@tmpa}% 
 \g@dim{#2}{\g@d@tmpb}% 
 \ifdim \g@d@tmpa<\g@d@tmpb
   #3=\g@d@tmpb
 \else
   #3=\g@d@tmpa
 \fi
 \relax
}
\def\g@maxD#1#2#3{%
 \g@max{#1}{#2}{\g@d@tmpc}%
 \edef\g@maxD@arg{\the\g@d@tmpc}%
 \edef\g@maxD@arg{\expandafter\g@del\g@maxD@arg}%
 \edef#3{\g@maxD@arg}%
}
%    \end{macrocode}
% \end{macro}\end{macro}
% \begin{macro}{\g@dist}\begin{macro}{\g@distD}
% Given two floats $f_1, f_2$ and a \TeX\ dimension $d_3$, the following
% macro performs $d_3 := f_1 - f_2$. 
%    \begin{macrocode}
\def\g@dist#1#2#3{%
 \g@dim{#1}{\g@d@tmpa}% f_1 := #1
 \g@dim{#2}{\g@d@tmpb}% f_2 := #2
 \ifdim \g@d@tmpa<\g@d@tmpb
   #3=\g@d@tmpb
   \advance#3 by-\g@d@tmpa
 \else
   #3=\g@d@tmpa
   \advance#3 by-\g@d@tmpb
 \fi
 \relax
}
%    \end{macrocode}
% Again, we have a $D$-version, where the result is given in double 
% representation.
%    \begin{macrocode}
\def\g@distD#1#2#3{%
 \g@dist{#1}{#2}{\g@d@tmpc}%
 \edef\g@distD@arg{\the\g@d@tmpc}%
 \edef\g@distD@arg{\expandafter\g@del\g@distD@arg}%
 \edef#3{\g@distD@arg}%
}
%    \end{macrocode}
% \end{macro}\end{macro}
%
% \begin{macro}{\g@updateArea}\begin{macro}{\g@update}
% While the macros that we have seen in this section so far are mainly used
% for elementary drawing purposes, we now define a slightly more sophisticated 
% macro.
% It is needed to update the leftmost $x$-values of the so-far matrix operation
% set (in terms of row operations). It is invoked after adding a new operation
% to the set.
%
% To update a float $f_1$ with respect to $f_2$ is defined to execute
% $f_1 := \max\{f_1, f_2\}$. This is what the following macro does.
%    \begin{macrocode}
\def\g@update#1#2{%
 \g@dim{#2}{\g@d@tmpe}
 \g@dim{#1}{\g@d@tmpb}
 \ifdim \g@d@tmpe>\g@d@tmpb
  \g@defdim{#1}{\g@d@tmpe}%
 \fi
}
%    \end{macrocode}
%
% The matrix dimensions are stored in floats named
% $\meta{name} + \meta{index}$ where \meta{name} spcifies the dimension 
% (e.g. ``cy'' for the current $y$ values of a \emph{c}olumn) and \meta{index}
% is the index of the row/column to which the float's value belongs.
%
% Now, the following macro iterates over $i\in\{|#3|,\dots,|#4|\}$ and updates
% all the floats with name $|#2| + i$ with respect to float |#1|.
%    \begin{macrocode}
\def\g@updateArea#1#2#3#4{\g@for#3\to#4\do{\g@update{#2##1}{#1}}}
%    \end{macrocode}
% \end{macro}\end{macro}
%
%
%
% \subsection{Macros for measurements}
%
% The macros defined in this section are used to measure the dimensions
% of a given matrix and store the measured values into floats.
%
% For each row~$i$ of the matrix, the $y$-position of the center of
% row~$i$ with respect to the bottom of the matrix is stored in a float
% named $|ry| + i$. Another float $|rx| + i$ is initialized to~$0$. This latter
% value will always contain the leftmost position at which a new row operation can
% start without intersecting previous operations that crossed row~$i$.
%
% For each row~$j$ of the matrix we similarly define the values
% $|cx| + i$ and $|cy| + i$. Note that $|cx| + i$ corresponds to $|ry| + i$
% and $|cy| + i$ corresponds to $|rx| + i$, since column operations grow
% vertically whereas row operations grow horizontally.
%
% \begin{macro}{\g@measureRows}
% We first consider row measuring. The following macro assumes that the current
% box is a |\vbox| that only contains a copy of the matrix, 
% i.e. one |\hbox| per row including all the intermediate glues and kerns and 
% whatever. You can initialize what we assume to have by saying
%
% \medskip
% {\obeylines |\vbox{\halign{|\dots|}}| (typeset the matrix)
% |\box0=\lastbox| (save the matrix)
% |\vbox{\unhcopy0\g@measureRows}| (measure the row's heights)
% |\box0| (restore the matrix)}
% \medskip
%
% \noindent Caution: The following macros will not work if the matrix was not 
% constructed with an |\halign| because special knowledge about the structure 
% of |\halign|'s result is used.
%
% It is assumed that |\g@d@tmp| initially contains the $y$-position of the
% matrix's bottom. It is further assumed that |\g@maxrow| contains the total
% number of rows. These two counters will be modified during the recursion.
%    \begin{macrocode}
\def\g@measureRows{%
 \setbox\g@trash\lastbox
 \ifnum\g@maxrow<0% base case: this box is not part of the matrix
 \else 
  \ifdim\ht\g@trash=0pt%
   \advance\g@d@tmp\lastskip\unskip
   \advance\g@d@tmp\lastkern\unkern
   \unpenalty
  \else
   \advance\g@d@tmp\dp\g@trash
   \advance\g@d@tmp\g@axisHeight
   \g@defdim{ry\the\g@maxrow}{\g@d@tmp}%
   \g@defdim{rx\the\g@maxrow}{\z@}%
   \advance\g@d@tmp-\g@axisHeight
   \advance\g@d@tmp\ht\g@trash
   \advance\g@maxrow-1%
  \fi
  \g@measureRows
 \fi
}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@measureCols}
% In fact, the row measurement is the easier case. The measurement of column
% widths is more complicated by two reasons: 1.\ The number of columns is
% unknown, and 2.\ we will meet the cells in reverse order.
%
% This is why column measurement is implemented in two main steps. First the
% width of each cell and the distance between two preceeding cells is 
% measured and stored into temporary floats $|ct| + \meta{index}$ (distance) and
% $|cy| + \meta{index}$ (width), where \meta{index} is counted from back to front.
% By the way, we count the number of columns.
%
% In the base case of the recursion we start a second recursion that will
% calculate the final results out of the intermediate results and that will
% arange the indexing properly.
%
% What input do we expect? It is assumed that the current box is an |\hbox|
% whose first item is an |\hbox| of width 100cm (to detect the base case),
% followed by a copy of the last row of the matrix to measure. See the 
% definition of |g@matrix| to see how such a situation can be constructed.
% 
% We further assume that |g@d@tmp| is initialized to 0.0pt.
%
%    \begin{macrocode}
\def\g@measureCols{%
 \setbox\g@trash\lastbox
 \ifdim \wd\g@trash=100cm%
  % base case. Invert the ordering and sum the dimensions.
  \g@defdouble{ct\the\g@maxcol}{0}%
  \g@defdouble{cy\the\g@maxcol}{0}%
  \global\g@maxcol\g@maxcol
  \g@c@tmp\g@maxcol
  \advance\g@c@tmp-1%
  \g@measureColsSucc
  \global\advance\g@maxcol-1%
 \else
  \ifdim \ht\g@trash=0pt%
   \advance\g@d@tmp\lastskip\unskip
   \advance\g@d@tmp\lastkern\unkern
   \unpenalty
  \else
   % use ct temporaryly to store the skip between
   % colnr + 1 and colnr.
   \g@defdim{ct\the\g@maxcol}{\g@d@tmp}%
   \g@d@tmp\z@
   % use cy temporaryly to store the cell's width.
   \g@defdim{cy\the\g@maxcol}{\wd\g@trash}%
   \advance\g@maxcol1%
  \fi
  \g@measureCols
 \fi
}
%    \end{macrocode}
% Now, the macro for the second step of the measurement algorithm is defined.
% This is easier, since we now already know the total number of columns that
% have been measured. Roughly speaking, we sum their width's from left to right
% to obtain the $x$-values of the horizontal center of each column. Furthermore,
% the $y$-values are now initialized to~$0$, and the order is inverted.
%
% Knowledge about the implementation of |g@matrix| is used!
%
%    \begin{macrocode}
\def\g@measureColsSucc{%
 \ifnum \g@c@tmp<0\else
  \g@c@tmpa=\g@maxcol
  \advance\g@c@tmpa-\g@c@tmp
  \advance\g@c@tmpa-1
  \g@dim{cy\the\g@c@tmp}{\g@d@tmpa}% width of this cell
  \g@dim{ct\the\g@c@tmp}{\g@d@tmpb}% glue right to this cell
  \advance\g@d@tmp.5\g@d@tmpa%
  \g@defdouble{cy\the\g@c@tmp}{0}%
  \g@defdim{cx\the\g@c@tmpa}{\g@d@tmp}%
  \advance\g@d@tmp.5\g@d@tmpa
  \advance\g@d@tmp\g@d@tmpb
  \ifnum \g@c@tmpa=0%
   \advance\g@d@tmp.5\g@tab
  \fi
  \advance\g@c@tmp-1
  \g@measureColsSucc
 \fi
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\g@measureAxis}
% This is an easier macro. It measures and defines some common lengths,
% e.g.\ the distance between bottom line and math axis, and the dimensions
% of math arrows which are used for drawing arrows in operations.
%    \begin{macrocode}
\def\g@measureAxis{%
 % 1. Where is the math axis relative to the ground line?
 \setbox\g@trash=\hbox{$\vcenter{\hbox to 5pt{}}$}%
 \global\g@axisHeight=\ht\g@trash
 % 2. What is the minimum width of a horizontal arrow?
 \setbox\g@trash=\hbox{$\leftarrow$}%
 \global\g@arrowwd\wd\g@trash
 % 3. What is the minimum height of a vertical arrow?
 \setbox\g@trash=\vbox{\g@vertline}%
 \global\g@arrowht=\ht\g@trash
 \global\advance\g@arrowht\dp\g@trash
 \global\advance\g@arrowht\lineskip
 % 4. What should be the thickness of ordinary lines?
 \global\g@linethickness=\fboxrule\relax
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\g@measureArea}
% The last marco of this subsection provides the measurement of 
% a set of floats. (Therefore, it is rather a calculus macro.)
%
% Assuming that |#4| is a float identifier and for all $i\in I:=\{|#2|,\dots,|#3|\}$
% $|#1|+i$ is a float identifier, the macro does
% \[
%  |#4| := \max_{i\in I}\{|#1| + i\}
% \]
%    \begin{macrocode}
\def\g@measureArea#1#2#3#4{%
 \g@defdim{#4}{\z@}%
 \g@for#2\to#3\do{%
  \g@max{#1##1}{#4}{\g@d@tmpe}%
  \g@defdim{#4}{\g@d@tmpe}%
 }%
}
%    \end{macrocode}
% \end{macro}
%
%
%
% \subsection{Macros for drawing purposes}
%
% This Section defines low level macros for drawing purposes within a
% |picture| environment by use of floats.
% 
% \begin{macro}{\g@vline}
% Let $f_1, f_2$ and~$f_3$ be floats. Then, 
% 
% \begin{example}
%  |\g@vline{|$f_1$|}{|$f_2$|}{|$f_3$|}|
% \switch\end{example}
% 
% \noindent draws a vertical line from $(f_1,f_2)$ to $(f_2, f_3)$.
%    \begin{macrocode}
\def\g@vline#1#2#3{%
 \g@minD{#2}{#3}{\min}
 \g@distD{#2}{#3}{\dist}
 \put(\g@double{#1},\min){\line(0,1){\dist}}
}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@vvline}
% Let $f_1, f_2$ and~$f_3$ be floats. Then,
% 
% \begin{example}
% |\g@vvline{|$f_1$|}{|$f_2$|}{|$f_3$|}|
% \switch\end{example}
% 
% \noindent draws a vertical line of length~$|f_3|$, starting at $(f_1,f_2)$, i.e.\
% a line from $(f_1,f_2)$ to $(f_1, f_2+f_3)$.
%    \begin{macrocode}
\def\g@vvline#1#2#3{%
 \put(\g@double{#1},\g@double{#2}){\line(0,1){\g@double{#3}}}
}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@varrow}
% Let $f_1, f_2$ and~$f_3$ be floats. Then,
%
% \begin{example}
% |\g@varrow{|$f_1$|}{|$f_2$|}{|$f_3$|}|
% \switch\end{example}
%
% \noindent draws an arrow from $(f_1, \max\{f_2,f_3\})$ to $(f_1, \min\{f_2,f_3\})$.
%    \begin{macrocode}
\def\g@varrow#1#2#3{%
 \g@dim{#2}{\g@d@tmpa}%
 \g@dim{#3}{\g@d@tmpb}%
 \advance\g@d@tmpb-\g@d@tmpa
 \g@cbox{#1}{#2}{\g@downarrow{\g@d@tmpb}}%
}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@hline}
% Let $f_1, f_2$ and~$f_3$ be floats. Then, 
% 
% \begin{example}
% |\g@hline{|$f_1$|}{|$f_2$|}{|$f_3$|}|
% \switch\end{example}
% 
% \noindent draws a horizontal line from $(f_1,f_2)$ to $(f_3,f_2)$.
%    \begin{macrocode}
\def\g@hline#1#2#3{%
 \g@minD{#1}{#3}{\min}%
 \g@distD{#1}{#3}{\dist}%
 \put(\min,\g@double{#2}){\line(1,0){\dist}}%
}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@hhline}
% Let $f_1, f_2$ and~$f_3$ be floats. Then,
% 
% \begin{example}
% |\g@hhline{|$f_1$|}{|$f_2$|}{|$f_3$|}|
% \switch\end{example}
% 
% \noindent draws a horizontal line of length~$|f_3|$, starting at $(f_1,f_2)$, 
% i.e.\ a line from $(f_1,f_2)$ to $(f_1+f_3, f_2)$.
%    \begin{macrocode}
\def\g@hhline#1#2#3{%
 \put(\g@double{#1},\g@double{#2}){\line(1,0){\g@double{#3}}}%
}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@harrow}
% Let $f_1, f_2$ and~$f_3$ be floats. Then,
%
% \begin{example}
% |\g@harrow{|$f_1$|}{|$f_2$|}{|$f_3$|}|
% \switch\end{example}
%
% \noindent draws an arrow from $(\max\{f_1,f_3\},f_2)$ to $(\min\{f_1,f_3\},f_2)$.
%    \begin{macrocode}
\def\g@harrow#1#2#3{%
 \g@dim{#1}{\g@d@tmpa}%
 \g@dim{#3}{\g@d@tmpb}%
 \advance\g@d@tmpb-\g@d@tmpa
 \advance\g@d@tmpb2\p@
 \g@rbox{#1}{#2}{\hbox to\g@d@tmpb{\leftarrowfill}}%
}
%    \end{macrocode}
% \end{macro}
%
% The remaining two macros allow to put arbitrary math material to a
% specified position. Those are used for typesetting so called labels within
% matrix operations, for example, the factor at an |\add| arrow.
%
% \begin{macro}{\g@rbox}
% The first one is intended to use for row operations.
% Assuming that |#1|, |#2| are float identifiers and |#3| is math material,
% we put |#3| into an |\hbox| and put that box to point $(|#1|,|#2|)$.
% 
% The box will be vertically aligned to |#2| (i.e., the math axis of |#3| will
% be at height |#2|), and horizontally start at |#1|.
%
% The macro puts the math material of |#3| into |\g@label| and just copies its content when
% using, so you can reuse |\g@label| (e.g.\ for measuring purposes).
%    \begin{macrocode}
\def\g@rbox#1#2#3{%
 \setbox\g@label=\hbox{$\relax#3\relax$}%
 \ht\g@label\z@\dp\g@label\z@
 \setbox\g@label=\hbox{$\mathstrut\box\g@label$}%  
 \put(\g@double{#1},\g@double{#2})%
 {\makebox(0,0)[l]{\kern-\p@\copy\g@label}}%
}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@cbox}
% The last macro of this section does the corresponding job for columns.
%
% Here, |#3| will be centered horizontally to |#1|, whereas |#2| denotes the
% height of the label's bottom.
%
% Again, you can reuse the constructed box, it remains in register |\g@label|.
%    \begin{macrocode}
\def\g@cbox#1#2#3{%
 \setbox\g@label=\hbox{$\relax#3\relax$}%
 \setbox\g@label=\hbox{\raise\dp\g@label\box\g@label}%
 \put(\g@double{#1},\g@double{#2})%
  {\makebox(0,0)[b]{\copy\g@label}}%
}
%    \end{macrocode}
% \end{macro}
%
%
%
% \subsection{Generic operation commands}
%
% Before |\halign| begins, the matrix construction macro defines, what to do
% if the matrix is finished. This is defined in |\g@endregion| (see the next
% section for further information).
%
% The |\rowops| and |\colops| commands are temporarily set to |\g@east| or 
% |\g@north|, respectively. Thus, when entering an operation part, the first
% thing to do is to invoke |\g@endregion| to do the things that have to be
% done when the matrix input finishes. After that, |\g@endregion| has to be
% redefined to avoid doing the same process two times. Fortunately, |\g@north|
% and |\g@east| can easily reuse |\g@endregion| and store there those things
% that have to be done at the end of a region.
%
% Hence, each switching to another part of the matrix input consists of three
% parts:
% \begin{enumerate}
%  \item Invoke |\g@endregion| to finish the current input part.
%  \item Redefine |\g@endregion| to do the stuff that has to be done at the end
%   of the region that is now starting. The result of the region is stored into
%   a special box register which is used in the |gmatrix| environment.
%  \item Initialize the new region.
% \end{enumerate}
% You can imagine that it is easy to define further regions (e.g.\ for operations
% to the right or below the matrix).
%
% The two predefined regions |\rowops| and |\colops| are very similar, so we will
% show just one of them in this documentation.
%
% \begin{macro}{\g@north}
% The |\g@north| macro is the generic version of |\colops|, its corresponding
% part is |\g@east|.
%  
%    \begin{macrocode}
\def\g@north{%
%    \end{macrocode}
% 1.\ Finish the current region
%    \begin{macrocode}
 \g@endregion
%    \end{macrocode}
% 2.\ Redefine |\g@endregion| and prevent |\colops| from being called again.
%    \begin{macrocode}
 \gdef\colops{\PackageError{gauss}
   {Two sets of column operations are specified in %
    just one matrix. This is not allowed.}}%
 \gdef\g@endregion{%
   \end{picture}\egroup
   \g@measureArea{cy}{0}{\the\g@maxcol}{sum}%
   \g@dim{sum}{\ht\g@northbox}%
   \global\setbox\g@northbox=\hbox{%
    \raise\colarrowsep\box\g@northbox}%
 }%
%    \end{macrocode}
% 3.\ Initialization of the |\colops| region: Define the operation macros to be 
% the corresponding private versions of this region (see below), set $|sum|:=0$ 
% and start the |picture| environment where the operations are painted in.
%    \begin{macrocode}
 \def\swap{\g@north@arrow11\colswapfromlabel\colswaptolabel}%
 \def\add{\g@north@arrow01\coladdfromlabel\coladdtolabel}%
 \let\mult\g@north@mult
 \g@defdim{sum}{\z@}%
 \global\setbox\g@northbox=\hbox\bgroup
   \begin{picture}(\g@double{w},0)(0,0)
     \linethickness{\g@linethickness}%
}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@north@mult}
% The multiplication macro is the simplest one because it affects only one single
% column.
%    \begin{macrocode}
\def\g@north@mult#1#2{%
 \ifx *#1
%    \end{macrocode}
% Reduce |*| to a set of numbers.
%    \begin{macrocode}
  \g@for0\to\g@maxcol\do{\g@north@mult{##1}{#2}}%
 \else
%    \end{macrocode}
% Now |#1| is a number. Is it an index?
%    \begin{macrocode}
  \g@checkBounds{c}{0}{#1}{\the\g@maxcol}%
  \ifg@indexCorrect
%    \end{macrocode}
% Yes, it is. Typeset the operation.
%    \begin{macrocode}
   \g@cbox{cx#1}{cy#1}{\colmultlabel{#2}}%
   \g@dim{cy#1}{\g@d@tmpc}%
   \advance\g@d@tmpc\ht\g@label
   \g@defdim{cy#1}{\g@d@tmpc}%
   \g@advance{cy#1}{\the\opskip}%
   \g@update{sum}{cx#1}%
  \fi
 \fi
}
%    \end{macrocode}
% \end{macro}
% \begin{macro}{\g@north@arrow}
% 
% The |\g@north@arrow| macro is a generalisation of |\swap| and |\add|. 
% It takes the following eight parameters:
%
% \begin{itemize}
% \item |#1|: 0 to make the `from' line non-arrowed, 1 to get an arrow tip
% \item |#2|: 0 to make the `to' line non-arrowed, 1 to get an arrow tip
% \item |#3|: macro for `from' label rendering
% \item |#4|: macro for `to' label rendering
% \item |#5|: [opt] label of the `from' row
% \item |#6|: [opt] label of the `to' row
% \item |#7|: index of the `from' row
% \item |#8|: index of the `to' row
% \end{itemize}
%
% If only one of the two optional arguments is given, then it is taken as |#5|
% and |#6| is taken empty. If both are missing, both are taken empty.
%
% In |\g@north| above, |\add| is defined to 
% \begin{example}
%  |\g@north@arrow01\coladdfromlabel\coladdtolabel|\kern-20em
% \switch\end{example}
% and |\swap| is defined as
% \begin{example}
%  |\g@north@arrow11\colswapfromlabel\colswaptolabel|\kern-20em
% \switch\end{example}
%
%    \begin{macrocode}
\def\g@north@arrow#1#2#3#4{%
 \@ifnextchar[%
 {\g@north@arrow@a{#1}{#2}{#3}{#4}}%
 {\g@north@arrow@b{#1}{#2}{#3}{#4}{}[]}%
}
\def\g@north@arrow@a#1#2#3#4[#5]{%
 \@ifnextchar[%
 {\g@north@arrow@b{#1}{#2}{#3}{#4}{#5}}%
 {\g@north@arrow@b{#1}{#2}{#3}{#4}{#5}[]}%
}
\def\g@north@arrow@b#1#2#3#4#5[#6]#7#8{%
 \ifx *#7
%    \end{macrocode}
% Reduce star indices to loops of number indices.
% |**| needs a special handling. 
%    \begin{macrocode}
  \ifx *#8
   \g@for0\to\g@maxcol\do{%
    \g@north@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{##1}{*}}%
  \else
%    \end{macrocode}
% Two loops rather than one because going from |#8| down
% to 0 looks better than going from 0 to |#8|
%    \begin{macrocode}
   \g@for#8\to0\do{%
    \g@north@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{##1}{#8}}%
   \g@for#8\to\g@maxcol\do{%
    \g@north@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{##1}{#8}}%
  \fi
 \else
  \ifx *#8
%    \end{macrocode}
% Reduce star indices to loops of number indices.
%    \begin{macrocode}
   \g@for#7\to0\do{%
    \g@north@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{#7}{##1}}%
   \g@for#7\to\g@maxcol\do{%
    \g@north@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{#7}{##1}}%
  \else
%    \end{macrocode}
% Now, |#7| and |#8| are numbers.
%    \begin{macrocode}
   \ifnum #7=#8\else
    \g@checkBounds{c}{0}{#7}{\the\g@maxcol}%
    \ifg@indexCorrect
     \g@checkBounds{c}{0}{#8}{\the\g@maxcol}%
     \ifg@indexCorrect
%    \end{macrocode}
% Now, |#7| and |#8| are different from each other, and
% both of them are legal indices.
% Store the current hights of the operations tower
% above column |#7| and |#8| into |tmp1| and |tmp2|,
% respectively.
%    \begin{macrocode}
      \g@defdouble{tmp1}{\g@double{cy#7}}%
      \g@defdouble{tmp2}{\g@double{cy#8}}%
%    \end{macrocode}
% Find out the height of the horizontal connection
% line.
% First increment |#7| and |#8| by the minimum amounts.
%    \begin{macrocode}
      \ifx0#1
       \g@advance{cy#7}{\the\colopminsize}%
      \else
       \g@advance{cy#7}{\the\g@arrowht}%
      \fi
      \ifx0#2
       \g@advance{cy#8}{\the\colopminsize}%
      \else
       \g@advance{cy#8}{\the\g@arrowht}%
      \fi      
%    \end{macrocode}
% Incorporate the columns between |#7| and |#8| into
% the height. Then |sum| denotes the level of the 
% horizontal line.
%    \begin{macrocode}
      \g@measureArea{cy}{#7}{#8}{sum}%
%    \end{macrocode}
% Draw arrows and/or vertical lines from |#7|'s and
% |#8|'s height up to |sum|.
%    \begin{macrocode}
      \ifx0#1
       \g@vline{cx#7}{tmp1}{sum}%
      \else
       \g@varrow{cx#7}{tmp1}{sum}%
      \fi
      \ifx0#2
       \g@vline{cx#8}{tmp2}{sum}%
      \else
       \g@varrow{cx#8}{tmp2}{sum}%
      \fi
%    \end{macrocode}
% Draw the horizontal line.
%    \begin{macrocode}
      \g@hline{cx#7}{sum}{cx#8}%
%    \end{macrocode}
% Insert space between the horizontal line and 
% the labels if at least one of the labels |#5| and |#6| is not empty.
% Typeset the labels into boxes and measure them. 
%    \begin{macrocode}
      \setbox\g@b@tmpa=\hbox{$#3{#5}$}%
      \setbox\g@b@tmpb=\hbox{$#4{#6}$}%
      \ifdim\ht\g@b@tmpa>\z@
       \g@advance{sum}{\the\labelskip}%
      \else
       \ifdim\ht\g@b@tmpb>\z@
        \g@advance{sum}{\the\labelskip}%
       \fi
      \fi
%    \end{macrocode}
% Draw the `from' label if there is one
%    \begin{macrocode}
      \g@d@tmpc\z@
      \ifdim\ht\g@b@tmpa>\z@
       \g@cbox{cx#7}{sum}{\kern-\p@\vcenter{\box\g@b@tmpa}}%
       \g@d@tmpc=\ht\g@label
      \fi
%    \end{macrocode}
% Draw the `to' label if there is one
%    \begin{macrocode}
      \ifdim\ht\g@b@tmpb>\z@
       \g@cbox{cx#8}{sum}{\kern-\p@\vcenter{\box\g@b@tmpb}}%
       \ifdim \ht\g@label>\g@d@tmpc
        \g@d@tmpc=\ht\g@label
       \fi
      \fi
%    \end{macrocode}
% Advance the sum by the maximum height of the two 
% labels and the desired space between two consecutive
% operations
%    \begin{macrocode}
      \g@advance{sum}{\the\g@d@tmpc}%
      \g@advance{sum}{\the\opskip}%
%    \end{macrocode}
% Update all column tower heights between |#7| and |#8| to
% |sum|.
%    \begin{macrocode}
      \g@updateArea{sum}{cy}{#7}{#8}%
%    \end{macrocode}
% That's it.
%    \begin{macrocode}
     \fi
    \fi
   \fi
  \fi
 \fi
}
%    \end{macrocode}
% \end{macro}
% 
% \begin{macro}{\g@east}
% \begin{macro}{\g@east@mult}
%
% The corresponding eastern macros are very similar to the 
% above defined northern versions. Maybe there is a way
% to define generic operation commands once for all regions,
% but this would at least lead to less comprehesive definitions.
% 
% We skip the definitions of |\g@east|, |\g@east@mult| and |\g@east@arrow|
% in this documentation.
%
\def\g@east{%
 \g@endregion 
 \def\swap{\g@east@arrow11\rowswapfromlabel\rowswaptolabel}
 \def\add{\g@east@arrow01\rowaddfromlabel\rowaddtolabel}
 \let\mult\g@east@mult
 \g@defdim{sum}{\z@}%
 \gdef\rowops{\PackageError{gauss}{Two sets of row operations were specified in %
  just one matrix. This is not allowed.}}
 \gdef\g@endregion{%
   \end{picture}\egroup
   \g@measureArea{rx}{0}{\the\g@maxrow}{sum}%
   \g@dim{sum}{\wd\g@eastbox}%
 }%
 \global\setbox\g@eastbox=\hbox\bgroup
  \begin{picture}(0,\g@double{h})(0,0)
   \linethickness{\g@linethickness}%
}
\def\g@east@mult#1#2{%
 \ifx *#1 
  \g@for0\to\g@maxrow\do{\g@east@mult{##1}{#2}}%
 \else
  \g@checkBounds{r}{0}{#1}{\the\g@maxrow}%
  \ifg@indexCorrect
   \g@rbox{rx#1}{ry#1}{\rowmultlabel{#2}}
   \g@dim{rx#1}{\g@d@tmpc}\advance\g@d@tmpc\wd\g@label
   \g@defdim{rx#1}{\g@d@tmpc}%
   \g@advance{rx#1}{\the\labelskip}%
   \g@update{sum}{rx#1}%
  \fi
 \fi
}
%
\def\g@east@arrow#1#2#3#4{%
 \@ifnextchar[%
 {\g@east@arrow@a{#1}{#2}{#3}{#4}}%
 {\g@east@arrow@b{#1}{#2}{#3}{#4}{}[]}%
}
\def\g@east@arrow@a#1#2#3#4[#5]{%
 \@ifnextchar[%
 {\g@east@arrow@b{#1}{#2}{#3}{#4}{#5}}%
 {\g@east@arrow@b{#1}{#2}{#3}{#4}{#5}[]}%
}
\def\g@east@arrow@b#1#2#3#4#5[#6]#7#8{%
 \ifx *#7
  \ifx *#8
   \g@for0\to\g@maxrow\do{\g@east@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{##1}{*}}%
  \else
   \g@for#8\to0\do{\g@east@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{##1}{#8}}%
   \g@for#8\to\g@maxrow\do{\g@east@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{##1}{#8}}%
  \fi
 \else
  \ifx *#8
   \g@for#7\to0\do{\g@east@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{#7}{##1}}%
   \g@for#7\to\g@maxrow\do{\g@east@arrow@b{#1}{#2}{#3}{#4}{#5}[#6]{#7}{##1}}%
  \else
   \ifnum #7=#8\else
    \g@checkBounds{r}{0}{#7}{\the\g@maxrow}%
    \ifg@indexCorrect
     \g@checkBounds{r}{0}{#8}{\the\g@maxrow}%
     \ifg@indexCorrect
      \g@defdouble{tmp1}{\g@double{rx#7}}%
      \g@defdouble{tmp2}{\g@double{rx#8}}%
      \ifx0#1
       \g@advance{rx#7}{\the\rowopminsize}%
      \else
       \g@advance{rx#7}{\the\g@arrowwd}%
      \fi
      \ifx0#2
       \g@advance{rx#8}{\the\rowopminsize}%
      \else
       \g@advance{rx#8}{\the\g@arrowwd}%
      \fi
      \g@measureArea{rx}{#7}{#8}{sum}%
      \ifx0#1
       \g@hline{tmp1}{ry#7}{sum}%
      \else
       \g@harrow{tmp1}{ry#7}{sum}%
      \fi
      \ifx0#2
       \g@hline{tmp2}{ry#8}{sum}%
      \else
       \g@harrow{tmp2}{ry#8}{sum}%
      \fi
      \g@vline{sum}{ry#7}{ry#8}%
      \setbox\g@b@tmpa=\hbox{$#3{#5}$}%
      \setbox\g@b@tmpb=\hbox{$#4{#6}$}%
      \ifdim\wd\g@b@tmpa>\z@
       \g@advance{sum}{\the\labelskip}%
      \else
       \ifdim\wd\g@b@tmpb>\z@
        \g@advance{sum}{\the\labelskip}%
       \fi 
      \fi 
      \g@d@tmpc\z@
      \ifdim\wd\g@b@tmpa>\z@
       \g@rbox{sum}{ry#7}{\kern-\p@\vcenter{\box\g@b@tmpa}}%
       \g@d@tmpc=\wd\g@label
      \fi
      \ifdim\wd\g@b@tmpb>\z@
       \g@rbox{sum}{ry#8}{\kern-\p@\vcenter{\box\g@b@tmpb}}%
       \ifdim \wd\g@label>\g@d@tmpc
        \g@d@tmpc=\wd\g@label
       \fi
      \fi
      \g@advance{sum}{\the\g@d@tmpc}%
      \g@advance{sum}{\the\opskip}%
      \g@updateArea{sum}{rx}{#7}{#8}%
     \fi
    \fi
   \fi
  \fi
 \fi
}
% \end{macro}\end{macro}
%
%
%
% \subsection{The \texttt{gmatrix} environment}
%
% |gmatrix| calls |#1matrix| where |matrix| is redefined to |g@matrix|.
% |g@matrix| typesets the matrix using |\halign| and stores the 
% operations into box registers |\g@northbox| and |\g@eastbox|, respectively.
% The matrix itself is stored into |\g@matrixbox|.
% 
% The ``real'' typesetting is done at the end of |gmatrix|.
%
% \begin{environment}{gmatrix}
% \dots and here is |gmatrix|:
%    \begin{macrocode}
\newenvironment{gmatrix}[1][]
{\unitlength=1pt\def\g@environment{#1matrix}%
 \begin{g@matrix}%
}{%
 \end{g@matrix}%
%    \end{macrocode}
% Delete temporarily the definition of |matrix|.
%    \begin{macrocode}
 \let\matrix\@empty
 \let\endmatrix\@empty
%    \end{macrocode}
% Find out sizes of the matrix. Set |\g@d@tmp| to the height of the matrix.
%    \begin{macrocode}
 \g@d@tmpa\ht\g@matrixbox \advance\g@d@tmpa\p@
 \g@d@tmpb\dp\g@matrixbox \advance\g@d@tmpb\p@
 \g@d@tmp\ht\g@northbox \ht\g@northbox\z@
 \dp\g@northbox\z@
 \ifdim \g@d@tmp>\z@
  \advance\g@d@tmp-\opskip
 \fi
 \advance\g@d@tmp.5\ht\g@matrixbox
 \advance\g@d@tmp.5\dp\g@matrixbox
%    \end{macrocode}
% Start the matrix environment to get the left delimiter.
%    \begin{macrocode}
 \begin{\g@environment}%
%    \end{macrocode}
% Typeset the column operations to the north of the matrix,
% and the matrix itself.
%    \begin{macrocode}
  \vcenter{\hbox{%
   \rlap{\raise\ht\g@matrixbox\box\g@northbox}% north
   % 1 additional pt above and below the matrix
   \rule\z@\g@d@tmpa\lower\g@d@tmpb\null
   \box\g@matrixbox% the matrix itself
  }}%
%    \end{macrocode}
% Close the matrix environment to get now the right delimiter.
%    \begin{macrocode}
 \end{\g@environment}%
%    \end{macrocode}
% Finally typeset the eastern operations.
% Insert vertical space of |\g@d@tmp| (the height
% of the matrix) and horizontal space of |\rowarrowsep| before.
%    \begin{macrocode}
 \rule\rowarrowsep\z@
 \rule\z@\g@d@tmp
 \g@dim{d}{\g@d@tmpa}%
 \vcenter{\hbox{\lower\g@d@tmpa\box\g@eastbox}}%
}
%    \end{macrocode}
% \end{environment}
% Here is the definition of |\g@endmatrix|. This is the initial |\g@endregion|
% which is defined within |\begin{gmatrix}| to finish the matrix input.
%    \begin{macrocode}
\def\g@endmatrix{%
   \mathstrut\crcr
  \egroup % end of \halign
 \egroup % end of \vbox, this contains the matrix
%    \end{macrocode}
% Save the matrix into matrixbox.
%    \begin{macrocode}
 \global\setbox\g@matrixbox\lastbox
%    \end{macrocode}
% Measure the matrix' dimensions.
%    \begin{macrocode}
 \g@measureAxis
 \setbox\g@trash=\vbox{%
  \unvcopy\g@matrixbox
%    \end{macrocode}
% Copy the last row of the matrix into |\g@eastbox| and reinsert it to the vbox.
%    \begin{macrocode}
  \global\setbox\g@eastbox=\lastbox
  \copy\g@eastbox
  \g@d@tmp\z@ {\g@measureRows}% measure rows
 }%
 \setbox\g@trash=\hbox{%
%    \end{macrocode}
% Insert a box of width 100cm to recognize the beginning of the hbox within the 
% measurement recursion.
%    \begin{macrocode}
  \hbox to 100cm{.\hfill.}%
  \unhbox\g@eastbox
  \g@d@tmp\z@ {\g@measureCols}% measure columns
 }%
%    \end{macrocode}
% Determine global dimensions of the matrix (total height, etc.).
%    \begin{macrocode}
 \g@d@tmpa=\ht\g@matrixbox\advance\g@d@tmpa\dp\g@matrixbox
 \g@defdim{h}{\g@d@tmpa}%
 \g@defdim{w}{\wd\g@matrixbox}%
 \g@defdim{d}{\dp\g@matrixbox}%
}%
%    \end{macrocode}
% \begin{environment}{g@matrix}
% Finally, we have the following definition of |g@matrix|:
%    \begin{macrocode}
\edef\g@prae{\hfil\noexpand\mathstrut$\relax}
\edef\g@post{\relax$\hfil} 
\newenvironment{g@matrix}
{\setbox\g@trash=\hbox\bgroup
  \global\g@maxrow@old\g@maxrow
  \global\g@maxcol@old\g@maxcol
  \global\g@maxrow0%
  \global\g@maxcol0%
  \let\rowops\g@east
  \let\colops\g@north
  \vbox\bgroup
   % count rows while typesetting
   \def\\{\mathstrut\cr\global\advance\g@maxrow1\relax}%
   \global\let\g@endregion\g@endmatrix
   \global\g@tab=2\arraycolsep
   \ialign\bgroup\g@prae##\g@post&&\kern\g@tab\g@prae##\g@post\cr
}{%
  \g@endregion
 \egroup % end of \hbox
 % enable nested gmatrixes (for DniQ :-)
 \global\g@maxrow\g@maxrow@old
 \global\g@maxcol\g@maxcol@old
 \global\let\g@endregion\g@endmatrix
 \global\let\rowops\g@east
 \global\let\colops\g@north
}
%    \end{macrocode}
% \end{environment}
% 
%
%
% \subsection{Public tools}
%
% \begin{macro}{\newmatrix}
% The |\newmatrix| command allows to define new matrix environments with
% special delimiters as described in Section~1.
%
%    \begin{macrocode}
\def\newmatrix#1#2#3{%
 \ifx g#3 \else
  \ifx {#3}{g@} \else
   \expandafter\ifx\csname#3matrix\endcsname\relax
    \newenvironment{#3matrix}%
     {\left#1\begin{matrix}}{\end{matrix}\right#2}%
   \else
    \renewenvironment{#3matrix}%
     {\left#1\begin{matrix}}{\end{matrix}\right#2}%
   \fi
  \fi
 \fi
}
%    \end{macrocode}
%
% For compatibility reasons, we redefine predefined matrix environments such
% as |pmatrix|. This is necessary to avoid problems that arise when dealing with
% earlier \AmS\TeX\ versions.
%
%    \begin{macrocode}
\newmatrix()p
\newmatrix[]b
\newmatrix\lbrace\rbrace B
\newmatrix\lvert\rvert v
\newmatrix\lVert\rVert V
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\rowmultlabel}\begin{macro}{\colmultlabel}
% \begin{macro}{\rowaddfromlabel}\begin{macro}{\coladdfromlabel}
% \begin{macro}{\rowaddtolabel}\begin{macro}{\coladdtolabel}
% \begin{macro}{\rowswapfromlabel}\begin{macro}{\colswapfromlabel}
% \begin{macro}{\rowswaptolabel}\begin{macro}{\colswaptolabel}
%  Labels of operations are typeset using the so-called fontifying macros
%  described in Section~\ref{ssec:atp}.
%  All of them take exaclty one argument, and they are called within math
%  mode. The user may redefine them to adjust the appearence of operations
%  according to his needs. The following is the standard definition:

%    \begin{macrocode}
\def\rowmultlabel#1{|\,#1}
\def\rowswapfromlabel#1{}
\def\rowswaptolabel#1{}
\def\rowaddfromlabel#1{\scriptstyle #1}
\def\rowaddtolabel#1{\scriptscriptstyle +}
\def\colmultlabel#1{%
 \underline{\hbox to 1.2em{$\hss\mathstrut{}#1\hss$}}%
}
\def\colswapfromlabel#1{}
\def\colswaptolabel#1{}
\def\coladdfromlabel#1{\scriptstyle #1}
\def\coladdtolabel#1{\scriptscriptstyle +}
%    \end{macrocode}
% \end{macro}\end{macro}
% \end{macro}\end{macro}
% \end{macro}\end{macro}
% \end{macro}\end{macro}
% \end{macro}\end{macro}
%
% \begin{macro}{\colarrowsep}
% \begin{macro}{\rowarrowsep}
% \begin{macro}{\labelskip}
% \begin{macro}{\opskip}
% \begin{macro}{\colopminsize}
% \begin{macro}{\rowopminsize}
% Finally, we define the public lengths of Section~\ref{ssec:atp}:
%    \begin{macrocode}
\newdimen\colarrowsep\colarrowsep=.5em
\newdimen\rowarrowsep\rowarrowsep=.5em
\newdimen\opskip\opskip=5pt
\newdimen\labelskip\labelskip=4pt
\newdimen\colopminsize\colopminsize=3pt
\newdimen\rowopminsize\rowopminsize=3pt
%    \end{macrocode}
% \end{macro}\end{macro}
% \end{macro}\end{macro}
% \end{macro}\end{macro}
%
% And that's all.
%    \begin{macrocode}
\makeatother
%    \end{macrocode}
% \CheckSum{1188}
% \Finale
\endinput