summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/mdwtools/syntax.dtx
blob: 1f251b736341918cd100b22bdd1c4f6c9b32e920 (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
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
% \begin{meta-comment}
%
% syntax.dtx
%
% Syntax typesetting package for LaTeX 2e
%
% (c) 1996 Mark Wooding
%
% \end{meta-comment}
%
% \begin{meta-comment} <general public licence>
%%
%% syntax package -- typesetting syntax descriptions
%% Copyright (c) 1996 Mark Wooding
%%
%% This program is free software; you can redistribute it and/or modify
%% it under the terms of the GNU General Public License as published by
%% the Free Software Foundation; either version 2 of the License, or
%% (at your option) any later version.
%%
%% This program is distributed in the hope that it will be useful,
%% but WITHOUT ANY WARRANTY; without even the implied warranty of
%% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
%% GNU General Public License for more details.
%%
%% You should have received a copy of the GNU General Public License
%% along with this program; if not, write to the Free Software
%% Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
%%
% \end{meta-comment}
%
% \begin{meta-comment} <Package preamble>
%<+package>\NeedsTeXFormat{LaTeX2e}
%<+package>\ProvidesPackage{syntax}
%<+package>                [1996/05/17 1.07 Syntax typesetting (MDW)]
% \end{meta-comment}
%
% \CheckSum{1439}
%% \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         \~}
%%
%
% \begin{meta-comment} <driver>
%
%<*driver>
%
% This hacking will remember the old default underscore character.  Even if
% T1 fonts are being used, it will get the grotty version.  Why is it that
% all of the encoding handling ends up looking like this?
%
\expandafter\let\expandafter\oldus\csname?\string\textunderscore\endcsname
%
\input{mdwtools}
\describespackage{syntax}
\DeclareRobustCommand\syn{\package{syntax}}
\mdwdoc
%</driver>
%
% \end{meta-comment}
%
% \section{User guide}
%
% \subsection{Introduction}
%
% The \syn\ package provides a number of commands and environments which
% extend \LaTeX\ and allow you to typeset good expositions of syntax.
%
% The package provides several different types of features: probably not all
% of these will be required by every document which needs the package:
% \begin{itemize}
% \item A system of abbreviated forms for typesetting syntactic items.
% \item An environment for typesetting BNF-type grammars
% \item A collection of environments for building syntax diagrams.
% \end{itemize}
%
% The package also includes some other features which, while not necessarily
% syntax-related, will probably come in handy for similar types of document:
% \begin{itemize}
% \item An abbreviated notation for verbatim text, similar to the
%       \package{shortvrb} package.
% \item A slightly different underscore character, which works as expected
%       in text and maths modes.
% \end{itemize}
%
% \subsection{The abbreviated verbatim notation}
%
% In documents describing programming languages and libraries, it can become
% tedious to type "\verb|...|" every time.  Like Frank Mittelbach's
% \package{shortvrb} package, \syn\ provides a way of setting up single-^^A
% character abbreviations.  The only real difference between the two is that
% the declarations provided by \syn\ obey \LaTeX's normal scoping rules.
%
% \DescribeMacro\shortverb
% You can set up a character as a `verbatim shorthand' character using the
% |\shortverb| command.  This takes a single argument, which should be a
% single-character control sequence containing the character you want to use.
% So, for example, the command
% \begin{listing}
%\shortverb{\|}
% \end{listing}
% would set up the `"|"' character to act as a verbatim delimiter.  While a
% |\shortverb| declaration is in force, any text surrounded by (in this case)
% vertical bar characters will be typeset as if using the normal |\verb|
% command.
%
% \DescribeEnv{shortverb}
% Since \LaTeX\ allows any declaration to be used as an environment, you can
% use a \env{shortverb} environment to delimit the text over which your
% character is active:
% \begin{listing}
%Some text...
%\begin{shortverb}{\|}
%...
%\end{shortverb}
% \end{listing}
%
% \DescribeMacro\unverb
% If you want to disable a |\shortverb| character without ending the scope
% of other declarations, you can use the |\unverb| command, passing it
% a character as a control sequence, in the same way as above.
%
% The default \TeX/\LaTeX\ underscore character is rather too short for
% use in identifiers.  For example:
%
% \begingroup \let\_=\oldus
% \begin{demo}{Old-style underscores}
%Typing long underscore-filled
%names, like big\_function\_name,
%is normally tedious.  The normal
%positioning of the underscore
%is wrong, too.
% \end{demo}
% \endgroup
%
% The \syn\ package redefines the |\_| command to draw a more attractive
% underscore character.  It also allows you to use the |_|~character
% directly to produce an underscore outside of maths mode: |_|~behaves
% as a subscript character as usual inside maths mode.
%
% \begin{demo}{New \syn\ underscores}
%You can use underscore-filled
%names, like big_function_name,
%simply and naturally.  Of
%course, subscripts still work
%normally in maths mode, e.g.,
%$x_i$.
% \end{demo}
%
% \subsection{Typesetting syntactic items}
% \begin{synshorts}
%
% The \syn\ package provides some simple commands for typesetting syntactic
% items.
%
% \DescribeMacro\synt
% Typing "\\synt{"<text>"}" typesets <text> as a \lq non-terminal',
% in italics and surrounded by angle brackets.  If you use "\\synt" a lot,
% you can use the incantation
% \begin{listing}
%\def\<#1>{\synt{#1}}
% \end{listing}
% to allow you to type "\\<"<text>">" as an alternative to 
% "\\synt{"<text>"}".
%
% \DescribeMacro\lit
% You can also display literal text, which the reader should type directly,
% using the "\\lit" command.
%
% \begin{demo}{Use of \cmd\lit}
%Type \lit{ls} to display a
%list of files.
% \end{demo}
%
% Note that the literal text appears in quotes.  To suppress the quotes,
% use the `*' variant.
%
% The "\\lit" command produces slightly better output than "\\verb" for
% running text, since the spaces are somewhat narrower.  However, "\\verb"
% allows you to type arbitrary characters, which are treated literally,
% whereas you must use commands such as "\\{" to use special characters
% within the argument to "\\lit".  Of course, you can use "\\lit" anywhere
% in the document: "\\verb" mustn't be used inside a command argument.
% \end{synshorts}
%
% \subsection{Abbreviated forms for syntactic items}
%
% It would be very tedious to require the use of commands like |\synt|
% when building syntax descriptions like BNF grammars.  It would also make
% your \LaTeX\ source hard to read.  Therefore, \syn\ provides some
% abbreviated forms which make typesetting syntax quicker and easier.
%
% Since the abbreviated forms use several characters which you may want to
% use in normal text, they aren't enabled by default.  They only work
% with special commands and environments provided by the \syn\ package.
%
% The abbreviated forms are shown in the table below:
%
% \begin{tab}[\synshorts]{ll}                                       \hline
% \bf Input        & \bf Output                                  \\ \hline
% "<some text>"    & <some text>                                 \\
% "`some text'"    & `some text'                                 \\
% "\"some text\""  & "some text"                                 \\ \hline
% \end{tab}
%
% Within one of these abbreviated forms, text is treated more-or-less
% verbatim:
% \begin{itemize}
%
% \item Any |$|, |%|, |^|, |&|, |{|, |}|, |~| or |#| characters are treated
%       literally: their normal special meanings are ignored.
%
% \item Other special characters, with the exception of |\|, are also treated
%       literally: this includes any characters made special by |\shortverb|.
%
% \end{itemize}
%
% However, the |\| character retains its meaning.  Since the brace
% characters are not recognised, most commands can't be used within
% abbreviated forms.  However, you can use special commands to type some
% of the remaining special characters:
%
% \begin{tab}[\synshorts]{ll}                                       \hline
% \bf Command   & \bf Result                                     \\ \hline
% "\\\\"        & A `\\' character                               \\
% "\\>"         & A `>' character                                \\
% "\\'"         & A `\'' character                               \\
% "\\\""        & A `"' character                                \\
% "\\\ "        & A `\ ' character (not a space)                 \\ \hline
% \end{tab}
%
% Note that |\\|, |\>|, |\"| and \verb*|\ | are only useful in a |\tt| font,
% i.e., inside |`...'| and |"..."| forms, since the characters don't exist
% in normal fonts.  The |\>|, |\"| and |\'| commands are only provided so
% you can use these characters within |<...>|, |"..."| and |`...'| forms
% respectively: in the other forms, there is no need to use the special
% command.
%
% In addition, when the above abbreviations are enabled, the character "|"
% is set to typeset a \syntax{|} symbol, which is conventionally used to
% separate alternatives in syntax descriptions.
%
% \DescribeMacro\syntax
% Normally, these abbreviated forms are enabled only within special
% environments, such as \env{grammar} and \env{syntdiag}.  To use them
% in running text, use the |\syntax| command.  The abbreviations are made
% active within the argument of the |\syntax| command.\footnote{^^A
%   The argument of the \cmd\syntax\ command may contain commands such
%   as \cmd\verb, which are normally not allowed within arguments.
% }  Note that you cannot use the |\syntax| command within the argument
% of another command.
%
% \DescribeMacro\synshorts
% \DescribeEnv{synshorts}
% You can also enable the syntax shortcuts using the |\synshorts| declaration
% or the \env{synshorts} environment.  This enables the syntax shortcuts
% until the scope of the declaration ends.
%
% \DescribeMacro\synshortsoff
% If syntax shortcuts are enabled, you can disable them using the
% |\synshortsoff| declaration.
%
% \subsection{The \env{grammar} environment}
%
% \DescribeEnv{grammar}
% For typesetting formal grammars, for example, of programming languages,
% the \syn\ package provides a \env{grammar} environment.  Within this
% environment, the abbreviated forms described above are enabled.
%
% Within the environment, separate production rules should be separated by
% blank lines.  You can use the normal |\\| command to perform line-breaking
% of a production rule.  Note that a production rule must begin with a
% nonterminal name enclosed in angle brackets (|<| \dots |>|), followed by
% whitespace, then some kind of production operator (usually `::=') and then
% some more whitespace.  You can control how this text is actually typeset,
% however.
%
% \DescribeMacro{\[[}
% \DescribeMacro{\]]}
% You can use syntax diagrams (see below) instead of a straight piece of BNF
% by enclosing it in a |\[[| \dots |\]]| pair.  Note that you can't mix 
% syntax diagrams and BNF in a production rule, and you will get something
% which looks very strange if you try.
%
% \DescribeMacro\alt
% In addition, a command |\alt| is provided for splitting long production
% rules over several lines: the |\alt| command starts a new line and places
% a \syntax{|} character slightly in the left margin.  This is useful when
% a symbol has many alternative productions.
%
% \begin{demo}[w]{The \env{grammar} environment}
%\begin{grammar}
%<statement> ::= <ident> `=' <expr>
%  \alt `for' <ident> `=' <expr> `to' <expr> `do' <statement>
%  \alt `{' <stat-list> `}'
%  \alt <empty>
%
%<stat-list> ::= <statement> `;' <stat-list> | <statement>
%\end{grammar}
% \end{demo}
%
% You can modify the appearance of grammars using three length parameters:
%
% \begin{description} \def\makelabel{\hskip\labelsep\cmd}
%
% \item [\grammarparsep] is the amount of space inserted between production
%       rules.  It is a rubber length whose default value is 8\,pt, with
%       1\,pt of stretch and shrink.
%
% \item [\grammarindent] is the amount by which the right hand side of a
%       production rule is indented from the left margin.  It is a rigid
%       length.  Its default value is 2\,em.
%
% \end{description}
%
% \DescribeMacro\grammarlabel
% You can also control how the `label' is typeset by redefining the
% |\grammarlabel| command.  The command is given two arguments: the name of
% the nonterminal (which was enclosed in angle brackets), and the `production
% operator'.  The command is expected to produce the label.  By default, it
% typesets the nonterminal name using |\synt| and the operator at opposite
% ends of the label, separated by an |\hfill|.
%
% \subsection{Syntax diagrams}
%
% A full formal BNF grammar can be somewhat overwhelming for less technical
% readers.  Documents aimed at such readers tend to display grammatical
% structures as \emph{syntax diagrams}.
%
% \DescribeEnv{syntdiag}
% A syntax diagram is always enclosed in a \env{syntdiag} environment.  You
% should think of the environment as enclosing a new sort of \LaTeX\ mode:
% trying to type normal text into a syntax diagram will result in very ugly
% output.  \LaTeX\ ignores spaces and return characters while in syntax
% diagram mode.
%
% The syntax of the environment is very simple:
%
% \begin{grammar}
% <synt-diag-env> ::= \[[
%   "\\begin{syntdiag}"
%   \begin{stack} \\ "[" <decls> "]" \end{stack}
%   <text>
%   "\\end{syntdiag}"
% \]]
% \end{grammar}
%
% The \<decls> contain any declarations you want to insert, to control
% the environment.  The parameters to tweak are described below.
%
% Within a syntax diagram, you can include syntactic items using the
% abbreviated forms described elsewhere.  The output from these forms is
% modified slightly in syntax diagram mode so that the diagram looks
% right.
%
% I probably ought to point out now that the syntax diagram typesetting
% commands produce beautiful-looking diagrams with all the rules and curves
% accurately positioned.  Some device drivers don't position these objects
% correctly in their output.  I've had particular trouble with |dvips|.  I'll
% say it again: it's not my fault!
%
% \DescribeEnv{syntdiag*}
% The \env{syntdiag} environment only works in paragraph mode, and it acts
% rather like a paragraph, splitting over several lines when appropriate.
% If you just want to typeset a snippet of a syntax diagram, you can
% use the starred environment \env{syntdiag$*$}.
%
% \begin{grammar}
% <synt-diag-star-env> ::= \[[
%   "\\begin{syntdiag*}"
%   \begin{stack} \\ "[" <decls> "]" \end{stack}
%   \begin{stack} \\ "[" <width> "]" \end{stack}
%   <text>
%   "\\end{syntdiag*}"
% \]]
% \end{grammar}
%
% When typesetting little demos like this, it's not normal to fully adorn
% the syntax diagram with the full double arrows
% (`\begin{syntdiag*}[\left{>>-}\right{-><}]\tok{$\cdots$}\end{syntdiag*}').
% The two declarations \syntax{"\\left{"<arrow>"}" and "\\right{"<arrow>"}"}
% allow you to choose the arrows on each side of the syntax diagram snippet.
% The possible values of \<arrow> are shown in the table-ette below:
%
% ^^A Time to remember what I learned about tables while writing mdwtab.
% ^^A Just for the embarassment factor, here's the number of attempts I
% ^^A took to get the table below to look right: __6.  Hmm... not as bad
% ^^A as I expected.  Most of them were fine-tuning things.
%
% \medskip					^^A Leave a vertical gap
% \hbox to\columnwidth{\hfil\vbox{\tabskip=0pt	^^A Centre it horizontally
% \sdsize \csname sd@setsize\endcsname		^^A Position syntdiag arrows
% \halign to .5\columnwidth{			^^A Set the table width
%   &\ttfamily\ignorespaces#\unskip\hfil\tabskip=0pt ^^A Typeset the name
%   &\quad\csname sd@arr@#\endcsname\hfil	^^A Typeset the arrow
%   &\setbox0=\hbox{#}\tabskip=0pt plus 1fil\cr	^^A Stretch between columns
%   >>-&>>-&   &>-&>-&   &->&->\cr
%   -><&-><&   &...&...& &-&-\cr
% }}\hfil}					^^A Close the boxing
% \medskip					^^A And leave another gap
%
% These declarations should be used only in the optional argument to the
% \env{syntdiag$*$} command.  The second optional argument to the
% environment, if specified, fixes the width of the syntax diagram snippet;
% if you omit this argument, the diagram is made just wide enough to
% fit everything in.
%
% \begin{figure}
% \begin{demo}[w]{Example of \env{syntdiag$*$}}
%\newcommand{\bs}[2]{%
%  \begin{minipage}{1.6in}%
%  \begin{syntdiag*}[\left{#1}\right{#2}][1.6in]%
%}
%\newcommand{\es}{\end{syntdiag*}\end{minipage}}
%
%\begin{center}
%\begin{tabular}{cl}                                      \\ \hline
%\bf Construction    & \bf Meaning                        \\ \hline
%\bs {>>-} {...} \es & Start of syntax diagram            \\
%\bs {...} {-><} \es & End of syntax diagram              \\
%\bs {>-}  {...} \es & Continued on next line             \\
%\bs {...} {->}  \es & Continued from previous line       \\ \hline
%\bs {...} {...}
%  \begin{stack} <option-a> \\ <option-b> \\ <option-c> \end{stack}
%\es                 & Alternatives: choose any one       \\
%\bs {...} {...}
%  \begin{rep} <repeat-me> \\ <separator> \end{rep}
%\es                 & One or more items, with separators \\ \hline
%\end{tabular}
%\end{center}
% \end{demo}
% \end{figure}
%
% \DescribeMacro\tok
% You can also include text using the |\tok| command.  The argument of this
% command is typeset in \LaTeX's LR~mode and inserted into the diagram. 
% Syntax abbreviations are allowed within the argument, so you can, for
% example, include textual descriptions like
% \begin{listing}
%\tok{any <char> except `"'}
% \end{listing}
%
% \DescribeEnv{stack}
% Within a syntax diagram, a choice between several different items is
% shown by stacking the alternatives vertically.  In \LaTeX, this is done
% by enclosing the items in a \env{stack} environment.  Each individual item
% is separated by |\\| commands, as in the \env{array} and \env{tabular}
% environments.  Each row may contain any syntax diagram material, including
% |\tok| commands and other \env{stack} environments.
%
% Note if you end a \env{stack} environment with a |\\| command, a blank
% row is added to the bottom of the stack, indicating that none of the items
% need be specified.
%
% \DescribeEnv{rep}
% Text which can be repeated is enclosed in a \env{rep} environment: the
% text is displayed with a backwards pointing arrow drawn over it, showing
% that it may be repeated.  Optionally, you can specify text to be
% displayed in the arrow, separating it from the main text with a |\\|
% command.
%
% Note that items on the backwards arrow of a \env{rep} construction should
% be displayed \emph{backwards}.  You must put the individual items in
% reverse order when building this part of your diagrams.  \syn\ will 
% correctly reverse the arrows on \env{rep} structures, but apart from
% this, you must cope on your own.  You are recommended to keep these parts
% of your diagrams as simple as possible to avoid confusing readers.
%
% \begin{demo}[w]{A syntax diagram}
%\begin{syntdiag}
%<ident> `('
%  \begin{rep} \begin{stack} \\
%    <type> \begin{stack} \\ <ident> \end{stack}
%  \end{stack} \\ `,' \end{rep}
%\begin{stack} \\ `...' \end{stack} `)'
%\end{syntdiag}
% \end{demo}
%
% \subsubsection{Line breaking in syntax diagrams}
%
% Syntax diagrams are automatically broken over lines and across pages.
% Lines are only broken between items on the outermost level of the diagram:
% i.e., not within \env{stack} or \env{rep} environments.
%
% You can force a line break at a particular place by using the |\\| command
% as usual.  This supports all the usual \LaTeX\ features: a `|*|' variant
% which prohibits page breaking, and an optional argument specifying the
% extra vertical space between lines.
%
% \subsubsection{Customising syntax diagrams}
%
% There are two basic styles of syntax diagrams supported:
%
% \begin{description}
%
% \item [square] Lines in the syntax diagram join at squared-off corners.
%       This appears to be the standard way of displaying syntax diagrams
%       in IBM manuals, and most other documents I've seen.
%
% \item [rounded] Lines curve around corners.  Also, no arrows are drawn
%       around repeating loops: the curving of the lines provides this
%       information instead.  This style is used in various texts on
%       Pascal, and appears to be more popular in academic circles.
%
% \end{description}
%
% You can specify the style you want to use for syntax diagrams by giving
% the style name as an option on the |\usepackage| command.  For example,
% to force rounded edges to be used, you could say
%
% \begin{listing}
%\usepackage[rounded]{syntax}
% \end{listing}
%
% \DescribeMacro\sdsize
% \DescribeMacro\sdlengths
% The \env{syntdiag} environment takes an option argument, which should
% contain declarations which are obeyed while the environment is set up.
% The default value of this argument is `|\sdsize\sdlengths|'.  The
% |\sdsize| command sets the default type size for the environment: this is
% normally |\small|.  |\sdlengths| sets the values of the length parameters
% used by the environment based on the current text size.  These parameters
% are described below.
%
% For example, if you wanted to reduce the type size of the diagrams still
% further, you could use the command
% \begin{listing}
%\begin{syntdiag}[\tiny\sdlengths]
% \end{listing}
%
% The following length parameters may be altered:
%
% \begin{description} \def\makelabel{\hskip\labelsep\cmd}
%
% \item [\sdstartspace] The length of the rule between the arrows which
%       begin each line of the syntax diagram and the first item on the line.
%       Note that most objects have some space on either side of them as
%       well.  This is a rubber length.  Its default value is 1\,em, although
%       it can shrink by up to 10\,pt.
%
% \item [\sdendspace] The length of the rule between the last item on a
%       line and the arrow at the very end.  Note that the final line also
%       has extra rubber space on the end.  This is a rubber length.  Its
%       default value is 1\,em, although it will shrink by up to 10\,pt.
%
% \item [\sdmidskip] The length of the rule on either side of a large
%       construction (either a \env{stack} or a \env{rep}).  It is a rubber
%       length.  Its default value is \smallf 1/2\,em, with a very small
%       amount of infinite stretch.
%
% \item [\sdtokskip] The length of the rule on either side of a |\tok|
%       item or syntax abbreviation.  It is a rubber length.  Its default
%       value is \smallf 1/4\,em, with a very small amount of inifnite
%       stretch.
%
% \item [\sdfinalskip] The length of the rule which finishes the last line
%       of a syntax diagram.  It is a rubber length.  Its default value is
%       \smallf 1/2\,em, with 10000\,fil of stretch, which will left-align
%       the items on the line.\footnote{^^A
%         This is a little \TeX nical.  The idea is that if a stray 1\,fil
%         of stretch is added to the end of the line, it won't be noticed.
%         However, the alignment of the text on the line can still be
%         modified using \cmd{\sd@rule}\cmd{\hfill}, if you're feeling
%         brave.
%       }
%
% \item [\sdrulewidth] Half the width of the rules used in the diagram.
%       It is a rigid length.  Its default value is 0.2\,pt.
%
% \item [\sdcirclediam] The diameter of the circle from which the quadrants
%       used in rounded-style diagrams are taken.  This must be a multiple
%       of 4\,pt, or else the lines on the diagram won't match up.
%
% \end{description}
%
% In addition, you should call |\sdsetstrut| passing it the total height
% (\({\rm height}+{\rm depth}\)) of a normal line of text at the current
% size.  Normally, the value of |\baselineskip| will be appropriate.
%
% You can also alter the appearance of \env{stack}s and \env{rep}s by using
% their optional positioning arguments.  By default, \env{stack}s descend
% below the main line of the diagram, and \env{rep}s extend above it.  
% Specifying an optional argument of |[b]| for either environment reverses
% this, putting \env{stack}s above and \env{rep}s below the line.
%
% \subsection{Changing the presentation styles}
%
% You can change the way in which the syntax items are typeset by altering
% some simple commands (using |\renewcommand|).  Each item (nonterminals,
% as typeset by |\synt|, and quoted and unquoted terminals, as typeset by
% |\lit| and |\lit*|) has two style commands associated with it, as shown
% in the table below.
%
% \begin{tab}{lll}						   \hline
% \bf Syntax item	& \bf Left command & \bf Right command	\\ \hline
% Nonterminals		& |\syntleft|	   & |\syntright|	\\
% Quoted terminals	& |\litleft|	   & |\litright|	\\
% Unquoted terminals	& |\ulitleft|	   & |\ulitright|	\\ \hline
% \end{tab}
%
% It's not too hard to see how this works.  For example, if you look at
% the implementation for |\syntleft| and |\syntright| in the implementation
% section, you'll notice that they're defined like this:
% \begin{listing}
%\newcommand{\syntleft}{$\langle$\normalfont\itshape}
%\newcommand{\syntright}{$\rangle$}
% \end{listing}
% I think this is fairly simple, if you understand things like font changing.
%
% Note that changing these style commands alters the appearance of all syntax
% objects of the appropriate types, as created by the |\synt| and |\lit|
% commands, in \env{grammar} environments, and in syntax diagrams.
%
%
% \section{Change history}
%
% \subsection*{Version 1.07}
%
% \begin{itemize}
% \item Fixed problem with underscore hacking in a \env{tabbing} environment.
% \end{itemize}
%
% \subsection*{Version 1.06}
%
% \begin{itemize}
% \item Added style hooks for syntax items.
% \item Improved colour handling in syntax diagrams, thanks to the |\doafter|
%       package.
% \item Fixed some nasty bugs in the \env{grammar} environment which confused
%       other lists and ruined the spacing.  The \env{grammar} handling is
%       now much tidier in general.
% \end{itemize}
%
% \subsection*{Version 1.05}
%
% \begin{itemize}
% \item Fixed `the bug' in the syntax diagram typesetting.  It now breaks
%       lines almost psychically, and doesn't break in the wrong places.
% \item Almost rewrote the \env{grammar} environment.  It now does lots of
%       the list handling itself, to allow more versatile typesetting of the
%       left hand sides.  There's lots of evil in there now.
% \item Added some more configurability.  In particular, two new settings
%       have been added to control \env{grammar} environments, and a neat
%       way of adding new syntax diagram structures has been introduced.
% \end{itemize}
%
% \subsection*{Version 1.04}
%
% \begin{itemize}
% \item Changed the vertical positioning of the rules, to make all the text
%       line up properly.  While the old version was elegant and simple, it
%       had the drawback of looking nasty.
% \item Allow line breaks at underscores, but don't if there's another one
%       afterwards.  Also, prevent losing following space if underscore is
%       written to a file.
% \end{itemize}
%
% \subsection*{Version 1.02}
%
% \begin{itemize}
% \item Added support for rounded corners in syntax diagrams.
% \item Changed lots of |\hskip| commands to |\kern|s, to prevent possible
%       line breaks.
% \end{itemize}
%
% \subsection*{Version 1.01}
%
% \begin{itemize}
% \item Allowed disabling of underscore active character, to avoid messing
%       up filenames.
% \item Added |\grammarparsep| and |\grammarindent| length parameters to
%       control the appearance of grammars.
% \end{itemize}
%
% \implementation
%
% \section{Implementation of \syn}
%
%    \begin{macrocode}
%<*package>
%    \end{macrocode}
%
% \subsection{Options handling}
%
% We define all the options we know about, and then see what's been put
% on the usepackage line.
%
% The options we provide currently are as follows:
%
% \begin{description}
% \item [rounded] draws neatly rounded edges on the diagram.
% \item [square] draws squared-off edges on the diagram.  This is the
%       default.
% \item [nounderscore] disables the undescore active character,  The |\_|
%       command still produces the nice version created here.
% \end{description}
%
%    \begin{macrocode}
\DeclareOption{rounded}{\sd@roundtrue}
\DeclareOption{square}{\sd@roundfalse}
\DeclareOption{nounderscore}{\@uscorefalse}
%    \end{macrocode}
%
% Now process the options:
%
%    \begin{macrocode}
\newif\ifsd@round
\newif\if@uscore\@uscoretrue
\ExecuteOptions{square}
\ProcessOptions
%    \end{macrocode}
%
% \subsection{Special character handling}
%
% A lot of the \syn\ package requires the use special active characters.
% These must be added to two lists: |\dospecials|, which is used by |\verb|
% and friends, and |\@sanitize|, which is used by |\index|.  The two macros
% here, |\addspecial| and |\remspecial|, provide these registration
% facilities.
%
% Two similar macros are found in Frank Mittelbach's \package{doc} package:
% these have the disadvantage of global operation.  My macros here are based
% on Frank's, which in turn appear to be based on Donald Knuth's list
% handling code presented in Appendix~D of \textit{The \TeX book}.
%
% Both these macros take a single argument: a single-character control
% sequence containing the special character to be added to or removed from
% the lists.
%
% \begin{macro}{\addspecial}
%
% This is reasonably straightforward.  We remove the sequence from the lists,
% in case it's already there, and add it in in the obvious way.  This
% requires a little bit of fun with |\expandafter|.
%
%    \begin{macrocode}
\def\addspecial#1{%
  \remspecial{#1}%
  \expandafter\def\expandafter\dospecials\expandafter{\dospecials\do#1}%
  \expandafter\def\expandafter\@santize\expandafter{%
    \@sanitize\@makeother#1}%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\remspecial}
%
% This is the difficult bit.  Since |\dospecials| and |\@sanitize| have the
% form of list macros, we can redefine |\do| and |\@makeother| to do the
% job for us.  We must be careful to put the old meaning of |\@makeother|
% back.  The current implementation assumes it knows what |\@makeother| does.
%
%    \begin{macrocode}
\def\remspecial#1{%
  \def\do##1{\ifnum`#1=`##1 \else\noexpand\do\noexpand##1\fi}%
  \edef\dospecials{\dospecials}%
  \def\@makeother##1{\ifnum`#1=`##1 \else%
    \noexpand\@makeother\noexpand##1\fi}%
  \edef\@sanitize{\@sanitize}%
  \def\@makeother##1{\catcode`##112}%
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsection{Underscore handling}
%
% When typing a lot of identifiers, it can be irksome to have to escape
% all `|_|' characters in the manuscript.  We make the underscore character
% active, so that it typesets an underscore in horizontal mode, and does
% its usual job as a subscript operator in maths mode.  Underscore must
% already be in the special character lists, because of its use as a
% subscript character, so this doesn't cause us a problem.
%
% \begin{macro}{\underscore}
%
% The |\underscore| macro typesets an underline character, using a horizontal
% rule.  This is positioned slightly below the baseline, and is also slightly
% wider than the default \TeX\ underscore.  This code is based on a similar
% implementation found in the \package{lgrind} package.
%
%    \begin{macrocode}
\def\underscore{%
  \leavevmode%
  \kern.06em%
  \vbox{%
    \hrule\@width.6em\@depth.4ex\@height-.34ex%
  }%
  \ifdim\fontdimen\@ne\font=\z@%
    \kern.06em%
  \fi%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\@uscore}
%
% This macro is called by the `|_|' active character to sort out what to do.
%
% If this is maths mode, we use the |\sb| macro, which is already defined
% to do subscripting.  Otherwise, we call |\textunderscore|, which picks the
% nicest underscore it can find.
%
% There's some extra cunningness here, because I'd like to be able to
% hyphenate after underscores usually, but not when there's another one
% following.  And then, because \env{tabbing} redefines |\_|, there's some
% more yukkiness to handle that: the usual |\@tabacckludge| mechanism doesn't
% cope with this particular case.
%
%    \begin{macrocode}
\let\usc@builtindischyphen\-
\def\@uscore.{%
  \ifmmode%
    \expandafter\@firstoftwo%
  \else%
    \expandafter\@secondoftwo%
  \fi%
  \sb%
  {\textunderscore\@ifnextchar_{}{\usc@builtindischyphen}}%
}
%    \end{macrocode}
%
% \end{macro}
%
% Now we set up the active character.  Note the |\protect|, which makes
% underscores work reasonably well in moving arguments.  Note also the way
% we end with a some funny stuff to prevent spaces being lost if this is
% written to a file.
%
%    \begin{macrocode}
\if@uscore
  \AtBeginDocument{%
    \catcode`\_\active%
    \begingroup%
    \lccode`\~`\_%
    \lowercase{\endgroup\def~{\protect\@uscore.}}%
  }
\fi
%    \end{macrocode}
%
% Finally, we redefine the |\_| macro to use our own |\underscore|, because
% it's prettier.  Actually, we don't: we just redefine the
% |\?\textunderscore| command (funny name, isn't it?).
%
%    \begin{macrocode}
\expandafter\let\csname?\string\textunderscore\endcsname\underscore
%    \end{macrocode}
%
% \subsection{Abbreviated verbatim notation}
%
%  In similar style to the \package{doc} package, we allow the user to set up
% characters which delimit verbatim text.  Unlike \package{doc}, we make
% such changes local to the current group.  This is performed through the
% |\shortverb| and |\unverb| commands.
%
% The implementations of these commands are based upon the |\MakeShortVerb|
% and |\DeleteShortVerb| commands of the \package{doc} package, although
% these versions have effect local to the current grouping level.  This
% prevents their redefinition of |\dospecials| from interfering with the
% grammar shortcuts, which require local changes only.
%
% The command |\shortverb| takes a single argument: a single-character
% control sequence defining which character to make into the verbatim text
% delimiter.  We store the old meaning of the active character in a control
% sequence called |\mn@\|\<char>.  Note that this control sequence
% contains a backslash character, which is a little odd.  We also define a
% command |\cc@\|\<char> which will return everything to normal.  This
% is used by the |\unverb| command.
%
% \begin{macro}{\shortverb}
%
% Here we build the control sequences we need to make everything work nicely.
% The active character is defined via |\lowercase|, using the |~| character:
% this is already made active by \TeX\@.
%
% The actual code requires lots of fiddling with |\expandafter| and friends.
%
%    \begin{macrocode}
\def\shortverb#1{%
%    \end{macrocode}
%
% First, we check to see if the command |\cc@\|\<char> has been defined.
%
%    \begin{macrocode}
  \@ifundefined{cc@\string#1}{%
%    \end{macrocode}
%
% If it hasn't been defined, we add the character to the specials list.
%
%    \begin{macrocode}
    \addspecial#1%
%    \end{macrocode}
%
% Now we set our character to be the lowercase version of |~|, which allows
% us to use it, even though we don't know what it is.
%
%    \begin{macrocode}
    \begingroup%
    \lccode`\~`#1%
%    \end{macrocode}
%
% Finally, we reach the tricky bit.  All of this is lowercased, so any
% occurrences of |~| are replaced by the user's special character.
%
%    \begin{macrocode}
    \lowercase{%
      \endgroup%
%    \end{macrocode}
%
% We remember the current meaning of the character, in case it has one.  We
% have to use |\csname| to build the rather strange name we use for this.
%
%    \begin{macrocode}
      \expandafter\let\csname mn@\string#1\endcsname~%
%    \end{macrocode}
%
% Now we build |\cc@\|\<char>.  This is done with |\edef|, since more
% of this needs to be expanded now than not.  In this way, the actual macros
% we create end up being very short.
%
%    \begin{macrocode}
      \expandafter\edef\csname cc@\string#1\endcsname{%
%    \end{macrocode}
%
% First, add a command to restore the character's old catcode.
%
%    \begin{macrocode}
        \catcode`\noexpand#1\the\catcode`#1%
%    \end{macrocode}
%
% Now we restore the character's old meaning, using the version we saved
% earlier.
%
%    \begin{macrocode}
        \let\noexpand~\expandafter\noexpand%
          \csname mn@\string#1\endcsname%
%    \end{macrocode}
%
% Now we remove the character from the specials lists.
%
%    \begin{macrocode}
        \noexpand\remspecial\noexpand#1%
%    \end{macrocode}
%
% Finally, we delete this macro, so that |\unverb| will generate a warning
% if the character is |\unverb|ed again.
%
%    \begin{macrocode}
        \let\csname cc@\string#1\endcsname\relax%
      }%
%    \end{macrocode}
%
% All of that's over now.  We set up the new definition of the character,
% in terms of |\verb|, and make the character active.  The nasty |\syn@ttspace|
% is there to make the spacing come out right.  It's all right really.  Honest.
%
%    \begin{macrocode}
      \def~{\verb~\syn@ttspace}%
    }%
    \catcode`#1\active%
%    \end{macrocode}
%
% If our magic control sequence already existed, we can assume that the
% character is already a verbatim delimiter, and raise a warning.
%
%    \begin{macrocode}
  }{%
    \PackageWarning{syntax}{Character `\expandafter\@gobble\string#1'
                            is already a verbatim\MessageBreak
                            delimiter}%
  }%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\unverb}
%
% This is actually terribly easy: we just use the |\cc@\|\<char> command
% we definied earlier, after making sure that it's been defined.
%
%    \begin{macrocode}
\def\unverb#1{%
  \@ifundefined{cc@\string#1}{%
    \PackageWarning{syntax}{Character `\expandafter\@gobble\string#1'
                            is not a verbatim\MessageBreak
                            delimiter}%
  }{%
    \csname cc@\string#1\endcsname%
  }%
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsection{Style hooks for syntax forms}
%
% To allow the appearance of syntax things to be configured, we provide some
% redefinable bits.
%
% The three types of objects (nonterminal symbols, and quoted and unquoted
% terminals) each have two macros associated with them: one which does the
% `left' bit of the typesetting, and one which does the `right' bit.  The
% items are typeset as LR~boxes.  I'll be extra good while defining these
% hooks, so that it's obvious what's going on; macho \TeX\ hacker things
% resume after this section.
%
% \begin{macro}{\syntleft}
% \begin{macro}{\syntright}
%
% I can't see why anyone would want to change the typesetting of
% nonterminals, although I'll provide the hooks for symmetry's sake.
%
%     \begin{macrocode}
\newcommand{\syntleft}{$\langle$\normalfont\itshape}
\newcommand{\syntright}{$\rangle$}
%    \end{macrocode}
%
% \end{macro}
% \end{macro}
%
% \begin{macro}{\ulitleft}
% \begin{macro}{\ulitright}
% \begin{macro}{\litleft}
% \begin{macro}{\litright}
%
% Now we can define the left and right parts of quoted and unquoted
% terminals.  US~readers may want to put double quotes around the quoted
% terminals, for example.
%
%    \begin{macrocode}
\newcommand{\ulitleft}{\normalfont\ttfamily\syn@ttspace\frenchspacing}
\newcommand{\ulitright}{}
\newcommand{\litleft}{`\bgroup\ulitleft}
\newcommand{\litright}{\ulitright\egroup'}
%    \end{macrocode}
%
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \subsection{Simple syntax typesetting}
%
% In general text, we allow access to our typesetting conventions through
% standard \LaTeX\ commands.
%
% \begin{macro}{\synt}
%
% The |\synt| macro typesets its argument as a syntactic quantity.  It puts
% the text of the argument in italics, and sets angle brackets around it.
% Breaking of a |\synt| object across lines is forbidden.
%
%    \begin{macrocode}
\def\synt#1{\mbox{\syntleft{#1\/}\syntright}}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\lit}
%
% The |\lit| macro typesets its argument as literal text, to be typed in.
% Normally, this means setting the text in |\tt| font, and putting quotes
% around it, although the quotes can be suppressed by using the $*$-variant.
%
% The |\syn@ttspace| macro sets up the spacing for the text nicely: |\tt|
% spaces tend to be a little wide.
%
%    \begin{macrocode}
\def\lit{\@ifstar{\lit@i\ulitleft\ulitright}{\lit@i\litleft\litright}}
\def\lit@i#1#2#3{\mbox{#1{#3\/}#2}}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\syn@ttspace}
%
% This sets up the |\spaceskip| value for |\tt| text.
%
%    \begin{macrocode}
\def\syn@ttspace@{\spaceskip.35em\@plus.2em\@minus.15em\relax}
%    \end{macrocode}
%
% However, this isn't always the right thing to do.
%
%    \begin{macrocode}
\def\ttthinspace{\let\syn@ttspace\syn@ttspace@}
\def\ttthickspace{\let\syn@ttspace\@empty}
%    \end{macrocode}
%
% I know what I like thoough.
%
%    \begin{macrocode}
\ttthinspace
%    \end{macrocode}
%
% \end{macro}
%
% \subsubsection{The shortcuts}
%
% The easy part is over now.  The next job is to set up the `grammar
% shortcuts' which allow easy changing of styles.
%
% We support four shortcuts:
% \begin{itemize}
% \item |`literal text'| typesets \syntax{`literal text'}
% \item |<non-terminal>| typesets \syntax{<non-terminal>}
% \item |"unquoted text"| typesets \syntax{"unquoted text"}
% \item \verb"|" typesets a \syntax{|} character
% \end{itemize}
% These are all implemented through active characters, which are enabled
% using the |\syntaxShortcuts| macro, described below.
%
% \begin{macro}{\readupto}
%
% \syntax{"\\readupto{"<char>"}{"<decls>"}{"<command>"}"} will read all
% characters up until the next occurrence of \<char>.  Normally, all
% special characters will be deactivated.  However, you can reactivate some
% characters, using the \<decls> argument, which is processed before the
% text is read.
%
% The code is borrowed fairly obviously from the \LaTeXe\ source for the
% |\verb| command.
%
%    \begin{macrocode}
\def\readupto#1#2#3{%
  \bgroup%
  \verb@eol@error%
  \let\do\@makeother\dospecials%
  #2%
  \catcode`#1\active%
  \lccode`\~`#1%
  \gdef\verb@balance@group{\verb@egroup%
     \@latex@error{\noexpand\verb illegal in command argument}\@ehc}%
  \def\@vhook{\verb@egroup#3}%
  \aftergroup\verb@balance@group%
  \lowercase{\let~\@vhook}%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\syn@assist}
%
% The |\syn@assist| macro is used for defining three of the shortcuts.  It
% is called as
%
% \begin{quote}
% \syntax{"\\syn@assist{"<left-decls>"}{"<actives>"}{"<delimeter>"}" \\
% \null \quad "{"<right-decls>"}{"<end-cmd>"}"}
% \end{quote}
%
% It creates an hbox, sets up the escape sequences for quoting our magic
% characters, and then typesets a box containing
%
% \begin{quote}
% \syntax{<left-decls>"{"<delimited-text>"\\/}"<right-decls>}
% \end{quote}
%
% The \<left-decls> and \<right-decls> can be |\relax| if they're not
% required.
%
% The \<actives> argument is passed to |\readupto|, to allow some special
% characters through.  By default, we re-enable |\|, and make `\verb*" "'
% typeset some space glue, rather than a space character.  A macro
% `\verb*"\ "' is defined to actually print a space character, which yield
% `\verb*" "' in the `|\tt|' font. 
%
% Finally, it defines a |\ch| command, which, given a single-character
% control sequence as its argument, typesets the character.  This is useful,
% since |`| has been made active when we set up these calls, so the
% direct |\char`\|\<char> doesn't work.
%
%    \begin{macrocode}
\def\syn@assist#1#2#3#4#5{%
%    \end{macrocode}
%
% First, we start the box, and open a group.  We use |\mbox| because it
% does all the messing with |\leavevmode| which is needed.
%
%    \begin{macrocode}
  \mbox\bgroup%
%    \end{macrocode}
%
% Next job is to set up the escape sequences.
%
%    \begin{macrocode}
  \chardef\\`\\%
  \chardef\>`\>%
  \chardef\'`\'%
  \chardef\"`\"%
  \chardef\ `\ %
%    \end{macrocode}
%
% Now to define |\ch|.  This is done the obvious way.
%
%    \begin{macrocode}
  \def\ch##1{\char`##1}%
%    \end{macrocode}
%
% For active characters, we do some fiddling with |\lccode|s.
%
%    \begin{macrocode}
  \def\act##1{%
    \catcode`##1\active%
    \begingroup%
    \lccode`\~`##1%
    \lowercase{\endgroup\def~}%
  }%
%    \end{macrocode}
%
% Finally, we do the real work of setting the text.  We use |\readupto| to
% actually find the text we want.
%
%    \begin{macrocode}
  #1%
  \begingroup%
  \readupto#3{%
    \catcode`\\0%
    \catcode`\ 10%
    #2%
  }{%
    \/\endgroup#4\egroup#5%
  }%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\syn@shorts}
%
% This macro actually defines the expansions for the active characters.
% We have to do this separately because |`| must be active when we use it
% in the |\def|, but we can't do that and use |\catcode| at the same time.
% The arguments are commands to do before and after the actual command.
% These are passed up from |\syntaxShortcuts|.
%
% All of the characters use |\syn@assist| in the obvious way except for
% \verb"|", which drops into maths mode instead.
%
% Note that when changing the catcodes, we must save |`| until last.
%
%    \begin{macrocode}
\begingroup
\catcode`\<\active
\catcode`\|\active
\catcode`\"\active
\catcode`\`\active
%
\gdef\syn@shorts#1#2{%
%    \end{macrocode}
%
% The `|<|' character must typeset its argument in italics.  We make `|_|'
% do the same as the `|\_|' command.
%
%    \begin{macrocode}
  \def<{%
    #1%
    \syn@assist%
      \syntleft%
      {\act_{\@uscore.}}%
      >%
      \syntright%
      {#2}%
  }%
%    \end{macrocode}
%
% The `|`|' and `|"|' characters should print its argument in |\tt| font.
% We change the `|\tt|' space glue to provide nicer spacing on the line.
%
%    \begin{macrocode}
  \def`{%
    #1%
    \syn@assist%
      \litleft%
      \relax%
      '%
      \litright%
      {#2}%
  }%
  \def"{%
    #1%
    \syn@assist%
      \ulitleft%
      \relax%
      "%
      \ulitright%
      {#2}%
  }%
%    \end{macrocode}
%
% Finally, the `\verb"|"' character is typeset by using the mysterious
% |\textbar| command.
%
%    \begin{macrocode}
  \def|{\textbar}%
%    \end{macrocode}
%
% We're finished here now.
%
%    \begin{macrocode}
}
%
\endgroup
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\syntaxShortcuts}
%
% This is a user-level command which enables the use of our shortcuts in the
% current group.  It uses |\addspecial|, defined below, to register the
% active characters, sets up their definitions and activates them.
%
% The two arguments are commands to be performed before and after the
% handling of the abbreviation.  In this way, you can further process the
% output.
%
% This command is not intended to be used directly by users: it should be
% used by other macros and packages which wish to take advantage of the
% facilities offered by this package.  We provide a |\synshorts| declaration
% (which may be used as an environment, of course) which is more `user
% palatable'.
%
%    \begin{macrocode}
\def\syntaxShortcuts#1#2{%
  \syn@shorts{#1}{#2}%
  \addspecial\`%
  \addspecial\<%
  \addspecial\|%
  \addspecial\"%
  \catcode`\|\active%
  \catcode`\<\active% 
  \catcode`\"\active%
  \catcode`\`\active%
}
%
\def\synshorts{\syntaxShortcuts\relax\relax}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\synshortsoff}
%
% This macro can be useful occasionally: it disables the syntax shortcuts,
% so you can type normal text for a while.
%
%    \begin{macrocode}
\def\synshortsoff{%
  \catcode`\|12%
  \catcode`\<12%
  \catcode`\"12%
  \catcode`\`12%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\syntax}
%
% The |\syntax| macro typesets its argument, allowing the use of our
% shortcuts within the argument.
%
% Actually, we go to some trouble to ensure that the argument to |\syntax|
% \emph{isn't} a real argument so we can change catcodes as we go.  We
% use the |\let\@let@token=| trick from \PlainTeX\ to do this.
%
%    \begin{macrocode}
\def\syntax#{\bgroup\syntaxShortcuts\relax\relax\let\@let@token}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{environment}{grammar}
%
% The \env{grammar} environment is the final object we have to define.  It
% allows typesetting of beautiful BNF grammars.
%
% First, we define the length parameters we need:
%
%    \begin{macrocode}
\newskip\grammarparsep
  \grammarparsep8\p@\@plus\p@\@minus\p@
\newdimen\grammarindent
  \grammarindent2em
%    \end{macrocode}
%
% Now define the default label typesetting.  This macro is designed to be
% replaced by a user, so we'll be extra-well-behaved and use genuine \LaTeX\
% commands.  Well, almost \dots
%
%    \begin{macrocode}
\newcommand{\grammarlabel}[2]{%
  \synt{#1} \hfill#2%
}
%    \end{macrocode}
%
% Now for a bit of hacking to make the item stuff work properly.  This gets
% done for every new paragraph that's started without an |\item| command.
%
% First, store the left hand side of the production in a box.  Then I'll
% end the paragraph, and insert some nasty glue to take up all the space,
% so no-one will ever notice that there was a paragraph break there.  The
% strut just makes sure that I know exactly how high the line is.
%
%    \begin{macrocode}
\def\gr@implitem<#1> #2 {%
  \sbox\z@{\hskip\labelsep\grammarlabel{#1}{#2}}%
  \strut\@@par%
  \vskip-\parskip%
  \vskip-\baselineskip%
%    \end{macrocode}
%
% The |\item| command will notice that I've inserted these funny glues and
% try to remove them: I'll stymie its efforts by inserting an invisible
% rule.  Then I'll insert the label using |\item| in the normal way.
%
%    \begin{macrocode}
  \hrule\@height\z@\@depth\z@\relax%
  \item[\unhbox\z@]%
%    \end{macrocode}
%
% Just before I go, I'll make \lit{<} back into an active character.
%
%    \begin{macrocode}
  \catcode`\<\active%
}
%    \end{macrocode}
%
% Now for the environment proper.  Deep down, it's a list environment, with
% some nasty tricks to stop anyone from noticing.
%
% The first job is to set up the list from the parameters I'm given.
%
%    \begin{macrocode}
\newenvironment{grammar}{%
  \list{}{%
    \labelwidth\grammarindent%
    \leftmargin\grammarindent%
    \advance\grammarindent\labelsep
    \itemindent\z@%
    \listparindent\z@%
    \parsep\grammarparsep%
  }%
%    \end{macrocode}
%
% We have major problems in |\raggedright| layouts, which try to use |\par|
% to start new lines.  We go back to normal |\\| newlines to try and bodge
% our way around these problems.
%
%    \begin{macrocode}
  \let\\\@normalcr
%    \end{macrocode}
%
% Now to enable the shortcuts.
%
%    \begin{macrocode}
  \syntaxShortcuts\relax\relax%
%    \end{macrocode}
%
% Now a little bit of magic.  The |\alt| macro moves us to a new line, and
% typesets a vertical bar in the margin.  This allows typesetting of
% multiline alternative productions in a pretty way.
%
%    \begin{macrocode}
  \def\alt{\\\llap{\textbar\quad}}%
%    \end{macrocode}
%
% Now for another bit of magic.  We set up some |\par| cleverness to spot
% the start of each production rule and format it in some cunning and
% user-defined way.
%
%    \begin{macrocode}
  \def\gr@setpar{%
    \def\par{%
      \parshape\@ne\@totalleftmargin\linewidth%
      \@@par%
      \catcode`\<12%
      \everypar{%
        \everypar{}%
        \catcode`\<\active%
        \gr@implitem%
      }%
    }%
  }%
  \gr@setpar%
  \par%
%    \end{macrocode}
%
% Now set up the |\[[| and |\]]| commands to do the right thing.  We have
% to check the next character to see if it's correct, otherwise we'll
% open a maths display as usual.
%
%    \begin{macrocode}
  \let\gr@leftsq\[%
  \let\gr@rightsq\]%
  \def\gr@endsyntdiag]{\end{syntdiag}\gr@setpar\par}%
  \def\[{\@ifnextchar[{\begin{syntdiag}\@gobble}\gr@leftsq}%
  \def\]{\@ifnextchar]\gr@endsyntdiag\gr@rightsq}%
%    \end{macrocode}
%
% Well, that's it for this side of the environment.
%
%    \begin{macrocode}
}{%
%    \end{macrocode}
%
% Closing the environment is a simple matter of tidying away the list.
%
%    \begin{macrocode}
  \@newlistfalse%
  \everypar{}%
  \endlist%
}
%    \end{macrocode}
%
% \end{environment}
%
% \subsection{Syntax diagrams}
%
% Now we come to the final and most complicated part of the package.
%
% Syntax diagrams are drawn using arrow characters from \LaTeX's line font,
% used in the \env{picture} environment, and rules.  The horizontal rules
% of the diagram are drawn along the baselines of the lines in which they
% are placed.  The text items in the diagram are placed in boxes and lowered
% below the main baseline.  Struts are added throughout to keep the vertical
% spacing consistent.
%
% The vertical structures (stacks and loops) are all implemented with \TeX's
% primitive |\halign| command.
%
% \subsubsection{User-configurable parameters}
%
% First, we allocate the \<dimen> and \<skip> arguments needed.  Fixed
% lengths, as the \LaTeX book calls them, are allocated as \<dimen>s, to
% take some of the load off of all the \<skip> registers.
%
%    \begin{macrocode}
\newskip\sdstartspace
\newskip\sdendspace
\newskip\sdmidskip
\newskip\sdtokskip
\newskip\sdfinalskip
\newdimen\sdrulewidth
\newdimen\sdcirclediam
\newdimen\sdindent
%    \end{macrocode}
%
% We need some \TeX\ \<dimen>s for our own purposes, to get everything in
% the right places.  We use labels for the `temporary' \TeX\ parameters
% which we use, to avoid wasting registers.
%
%    \begin{macrocode}
\dimendef\sd@lower\z@
\dimendef\sd@upper\tw@
\dimendef\sd@mid4
\dimendef\sd@topcirc6
\dimendef\sd@botcirc8
%    \end{macrocode}
%
% \begin{macro}{\sd@setsize}
% When the text size for syntax diagrams changes, it's necessary to work out
% the height for various rules in the diagram.
%
%    \begin{macrocode}
\def\sd@setsize{%
  \sd@mid\ht\strutbox%
  \advance\sd@mid-\dp\strutbox%
  \sd@mid.5\sd@mid%
  \sd@upper\sdrulewidth%
    \advance\sd@upper\sd@mid%
  \sd@lower\sdrulewidth%
    \advance\sd@lower-\sd@mid%
  \sd@topcirc-.5\sdcirclediam%
    \advance\sd@topcirc\sd@mid%
  \sd@botcirc-.5\sdcirclediam%
    \advance\sd@botcirc-\sd@mid%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sdsize}
%
% You can set the default type size used by syntax diagrams by redefining
% the |\sdsize| command, using the |\renewcommand| command.
%
% By default, syntax diagrams are set slightly smaller than the main body
% text.\footnote{^^A
%   I've used pure \LaTeX\ commands for this and the \cmd\sdlengths\ macro,
%   to try and illustrate how these values might be changed by a user.  The
%   rest of the code is almost obfuscted in its use of raw \TeX\ features,
%   in an attempt to dissuade more na\"\i ve users from fiddling with it.
%   I suppose this is what you get when you let assembler hackers loose with
%   something like \LaTeX.
% }
%
%    \begin{macrocode}
\newcommand{\sdsize}{%
  \small%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sdlengths}
%
% Finally, the default length parameters are set in the |\sdlengths| command.
% You can redefine the command using |\renewcommand|.
%
% We set up the length parameters here.
%
%    \begin{macrocode}
\newcommand{\sdlengths}{%
  \setlength{\sdstartspace}{1em minus 10pt}%
  \setlength{\sdendspace}{1em minus 10pt}%
  \setlength{\sdmidskip}{0.5em plus 0.0001fil}%
  \setlength{\sdtokskip}{0.25em plus 0.0001fil}%
  \setlength{\sdfinalskip}{0.5em plus 10000fil}%
  \setlength{\sdrulewidth}{0.2pt}%
  \setlength{\sdcirclediam}{8pt}%
  \setlength{\sdindent}{0pt}%
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsubsection{Other declarations}
%
% We define four switches.  The table shows what they're used for.
%
% \begin{table}
% \begin{tab}{lp{3in}}                                              \hline
%
% \bf Switch        & \bf Meaning                                \\ \hline
%
% |\ifsd@base|      & We are at `base level' in the diagram:
%                     i.e., not in any other sorts of
%                     constructions.  This is used to decide
%                     whether to allow line breaking.            \\[2pt]
%
% |\ifsd@top|       & The current loop construct is being
%                     typeset with the loop arrow above the
%                     baseline.                                  \\[2pt]
%
% |\ifsd@toplayer|  & We are typesetting the top layer of
%                     a stack.  This is used to ensure that
%                     the vertical rules on either side are
%                     typeset at the right height.               \\[2pt]
%
% |\ifsd@backwards| & We're typesetting backwards, because
%                     we're in the middle of a loop arrow.
%                     the only difference this makes is that
%                     any subloops have the arrow on the
%                     side.                                      \\ \hline
%
% \end{tab}
% \caption{Syntax diagram switches}
% \end{table}
%
%    \begin{macrocode}
\newif\ifsd@base
\newif\ifsd@top
\newif\ifsd@toplayer
\newif\ifsd@backwards
%    \end{macrocode}
%
% \begin{macro}{\sd@err}
%
% We output our errors through this macro, which saves a little typing.
%
%    \begin{macrocode}
\def\sd@err{\PackageError{syntax}}
%    \end{macrocode}
%
% \end{macro}
%
% \subsubsection{Arrow-drawing}
%
% We need to draw some arrows.  \LaTeX\ tries to make this as awkward as
% possible, so we have to start moving the arrows around in boxes quite a
% lot.
%
% The left and right pointing arrows are fairly simple: we just add some
% horizontal spacing to prevent the width of the arrow looking odd.
%
%    \begin{macrocode}
\def\sd@arrow{%
  \ht\tw@\z@%
  \dp\tw@\z@%
  \raise\sd@mid\box\tw@%
  \egroup%
}
\def\sd@rightarr{%
  \bgroup%
  \setbox\tw@\hbox{\kern-6\p@\@linefnt\char'55}%
  \sd@arrow%
}
\def\sd@leftarr{%
  \bgroup%
  \raise\sd@mid\hbox{\@linefnt\char'33\kern-6\p@}%
  \sd@arrow%
}
%    \end{macrocode}
%
% The up arrow is very strange.  We need to bring the arrow down to base
% level, and smash its height.
%
%    \begin{macrocode}
\def\sd@uparr{%
  \bgroup%
  \setbox\tw@\hb@xt@\z@{\kern-\sdrulewidth\@linefnt\char'66\hss}%
  \setbox\tw@\hbox{\lower10\p@\box\tw@}%
  \sd@arrow%
}
%    \end{macrocode}
%
% The down arrow is similar, although it's already at the right height.
% Thus, we can just smash the box.
%
%    \begin{macrocode}
\def\sd@downarr{%
  \bgroup%
  \setbox\tw@\hb@xt@\z@{\kern-\sdrulewidth\@linefnt\char'77\hss}%
  \sd@arrow%
}
%    \end{macrocode}
%
% \subsubsection{Drawing curves}
%
% If the user has selected curved edges, we use the \LaTeX\ features provided
% to obtain the curves.  These are drawn slightly oddly to make it easier
% to fit them into the diagram.
%
% Some explanation about the \LaTeX\ circle font is probably called for
% before we go any further.  The font consists of sets of four quadrants
% of a particular size (and some other characters, which aren't important
% at the moment).  Each collection of quadrants fit together to form a
% perfect circle of a given diameter.  The individual quadrant characters
% have strange bounding boxes, as described in the files \textit{lcircle.mf}
% and \textit{ltpict.dtx}, and also in Appendix~D of \textit{The \TeX book}.
% Our job here is to make these quadrants useful in the context of
% drawing syntax diagrams.
%
% \begin{macro}{\sd@circ}
% First, we define |\sd@circ|, which performs the common parts of the four
% routines.  Since the characters in the circle font are grouped together,
% we can pick out a particular corner piece by specifying its index into
% the group for the required size.  The |\sd@circ| routine will pick out
% the required character, given this index as an argument, and put it in
% box~2, after fiddling with the sizes a little:
% \begin{itemize}
%
% \item We clear the width to zero.  The individual routines then add a kern
%       of the correct amount, so that the quadrant appears in the right
%       place.
%
% \item The piece is lowered by half the rule width.  This positions the
%       top and bottom pieces of the circle to be half way over the baseline,
%       which is the correct position for the rest of the diagram.
%
% \end{itemize}
%
% Finally, we make sure we're in horizontal mode: horrific results occur
% if this is not the case.  I'm sure I don't need to explain this any more
% graphically.
%
%    \begin{macrocode}
\def\sd@circ#1{%
  \@getcirc\sdcirclediam%
  \advance\@tempcnta#1%
  \setbox\tw@\hbox{\lower\sdrulewidth%
    \hbox{\@circlefnt\char\@tempcnta}}%
  \wd\tw@\z@%
  \leavevmode%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sd@tlcirc}
% \begin{macro}{\sd@trcirc}
% \begin{macro}{\sd@blcirc}
% \begin{macro}{\sd@brcirc}
%
% These are the macros which actually draw quadrants of circles.  They all
% call |\sd@circ|, passing an appropriate index, and then fiddle with the
% box sizes and apply kerning specific to the quadrant positioning.
%
% The exact requirements for positioning are as follows:
%
% \begin{itemize}
%
% \item The horizontal parts of the arcs must lie along the baseline (i.e.,
%       half the line must be above the baseline, and half must be below).
%       This is consistent with the horizontal rules used in the diagram.
%
% \item The vertical parts must overlap vertical rules on either side, so
%       that a |\vrule\sd@|\textit{xx}|circ| makes the arc appear to be
%       a real curve in the line.  The requirements are actually somewhat
%       inconsistent; for example, the \env{stack} environment uses curves
%       \emph{before} the |\vrule|s.  Special requirements like this are
%       handled as special cases later.
%
% \item The height and width of the arc are at least roughly correct.
%
% \end{itemize}
%
%    \begin{macrocode}
\def\sd@tlcirc{{%
  \sd@circ3%
  \ht\tw@\sdrulewidth%
  \dp\tw@.5\sdcirclediam%
  \kern-\tw@\sdrulewidth%
  \raise\sd@mid\box\tw@%
  \kern.5\sdcirclediam%
}}
%    \end{macrocode}
%
%    \begin{macrocode}
\def\sd@trcirc{{%
  \sd@circ0%
  \ht\tw@\sdrulewidth%
  \dp\tw@.5\sdcirclediam%
  \kern.5\sdcirclediam%
  \raise\sd@mid\box\tw@%
}}
%    \end{macrocode}
%
%    \begin{macrocode}
\def\sd@blcirc{{%
  \sd@circ2%
  \ht\tw@.5\sdcirclediam%
  \dp\tw@\sdrulewidth%
  \kern-\tw@\sdrulewidth%
  \raise\sd@mid\box\tw@%
  \kern.5\sdcirclediam%
}}
%    \end{macrocode}
%
%    \begin{macrocode}
\def\sd@brcirc{{%
  \sd@circ1%
  \ht\tw@.5\sdcirclediam%
  \dp\tw@\sdrulewidth%
  \kern.5\sdcirclediam%
  \raise\sd@mid\box\tw@%
}}
%    \end{macrocode}
%
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\sd@llc}
% \begin{macro}{\sd@rlc}
%
% In the \env{rep} environment, we need to be able to draw arcs with
% horizontal lines running through them.  The two macros here do the job
% nicely.  |\sd@llc| (which is short for left overlapping circle) is
% analogous to |\llap|: it puts its argument in a box of zero width, sticking
% out to the left.  However, it also draws a rule along the baseline.  This
% is important, as it prevents text from overprinting the arc.  |\sd@rlc|
% is very similar, just the other way around.
%
%    \begin{macrocode}
\def\sd@llc#1{%
  \hb@xt@.5\sdcirclediam{%
    \sd@rule\hskip.5\sdcirclediam%
    \hss%
    #1%
  }%
}
%    \end{macrocode}
%
%    \begin{macrocode}
\def\sd@rlc#1{%
  \hb@xt@.5\sdcirclediam{%
    #1%
    \hss%
    \sd@rule\hskip.5\sdcirclediam%
  }%
}
%    \end{macrocode}
%
% \end{macro}
% \end{macro}
%
% \subsubsection{Drawing rules}
%
% It's important to draw the rules \emph{along} the baseline, rather than
% above it: hence, the depth of the rule must be equal to the height.
%
% \begin{macro}{\sd@rule}
%
% We use rule leaders instead of glue through most of the syntax diagrams.
% The command \syntax{"\\sd@rule"<skip>} draws a rule of the correct
% dimensions, which has the behaviour of an \syntax{"\\hskip"<skip>}.
%
%    \begin{macrocode}
\def\sd@rule{\leaders\hrule\@height\sd@upper\@depth\sd@lower}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sd@gap}
%
% The gap between elements is added using this macro.  It will allow a
% line break if we're at the top level of the diagram, using a rather
% strange discretionary.
%
% This is called as \syntax{"\\sd@gap{"<skip-register>"}"}.
%
%    \begin{macrocode}
\def\sd@gap#1{%
%    \end{macrocode}
%
% First, we see if we're at the top level.  Within constructs, we avoid the
% overhead of a |\discretionary|.  We put half of the width of the skip on
% each side of the discretionary break.
%
%    \begin{macrocode}
  \ifsd@base%
    \skip@#1%
      \divide\skip\z@\tw@%
    \nobreak\sd@rule\hskip\skip@%
    \discretionary{%
      \sd@qarrow{->}%
    }{%
      \hbox{%
        \sd@qarrow{>-}%
        \sd@rule\hskip\sdstartspace%
        \sd@rule\hskip3.5\p@%
      }%
    }{%
    }%
    \nobreak\sd@rule\hskip\skip@%
%    \end{macrocode}
%
% If we're not at the base level, we just put in a rule of the correct
% width.
%
%    \begin{macrocode}
  \else%
    \sd@rule\hskip#1%
  \fi%
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsubsection{The \protect\env{syntdiag} environment}
%
% All syntax diagrams are contained within a \env{syntdiag} environment.
%
% \begin{environment}{syntdiag}
%
% The only argument is a collection of declarations, which by
% default is
%
% \begin{listing}
%\sdsize\sdlengths
% \end{listing}
%
% However, if the optional argument is not specified, \TeX\ reads the first
% character of the environment, which may not be catcoded correctly.  We set
% up the catcodes first, using the |\syntaxShortcuts| command, and then read
% the argument.  We don't use |\newcommand|, because that would involve
% creating yet \emph{another} macro.  Time to fiddle with |\@ifnextchar|
% \dots
%
%    \begin{macrocode}
\def\syntdiag{%
  \syntaxShortcuts\sd@tok@i\sd@tok@ii%
  \@ifnextchar[\syntdiag@i{\syntdiag@i[]}%
}
%    \end{macrocode}
%
% Now we actually do the job we're meant to.
%
%    \begin{macrocode}
\def\syntdiag@i[#1]{%
%    \end{macrocode}
%
% The first thing to do is execute the user's declarations.  We then set
% up things for the font size.
%
%    \begin{macrocode}
  \sdsize\sdlengths%
  #1%
  \sd@setsize%
%    \end{macrocode}
%
% Next, we start a list, to change the text layout.
%
%    \begin{macrocode}
  \list{}{%
    \leftmargin\sdindent%
    \rightmargin\leftmargin%
    \labelsep\z@%
    \labelwidth\z@%
  }%
  \item[]%
%    \end{macrocode}
%
% We reconfigure the paragraph format quite a lot now.  We clear
% |\parfillskip| to avoid any justification at the end of the paragraph.
% We also turn off paragraph indentation.
%
%    \begin{macrocode}
  \parfillskip\z@%
  \noindent%
%    \end{macrocode}
%
% Next, we add in the arrows on the beginning of the line, and a bit of
% glue.
%
%    \begin{macrocode}
  \sd@qarrow{>>-}%
  \nobreak\sd@rule\hskip\sdstartspace%
%    \end{macrocode}
%
% This is the base level of the diagram, so we enable line breaking.
%
%    \begin{macrocode}
  \sd@basetrue%
%    \end{macrocode}
%
% Since the objects being broken are rather large, we enable sloppy line
% breaking.  We also try to avoid page breaks in mid-diagram, by upping the
% |\interlinepenalty|.
%
%    \begin{macrocode}
  \sloppy%
  \interlinepenalty100%
  \hyphenpenalty0%
%    \end{macrocode}
%
% We handle all the spacing within the environment, so we make \TeX\ ignore
% spaces and newlines.
%
%    \begin{macrocode}
  \catcode`\ 9%
  \catcode`\^^M9%
%    \end{macrocode}
%
% We now have to change the behaviour of |\\| to line-break syntax diagrams.
%
%    \begin{macrocode}
  \let\\\sd@newline%
  \ignorespaces%
}
%    \end{macrocode}
%
% When we end the diagram, we just have to add in the final fillskip, and
% double arrow.
%
%    \begin{macrocode}
\def\endsyntdiag{%
  \unskip%
  \nobreak\sd@rule\hskip\sdmidskip%
  \sd@rule\hskip\sdfinalskip%
  \sd@qarrow{-><}%
  \endlist%
}
%    \end{macrocode}
%
% \end{environment}
%
% \begin{environment}{syntdiag*}
%
% The starred form of \env{syntdiag} typesets a syntax diagram in LR-mode;
% this is useful if you're describing parts of syntax diagrams, for example.
%
% This is in fact really easy.  The first bit which checks for an optional
% argument is almost identical to the non-$*$ version.
%
%    \begin{macrocode}
\@namedef{syntdiag*}{%
  \syntaxShortcuts\sd@tok@i\sd@tok@ii%
  \@ifnextchar[\syntdiag@s@i{\syntdiag@s@i[]}%
}
%    \end{macrocode}
%
% Handle another optional argument giving the width of the box to fill.
%
%    \begin{macrocode}
\def\syntdiag@s@i[#1]{%
  \@ifnextchar[{\syntdiag@s@ii{#1}}{\syntdiag@s@iii{#1}{\hbox}}%
}
\def\syntdiag@s@ii#1[#2]{\syntdiag@s@iii{#1}{\hb@xt@#2}}
%    \end{macrocode}
%
% Now to actually start the display.  This is mostly simple.  Just to make
% sure about the LR-ness of the typesetting, I'll put everything in an hbox.
%
%    \begin{macrocode}
\def\syntdiag@s@iii#1#2{%
  \leavevmode%
  #2\bgroup%
%    \end{macrocode}
%
% Now configure the typesetting according to the user's wishes.
%
%    \begin{macrocode}
  \let\@@left\left%
  \let\@@right\right%
  \def\left##1{\def\sd@startarr{##1}}%
  \def\right##1{\def\sd@endarr{##1}}%
  \left{>-}\right{->}%
  \sdsize\sdlengths%
  #1%
  \sd@setsize%
  \let\left\@@left%
  \let\right\@@right%
%    \end{macrocode}
%
% Put in the initial double-arrow.
%
%    \begin{macrocode}
  \sd@qarrow\sd@startarr%
  \sd@rule\hskip\sdmidskip%
%    \end{macrocode}
%
% We're in horizontal mode, so don't bother with linebreaking.
%
%    \begin{macrocode}
  \sd@basefalse%
%    \end{macrocode}
%
% Finally, disable spaces and things.
%
%    \begin{macrocode}
  \catcode`\ 9%
  \catcode`\^^M9%
  \ignorespaces%
}
%    \end{macrocode}
%
% Ending the environment is very similar.
%
%    \begin{macrocode}
\@namedef{endsyntdiag*}{%
  \unskip%
  \sd@rule\hskip\sdmidskip%
  \sd@rule\hskip\sdfinalskip%
  \sd@qarrow\sd@endarr%
  \egroup%
}
%    \end{macrocode}
%
% \end{environment}
%
% \begin{macro}{\sd@qarrow}
%
% This typesets the various left and right arrows required in syntax
% diagrams.  The argument is one of \syntax{`>>-', `->', `>-' or `-><'}.
%
%    \begin{macrocode}
\def\sd@qarrow#1{%
  \begingroup%
  \lccode`\~=`\<\lowercase{\def~{<}}%
  \hbox{\csname sd@arr@#1\endcsname}%
  \endgroup%
}
\@namedef{sd@arr@>>-}{\sd@rightarr\kern-.5\p@\sd@rightarr\kern-\p@}
\@namedef{sd@arr@>-}{\sd@rightarr\kern-\p@}
\@namedef{sd@arr@->}{\sd@rightarr}
\@namedef{sd@arr@-><}{\sd@rightarr\kern-\p@\sd@leftarr}
\@namedef{sd@arr@...}{$\cdots$}
\@namedef{sd@arr@-}{}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sd@newline}
%
% The line breaking within a syntax diagram is controlled by the
% |\sd@newline| command, to which |\\| is assigned.
%
% We support all the standard \LaTeX\ features here.  The line breaking
% involves adding a fill skip and arrow, moving to the next line, adding
% an arrow and a rule, and continuing.
%
%    \begin{macrocode}
\def\sd@newline{\@ifstar{\vadjust{\penalty\@M}\sd@nl@i}\sd@nl@i}
\def\sd@nl@i{\@ifnextchar[\sd@nl@ii\sd@nl@iii}
\def\sd@nl@ii[#1]{\vspace{#1}\sd@nl@iii}
\def\sd@nl@iii{%
  \nobreak\sd@rule\hskip\sdmidskip%
  \sd@rule\hskip\sdfinalskip%
  \kern-3\p@%
  \sd@rightarr%
  \newline%
  \sd@rightarr%
  \nobreak\sd@rule\hskip\sdstartspace%
  \sd@rule\hskip3.5\p@%
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsubsection{Putting things in the right place}
%
% Syntax diagrams have fairly stiff requirements on the positioning of text
% relative to the diagram's rules.  To help people (and me) to write
% extensions to the syntax diagram typesetting which automatically put things
% in the right place, I provide some simple macros.
%
% \begin{environment}{sdbox}
%
% By placing some text in the \env{sdbox} environment, it will be read into a
% box and then output at the correct height for the syntax diagram.  Note
% that stuff in the box is set in horizontal (LR) mode, so you'll have to use
% a \env{minipage} if you want formatted text.  The macro also supplies rules
% on either side of the box, with a length given in the environment's
% argument.
%
% Macro writers are given explicit permission to use this environment through
% the |\sdbox| and |\endsdbox| commands if this makes life easier.
%
% The calculation in the |\endsdbox| macro works out how to centre the box
% vertically over the baseline.  If the box's height is~$h$, and its depth
% is~$d$, then its centre-line is $(h+d)/2$ from the bottom of the box.
% Since the baseline is already $d$ from the bottom, we need to lower the box
% by $(h+d)/2 - d$, or $h/2-d/2$.
%
%    \begin{macrocode}
\def\sdbox#1{%
  \@tempskipa#1\relax%
  \sd@gap\@tempskipa%
  \setbox\z@\hbox\bgroup%
    \begingroup%
    \catcode`\ 10%
    \catcode`\^^M5%
    \synshortsoff%
}
\def\endsdbox{%
    \endgroup%
  \egroup%
  \@tempdima\ht\z@%
  \advance\@tempdima-\dp\z@%
  \advance\@tempdima-\tw@\sd@mid%
  \lower.5\@tempdima\box\z@%
  \sd@gap\@tempskipa%
}
%    \end{macrocode}
%
% \end{environment}
%
% \subsubsection{Typesetting syntactic items}
%
% Using the hooks built into the syntax abbreviations above, we typeset
% the text into a box, and write it out, centred over the baseline.  A strut
% helps to keep the actual text baselines level for short pieces of text.
%
% \begin{macro}{\sd@tok@i}
%
% The preamble for a syntax abbreviation.  We start a box, and set the
% space and return characters to work again.  A strut is added to the box to
% ensure correct vertical spacing for normal text.
%
%    \begin{macrocode}
\def\sd@tok@i{%
  \sdbox\sdtokskip%
  \strut%
  \space%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sd@tok@ii}
%
%    \begin{macrocode}
\def\sd@tok@ii{%
  \space%
  \endsdbox%
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsubsection{Inserting other pieces of text}
%
% Arbitrary text may be put into a syntax diagram through the use of the
% |\tok| macro.  Its `argument' is typeset in the same way as a syntactic
% item (centred over the baseline).  The implementation goes to some effort
% to ensure that the text is not actually an argument, to allow category
% codes to change while the text is being typeset.
%
% \begin{macro}{\tok}
%
% We start a box, and make space and return do their normal jobs.  We use
% |\aftergroup| to regain control once the box is finished.  |\doafter| is
% used to get control after the group finishes.
%
%    \begin{macrocode}
\def\tok#{%
  \sdbox\sdtokskip%
  \strut%
  \enspace%
  \syntaxShortcuts\relax\relax%
  \doafter\sd@tok%
}
%    \end{macrocode}
%
% The |\sd@tok| macro is similar to |\sd@tok@ii| above.
%
%    \begin{macrocode}
\def\sd@tok{%
  \enspace%
  \endsdbox%
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsubsection{The \protect\env{stack} environment}
%
% The \env{stack} environment is used to present alternatives in a syntax
% diagram.  The alternatives are separated by |\\| commands.
%
% \begin{macro}{\stack}
%
% The optional positioning argument is handled using \LaTeX's |\newcommand|
% mechanism.
%
%    \begin{macrocode}
\newcommand\stack[1][t]{%
%    \end{macrocode}
%
% First, we add some horizontal space.
%
%    \begin{macrocode}
  \sd@gap\sdmidskip%
%    \end{macrocode}
%
% We're within a complex construction, so we need to clear the |\ifsd@base|
% flag.
%
%    \begin{macrocode}
  \begingroup\sd@basefalse%
%    \end{macrocode}
%
% The top and bottom rows of the stack are different to the others, since
% the vertical rules mustn't extend all the way up the side of the item.
% The bottom row is handled separately by |\endstack| below.  The top row
% must be handled via a flag, |\ifsd@toplayer|.
%
% Initially, the flag must be set true.
%
%    \begin{macrocode}
  \sd@toplayertrue%
%    \end{macrocode}
%
% We set the |\\| command to separate the items in the |\halign|.
%
%    \begin{macrocode}
  \let\\\sd@stackcr%
%    \end{macrocode}
%
% The actual structure must be set in vertical mode, so we must place it
% in a box.  The position argument determines whether this must be a
% |\vbox| or a |\vtop|.  We also insert a bit of rounding if the options say
% we must.
%
%    \begin{macrocode}
  \if#1t%
    \let\@tempa\vtop%
    \sd@toptrue%
    \ifsd@round\llap{\sd@trcirc\kern\tw@\sdrulewidth}\fi%
  \else\if#1b%
    \let\@tempa\vbox%
    \sd@topfalse%
    \ifsd@round\llap{\sd@brcirc\kern\tw@\sdrulewidth}\fi%
  \else%
    \sd@err{Bad position argument passed to stack}%
           {The positioning argument must be one of `t' or `b'.  I%
            have^^Jassumed you meant to type `t'.}%
    \let\@tempa\vtop%
  \fi\fi%
%    \end{macrocode}
%
% Now we start the box, which we will complete at the end of the environment.
%
%    \begin{macrocode}
  \@tempa\bgroup%
%    \end{macrocode}
%
% We must remove any extra space between rows of the table, since the rules
% will not join up correctly.  We can use |\offinterlineskip| safely, since
% each individual row contains a strut.
%
%    \begin{macrocode}
  \offinterlineskip%
%    \end{macrocode}
%
% Now we can start the alignment.  We actually use \PlainTeX's |\ialign|
% macro, which also clears |\tabskip| for us.
%
%    \begin{macrocode}
  \ialign\bgroup%
%    \end{macrocode}
%
% The preamble is trivial, since we must do all of the work ourselves
%
%    \begin{macrocode}
    ##\cr%
%    \end{macrocode}
%
% We can now start putting the text into a box ready for typesetting later.
% The strut makes the vertical spacing correct.
%
%    \begin{macrocode}
  \setbox\z@\hbox\bgroup%
    \strut%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\endstack}
%
% The first part of this is similar to the |\sd@stackcr| macro below, except
% that the vertical rules are different.  We don't support rounded edges
% on single-row stacks, although this isn't a great loss to humanity.
%
%    \begin{macrocode}
\def\endstack{%
  \egroup%
  \ifsd@toplayer%
    \sd@dostack\sd@upper\sd@lower\relax\relax%
  \else%
    \ifsd@round%
      \ifsd@top%
        \sd@dostack{\ht\z@}\sd@botcirc\sd@blcirc\sd@brcirc%
      \else%
        \sd@dostack{\ht\z@}\sd@botcirc\relax\relax%
      \fi%
    \else%
      \sd@dostack{\ht\z@}\sd@lower\relax\relax%
    \fi%
  \fi%
%    \end{macrocode}
%
% We now close the |\halign| and the vbox we created.
%
%    \begin{macrocode}
  \egroup%
  \egroup%
%    \end{macrocode}
%
% Deal with any rounding we started off.
%
%    \begin{macrocode}
  \ifsd@round%
    \ifsd@top
      \rlap{\kern\tw@\sdrulewidth\sd@tlcirc}%
    \else%
      \rlap{\kern\tw@\sdrulewidth\sd@blcirc}%
    \fi%
  \fi%
%    \end{macrocode}
%
% Finally, we add some horizontal glue to space the diagram out.
%
%    \begin{macrocode}
  \endgroup\sd@gap\sdmidskip%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sd@stackcr}
%
% The |\\| command is set to this macro during a \env{stack} environment.
%
%    \begin{macrocode}
\def\sd@stackcr{%
%    \end{macrocode}
%
% The first job is to close the box containing the previous item.
%
%    \begin{macrocode}
  \egroup%
%    \end{macrocode}
%
% Now we typeset the vertical rules differently depending on whether this is
% the first item in the stack.  This looks quite terrifying initially, but
% it's just an enumeration of the possible cases for the different values
% of |\ifsd@toplayer|, |\ifsd@top| and |\ifsd@round|, putting in appropriate
% rules and arcs in the right places.
%
%    \begin{macrocode}
  \ifsd@toplayer%
    \ifsd@round%
      \ifsd@top%
        \sd@dostack\sd@topcirc{\dp\z@}\relax\relax%
      \else%
        \sd@dostack\sd@topcirc{\dp\z@}\sd@tlcirc\sd@trcirc%
      \fi%
    \else%
      \sd@dostack\sd@upper{\dp\z@}\relax\relax%
    \fi%
  \else%
    \ifsd@round%
      \ifsd@top%
        \sd@dostack{\ht\z@}{\dp\z@}\sd@blcirc\sd@brcirc%
      \else%
        \sd@dostack{\ht\z@}{\dp\z@}\sd@tlcirc\sd@trcirc%
      \fi%
    \else%
      \sd@dostack{\ht\z@}{\dp\z@}\relax\relax%
    \fi%
  \fi%
%    \end{macrocode}
%
% The next item won't be the first, so we clear the flag.
%
%    \begin{macrocode}
  \sd@toplayerfalse%
%    \end{macrocode}
%
% Now we have to set up the next cell.  We put the text into a box again.
%
%    \begin{macrocode}
  \setbox\z@\hbox\bgroup%
    \strut%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sd@dostack}
%
% Actually typesetting the text in a cell is performed here.  The macro is
% called as
% \begin{quote}\synshorts
% "\\sd@dostack{"<height>"}{"<depth>"}{"<left-arc>"}{"<right-arc>"}"
% \end{quote}
% where \<height> and \<depth> are the height and depth of the vertical
% rules to put around the item, and \<left-arc> and \<right-arc> are
% commands to draw rounded edges on the left and right hand sides of the
% item.
%
% The values for the height and depth are quite often going to be the height
% and depth of box~0.  Since we empty box~0 in the course of typesetting the
% row, we need to cache the sizes on entry.
%
%    \begin{macrocode}
\def\sd@dostack#1#2#3#4{%
  \@tempdima#1%
  \@tempdimb#2%
  \kern-\tw@\sdrulewidth%
  \vrule\@height\@tempdima\@depth\@tempdimb\@width\tw@\sdrulewidth%
  #3%
  \sd@rule\hfill%
  \sd@gap\sdtokskip%
  \unhbox\z@%
  \sd@gap\sdtokskip%
  \sd@rule\hfill%
  #4%
  \vrule\@height\@tempdima\@depth\@tempdimb\@width\tw@\sdrulewidth%
  \kern-\tw@\sdrulewidth%
  \cr%
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsubsection{The \protect\env{rep} environment}
%
% The \env{rep} environment is used for typesetting loops in the diagram.
% Again, we use |\halign| for the typesetting.  Loops are simpler than
% stacks, however, since there are always two rows.  We store both rows in
% box registers, and build the loop at the end.
%
% \begin{macro}{\rep}
%
% Again, we use |\newcommand| to process the optional argument.
%
%    \begin{macrocode}
\newcommand\rep[1][t]{%
%    \end{macrocode}
%
% First, leave a gap on the left side.
%
%    \begin{macrocode}
  \sd@gap\sdmidskip%
%    \end{macrocode}
%
% We're not at base level any more, so disable linebreaking.
%
%    \begin{macrocode}
  \begingroup\sd@basefalse%
%    \end{macrocode}
%
% Remember we're going backwards now.
%
%    \begin{macrocode}
  \ifsd@backwards\sd@backwardsfalse\else\sd@backwardstrue\fi%
%    \end{macrocode}
%
% Define |\\| to separate the two parts of the loop.
%
%    \begin{macrocode}
   \let\\\sd@loop%
%    \end{macrocode}
%
% Now check the argument, and use the appropriate type of box.  In addition
% to changing the typesetting, we must remember which way up to typeset the
% loop, since the end code must always put the first argument on the
% baseline, with the loop either above or below.
%
%    \begin{macrocode}
  \if#1t%
    \let\@tempa\vbox%
    \sd@toptrue%
  \else\if#1b%
    \let\@tempa\vtop%
    \sd@topfalse%
  \else%
    \sd@err{Bad position argument passed to loop}%
           {The positioning argument must be `t' or `b'.  I have^^J%
            assumed you meant to type `t'.}%
    \let\@tempa\vbox%
    \sd@toptrue%
  \fi\fi%
%    \end{macrocode}
%
% Now we start the box.
%
%    \begin{macrocode}
  \@tempa\bgroup%
%    \end{macrocode}
%
% The loop is by default empty, apart from a strut.  This is put into box~1.
%
%    \begin{macrocode}
  \setbox\tw@\copy\strutbox%
%    \end{macrocode}
%
% Now start typesetting the main text in box~0.
%
%    \begin{macrocode}
  \setbox\z@\hbox\bgroup\strut%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\endrep}
%
% The final code must first close whatever box was open.
%
%    \begin{macrocode}
\def\endrep{%
  \egroup%
%    \end{macrocode}
%
% Now we typeset the loop, depending on which way up it was meant to be.
% Again, this terrifying piece of code is a simple list of possibile values
% of our various flags.
%
%    \begin{macrocode}
  \ifsd@top%
    \ifsd@round%
      \sd@doloop\tw@\z@\relax\relax%
        \sd@tlcirc\sd@trcirc{\sd@rlc\sd@blcirc}{\sd@llc\sd@brcirc}%
    \else%
      \sd@doloop\tw@\z@\relax\sd@downarr\relax\relax\relax\relax%
    \fi%
  \else%
    \ifsd@round%
      \sd@doloop\z@\tw@\relax\relax%
        {\sd@rlc\sd@tlcirc}{\sd@llc\sd@trcirc}\sd@blcirc\sd@brcirc%
    \else%
      \sd@doloop\z@\tw@\sd@uparr\relax\relax\relax\relax\relax%
    \fi%
  \fi%
%    \end{macrocode}
%
% Close the vbox we opened.
%
%    \begin{macrocode}
  \egroup%
%    \end{macrocode}
%
% Finally, we leave a gap before the next structure.
%
%    \begin{macrocode}
  \endgroup\sd@gap\sdmidskip%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sd@loop}
%
% This macro handles the |\\| command within a loop environment.  We close
% the current box, and start filling in box~1.  We also redefine |\\| to
% raise an error when the |\\| command is used again.
%
%    \begin{macrocode}
\def\sd@loop{%
  \egroup%
  \def\\{\sd@err{Too many \string\\\space commands in loop}\@ehc}%
  \setbox\tw@\hbox\bgroup\strut%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sd@doloop}
%
% This is the macro which actually creates the |\halign| for the loop.  It
% is called with four arguments, as:
% \begin{quote}\synshorts
% "\\sd@doloop{"<top-box>"}{"<bottom-box>"}"^^A
%                "{"<top-arrow>"}{"<btm-arrow>"}" \\
% \hbox{}\quad "{"<top-left-arc>"}{"<top-right-arc>"}"^^A
%                "{"<bottom-left-arc>"}{"<btm-right-arc>"}"^^A
% \kern-1in ^^A It may be overfull, but it looks OK to me ;-)
% \end{quote}
%
% The two \<box> arguments give the numbers of boxes to extract in the top
% and bottom rows of the alignment.  The \<arrow> arguments specify
% characters to typeset at the end of the top and bottom rows for arrows.
% The various \<arc> arguments are commands which typeset arcs around the
% various parts of the items.
%
% We calculate the height and depth of the two boxes, and store them in
% \<dimen> registers, because the boxes are emptied before the right-hand
% rules are typeset.
%
% Actually, the two rows of the alignment are typeset in a different macro:
% we just pass the correct information on.
%
%    \begin{macrocode}
\def\sd@doloop#1#2#3#4#5#6#7#8{%
  \@tempdima\dp#1\relax%
  \@tempdimb\ht#2\relax%
  \offinterlineskip%
  \ialign{%
    ##\cr%
    \ifsd@round%
      \sd@doloop@i#1#3\sd@topcirc\@tempdima{#5}{#6}%
      \sd@doloop@i#2#4\@tempdimb\sd@botcirc{#7}{#8}%
    \else%
      \sd@doloop@i#1#3\sd@upper\@tempdima{#5}{#6}%
      \sd@doloop@i#2#4\@tempdimb\sd@lower{#7}{#8}%
    \fi%
  }%
}
%    \end{macrocode}
%
% \end{macro}
%
% \begin{macro}{\sd@doloop@i}
%
% Here we do the actual job of typesetting the rows of a loop alignment.
% The four arguments are:
% \begin{quote}\synshorts
% "\\sd@doloop@i{"<box>"}{"<arrow>"}"^^A
%              "{"<rule-height>"}{"<rule-depth>"}" \\
% \hbox{}\quad "{"<left-arc>"}{"<right-arc>"}"^^A
% \end{quote}
%
% The arrow position is determined by the |\ifsd@backwards| flag.  The rest
% is fairly simple.
%
%    \begin{macrocode}
\def\sd@doloop@i#1#2#3#4#5#6{%
  \ifsd@backwards#2\fi%
  \kern-\tw@\sdrulewidth%
  \vrule\@height#3\@depth#4\@width\tw@\sdrulewidth%
  #5%
  \sd@rule\hfill%
  \sd@gap\sdtokskip%
  \unhbox#1%
  \sd@gap\sdtokskip%
  \sd@rule\hfill%
  #6%
  \vrule\@height#3\@depth#4\@width\tw@\sdrulewidth%
  \ifsd@backwards\else#2\fi%
  \kern-\tw@\sdrulewidth%
  \cr%
}
%    \end{macrocode}
%
% \end{macro}
%
% \subsection{The end}
%
% Phew!  That's all of it completed.  I hope this collection of commands
% and environments is of some help to someone.
%
%    \begin{macrocode}
%</package>
%    \end{macrocode}
%
% \hfill Mark Wooding, \today
%
% \Finale
%
\endinput