File: input.c

package info (click to toggle)
halibut 1.2-2
  • links: PTS, VCS
  • area: main
  • in suites: bullseye, buster, sid
  • size: 4,084 kB
  • sloc: ansic: 58,429; makefile: 294; perl: 194; lisp: 76; sh: 21
file content (1676 lines) | stat: -rw-r--r-- 44,799 bytes parent folder | download | duplicates (2)
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
/*
 * input.c: read the source form
 */

#include <stdio.h>
#include <assert.h>
#include <time.h>
#include "halibut.h"

#define TAB_STOP 8		       /* for column number tracking */

static void setpos(input *in, char *fname) {
    in->pos.filename = fname;
    in->pos.line = 1;
    in->pos.col = (in->reportcols ? 1 : -1);
}

static void unget(input *in, int c, filepos *pos) {
    if (in->npushback >= in->pushbacksize) {
	in->pushbacksize = in->npushback + 16;
	in->pushback = sresize(in->pushback, in->pushbacksize, pushback);
    }
    in->pushback[in->npushback].chr = c;
    in->pushback[in->npushback].pos = *pos;   /* structure copy */
    in->npushback++;
}

/* ---------------------------------------------------------------------- */
/*
 * Macro subsystem
 */
typedef struct macro_Tag macro;
struct macro_Tag {
    wchar_t *name, *text;
};
struct macrostack_Tag {
    macrostack *next;
    wchar_t *text;
    int ptr, npushback;
    filepos pos;
};
static int macrocmp(void *av, void *bv) {
    macro *a = (macro *)av, *b = (macro *)bv;
    return ustrcmp(a->name, b->name);
}
static void macrodef(tree234 *macros, wchar_t *name, wchar_t *text,
		     filepos fpos) {
    macro *m = snew(macro);
    m->name = name;
    m->text = text;
    if (add234(macros, m) != m) {
	err_macroexists(&fpos, name);
	sfree(name);
	sfree(text);
    }
}
static int macrolookup(tree234 *macros, input *in, wchar_t *name,
		       filepos *pos) {
    macro m, *gotit;
    m.name = name;
    gotit = find234(macros, &m, NULL);
    if (gotit) {
	macrostack *expansion = snew(macrostack);
	expansion->next = in->stack;
	expansion->text = gotit->text;
	expansion->pos = *pos;	       /* structure copy */
	expansion->ptr = 0;
	expansion->npushback = in->npushback;
	in->stack = expansion;
	return TRUE;
    } else
	return FALSE;
}
static void macrocleanup(tree234 *macros) {
    int ti;
    macro *m;
    for (ti = 0; (m = (macro *)index234(macros, ti)) != NULL; ti++) {
	sfree(m->name);
	sfree(m->text);
	sfree(m);
    }
    freetree234(macros);
}

static void input_configure(input *in, paragraph *cfg) {
    assert(cfg->type == para_Config);

    if (!ustricmp(cfg->keyword, L"input-charset")) {
	in->charset = charset_from_ustr(&cfg->fpos, uadv(cfg->keyword));
    }
}

/*
 * Can return EOF
 */
static int get(input *in, filepos *pos, rdstringc *rsc) {
    int pushbackpt = in->stack ? in->stack->npushback : 0;
    if (in->npushback > pushbackpt) {
	--in->npushback;
	if (pos)
	    *pos = in->pushback[in->npushback].pos;   /* structure copy */
	return in->pushback[in->npushback].chr;
    }
    else if (in->stack) {
	wchar_t c = in->stack->text[in->stack->ptr];
        if (pos)
            *pos = in->stack->pos;
	if (in->stack->text[++in->stack->ptr] == L'\0') {
	    macrostack *tmp = in->stack;
	    in->stack = tmp->next;
	    sfree(tmp);
	}
	return c;
    }
    else if (in->currfp) {

	while (in->wcpos >= in->nwc) {

	    int c = getc(in->currfp);

	    if (c == EOF) {
		if (in->wantclose)
		    fclose(in->currfp);
		in->currfp = NULL;
		return EOF;
	    }

	    if (rsc)
		rdaddc(rsc, c);

	    /* Track line numbers, for error reporting */
	    if (pos)
		*pos = in->pos;
	    if (in->reportcols) {
		switch (c) {
		  case '\t':
		    in->pos.col = 1 + (in->pos.col + TAB_STOP-1) % TAB_STOP;
		    break;
		  case '\n':
		    in->pos.col = 1;
		    in->pos.line++;
		    break;
		  default:
		    in->pos.col++;
		    break;
		}
	    } else {
		in->pos.col = -1;
		if (c == '\n')
		    in->pos.line++;
	    }

	    /*
	     * Do input character set translation, so that we return
	     * Unicode.
	     */
	    {
		char buf[1];
		char const *p;
		int inlen;

		buf[0] = (char)c;
		p = buf;
		inlen = 1;

		in->nwc = charset_to_unicode(&p, &inlen,
					     in->wc, lenof(in->wc),
					     in->charset, &in->csstate,
					     NULL, 0);
		assert(p == buf+1 && inlen == 0);

		in->wcpos = 0;
	    }
	}

	return in->wc[in->wcpos++];

    } else
	return EOF;
}

/*
 * Lexical analysis of source files.
 */
typedef struct token_Tag token;
struct token_Tag {
    int type;
    int cmd, aux;
    wchar_t *text;
    char *origtext;
    filepos pos;
};
enum {
    tok_eof,			       /* end of file */
    tok_eop,			       /* end of paragraph */
    tok_white,			       /* whitespace */
    tok_word,			       /* a word or word fragment */
    tok_cmd,			       /* \command */
    tok_lbrace,			       /* { */
    tok_rbrace			       /* } */
};

/* Halibut command keywords. */
enum {
    c__invalid,			       /* invalid command */
    c__comment,			       /* comment command (\#) */
    c__escaped,			       /* escaped character */
    c__nop,			       /* no-op */
    c__nbsp,			       /* nonbreaking space */
    c_A,			       /* appendix heading */
    c_B,			       /* bibliography entry */
    c_BR,			       /* bibliography rewrite */
    c_C,			       /* chapter heading */
    c_H,			       /* heading */
    c_I,			       /* invisible index mark */
    c_IM,			       /* index merge/rewrite */
    c_K,			       /* capitalised cross-reference */
    c_S,			       /* aux field is 0, 1, 2, ... */
    c_U,			       /* unnumbered-chapter heading */
    c_W,			       /* Web hyperlink */
    c_b,			       /* bulletted list */
    c_c,			       /* code */
    c_cfg,			       /* configuration directive */
    c_copyright,		       /* copyright statement */
    c_cq,			       /* quoted code (sugar for \q{\cw{x}}) */
    c_cw,			       /* weak code */
    c_date,			       /* document processing date */
    c_dd,			       /* description list: description */
    c_define,			       /* macro definition */
    c_dt,			       /* description list: described thing */
    c_e,			       /* emphasis */
    c_i,			       /* visible index mark */
    c_ii,			       /* uncapitalised visible index mark */
    c_k,			       /* uncapitalised cross-reference */
    c_lcont,			       /* continuation para(s) for list item */
    c_n,			       /* numbered list */
    c_nocite,			       /* bibliography trickery */
    c_preamble,			       /* (obsolete) preamble text */
    c_q,			       /* quote marks */
    c_quote,			       /* block-quoted paragraphs */
    c_rule,			       /* horizontal rule */
    c_s,			       /* strong */
    c_title,			       /* document title */
    c_u,			       /* aux field is char code */
    c_versionid			       /* document RCS id */
};

/* Perhaps whitespace should be defined in a more Unicode-friendly way? */
#define iswhite(c) ( (c)==32 || (c)==9 || (c)==13 || (c)==10 )
#define isnl(c) ( (c)==10 )
#define isdec(c) ( ((c)>='0'&&(c)<='9') )
#define fromdec(c) ( (c)-'0' )
#define ishex(c) ( ((c)>='0'&&(c)<='9') || ((c)>='A'&&(c)<='F') || ((c)>='a'&&(c)<='f'))
#define fromhex(c) ( (c)<='9' ? (c)-'0' : ((c)&0xDF) - ('A'-10) )
#define iscmd(c) ( ((c)>='0'&&(c)<='9') || ((c)>='A'&&(c)<='Z') || ((c)>='a'&&(c)<='z'))

/*
 * Keyword comparison function. Like strcmp, but between a wchar_t *
 * and a char *.
 */
static int kwcmp(wchar_t const *p, char const *q) {
    int i;
    do {
	i = *p - *q;
    } while (*p++ && *q++ && !i);
    return i;
}

/*
 * Match a keyword.
 */
static void match_kw(token *tok) {
    /*
     * FIXME. The ids are explicit in here so as to allow long-name
     * equivalents to the various very short keywords.
     */
    static const struct { char const *name; int id; } keywords[] = {
	{"#", c__comment},	       /* comment command (\#) */
	{"-", c__escaped},	       /* nonbreaking hyphen */
	{".", c__nop},		       /* no-op */
	{"A", c_A},		       /* appendix heading */
	{"B", c_B},		       /* bibliography entry */
	{"BR", c_BR},		       /* bibliography rewrite */
	{"C", c_C},		       /* chapter heading */
	{"H", c_H},		       /* heading */
	{"I", c_I},		       /* invisible index mark */
	{"IM", c_IM},		       /* index merge/rewrite */
	{"K", c_K},		       /* capitalised cross-reference */
	{"U", c_U},		       /* unnumbered-chapter heading */
	{"W", c_W},		       /* Web hyperlink */
	{"\\", c__escaped},	       /* escaped backslash (\\) */
	{"_", c__nbsp},		       /* nonbreaking space (\_) */
	{"b", c_b},		       /* bulletted list */
	{"c", c_c},		       /* code */
	{"cfg", c_cfg},		       /* configuration directive */
	{"copyright", c_copyright},    /* copyright statement */
	{"cq", c_cq},		       /* quoted code (sugar for \q{\cw{x}}) */
	{"cw", c_cw},		       /* weak code */
	{"date", c_date},	       /* document processing date */
	{"dd", c_dd},		       /* description list: description */
	{"define", c_define},	       /* macro definition */
	{"dt", c_dt},		       /* description list: described thing */
	{"e", c_e},		       /* emphasis */
	{"i", c_i},		       /* visible index mark */
	{"ii", c_ii},		       /* uncapitalised visible index mark */
	{"k", c_k},		       /* uncapitalised cross-reference */
	{"lcont", c_lcont},	       /* continuation para(s) for list item */
	{"n", c_n},		       /* numbered list */
	{"nocite", c_nocite},	       /* bibliography trickery */
	{"preamble", c_preamble},      /* (obsolete) preamble text */
	{"q", c_q},		       /* quote marks */
	{"quote", c_quote},	       /* block-quoted paragraphs */
	{"rule", c_rule},	       /* horizontal rule */
	{"s", c_s},		       /* strong */
	{"title", c_title},	       /* document title */
	{"versionid", c_versionid},    /* document RCS id */
	{"{", c__escaped},	       /* escaped lbrace (\{) */
	{"}", c__escaped},	       /* escaped rbrace (\}) */
    };
    int i, j, k, c;

    /*
     * Special cases: \S{0,1,2,...} and \uABCD. If the syntax
     * doesn't match correctly, we just fall through to the
     * binary-search phase.
     */
    if (tok->text[0] == 'S') {
	/* We expect numeric characters thereafter. */
	wchar_t *p = tok->text+1;
	int n;
	if (!*p)
	    n = 1;
	else {
	    n = 0;
	    while (*p && isdec(*p)) {
		n = 10 * n + fromdec(*p);
		p++;
	    }
	}
	if (!*p) {
	    tok->cmd = c_S;
	    tok->aux = n;
	    return;
	}
    } else if (tok->text[0] == 'u') {
	/* We expect hex characters thereafter. */
	wchar_t *p = tok->text+1;
	int n = 0;
	while (*p && ishex(*p)) {
	    n = 16 * n + fromhex(*p);
	    p++;
	}
	if (!*p) {
	    tok->cmd = c_u;
	    tok->aux = n;
	    return;
	}
    }

    i = -1;
    j = sizeof(keywords)/sizeof(*keywords);
    while (j-i > 1) {
	k = (i+j)/2;
	c = kwcmp(tok->text, keywords[k].name);
	if (c < 0)
	    j = k;
	else if (c > 0)
	    i = k;
	else /* c == 0 */ {
	    tok->cmd = keywords[k].id;
	    return;
	}
    }

    tok->cmd = c__invalid;
}


/*
 * Read a token from the input file, in the normal way (`normal' in
 * the sense that code paragraphs work a different way).
 */
token get_token(input *in) {
    int c;
    int nls;
    int prevpos;
    token ret;
    rdstring rs = { 0, 0, NULL };
    rdstringc rsc = { 0, 0, NULL };
    filepos cpos;

    ret.text = NULL;		       /* default */
    ret.origtext = NULL;	       /* default */
    if (in->pushback_chars) {
	rdaddsc(&rsc, in->pushback_chars);
	sfree(in->pushback_chars);
	in->pushback_chars = NULL;
    }
    c = get(in, &cpos, &rsc);
    ret.pos = cpos;
    if (iswhite(c)) {		       /* tok_white or tok_eop */
	nls = 0;
	prevpos = 0;
	do {
	    if (isnl(c))
		nls++;
	    prevpos = rsc.pos;
	} while ((c = get(in, &cpos, &rsc)) != EOF && iswhite(c));
	if (c == EOF) {
	    ret.type = tok_eof;
	    sfree(rsc.text);
	    return ret;
	}
	if (rsc.text) {
	    in->pushback_chars = dupstr(rsc.text + prevpos);
	    sfree(rsc.text);
	}
	unget(in, c, &cpos);
	ret.type = (nls > 1 ? tok_eop : tok_white);
	return ret;
    } else if (c == EOF) {	       /* tok_eof */
	ret.type = tok_eof;
	sfree(rsc.text);
	return ret;
    } else if (c == '\\') {	       /* tok_cmd */
	rsc.pos = prevpos = 0;
	c = get(in, &cpos, &rsc);
	if (c == '-' || c == '\\' || c == '_' ||
	    c == '#' || c == '{' || c == '}' || c == '.') {
	    /* single-char command */
	    rdadd(&rs, c);
	    prevpos = rsc.pos;
	} else if (c == 'u') {
	    int len = 0;
	    do {
		rdadd(&rs, c);
		len++;
		prevpos = rsc.pos;
		c = get(in, &cpos, &rsc);
	    } while (ishex(c) && len < 5);
	    unget(in, c, &cpos);
	} else if (iscmd(c)) {
	    do {
		rdadd(&rs, c);
		prevpos = rsc.pos;
		c = get(in, &cpos, &rsc);
	    } while (iscmd(c));
	    unget(in, c, &cpos);
	}
	/*
	 * Now match the command against the list of available
	 * ones.
	 */
	ret.type = tok_cmd;
	ret.text = ustrdup(rs.text);
	if (rsc.text) {
	    in->pushback_chars = dupstr(rsc.text + prevpos);
	    rsc.text[prevpos] = '\0';
	    ret.origtext = dupstr(rsc.text);
	} else {
	    ret.origtext = dupstr("");
	}
	match_kw(&ret);
	sfree(rs.text);
	sfree(rsc.text);
	return ret;
    } else if (c == '{') {	       /* tok_lbrace */
	ret.type = tok_lbrace;
	sfree(rsc.text);
	return ret;
    } else if (c == '}') {	       /* tok_rbrace */
	ret.type = tok_rbrace;
	sfree(rsc.text);
	return ret;
    } else {			       /* tok_word */
	/*
	 * Read a word: the longest possible contiguous sequence of
	 * things other than whitespace, backslash, braces and
	 * hyphen. A hyphen terminates the word but is returned as
	 * part of it; everything else is pushed back for the next
	 * token. The `aux' field contains TRUE if the word ends in
	 * a hyphen.
	 */
	ret.aux = FALSE;	       /* assumed for now */
	prevpos = 0;
	while (1) {
	    if (iswhite(c) || c=='{' || c=='}' || c=='\\' || c==EOF) {
		/* Put back the character that caused termination */
		unget(in, c, &cpos);
		break;
	    } else {
		rdadd(&rs, c);
		if (c == '-') {
		    prevpos = rsc.pos;
		    ret.aux = TRUE;
		    break;	       /* hyphen terminates word */
		}
	    }
	    prevpos = rsc.pos;
	    c = get(in, &cpos, &rsc);
	}
	ret.type = tok_word;
	ret.text = ustrdup(rs.text);
	if (rsc.text) {
	    in->pushback_chars = dupstr(rsc.text + prevpos);
	    rsc.text[prevpos] = '\0';
	    ret.origtext = dupstr(rsc.text);
	} else {
	    ret.origtext = dupstr("");
	}
	sfree(rs.text);
	sfree(rsc.text);
	return ret;
    }
}

/*
 * Determine whether the next input character is an open brace (for
 * telling code paragraphs from paragraphs which merely start with
 * code).
 */
int isbrace(input *in) {
    int c;
    filepos cpos;

    c = get(in, &cpos, NULL);
    unget(in, c, &cpos);
    return (c == '{');
}

/*
 * Read the rest of a line that starts `\c'. Including nothing at
 * all (tok_word with empty text).
 */
token get_codepar_token(input *in) {
    int c;
    token ret;
    rdstring rs = { 0, 0, NULL };
    filepos cpos;

    ret.type = tok_word;
    ret.origtext = NULL;
    c = get(in, &cpos, NULL);	       /* expect (and discard) one space */
    ret.pos = cpos;
    if (c == ' ') {
	c = get(in, &cpos, NULL);
	ret.pos = cpos;
    }
    while (!isnl(c) && c != EOF) {
	int c2 = c;
	c = get(in, &cpos, NULL);
	/* Discard \r just before \n. */
	if (c2 != 13 || !isnl(c))
	    rdadd(&rs, c2);
    }
    unget(in, c, &cpos);
    ret.text = ustrdup(rs.text);
    sfree(rs.text);
    return ret;
}

/*
 * Adds a new word to a linked list
 */
static word *addword(word newword, word ***hptrptr) {
    word *mnewword;
    if (!hptrptr)
	return NULL;
    mnewword = snew(word);
    *mnewword = newword;	       /* structure copy */
    mnewword->next = NULL;
    **hptrptr = mnewword;
    *hptrptr = &mnewword->next;
    return mnewword;
}

/*
 * Adds a new paragraph to a linked list
 */
static paragraph *addpara(paragraph newpara, paragraph ***hptrptr) {
    paragraph *mnewpara = snew(paragraph);
    *mnewpara = newpara;	       /* structure copy */
    mnewpara->next = NULL;
    **hptrptr = mnewpara;
    *hptrptr = &mnewpara->next;
    return mnewpara;
}

/*
 * Destructor before token is reassigned; should catch most memory
 * leaks
 */
#define dtor(t) ( sfree(t.text), sfree(t.origtext) )

/*
 * Reads a single file (ie until get() returns EOF)
 */
static void read_file(paragraph ***ret, input *in, indexdata *idx,
		      tree234 *macros) {
    token t;
    paragraph par;
    word wd, **whptr, **idximplicit;
    wchar_t utext[2], *wdtext;
    int style, spcstyle;
    int already;
    int iswhite, seenwhite;
    int type;
    int prev_para_type;
    struct stack_item {
	enum {
	    stack_nop = 0,	       /* do nothing (for error recovery) */
	    stack_ualt = 1,	       /* \u alternative */
	    stack_style = 2,	       /* \e, \c, \cw */
	    stack_idx = 4,	       /* \I, \i, \ii */
	    stack_hyper = 8,	       /* \W */
	    stack_quote = 16	       /* \q */
	} type;
	word **whptr;		       /* to restore from \u alternatives */
	word **idximplicit;	       /* to restore from \u alternatives */
	filepos fpos;
	int in_code;
    } *sitem;
    stack parsestk;
    struct crossparaitem {
 	int type;		       /* currently c_lcont, c_quote or -1 */
	int seen_lcont, seen_quote;
    };
    stack crossparastk;
    word *indexword, *uword, *iword;
    word *idxwordlist;
    rdstring indexstr;
    int index_downcase, index_visible, indexing;
    const rdstring nullrs = { 0, 0, NULL };
    wchar_t uchr;

    t.text = NULL;
    t.origtext = NULL;
    already = FALSE;

    crossparastk = stk_new();

    /*
     * Loop on each paragraph.
     */
    while (1) {
	int start_cmd = c__invalid;
	par.words = NULL;
	par.keyword = NULL;
	par.origkeyword = NULL;
	whptr = &par.words;

	/*
	 * Get a token.
	 */
	do {
	    if (!already) {
		dtor(t), t = get_token(in);
	    }
	    already = FALSE;
	} while (t.type == tok_eop);
	if (t.type == tok_eof)
	    break;

	/*
	 * Parse code paragraphs separately.
	 */
	if (t.type == tok_cmd && t.cmd == c_c && !isbrace(in)) {
	    int wtype = word_WeakCode;

	    par.type = para_Code;
	    par.fpos = t.pos;
	    while (1) {
		dtor(t), t = get_codepar_token(in);
		wd.type = wtype;
		wd.breaks = FALSE;     /* shouldn't need this... */
		wd.text = ustrdup(t.text);
		wd.alt = NULL;
		wd.fpos = t.pos;
		addword(wd, &whptr);
		dtor(t), t = get_token(in);
		if (t.type == tok_white) {
		    /*
		     * The newline after a code-paragraph line
		     */
		    dtor(t), t = get_token(in);
		}
		if (t.type == tok_eop || t.type == tok_eof ||
		    t.type == tok_rbrace) { /* might be } terminating \lcont */
		    if (t.type == tok_rbrace)
			already = TRUE;
		    break;
		} else if (t.type == tok_cmd && t.cmd == c_c) {
		    wtype = word_WeakCode;
		} else if (t.type == tok_cmd && t.cmd == c_e &&
			   wtype == word_WeakCode) {
		    wtype = word_Emph;
		} else if (t.type == tok_cmd && t.cmd == c_s &&
			   wtype == word_WeakCode) {
		    wtype = word_Strong;
		} else {
		    err_brokencodepara(&t.pos);
		    prev_para_type = par.type;
		    addpara(par, ret);
		    while (t.type != tok_eop)   /* error recovery: */
			dtor(t), t = get_token(in);   /* eat rest of paragraph */
		    goto codeparabroken;   /* ick, but such is life */
		}
	    }
	    prev_para_type = par.type;
	    addpara(par, ret);
	    codeparabroken:
	    continue;
	}

	/*
	 * Spot the special commands that define a grouping of more
	 * than one paragraph, and also the closing braces that
	 * finish them.
	 */
	if (t.type == tok_cmd &&
	    (t.cmd == c_lcont || t.cmd == c_quote)) {
	    struct crossparaitem *sitem, *stop;
	    int cmd = t.cmd;

	    /*
	     * Expect, and swallow, an open brace.
	     */
	    dtor(t), t = get_token(in);
	    if (t.type != tok_lbrace) {
		err_explbr(&t.pos);
		continue;
	    }

	    /*
	     * Also expect, and swallow, any whitespace after that
	     * (a newline before a code paragraph wouldn't be
	     * surprising).
	     */
	    do {
		dtor(t), t = get_token(in);
	    } while (t.type == tok_white);
	    already = TRUE;

	    if (cmd == c_lcont) {
		/*
		 * \lcont causes a continuation of a list item into
		 * multiple paragraphs (which may in turn contain
		 * nested lists, code paras etc). Hence, the previous
		 * paragraph must be of a list type.
		 */
		sitem = snew(struct crossparaitem);
		stop = (struct crossparaitem *)stk_top(crossparastk);
		if (stop)
		    *sitem = *stop;
		else
		    sitem->seen_quote = sitem->seen_lcont = 0;

		if (prev_para_type == para_Bullet ||
		    prev_para_type == para_NumberedList ||
		    prev_para_type == para_Description) {
		    sitem->type = c_lcont;
		    sitem->seen_lcont = 1;
		    par.type = para_LcontPush;
		    prev_para_type = par.type;
		    addpara(par, ret);
		} else {
		    /*
		     * Push a null item on the cross-para stack so that
		     * when we see the corresponding closing brace we
		     * don't give a cascade error.
		     */
		    sitem->type = -1;
		    err_misplacedlcont(&t.pos);
		}
	    } else {
		/*
		 * \quote causes a group of paragraphs to be
		 * block-quoted (typically they will be indented a
		 * bit).
		 */
		sitem = snew(struct crossparaitem);
		stop = (struct crossparaitem *)stk_top(crossparastk);
		if (stop)
		    *sitem = *stop;
		else
		    sitem->seen_quote = sitem->seen_lcont = 0;
		sitem->type = c_quote;
		sitem->seen_quote = 1;
		par.type = para_QuotePush;
		prev_para_type = par.type;
		addpara(par, ret);
	    }
	    stk_push(crossparastk, sitem);
	    continue;
	} else if (t.type == tok_rbrace) {
	    struct crossparaitem *sitem = stk_pop(crossparastk);
	    if (!sitem)
		err_unexbrace(&t.pos);
	    else {
		switch (sitem->type) {
		  case c_lcont:
		    par.type = para_LcontPop;
		    prev_para_type = par.type;
		    addpara(par, ret);
		    break;
		  case c_quote:
		    par.type = para_QuotePop;
		    prev_para_type = par.type;
		    addpara(par, ret);
		    break;
		}
		sfree(sitem);
	    }
	    continue;
	}

	while (t.type == tok_cmd &&
	       macrolookup(macros, in, t.text, &t.pos)) {
	    dtor(t), t = get_token(in);
	}

	/*
	 * This token begins a paragraph. See if it's one of the
	 * special commands that define a paragraph type.
	 *
	 * (note that \# is special in a way, and \nocite takes no
	 * text)
	 */
	par.type = para_Normal;
	if (t.type == tok_cmd) {
	    int needkw;
	    int is_macro = FALSE;

	    par.fpos = t.pos;
	    switch (t.cmd) {
	      default:
		needkw = -1;
		break;
	      case c__invalid:
		err_badparatype(t.text, &t.pos);
		needkw = 4;
		break;
	      case c__comment:
		if (isbrace(in)) {
		    needkw = -1;
		    break;	       /* `\#{': isn't a comment para */
		}
		do {
		    dtor(t), t = get_token(in);
		} while (t.type != tok_eop && t.type != tok_eof);
		continue;	       /* next paragraph */
		/*
		 * `needkw' values:
		 *
		 *   1 -- exactly one keyword
		 *   2 -- at least one keyword
		 *   4 -- any number of keywords including zero
		 *   8 -- at least one keyword and then nothing else
		 *  16 -- nothing at all! no keywords, no body
		 *  32 -- no keywords at all
		 */
	      case c_A: needkw = 2; par.type = para_Appendix; break;
	      case c_B: needkw = 2; par.type = para_Biblio; break;
	      case c_BR: needkw = 1; par.type = para_BR;
		start_cmd = c_BR; break;
	      case c_C: needkw = 2; par.type = para_Chapter; break;
	      case c_H: needkw = 2; par.type = para_Heading;
		par.aux = 0;
		break;
	      case c_IM: needkw = 2; par.type = para_IM;
		start_cmd = c_IM; break;
	      case c_S: needkw = 2; par.type = para_Subsect;
		par.aux = t.aux; break;
	      case c_U: needkw = 32; par.type = para_UnnumberedChapter; break;
		/* For \b and \n the keyword is optional */
	      case c_b: needkw = 4; par.type = para_Bullet; break;
	      case c_dt: needkw = 4; par.type = para_DescribedThing; break;
	      case c_dd: needkw = 4; par.type = para_Description; break;
	      case c_n: needkw = 4; par.type = para_NumberedList; break;
	      case c_cfg: needkw = 8; par.type = para_Config;
		start_cmd = c_cfg; break;
	      case c_copyright: needkw = 32; par.type = para_Copyright; break;
	      case c_define: is_macro = TRUE; needkw = 1; break;
		/* For \nocite the keyword is _everything_ */
	      case c_nocite: needkw = 8; par.type = para_NoCite; break;
	      case c_preamble: needkw = 32; par.type = para_Normal; break;
	      case c_rule: needkw = 16; par.type = para_Rule; break;
	      case c_title: needkw = 32; par.type = para_Title; break;
	      case c_versionid: needkw = 32; par.type = para_VersionID; break;
	    }

	    if (par.type == para_Chapter ||
		par.type == para_Heading ||
		par.type == para_Subsect ||
		par.type == para_Appendix ||
		par.type == para_UnnumberedChapter) {
		struct crossparaitem *sitem = stk_top(crossparastk);
		if (sitem && (sitem->seen_lcont || sitem->seen_quote)) {
		    err_sectmarkerinblock(			  &t.pos,
			  (sitem->seen_lcont ? "lcont" : "quote"));
		}
	    }

	    if (needkw > 0) {
		rdstring rs = { 0, 0, NULL };
		rdstringc rsc = { 0, 0, NULL };
		int nkeys = 0;
		filepos fp;

		/* Get keywords. */
		dtor(t), t = get_token(in);
		fp = t.pos;
		while (t.type == tok_lbrace ||
		       (t.type == tok_white && (needkw & 24))) {
		    /*
		     * In paragraph types which can't accept any
		     * body text (such as \cfg), we are lenient
		     * about whitespace between keywords. This is
		     * important for \cfg in particular since it
		     * can often have many keywords which are long
		     * pieces of text, so it's useful to permit the
		     * user to wrap the line between them.
		     */
		    if (t.type == tok_white) {
			dtor(t), t = get_token(in); /* eat the space */
			continue;
		    }
		    /* This is a keyword. */
		    nkeys++;
		    /* FIXME: there will be bugs if anyone specifies an
		     * empty keyword (\foo{}), so trap this case. */
		    while (dtor(t), t = get_token(in),
			   t.type == tok_word || 
			   t.type == tok_white ||
			   (t.type == tok_cmd && t.cmd == c__nbsp) ||
			   (t.type == tok_cmd && t.cmd == c__escaped) ||
			   (t.type == tok_cmd && t.cmd == c_u)) {
			if (t.type == tok_white ||
			    (t.type == tok_cmd && t.cmd == c__nbsp)) {
			    rdadd(&rs, ' ');
			    rdaddc(&rsc, ' ');
			} else if (t.type == tok_cmd && t.cmd == c_u) {
			    rdadd(&rs, t.aux);
			    rdaddc(&rsc, '\\');
			    rdaddsc(&rsc, t.origtext);
			} else {
			    rdadds(&rs, t.text);
			    rdaddsc(&rsc, t.origtext);
			}
		    }
		    if (t.type != tok_rbrace) {
			err_kwunclosed(&t.pos);
			continue;
		    }
		    rdadd(&rs, 0);     /* add string terminator */
		    rdaddc(&rsc, 0);   /* add string terminator */
		    dtor(t), t = get_token(in); /* eat right brace */
		}

		rdadd(&rs, 0);	       /* add string terminator */
		rdaddc(&rsc, 0);       /* add string terminator */

		/* See whether we have the right number of keywords. */
		if ((needkw & 48) && nkeys > 0)
		    err_kwillegal(&fp);
		if ((needkw & 11) && nkeys == 0)
		    err_kwexpected(&fp);
		if ((needkw & 5) && nkeys > 1)
		    err_kwtoomany(&fp);

		if (is_macro) {
		    /*
		     * Macro definition. Get the rest of the line
		     * as a code-paragraph token, repeatedly until
		     * there's nothing more left of it. Separate
		     * with newlines.
		     */
		    rdstring macrotext = { 0, 0, NULL };
		    while (1) {
			dtor(t), t = get_codepar_token(in);
			if (macrotext.pos > 0)
			    rdadd(&macrotext, L'\n');
			rdadds(&macrotext, t.text);
			dtor(t), t = get_token(in);
			if (t.type == tok_eop || t.type == tok_eof)
                            break;
		    }
		    macrodef(macros, rs.text, macrotext.text, fp);
		    continue;	       /* next paragraph */
		}

		par.keyword = rdtrim(&rs);
		par.origkeyword = rdtrimc(&rsc);

		/* Move to EOP in case of needkw==8 or 16 (no body) */
		if (needkw & 24) {
		    /* We allow whitespace even when we expect no para body */
		    while (t.type == tok_white)
			dtor(t), t = get_token(in);
		    if (t.type != tok_eop && t.type != tok_eof &&
			(start_cmd == c__invalid ||
			 t.type != tok_cmd || t.cmd != start_cmd)) {
			err_bodyillegal(&t.pos);
			/* Error recovery: eat the rest of the paragraph */
			while (t.type != tok_eop && t.type != tok_eof &&
			       (start_cmd == c__invalid ||
				t.type != tok_cmd || t.cmd != start_cmd))
			    dtor(t), t = get_token(in);
		    }
		    if (t.type == tok_cmd)
			already = TRUE;/* inhibit get_token at top of loop */
		    prev_para_type = par.type;
		    addpara(par, ret);

		    if (par.type == para_Config) {
			input_configure(in, &par);
		    }
		    continue;	       /* next paragraph */
		}
	    }
	}		  

	/*
	 * Now read the actual paragraph, word by word, adding to
	 * the paragraph list.
	 *
	 * Mid-paragraph commands:
	 *
	 *  \K \k
	 *  \c \cw \cq
	 *  \e
	 *  \i \ii
	 *  \I
         *  \q
	 *  \u
	 *  \W
	 *  \date
	 *  \\ \{ \}
	 */
	parsestk = stk_new();
	style = word_Normal;
	spcstyle = word_WhiteSpace;
	indexing = FALSE;
	seenwhite = TRUE;
	while (t.type != tok_eop && t.type != tok_eof) {
	    iswhite = FALSE;
	    already = FALSE;

	    /* Handle implicit paragraph breaks after \IM, \BR etc */
	    if (start_cmd != c__invalid &&
		t.type == tok_cmd && t.cmd == start_cmd) {
		already = TRUE;	       /* inhibit get_token at top of loop */
		break;
	    }

	    if (t.type == tok_cmd && t.cmd == c__nop) {
		dtor(t), t = get_token(in);
		continue;	       /* do nothing! */
	    }

	    if (t.type == tok_cmd && t.cmd == c__escaped) {
		t.type = tok_word;     /* nice and simple */
		t.aux = 0;	       /* even if `\-' - nonbreaking! */
	    }
	    if (t.type == tok_cmd && t.cmd == c__nbsp) {
		t.type = tok_word;     /* nice and simple */
		sfree(t.text);
		t.text = ustrdup(L" ");  /* text is ` ' not `_' */
		t.aux = 0;	       /* (nonbreaking) */
	    }
	    switch (t.type) {
	      case tok_white:
		if (whptr == &par.words)
		    break;	       /* strip whitespace at start of para */
		wd.text = NULL;
		wd.type = spcstyle;
		wd.alt = NULL;
		wd.aux = 0;
		wd.fpos = t.pos;
		wd.breaks = FALSE;

		/*
		 * Inhibit use of whitespace if it's (probably the
		 * newline) before a repeat \IM / \BR type
		 * directive.
		 */
		if (start_cmd != c__invalid) {
		    dtor(t), t = get_token(in);
		    already = TRUE;
		    if (t.type == tok_cmd && t.cmd == start_cmd)
			break;
		}

		if (indexing)
		    rdadd(&indexstr, ' ');
		if (!indexing || index_visible)
		    addword(wd, &whptr);
		if (indexing)
		    addword(wd, &idximplicit);
		iswhite = TRUE;
		break;
	      case tok_word:
		if (indexing)
		    rdadds(&indexstr, t.text);
		wd.type = style;
		wd.alt = NULL;
		wd.aux = 0;
		wd.fpos = t.pos;
		wd.breaks = t.aux;
		if (!indexing || index_visible) {
		    wd.text = ustrdup(t.text);
		    addword(wd, &whptr);
		}
		if (indexing) {
		    wd.text = ustrdup(t.text);
		    addword(wd, &idximplicit);
		}
		break;
	      case tok_lbrace:
		err_unexbrace(&t.pos);
		/* Error recovery: push nop */
		sitem = snew(struct stack_item);
		sitem->type = stack_nop;
		sitem->fpos = t.pos;
		stk_push(parsestk, sitem);
		break;
	      case tok_rbrace:
		sitem = stk_pop(parsestk);
		if (!sitem) {
		    /*
		     * This closing brace could have been an
		     * indication that the cross-paragraph stack
		     * wants popping. Accordingly, we treat it here
		     * as an indication that the paragraph is over.
		     */
		    already = TRUE;
		    goto finished_para;
		} else {
		    if (sitem->type & stack_ualt) {
			whptr = sitem->whptr;
			idximplicit = sitem->idximplicit;
		    }
		    if (sitem->type & stack_style) {
			style = word_Normal;
			spcstyle = word_WhiteSpace;
		    }
		    if (sitem->type & stack_idx) {
			indexword->text = ustrdup(indexstr.text);
			if (index_downcase) {
			    word *w;

			    ustrlow(indexword->text);
			    ustrlow(indexstr.text);

			    for (w = idxwordlist; w; w = w->next)
				if (w->text)
				    ustrlow(w->text);
			}
			indexing = FALSE;
			rdadd(&indexstr, L'\0');
			index_merge(idx, FALSE, indexstr.text,
				    idxwordlist, &sitem->fpos);
			sfree(indexstr.text);
		    }
		    if (sitem->type & stack_hyper) {
			wd.text = NULL;
			wd.type = word_HyperEnd;
			wd.alt = NULL;
			wd.aux = 0;
			wd.fpos = t.pos;
			wd.breaks = FALSE;
			if (!indexing || index_visible)
			    addword(wd, &whptr);
			if (indexing)
			    addword(wd, &idximplicit);
		    }
		    if (sitem->type & stack_quote) {
			wd.text = NULL;
			wd.type = toquotestyle(style);
			wd.alt = NULL;
			wd.aux = quote_Close;
			wd.fpos = t.pos;
			wd.breaks = FALSE;
			if (!indexing || index_visible)
			    addword(wd, &whptr);
			if (indexing) {
			    rdadd(&indexstr, L'"');
			    addword(wd, &idximplicit);
			}
		    }
		}
		sfree(sitem);
		break;
	      case tok_cmd:
		switch (t.cmd) {
		  case c__comment:
		    /*
		     * In-paragraph comment: \#{ balanced braces }
		     *
		     * Anything goes here; even tok_eop. We should
		     * eat whitespace after the close brace _if_
		     * there was whitespace before the \#.
		     */
		    dtor(t), t = get_token(in);
		    if (t.type != tok_lbrace) {
			err_explbr(&t.pos);
		    } else {
			int braces = 1;
			while (braces > 0) {
			    dtor(t), t = get_token(in);
			    if (t.type == tok_lbrace)
				braces++;
			    else if (t.type == tok_rbrace)
				braces--;
			    else if (t.type == tok_eof) {
				err_commenteof(&t.pos);
				break;
			    }
			}
		    }
		    if (seenwhite) {
			already = TRUE;
			dtor(t), t = get_token(in);
			if (t.type == tok_white) {
			    iswhite = TRUE;
			    already = FALSE;
			}
		    }
		    break;
		  case c_q:
                  case c_cq:
                    type = t.cmd;
		    dtor(t), t = get_token(in);
		    if (t.type != tok_lbrace) {
			err_explbr(&t.pos);
		    } else {
			/*
			 * Enforce that \q may not be used anywhere
			 * within \c. (It shouldn't be necessary
			 * since the whole point of \c should be
			 * that the user wants to exercise exact
			 * control over the glyphs used, and
			 * forbidding it has the useful effect of
			 * relieving some backends of having to
			 * make difficult decisions.)
			 */
			int stype;

			if (style != word_Code && style != word_WeakCode) {
			    wd.text = NULL;
			    wd.type = toquotestyle(style);
			    wd.alt = NULL;
			    wd.aux = quote_Open;
			    wd.fpos = t.pos;
			    wd.breaks = FALSE;
			    if (!indexing || index_visible)
				addword(wd, &whptr);
			    if (indexing) {
				rdadd(&indexstr, L'"');
				addword(wd, &idximplicit);
			    }
			    stype = stack_quote;
			} else {
			    err_codequote(&t.pos);
			    stype = stack_nop;
			}
			sitem = snew(struct stack_item);
			sitem->fpos = t.pos;
			sitem->type = stype;
                        if (type == c_cq) {
                            if (style != word_Normal) {
                                err_nestedstyles(&t.pos);
                            } else {
                                style = word_WeakCode;
                                spcstyle = tospacestyle(style);
                                sitem->type |= stack_style;
                            }
                        }
			stk_push(parsestk, sitem);
		    }
		    break;
		  case c_K:
		  case c_k:
		  case c_W:
		  case c_date:
		    /*
		     * Keyword, hyperlink, or \date. We expect a
		     * left brace, some text, and then a right
		     * brace. No nesting; no arguments.
		     */
		    wd.fpos = t.pos;
		    wd.breaks = FALSE;
		    if (t.cmd == c_K)
			wd.type = word_UpperXref;
		    else if (t.cmd == c_k)
			wd.type = word_LowerXref;
		    else if (t.cmd == c_W)
			wd.type = word_HyperLink;
		    else
			wd.type = word_Normal;
		    dtor(t), t = get_token(in);
		    if (t.type != tok_lbrace) {
			if (wd.type == word_Normal) {
			    time_t thetime = current_time();
			    struct tm *broken = localtime(&thetime);
			    already = TRUE;
			    wdtext = ustrftime(NULL, broken);
			    wd.type = style;
			} else {
			    err_explbr(&t.pos);
			    wdtext = NULL;
			}
		    } else {
			rdstring rs = { 0, 0, NULL };
			while (dtor(t), t = get_token(in),
			       t.type == tok_word || t.type == tok_white) {
			    if (t.type == tok_white)
				rdadd(&rs, ' ');
			    else
				rdadds(&rs, t.text);
			}
			if (wd.type == word_Normal) {
			    time_t thetime = current_time();
			    struct tm *broken = localtime(&thetime);
			    wdtext = ustrftime(rs.text, broken);
			    wd.type = style;
			} else {
			    wdtext = ustrdup(rs.text);
			}
			sfree(rs.text);
			if (t.type != tok_rbrace) {
			    err_kwexprbr(&t.pos);
			}
		    }
		    wd.alt = NULL;
		    wd.aux = 0;
		    if (!indexing || index_visible) {
			wd.text = ustrdup(wdtext);
			addword(wd, &whptr);
		    }
		    if (indexing) {
			wd.text = ustrdup(wdtext);
			addword(wd, &idximplicit);
		    }
		    sfree(wdtext);
		    if (wd.type == word_HyperLink) {
			/*
			 * Hyperlinks are different: they then
			 * expect another left brace, to begin
			 * delimiting the text marked by the link.
			 */
			dtor(t), t = get_token(in);
			sitem = snew(struct stack_item);
			sitem->fpos = wd.fpos;
			sitem->type = stack_hyper;
			/*
			 * Special cases: \W{}\i, \W{}\ii
			 */
			if (t.type == tok_cmd &&
			    (t.cmd == c_i || t.cmd == c_ii)) {
			    if (indexing) {
				err_nestedindex(&t.pos);
			    } else {
				/* Add an index-reference word with no
				 * text as yet */
				wd.type = word_IndexRef;
				wd.text = NULL;
				wd.alt = NULL;
				wd.aux = 0;
				wd.breaks = FALSE;
				indexword = addword(wd, &whptr);
				/* Set up a rdstring to read the
				 * index text */
				indexstr = nullrs;
				/* Flags so that we do the Right
				 * Things with text */
				index_visible = (type != c_I);
				index_downcase = (type == c_ii);
				indexing = TRUE;
				idxwordlist = NULL;
				idximplicit = &idxwordlist;

				sitem->type |= stack_idx;
			    }
			    dtor(t), t = get_token(in);
			}
			/*
			 * Special cases: \W{}\c, \W{}\e, \W{}\s, \W{}\cw
			 */
			if (t.type == tok_cmd &&
			    (t.cmd == c_e || t.cmd == c_s ||
                             t.cmd == c_c || t.cmd == c_cw)) {
			    if (style != word_Normal)
				err_nestedstyles(&t.pos);
			    else {
				style = (t.cmd == c_c ? word_Code :
					 t.cmd == c_cw ? word_WeakCode :
					 t.cmd == c_s ? word_Strong :
					 word_Emph);
				spcstyle = tospacestyle(style);
				sitem->type |= stack_style;
			    }
			    dtor(t), t = get_token(in);
			}
			if (t.type != tok_lbrace) {
			    err_explbr(&t.pos);
			    sfree(sitem);
			} else {
			    stk_push(parsestk, sitem);
			}
		    }
		    break;
		  case c_c:
		  case c_cw:
		  case c_e:
		  case c_s:
		    type = t.cmd;
		    if (style != word_Normal) {
			err_nestedstyles(&t.pos);
			/* Error recovery: eat lbrace, push nop. */
			dtor(t), t = get_token(in);
			sitem = snew(struct stack_item);
			sitem->fpos = t.pos;
			sitem->type = stack_nop;
			stk_push(parsestk, sitem);
		    }
		    dtor(t), t = get_token(in);
		    if (t.type != tok_lbrace) {
			err_explbr(&t.pos);
		    } else {
			style = (type == c_c ? word_Code :
				 type == c_cw ? word_WeakCode :
				 type == c_s ? word_Strong :
				 word_Emph);
			spcstyle = tospacestyle(style);
			sitem = snew(struct stack_item);
			sitem->fpos = t.pos;
			sitem->type = stack_style;
			stk_push(parsestk, sitem);
		    }
		    break;
		  case c_i:
		  case c_ii:
		  case c_I:
		    type = t.cmd;
		    if (indexing) {
			err_nestedindex(&t.pos);
			/* Error recovery: eat lbrace, push nop. */
			dtor(t), t = get_token(in);
			sitem = snew(struct stack_item);
			sitem->fpos = t.pos;
			sitem->type = stack_nop;
			stk_push(parsestk, sitem);
		    }
		    sitem = snew(struct stack_item);
		    sitem->fpos = t.pos;
		    sitem->type = stack_idx;
		    dtor(t), t = get_token(in);
		    /*
		     * Special cases: \i\c, \i\e, \i\s, \i\cw
		     */
		    wd.fpos = t.pos;
		    if (t.type == tok_cmd &&
			(t.cmd == c_e || t.cmd == c_s ||
                         t.cmd == c_c || t.cmd == c_cw)) {
			if (style != word_Normal)
			    err_nestedstyles(&t.pos);
			else {
			    style = (t.cmd == c_c ? word_Code :
				     t.cmd == c_cw ? word_WeakCode :
				     t.cmd == c_s ? word_Strong :
				     word_Emph);
			    spcstyle = tospacestyle(style);
			    sitem->type |= stack_style;
			}
			dtor(t), t = get_token(in);
		    }
		    if (t.type != tok_lbrace) {
			sfree(sitem);
			err_explbr(&t.pos);
		    } else {
			/* Add an index-reference word with no text as yet */
			wd.type = word_IndexRef;
			wd.text = NULL;
			wd.alt = NULL;
			wd.aux = 0;
			wd.breaks = FALSE;
			indexword = addword(wd, &whptr);
			/* Set up a rdstring to read the index text */
			indexstr = nullrs;
			/* Flags so that we do the Right Things with text */
			index_visible = (type != c_I);
			index_downcase = (type == c_ii);
			indexing = TRUE;
			idxwordlist = NULL;
			idximplicit = &idxwordlist;
			/* Stack item to close the indexing on exit */
			stk_push(parsestk, sitem);
		    }
		    break;
		  case c_u:
		    uchr = t.aux;
		    utext[0] = uchr; utext[1] = 0;
		    wd.type = style;
		    wd.breaks = FALSE;
		    wd.alt = NULL;
		    wd.aux = 0;
		    wd.fpos = t.pos;
		    if (!indexing || index_visible) {
			wd.text = ustrdup(utext);
			uword = addword(wd, &whptr);
		    } else
			uword = NULL;
		    if (indexing) {
			wd.text = ustrdup(utext);
			iword = addword(wd, &idximplicit);
		    } else
			iword = NULL;
		    dtor(t), t = get_token(in);
		    if (t.type == tok_lbrace) {
			/*
			 * \u with a left brace. Until the brace
			 * closes, all further words go on a
			 * sidetrack from the main thread of the
			 * paragraph.
			 */
			sitem = snew(struct stack_item);
			sitem->fpos = t.pos;
			sitem->type = stack_ualt;
			sitem->whptr = whptr;
			sitem->idximplicit = idximplicit;
			stk_push(parsestk, sitem);
			whptr = uword ? &uword->alt : NULL;
			idximplicit = iword ? &iword->alt : NULL;
		    } else {
			if (indexing)
			    rdadd(&indexstr, uchr);
			already = TRUE;
		    }
		    break;
		  default:
		    if (!macrolookup(macros, in, t.text, &t.pos))
			err_badmidcmd(t.text, &t.pos);
		    break;
		}
	    }
	    if (!already)
		dtor(t), t = get_token(in);
	    seenwhite = iswhite;
	}
	finished_para:
	/* Check the stack is empty */
	if (stk_top(parsestk)) {
	    while ((sitem = stk_pop(parsestk)))
		sfree(sitem);
	    err_missingrbrace(&t.pos);
	}
	stk_free(parsestk);
	prev_para_type = par.type;
	/*
	 * Before we add the paragraph to the output list, we
	 * should check that there was any text in it at all; there
	 * might not be if (for example) the paragraph contained
	 * nothing but an unrecognised command sequence, and if we
	 * put an empty paragraph on the list it may confuse the
	 * back ends later on.
	 */
	if (par.words) {
	    addpara(par, ret);
	}
	if (t.type == tok_eof)
	    already = TRUE;
    }

    if (stk_top(crossparastk)) {
	void *p;

	err_missingrbrace2(&t.pos);
	while ((p = stk_pop(crossparastk)))
	    sfree(p);
    }

    /*
     * We break to here rather than returning, because otherwise
     * this cleanup doesn't happen.
     */
    dtor(t);

    stk_free(crossparastk);
}

struct {
    char const *magic;
    size_t nmagic;
    int binary;
    void (*reader)(input *);
} magics[] = {
    { "%!FontType1-",     12, FALSE, &read_pfa_file },
    { "%!PS-AdobeFont-",  15, FALSE, &read_pfa_file },
    { "\x80\x01",          2, TRUE,  &read_pfb_file },
    { "StartFontMetrics", 16, FALSE, &read_afm_file },
    { "\x00\x01\x00\x00",  4, TRUE,  &read_sfnt_file },
    { "true",		   4, TRUE,  &read_sfnt_file },
};

paragraph *read_input(input *in, indexdata *idx) {
    paragraph *head = NULL;
    paragraph **hptr = &head;
    tree234 *macros;
    char mag[16];
    size_t len, i;
    int binary;
    void (*reader)(input *);

    macros = newtree234(macrocmp);

    while (in->currindex < in->nfiles) {
	setpos(in, in->filenames[in->currindex]);
	in->charset = in->defcharset;
	in->csstate = charset_init_state;
	in->wcpos = in->nwc = 0;
	in->pushback_chars = NULL;

	if (!in->filenames[in->currindex]) {
	    in->currfp = stdin;
	    in->wantclose = FALSE;     /* don't fclose stdin */
	    /*
	     * When reading standard input, we always expect to see
	     * an actual Halibut file and not any of the unusual
	     * input types like fonts.
	     */
	    reader = NULL;
	} else {
	    /*
	     * Open the file in binary mode to look for magic
	     * numbers. We'll switch to text mode if we find we're
	     * looking at a text file type.
	     */
	    in->currfp = fopen(in->filenames[in->currindex], "rb");
	    binary = FALSE; /* default to Halibut source, which is text */
	    if (in->currfp) {
		in->wantclose = TRUE;
		reader = NULL;
		len = fread(mag, 1, sizeof(mag), in->currfp);
		for (i = 0; i < lenof(magics); i++) {
		    if (len >= magics[i].nmagic &&
			memcmp(mag, magics[i].magic, magics[i].nmagic) == 0) {
			reader = magics[i].reader;
			binary = magics[i].binary;
			break;
		    }
		}
		rewind(in->currfp);
	    }
	    if (!binary) {
                if (in->currfp)
                    fclose(in->currfp);
		in->currfp = fopen(in->filenames[in->currindex], "r");
	    }
	}
	if (in->currfp) {
	    if (reader == NULL) {
		read_file(&hptr, in, idx, macros);
	    } else {
		(*reader)(in);
	    }
	} else {
            err_cantopen(in->filenames[in->currindex]);
        }
	in->currindex++;
    }

    macrocleanup(macros);

    return head;
}