File: app.c

package info (click to toggle)
binutils 2.31.1-16
  • links: PTS, VCS
  • area: main
  • in suites: buster
  • size: 309,412 kB
  • sloc: ansic: 1,161,194; asm: 638,508; cpp: 128,829; exp: 68,580; makefile: 55,828; sh: 22,360; yacc: 14,238; lisp: 13,272; perl: 2,111; ada: 1,681; lex: 1,652; pascal: 1,446; cs: 879; sed: 195; python: 154; xml: 95; awk: 25
file content (1499 lines) | stat: -rw-r--r-- 36,186 bytes parent folder | download | duplicates (5)
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
/* This is the Assembler Pre-Processor
   Copyright (C) 1987-2018 Free Software Foundation, Inc.

   This file is part of GAS, the GNU Assembler.

   GAS 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 3, or (at your option)
   any later version.

   GAS 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 GAS; see the file COPYING.  If not, write to the Free
   Software Foundation, 51 Franklin Street - Fifth Floor, Boston, MA
   02110-1301, USA.  */

/* Modified by Allen Wirfs-Brock, Instantiations Inc 2/90.  */
/* App, the assembler pre-processor.  This pre-processor strips out
   excess spaces, turns single-quoted characters into a decimal
   constant, and turns the # in # <number> <filename> <garbage> into a
   .linefile.  This needs better error-handling.  */

#include "as.h"

#if (__STDC__ != 1)
#ifndef const
#define const  /* empty */
#endif
#endif

#ifdef H_TICK_HEX
int enable_h_tick_hex = 0;
#endif

#ifdef TC_M68K
/* Whether we are scrubbing in m68k MRI mode.  This is different from
   flag_m68k_mri, because the two flags will be affected by the .mri
   pseudo-op at different times.  */
static int scrub_m68k_mri;

/* The pseudo-op which switches in and out of MRI mode.  See the
   comment in do_scrub_chars.  */
static const char mri_pseudo[] = ".mri 0";
#else
#define scrub_m68k_mri 0
#endif

#if defined TC_ARM && defined OBJ_ELF
/* The pseudo-op for which we need to special-case `@' characters.
   See the comment in do_scrub_chars.  */
static const char   symver_pseudo[] = ".symver";
static const char * symver_state;
#endif
#ifdef TC_ARM
static char last_char;
#endif

static char lex[256];
static const char symbol_chars[] =
"$._ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789";

#define LEX_IS_SYMBOL_COMPONENT		1
#define LEX_IS_WHITESPACE		2
#define LEX_IS_LINE_SEPARATOR		3
#define LEX_IS_COMMENT_START		4
#define LEX_IS_LINE_COMMENT_START	5
#define	LEX_IS_TWOCHAR_COMMENT_1ST	6
#define	LEX_IS_STRINGQUOTE		8
#define	LEX_IS_COLON			9
#define	LEX_IS_NEWLINE			10
#define	LEX_IS_ONECHAR_QUOTE		11
#ifdef TC_V850
#define LEX_IS_DOUBLEDASH_1ST		12
#endif
#ifdef TC_M32R
#define DOUBLEBAR_PARALLEL
#endif
#ifdef DOUBLEBAR_PARALLEL
#define LEX_IS_DOUBLEBAR_1ST		13
#endif
#define LEX_IS_PARALLEL_SEPARATOR	14
#ifdef H_TICK_HEX
#define LEX_IS_H			15
#endif
#define IS_SYMBOL_COMPONENT(c)		(lex[c] == LEX_IS_SYMBOL_COMPONENT)
#define IS_WHITESPACE(c)		(lex[c] == LEX_IS_WHITESPACE)
#define IS_LINE_SEPARATOR(c)		(lex[c] == LEX_IS_LINE_SEPARATOR)
#define IS_PARALLEL_SEPARATOR(c)	(lex[c] == LEX_IS_PARALLEL_SEPARATOR)
#define IS_COMMENT(c)			(lex[c] == LEX_IS_COMMENT_START)
#define IS_LINE_COMMENT(c)		(lex[c] == LEX_IS_LINE_COMMENT_START)
#define	IS_NEWLINE(c)			(lex[c] == LEX_IS_NEWLINE)

static int process_escape (int);

/* FIXME-soon: The entire lexer/parser thingy should be
   built statically at compile time rather than dynamically
   each and every time the assembler is run.  xoxorich.  */

void
do_scrub_begin (int m68k_mri ATTRIBUTE_UNUSED)
{
  const char *p;
  int c;

  lex[' '] = LEX_IS_WHITESPACE;
  lex['\t'] = LEX_IS_WHITESPACE;
  lex['\r'] = LEX_IS_WHITESPACE;
  lex['\n'] = LEX_IS_NEWLINE;
  lex[':'] = LEX_IS_COLON;

#ifdef TC_M68K
  scrub_m68k_mri = m68k_mri;

  if (! m68k_mri)
#endif
    {
      lex['"'] = LEX_IS_STRINGQUOTE;

#if ! defined (TC_HPPA)
      lex['\''] = LEX_IS_ONECHAR_QUOTE;
#endif

#ifdef SINGLE_QUOTE_STRINGS
      lex['\''] = LEX_IS_STRINGQUOTE;
#endif
    }

  /* Note: if any other character can be LEX_IS_STRINGQUOTE, the loop
     in state 5 of do_scrub_chars must be changed.  */

  /* Note that these override the previous defaults, e.g. if ';' is a
     comment char, then it isn't a line separator.  */
  for (p = symbol_chars; *p; ++p)
    lex[(unsigned char) *p] = LEX_IS_SYMBOL_COMPONENT;

  for (c = 128; c < 256; ++c)
    lex[c] = LEX_IS_SYMBOL_COMPONENT;

#ifdef tc_symbol_chars
  /* This macro permits the processor to specify all characters which
     may appears in an operand.  This will prevent the scrubber from
     discarding meaningful whitespace in certain cases.  The i386
     backend uses this to support prefixes, which can confuse the
     scrubber as to whether it is parsing operands or opcodes.  */
  for (p = tc_symbol_chars; *p; ++p)
    lex[(unsigned char) *p] = LEX_IS_SYMBOL_COMPONENT;
#endif

  /* The m68k backend wants to be able to change comment_chars.  */
#ifndef tc_comment_chars
#define tc_comment_chars comment_chars
#endif
  for (p = tc_comment_chars; *p; p++)
    lex[(unsigned char) *p] = LEX_IS_COMMENT_START;

  for (p = line_comment_chars; *p; p++)
    lex[(unsigned char) *p] = LEX_IS_LINE_COMMENT_START;

#ifndef tc_line_separator_chars
#define tc_line_separator_chars line_separator_chars
#endif
  for (p = tc_line_separator_chars; *p; p++)
    lex[(unsigned char) *p] = LEX_IS_LINE_SEPARATOR;

#ifdef tc_parallel_separator_chars
  /* This macro permits the processor to specify all characters which
     separate parallel insns on the same line.  */
  for (p = tc_parallel_separator_chars; *p; p++)
    lex[(unsigned char) *p] = LEX_IS_PARALLEL_SEPARATOR;
#endif

  /* Only allow slash-star comments if slash is not in use.
     FIXME: This isn't right.  We should always permit them.  */
  if (lex['/'] == 0)
    lex['/'] = LEX_IS_TWOCHAR_COMMENT_1ST;

#ifdef TC_M68K
  if (m68k_mri)
    {
      lex['\''] = LEX_IS_STRINGQUOTE;
      lex[';'] = LEX_IS_COMMENT_START;
      lex['*'] = LEX_IS_LINE_COMMENT_START;
      /* The MRI documentation says '!' is LEX_IS_COMMENT_START, but
	 then it can't be used in an expression.  */
      lex['!'] = LEX_IS_LINE_COMMENT_START;
    }
#endif

#ifdef TC_V850
  lex['-'] = LEX_IS_DOUBLEDASH_1ST;
#endif
#ifdef DOUBLEBAR_PARALLEL
  lex['|'] = LEX_IS_DOUBLEBAR_1ST;
#endif
#ifdef TC_D30V
  /* Must do this is we want VLIW instruction with "->" or "<-".  */
  lex['-'] = LEX_IS_SYMBOL_COMPONENT;
#endif

#ifdef H_TICK_HEX
  if (enable_h_tick_hex)
    {
      lex['h'] = LEX_IS_H;
      lex['H'] = LEX_IS_H;
    }
#endif
}

/* Saved state of the scrubber.  */
static int state;
static int old_state;
static const char *out_string;
static char out_buf[20];
static int add_newlines;
static char *saved_input;
static size_t saved_input_len;
static char input_buffer[32 * 1024];
static const char *mri_state;
static char mri_last_ch;

/* Data structure for saving the state of app across #include's.  Note that
   app is called asynchronously to the parsing of the .include's, so our
   state at the time .include is interpreted is completely unrelated.
   That's why we have to save it all.  */

struct app_save
{
  int          state;
  int          old_state;
  const char * out_string;
  char         out_buf[sizeof (out_buf)];
  int          add_newlines;
  char *       saved_input;
  size_t       saved_input_len;
#ifdef TC_M68K
  int          scrub_m68k_mri;
#endif
  const char * mri_state;
  char         mri_last_ch;
#if defined TC_ARM && defined OBJ_ELF
  const char * symver_state;
#endif
#ifdef TC_ARM
  char last_char;
#endif
};

char *
app_push (void)
{
  struct app_save *saved;

  saved = XNEW (struct app_save);
  saved->state = state;
  saved->old_state = old_state;
  saved->out_string = out_string;
  memcpy (saved->out_buf, out_buf, sizeof (out_buf));
  saved->add_newlines = add_newlines;
  if (saved_input == NULL)
    saved->saved_input = NULL;
  else
    {
      saved->saved_input = XNEWVEC (char, saved_input_len);
      memcpy (saved->saved_input, saved_input, saved_input_len);
      saved->saved_input_len = saved_input_len;
    }
#ifdef TC_M68K
  saved->scrub_m68k_mri = scrub_m68k_mri;
#endif
  saved->mri_state = mri_state;
  saved->mri_last_ch = mri_last_ch;
#if defined TC_ARM && defined OBJ_ELF
  saved->symver_state = symver_state;
#endif
#ifdef TC_ARM
  saved->last_char = last_char;
#endif

  /* do_scrub_begin() is not useful, just wastes time.  */

  state = 0;
  saved_input = NULL;
  add_newlines = 0;

  return (char *) saved;
}

void
app_pop (char *arg)
{
  struct app_save *saved = (struct app_save *) arg;

  /* There is no do_scrub_end ().  */
  state = saved->state;
  old_state = saved->old_state;
  out_string = saved->out_string;
  memcpy (out_buf, saved->out_buf, sizeof (out_buf));
  add_newlines = saved->add_newlines;
  if (saved->saved_input == NULL)
    saved_input = NULL;
  else
    {
      gas_assert (saved->saved_input_len <= sizeof (input_buffer));
      memcpy (input_buffer, saved->saved_input, saved->saved_input_len);
      saved_input = input_buffer;
      saved_input_len = saved->saved_input_len;
      free (saved->saved_input);
    }
#ifdef TC_M68K
  scrub_m68k_mri = saved->scrub_m68k_mri;
#endif
  mri_state = saved->mri_state;
  mri_last_ch = saved->mri_last_ch;
#if defined TC_ARM && defined OBJ_ELF
  symver_state = saved->symver_state;
#endif
#ifdef TC_ARM
  last_char = saved->last_char;
#endif

  free (arg);
}

/* @@ This assumes that \n &c are the same on host and target.  This is not
   necessarily true.  */

static int
process_escape (int ch)
{
  switch (ch)
    {
    case 'b':
      return '\b';
    case 'f':
      return '\f';
    case 'n':
      return '\n';
    case 'r':
      return '\r';
    case 't':
      return '\t';
    case '\'':
      return '\'';
    case '"':
      return '\"';
    default:
      return ch;
    }
}

/* This function is called to process input characters.  The GET
   parameter is used to retrieve more input characters.  GET should
   set its parameter to point to a buffer, and return the length of
   the buffer; it should return 0 at end of file.  The scrubbed output
   characters are put into the buffer starting at TOSTART; the TOSTART
   buffer is TOLEN bytes in length.  The function returns the number
   of scrubbed characters put into TOSTART.  This will be TOLEN unless
   end of file was seen.  This function is arranged as a state
   machine, and saves its state so that it may return at any point.
   This is the way the old code used to work.  */

size_t
do_scrub_chars (size_t (*get) (char *, size_t), char *tostart, size_t tolen)
{
  char *to = tostart;
  char *toend = tostart + tolen;
  char *from;
  char *fromend;
  size_t fromlen;
  int ch, ch2 = 0;
  /* Character that started the string we're working on.  */
  static char quotechar;

  /*State 0: beginning of normal line
	  1: After first whitespace on line (flush more white)
	  2: After first non-white (opcode) on line (keep 1white)
	  3: after second white on line (into operands) (flush white)
	  4: after putting out a .linefile, put out digits
	  5: parsing a string, then go to old-state
	  6: putting out \ escape in a "d string.
	  7: no longer used
	  8: no longer used
	  9: After seeing symbol char in state 3 (keep 1white after symchar)
	 10: After seeing whitespace in state 9 (keep white before symchar)
	 11: After seeing a symbol character in state 0 (eg a label definition)
	 -1: output string in out_string and go to the state in old_state
	 -2: flush text until a '*' '/' is seen, then go to state old_state
#ifdef TC_V850
	 12: After seeing a dash, looking for a second dash as a start
	     of comment.
#endif
#ifdef DOUBLEBAR_PARALLEL
	 13: After seeing a vertical bar, looking for a second
	     vertical bar as a parallel expression separator.
#endif
#ifdef TC_PREDICATE_START_CHAR
	 14: After seeing a predicate start character at state 0, looking
	     for a predicate end character as predicate.
	 15: After seeing a predicate start character at state 1, looking
	     for a predicate end character as predicate.
#endif
#ifdef TC_Z80
	 16: After seeing an 'a' or an 'A' at the start of a symbol
	 17: After seeing an 'f' or an 'F' in state 16
#endif
	  */

  /* I added states 9 and 10 because the MIPS ECOFF assembler uses
     constructs like ``.loc 1 20''.  This was turning into ``.loc
     120''.  States 9 and 10 ensure that a space is never dropped in
     between characters which could appear in an identifier.  Ian
     Taylor, ian@cygnus.com.

     I added state 11 so that something like "Lfoo add %r25,%r26,%r27" works
     correctly on the PA (and any other target where colons are optional).
     Jeff Law, law@cs.utah.edu.

     I added state 13 so that something like "cmp r1, r2 || trap #1" does not
     get squashed into "cmp r1,r2||trap#1", with the all important space
     between the 'trap' and the '#1' being eliminated.  nickc@cygnus.com  */

  /* This macro gets the next input character.  */

#define GET()							\
  (from < fromend						\
   ? * (unsigned char *) (from++)				\
   : (saved_input = NULL,					\
      fromlen = (*get) (input_buffer, sizeof input_buffer),	\
      from = input_buffer,					\
      fromend = from + fromlen,					\
      (fromlen == 0						\
       ? EOF							\
       : * (unsigned char *) (from++))))

  /* This macro pushes a character back on the input stream.  */

#define UNGET(uch) (*--from = (uch))

  /* This macro puts a character into the output buffer.  If this
     character fills the output buffer, this macro jumps to the label
     TOFULL.  We use this rather ugly approach because we need to
     handle two different termination conditions: EOF on the input
     stream, and a full output buffer.  It would be simpler if we
     always read in the entire input stream before processing it, but
     I don't want to make such a significant change to the assembler's
     memory usage.  */

#define PUT(pch)				\
  do						\
    {						\
      *to++ = (pch);				\
      if (to >= toend)				\
	goto tofull;				\
    }						\
  while (0)

  if (saved_input != NULL)
    {
      from = saved_input;
      fromend = from + saved_input_len;
    }
  else
    {
      fromlen = (*get) (input_buffer, sizeof input_buffer);
      if (fromlen == 0)
	return 0;
      from = input_buffer;
      fromend = from + fromlen;
    }

  while (1)
    {
      /* The cases in this switch end with continue, in order to
	 branch back to the top of this while loop and generate the
	 next output character in the appropriate state.  */
      switch (state)
	{
	case -1:
	  ch = *out_string++;
	  if (*out_string == '\0')
	    {
	      state = old_state;
	      old_state = 3;
	    }
	  PUT (ch);
	  continue;

	case -2:
	  for (;;)
	    {
	      do
		{
		  ch = GET ();

		  if (ch == EOF)
		    {
		      as_warn (_("end of file in comment"));
		      goto fromeof;
		    }

		  if (ch == '\n')
		    PUT ('\n');
		}
	      while (ch != '*');

	      while ((ch = GET ()) == '*')
		;

	      if (ch == EOF)
		{
		  as_warn (_("end of file in comment"));
		  goto fromeof;
		}

	      if (ch == '/')
		break;

	      UNGET (ch);
	    }

	  state = old_state;
	  UNGET (' ');
	  continue;

	case 4:
	  ch = GET ();
	  if (ch == EOF)
	    goto fromeof;
	  else if (ch >= '0' && ch <= '9')
	    PUT (ch);
	  else
	    {
	      while (ch != EOF && IS_WHITESPACE (ch))
		ch = GET ();
	      if (ch == '"')
		{
		  quotechar = ch;
		  state = 5;
		  old_state = 3;
		  PUT (ch);
		}
	      else
		{
		  while (ch != EOF && ch != '\n')
		    ch = GET ();
		  state = 0;
		  PUT (ch);
		}
	    }
	  continue;

	case 5:
	  /* We are going to copy everything up to a quote character,
	     with special handling for a backslash.  We try to
	     optimize the copying in the simple case without using the
	     GET and PUT macros.  */
	  {
	    char *s;
	    ptrdiff_t len;

	    for (s = from; s < fromend; s++)
	      {
		ch = *s;
		if (ch == '\\'
		    || ch == quotechar
		    || ch == '\n')
		  break;
	      }
	    len = s - from;
	    if (len > toend - to)
	      len = toend - to;
	    if (len > 0)
	      {
		memcpy (to, from, len);
		to += len;
		from += len;
		if (to >= toend)
		  goto tofull;
	      }
	  }

	  ch = GET ();
	  if (ch == EOF)
	    {
	      /* This buffer is here specifically so
		 that the UNGET below will work.  */
	      static char one_char_buf[1];

	      as_warn (_("end of file in string; '%c' inserted"), quotechar);
	      state = old_state;
	      from = fromend = one_char_buf + 1;
	      fromlen = 1;
	      UNGET ('\n');
	      PUT (quotechar);
	    }
	  else if (ch == quotechar)
	    {
	      state = old_state;
	      PUT (ch);
	    }
#ifndef NO_STRING_ESCAPES
	  else if (ch == '\\')
	    {
	      state = 6;
	      PUT (ch);
	    }
#endif
	  else if (scrub_m68k_mri && ch == '\n')
	    {
	      /* Just quietly terminate the string.  This permits lines like
		   bne	label	loop if we haven't reach end yet.  */
	      state = old_state;
	      UNGET (ch);
	      PUT ('\'');
	    }
	  else
	    {
	      PUT (ch);
	    }
	  continue;

	case 6:
	  state = 5;
	  ch = GET ();
	  switch (ch)
	    {
	      /* Handle strings broken across lines, by turning '\n' into
		 '\\' and 'n'.  */
	    case '\n':
	      UNGET ('n');
	      add_newlines++;
	      PUT ('\\');
	      continue;

	    case EOF:
	      as_warn (_("end of file in string; '%c' inserted"), quotechar);
	      PUT (quotechar);
	      continue;

	    case '"':
	    case '\\':
	    case 'b':
	    case 'f':
	    case 'n':
	    case 'r':
	    case 't':
	    case 'v':
	    case 'x':
	    case 'X':
	    case '0':
	    case '1':
	    case '2':
	    case '3':
	    case '4':
	    case '5':
	    case '6':
	    case '7':
	      break;

	    default:
#ifdef ONLY_STANDARD_ESCAPES
	      as_warn (_("unknown escape '\\%c' in string; ignored"), ch);
#endif
	      break;
	    }
	  PUT (ch);
	  continue;

#ifdef DOUBLEBAR_PARALLEL
	case 13:
	  ch = GET ();
	  if (ch != '|')
	    abort ();

	  /* Reset back to state 1 and pretend that we are parsing a
	     line from just after the first white space.  */
	  state = 1;
	  PUT ('|');
#ifdef TC_TIC6X
	  /* "||^" is used for SPMASKed instructions.  */
	  ch = GET ();
	  if (ch == EOF)
	    goto fromeof;
	  else if (ch == '^')
	    PUT ('^');
	  else
	    UNGET (ch);
#endif
	  continue;
#endif
#ifdef TC_Z80
	case 16:
	  /* We have seen an 'a' at the start of a symbol, look for an 'f'.  */
	  ch = GET ();
	  if (ch == 'f' || ch == 'F')
	    {
	      state = 17;
	      PUT (ch);
	    }
	  else
	    {
	      state = 9;
	      break;
	    }
	  /* Fall through.  */
	case 17:
	  /* We have seen "af" at the start of a symbol,
	     a ' here is a part of that symbol.  */
	  ch = GET ();
	  state = 9;
	  if (ch == '\'')
	    /* Change to avoid warning about unclosed string.  */
	    PUT ('`');
	  else if (ch != EOF)
	    UNGET (ch);
	  break;
#endif
	}

      /* OK, we are somewhere in states 0 through 4 or 9 through 11.  */

      /* flushchar: */
      ch = GET ();

#ifdef TC_PREDICATE_START_CHAR
      if (ch == TC_PREDICATE_START_CHAR && (state == 0 || state == 1))
	{
	  state += 14;
	  PUT (ch);
	  continue;
	}
      else if (state == 14 || state == 15)
	{
	  if (ch == TC_PREDICATE_END_CHAR)
	    {
	      state -= 14;
	      PUT (ch);
	      ch = GET ();
	    }
	  else
	    {
	      PUT (ch);
	      continue;
	    }
	}
#endif

    recycle:

#if defined TC_ARM && defined OBJ_ELF
      /* We need to watch out for .symver directives.  See the comment later
	 in this function.  */
      if (symver_state == NULL)
	{
	  if ((state == 0 || state == 1) && ch == symver_pseudo[0])
	    symver_state = symver_pseudo + 1;
	}
      else
	{
	  /* We advance to the next state if we find the right
	     character.  */
	  if (ch != '\0' && (*symver_state == ch))
	    ++symver_state;
	  else if (*symver_state != '\0')
	    /* We did not get the expected character, or we didn't
	       get a valid terminating character after seeing the
	       entire pseudo-op, so we must go back to the beginning.  */
	    symver_state = NULL;
	  else
	    {
	      /* We've read the entire pseudo-op.  If this is the end
		 of the line, go back to the beginning.  */
	      if (IS_NEWLINE (ch))
		symver_state = NULL;
	    }
	}
#endif /* TC_ARM && OBJ_ELF */

#ifdef TC_M68K
      /* We want to have pseudo-ops which control whether we are in
	 MRI mode or not.  Unfortunately, since m68k MRI mode affects
	 the scrubber, that means that we need a special purpose
	 recognizer here.  */
      if (mri_state == NULL)
	{
	  if ((state == 0 || state == 1)
	      && ch == mri_pseudo[0])
	    mri_state = mri_pseudo + 1;
	}
      else
	{
	  /* We advance to the next state if we find the right
	     character, or if we need a space character and we get any
	     whitespace character, or if we need a '0' and we get a
	     '1' (this is so that we only need one state to handle
	     ``.mri 0'' and ``.mri 1'').  */
	  if (ch != '\0'
	      && (*mri_state == ch
		  || (*mri_state == ' '
		      && lex[ch] == LEX_IS_WHITESPACE)
		  || (*mri_state == '0'
		      && ch == '1')))
	    {
	      mri_last_ch = ch;
	      ++mri_state;
	    }
	  else if (*mri_state != '\0'
		   || (lex[ch] != LEX_IS_WHITESPACE
		       && lex[ch] != LEX_IS_NEWLINE))
	    {
	      /* We did not get the expected character, or we didn't
		 get a valid terminating character after seeing the
		 entire pseudo-op, so we must go back to the
		 beginning.  */
	      mri_state = NULL;
	    }
	  else
	    {
	      /* We've read the entire pseudo-op.  mips_last_ch is
		 either '0' or '1' indicating whether to enter or
		 leave MRI mode.  */
	      do_scrub_begin (mri_last_ch == '1');
	      mri_state = NULL;

	      /* We continue handling the character as usual.  The
		 main gas reader must also handle the .mri pseudo-op
		 to control expression parsing and the like.  */
	    }
	}
#endif

      if (ch == EOF)
	{
	  if (state != 0)
	    {
	      as_warn (_("end of file not at end of a line; newline inserted"));
	      state = 0;
	      PUT ('\n');
	    }
	  goto fromeof;
	}

      switch (lex[ch])
	{
	case LEX_IS_WHITESPACE:
	  do
	    {
	      ch = GET ();
	    }
	  while (ch != EOF && IS_WHITESPACE (ch));
	  if (ch == EOF)
	    goto fromeof;

	  if (state == 0)
	    {
	      /* Preserve a single whitespace character at the
		 beginning of a line.  */
	      state = 1;
	      UNGET (ch);
	      PUT (' ');
	      break;
	    }

#ifdef KEEP_WHITE_AROUND_COLON
	  if (lex[ch] == LEX_IS_COLON)
	    {
	      /* Only keep this white if there's no white *after* the
		 colon.  */
	      ch2 = GET ();
	      if (ch2 != EOF)
		UNGET (ch2);
	      if (!IS_WHITESPACE (ch2))
		{
		  state = 9;
		  UNGET (ch);
		  PUT (' ');
		  break;
		}
	    }
#endif
	  if (IS_COMMENT (ch)
	      || ch == '/'
	      || IS_LINE_SEPARATOR (ch)
	      || IS_PARALLEL_SEPARATOR (ch))
	    {
	      if (scrub_m68k_mri)
		{
		  /* In MRI mode, we keep these spaces.  */
		  UNGET (ch);
		  PUT (' ');
		  break;
		}
	      goto recycle;
	    }

	  /* If we're in state 2 or 11, we've seen a non-white
	     character followed by whitespace.  If the next character
	     is ':', this is whitespace after a label name which we
	     normally must ignore.  In MRI mode, though, spaces are
	     not permitted between the label and the colon.  */
	  if ((state == 2 || state == 11)
	      && lex[ch] == LEX_IS_COLON
	      && ! scrub_m68k_mri)
	    {
	      state = 1;
	      PUT (ch);
	      break;
	    }

	  switch (state)
	    {
	    case 1:
	      /* We can arrive here if we leave a leading whitespace
		 character at the beginning of a line.  */
	      goto recycle;
	    case 2:
	      state = 3;
	      if (to + 1 < toend)
		{
		  /* Optimize common case by skipping UNGET/GET.  */
		  PUT (' ');	/* Sp after opco */
		  goto recycle;
		}
	      UNGET (ch);
	      PUT (' ');
	      break;
	    case 3:
#ifndef TC_KEEP_OPERAND_SPACES
	      /* For TI C6X, we keep these spaces as they may separate
		 functional unit specifiers from operands.  */
	      if (scrub_m68k_mri)
#endif
		{
		  /* In MRI mode, we keep these spaces.  */
		  UNGET (ch);
		  PUT (' ');
		  break;
		}
	      goto recycle;	/* Sp in operands */
	    case 9:
	    case 10:
#ifndef TC_KEEP_OPERAND_SPACES
	      if (scrub_m68k_mri)
#endif
		{
		  /* In MRI mode, we keep these spaces.  */
		  state = 3;
		  UNGET (ch);
		  PUT (' ');
		  break;
		}
	      state = 10;	/* Sp after symbol char */
	      goto recycle;
	    case 11:
	      if (LABELS_WITHOUT_COLONS || flag_m68k_mri)
		state = 1;
	      else
		{
		  /* We know that ch is not ':', since we tested that
		     case above.  Therefore this is not a label, so it
		     must be the opcode, and we've just seen the
		     whitespace after it.  */
		  state = 3;
		}
	      UNGET (ch);
	      PUT (' ');	/* Sp after label definition.  */
	      break;
	    default:
	      BAD_CASE (state);
	    }
	  break;

	case LEX_IS_TWOCHAR_COMMENT_1ST:
	  ch2 = GET ();
	  if (ch2 == '*')
	    {
	      for (;;)
		{
		  do
		    {
		      ch2 = GET ();
		      if (ch2 != EOF && IS_NEWLINE (ch2))
			add_newlines++;
		    }
		  while (ch2 != EOF && ch2 != '*');

		  while (ch2 == '*')
		    ch2 = GET ();

		  if (ch2 == EOF || ch2 == '/')
		    break;

		  /* This UNGET will ensure that we count newlines
		     correctly.  */
		  UNGET (ch2);
		}

	      if (ch2 == EOF)
		as_warn (_("end of file in multiline comment"));

	      ch = ' ';
	      goto recycle;
	    }
#ifdef DOUBLESLASH_LINE_COMMENTS
	  else if (ch2 == '/')
	    {
	      do
		{
		  ch = GET ();
		}
	      while (ch != EOF && !IS_NEWLINE (ch));
	      if (ch == EOF)
		as_warn ("end of file in comment; newline inserted");
	      state = 0;
	      PUT ('\n');
	      break;
	    }
#endif
	  else
	    {
	      if (ch2 != EOF)
		UNGET (ch2);
	      if (state == 9 || state == 10)
		state = 3;
	      PUT (ch);
	    }
	  break;

	case LEX_IS_STRINGQUOTE:
	  quotechar = ch;
	  if (state == 10)
	    {
	      /* Preserve the whitespace in foo "bar".  */
	      UNGET (ch);
	      state = 3;
	      PUT (' ');

	      /* PUT didn't jump out.  We could just break, but we
		 know what will happen, so optimize a bit.  */
	      ch = GET ();
	      old_state = 3;
	    }
	  else if (state == 9)
	    old_state = 3;
	  else
	    old_state = state;
	  state = 5;
	  PUT (ch);
	  break;

	case LEX_IS_ONECHAR_QUOTE:
#ifdef H_TICK_HEX
	  if (state == 9 && enable_h_tick_hex)
	    {
	      char c;

	      c = GET ();
	      as_warn ("'%c found after symbol", c);
	      UNGET (c);
	    }
#endif
	  if (state == 10)
	    {
	      /* Preserve the whitespace in foo 'b'.  */
	      UNGET (ch);
	      state = 3;
	      PUT (' ');
	      break;
	    }
	  ch = GET ();
	  if (ch == EOF)
	    {
	      as_warn (_("end of file after a one-character quote; \\0 inserted"));
	      ch = 0;
	    }
	  if (ch == '\\')
	    {
	      ch = GET ();
	      if (ch == EOF)
		{
		  as_warn (_("end of file in escape character"));
		  ch = '\\';
		}
	      else
		ch = process_escape (ch);
	    }
	  sprintf (out_buf, "%d", (int) (unsigned char) ch);

	  /* None of these 'x constants for us.  We want 'x'.  */
	  if ((ch = GET ()) != '\'')
	    {
#ifdef REQUIRE_CHAR_CLOSE_QUOTE
	      as_warn (_("missing close quote; (assumed)"));
#else
	      if (ch != EOF)
		UNGET (ch);
#endif
	    }
	  if (strlen (out_buf) == 1)
	    {
	      PUT (out_buf[0]);
	      break;
	    }
	  if (state == 9)
	    old_state = 3;
	  else
	    old_state = state;
	  state = -1;
	  out_string = out_buf;
	  PUT (*out_string++);
	  break;

	case LEX_IS_COLON:
#ifdef KEEP_WHITE_AROUND_COLON
	  state = 9;
#else
	  if (state == 9 || state == 10)
	    state = 3;
	  else if (state != 3)
	    state = 1;
#endif
	  PUT (ch);
	  break;

	case LEX_IS_NEWLINE:
	  /* Roll out a bunch of newlines from inside comments, etc.  */
	  if (add_newlines)
	    {
	      --add_newlines;
	      UNGET (ch);
	    }
	  /* Fall through.  */

	case LEX_IS_LINE_SEPARATOR:
	  state = 0;
	  PUT (ch);
	  break;

	case LEX_IS_PARALLEL_SEPARATOR:
	  state = 1;
	  PUT (ch);
	  break;

#ifdef TC_V850
	case LEX_IS_DOUBLEDASH_1ST:
	  ch2 = GET ();
	  if (ch2 != '-')
	    {
	      if (ch2 != EOF)
		UNGET (ch2);
	      goto de_fault;
	    }
	  /* Read and skip to end of line.  */
	  do
	    {
	      ch = GET ();
	    }
	  while (ch != EOF && ch != '\n');

	  if (ch == EOF)
	    as_warn (_("end of file in comment; newline inserted"));

	  state = 0;
	  PUT ('\n');
	  break;
#endif
#ifdef DOUBLEBAR_PARALLEL
	case LEX_IS_DOUBLEBAR_1ST:
	  ch2 = GET ();
	  if (ch2 != EOF)
	    UNGET (ch2);
	  if (ch2 != '|')
	    goto de_fault;

	  /* Handle '||' in two states as invoking PUT twice might
	     result in the first one jumping out of this loop.  We'd
	     then lose track of the state and one '|' char.  */
	  state = 13;
	  PUT ('|');
	  break;
#endif
	case LEX_IS_LINE_COMMENT_START:
	  /* FIXME-someday: The two character comment stuff was badly
	     thought out.  On i386, we want '/' as line comment start
	     AND we want C style comments.  hence this hack.  The
	     whole lexical process should be reworked.  xoxorich.  */
	  if (ch == '/')
	    {
	      ch2 = GET ();
	      if (ch2 == '*')
		{
		  old_state = 3;
		  state = -2;
		  break;
		}
	      else if (ch2 != EOF)
		{
		  UNGET (ch2);
		}
	    }

	  if (state == 0 || state == 1)	/* Only comment at start of line.  */
	    {
	      int startch;

	      startch = ch;

	      do
		{
		  ch = GET ();
		}
	      while (ch != EOF && IS_WHITESPACE (ch));

	      if (ch == EOF)
		{
		  as_warn (_("end of file in comment; newline inserted"));
		  PUT ('\n');
		  break;
		}

	      if (ch < '0' || ch > '9' || state != 0 || startch != '#')
		{
		  /* Not a cpp line.  */
		  while (ch != EOF && !IS_NEWLINE (ch))
		    ch = GET ();
		  if (ch == EOF)
		    {
		      as_warn (_("end of file in comment; newline inserted"));
		      PUT ('\n');
		    }
		  else /* IS_NEWLINE (ch) */
		    {
		      /* To process non-zero add_newlines.  */
		      UNGET (ch);
		    }
		  state = 0;
		  break;
		}
	      /* Looks like `# 123 "filename"' from cpp.  */
	      UNGET (ch);
	      old_state = 4;
	      state = -1;
	      if (scrub_m68k_mri)
		out_string = "\tlinefile ";
	      else
		out_string = "\t.linefile ";
	      PUT (*out_string++);
	      break;
	    }

#ifdef TC_D10V
	  /* All insns end in a char for which LEX_IS_SYMBOL_COMPONENT is true.
	     Trap is the only short insn that has a first operand that is
	     neither register nor label.
	     We must prevent exef0f ||trap #1 to degenerate to exef0f ||trap#1 .
	     We can't make '#' LEX_IS_SYMBOL_COMPONENT because it is
	     already LEX_IS_LINE_COMMENT_START.  However, it is the
	     only character in line_comment_chars for d10v, hence we
	     can recognize it as such.  */
	  /* An alternative approach would be to reset the state to 1 when
	     we see '||', '<'- or '->', but that seems to be overkill.  */
	  if (state == 10)
	    PUT (' ');
#endif
	  /* We have a line comment character which is not at the
	     start of a line.  If this is also a normal comment
	     character, fall through.  Otherwise treat it as a default
	     character.  */
	  if (strchr (tc_comment_chars, ch) == NULL
	      && (! scrub_m68k_mri
		  || (ch != '!' && ch != '*')))
	    goto de_fault;
	  if (scrub_m68k_mri
	      && (ch == '!' || ch == '*' || ch == '#')
	      && state != 1
	      && state != 10)
	    goto de_fault;
	  /* Fall through.  */
	case LEX_IS_COMMENT_START:
#if defined TC_ARM && defined OBJ_ELF
	  /* On the ARM, `@' is the comment character.
	     Unfortunately this is also a special character in ELF .symver
	     directives (and .type, though we deal with those another way).
	     So we check if this line is such a directive, and treat
	     the character as default if so.  This is a hack.  */
	  if ((symver_state != NULL) && (*symver_state == 0))
	    goto de_fault;
#endif

#ifdef TC_ARM
	  /* For the ARM, care is needed not to damage occurrences of \@
	     by stripping the @ onwards.  Yuck.  */
	  if ((to > tostart ? to[-1] : last_char) == '\\')
	    /* Do not treat the @ as a start-of-comment.  */
	    goto de_fault;
#endif

#ifdef WARN_COMMENTS
	  if (!found_comment)
	    found_comment_file = as_where (&found_comment);
#endif
	  do
	    {
	      ch = GET ();
	    }
	  while (ch != EOF && !IS_NEWLINE (ch));
	  if (ch == EOF)
	    as_warn (_("end of file in comment; newline inserted"));
	  state = 0;
	  PUT ('\n');
	  break;

#ifdef H_TICK_HEX
	case LEX_IS_H:
	  /* Look for strings like H'[0-9A-Fa-f] and if found, replace
	     the H' with 0x to make them gas-style hex characters.  */
	  if (enable_h_tick_hex)
	    {
	      char quot;

	      quot = GET ();
	      if (quot == '\'')
		{
		  UNGET ('x');
		  ch = '0';
		}
	      else
		UNGET (quot);
	    }
#endif
	  /* Fall through.  */

	case LEX_IS_SYMBOL_COMPONENT:
	  if (state == 10)
	    {
	      /* This is a symbol character following another symbol
		 character, with whitespace in between.  We skipped
		 the whitespace earlier, so output it now.  */
	      UNGET (ch);
	      state = 3;
	      PUT (' ');
	      break;
	    }

#ifdef TC_Z80
	  /* "af'" is a symbol containing '\''.  */
	  if (state == 3 && (ch == 'a' || ch == 'A'))
	    {
	      state = 16;
	      PUT (ch);
	      ch = GET ();
	      if (ch == 'f' || ch == 'F')
		{
		  state = 17;
		  PUT (ch);
		  break;
		}
	      else
		{
		  state = 9;
		  if (ch == EOF || !IS_SYMBOL_COMPONENT (ch))
		    {
		      if (ch != EOF)
			UNGET (ch);
		      break;
		    }
		}
	    }
#endif
	  if (state == 3)
	    state = 9;

	  /* This is a common case.  Quickly copy CH and all the
	     following symbol component or normal characters.  */
	  if (to + 1 < toend
	      && mri_state == NULL
#if defined TC_ARM && defined OBJ_ELF
	      && symver_state == NULL
#endif
	      )
	    {
	      char *s;
	      ptrdiff_t len;

	      for (s = from; s < fromend; s++)
		{
		  int type;

		  ch2 = *(unsigned char *) s;
		  type = lex[ch2];
		  if (type != 0
		      && type != LEX_IS_SYMBOL_COMPONENT)
		    break;
		}

	      if (s > from)
		/* Handle the last character normally, for
		   simplicity.  */
		--s;

	      len = s - from;

	      if (len > (toend - to) - 1)
		len = (toend - to) - 1;

	      if (len > 0)
		{
		  PUT (ch);
		  memcpy (to, from, len);
		  to += len;
		  from += len;
		  if (to >= toend)
		    goto tofull;
		  ch = GET ();
		}
	    }

	  /* Fall through.  */
	default:
	de_fault:
	  /* Some relatively `normal' character.  */
	  if (state == 0)
	    {
	      state = 11;	/* Now seeing label definition.  */
	    }
	  else if (state == 1)
	    {
	      state = 2;	/* Ditto.  */
	    }
	  else if (state == 9)
	    {
	      if (!IS_SYMBOL_COMPONENT (ch))
		state = 3;
	    }
	  else if (state == 10)
	    {
	      if (ch == '\\')
		{
		  /* Special handling for backslash: a backslash may
		     be the beginning of a formal parameter (of a
		     macro) following another symbol character, with
		     whitespace in between.  If that is the case, we
		     output a space before the parameter.  Strictly
		     speaking, correct handling depends upon what the
		     macro parameter expands into; if the parameter
		     expands into something which does not start with
		     an operand character, then we don't want to keep
		     the space.  We don't have enough information to
		     make the right choice, so here we are making the
		     choice which is more likely to be correct.  */
		  if (to + 1 >= toend)
		    {
		      /* If we're near the end of the buffer, save the
		         character for the next time round.  Otherwise
		         we'll lose our state.  */
		      UNGET (ch);
		      goto tofull;
		    }
		  *to++ = ' ';
		}

	      state = 3;
	    }
	  PUT (ch);
	  break;
	}
    }

  /*NOTREACHED*/

 fromeof:
  /* We have reached the end of the input.  */
#ifdef TC_ARM
  if (to > tostart)
    last_char = to[-1];
#endif
  return to - tostart;

 tofull:
  /* The output buffer is full.  Save any input we have not yet
     processed.  */
  if (fromend > from)
    {
      saved_input = from;
      saved_input_len = fromend - from;
    }
  else
    saved_input = NULL;

#ifdef TC_ARM
  if (to > tostart)
    last_char = to[-1];
#endif
  return to - tostart;
}