forked from gcc-mirror/gcc
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcharset.c
1672 lines (1449 loc) · 51 KB
/
charset.c
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
/* CPP Library - charsets
Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2006
Free Software Foundation, Inc.
Broken out of c-lex.c Apr 2003, adding valid C99 UCN ranges.
This program is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2, or (at your option) any
later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
#include "config.h"
#include "system.h"
#include "cpplib.h"
#include "internal.h"
/* Character set handling for C-family languages.
Terminological note: In what follows, "charset" or "character set"
will be taken to mean both an abstract set of characters and an
encoding for that set.
The C99 standard discusses two character sets: source and execution.
The source character set is used for internal processing in translation
phases 1 through 4; the execution character set is used thereafter.
Both are required by 5.2.1.2p1 to be multibyte encodings, not wide
character encodings (see 3.7.2, 3.7.3 for the standardese meanings
of these terms). Furthermore, the "basic character set" (listed in
5.2.1p3) is to be encoded in each with values one byte wide, and is
to appear in the initial shift state.
It is not explicitly mentioned, but there is also a "wide execution
character set" used to encode wide character constants and wide
string literals; this is supposed to be the result of applying the
standard library function mbstowcs() to an equivalent narrow string
(6.4.5p5). However, the behavior of hexadecimal and octal
\-escapes is at odds with this; they are supposed to be translated
directly to wchar_t values (6.4.4.4p5,6).
The source character set is not necessarily the character set used
to encode physical source files on disk; translation phase 1 converts
from whatever that encoding is to the source character set.
The presence of universal character names in C99 (6.4.3 et seq.)
forces the source character set to be isomorphic to ISO 10646,
that is, Unicode. There is no such constraint on the execution
character set; note also that the conversion from source to
execution character set does not occur for identifiers (5.1.1.2p1#5).
For convenience of implementation, the source character set's
encoding of the basic character set should be identical to the
execution character set OF THE HOST SYSTEM's encoding of the basic
character set, and it should not be a state-dependent encoding.
cpplib uses UTF-8 or UTF-EBCDIC for the source character set,
depending on whether the host is based on ASCII or EBCDIC (see
respectively Unicode section 2.3/ISO10646 Amendment 2, and Unicode
Technical Report #16). With limited exceptions, it relies on the
system library's iconv() primitive to do charset conversion
(specified in SUSv2). */
#if !HAVE_ICONV
/* Make certain that the uses of iconv(), iconv_open(), iconv_close()
below, which are guarded only by if statements with compile-time
constant conditions, do not cause link errors. */
#define iconv_open(x, y) (errno = EINVAL, (iconv_t)-1)
#define iconv(a,b,c,d,e) (errno = EINVAL, (size_t)-1)
#define iconv_close(x) (void)0
#define ICONV_CONST
#endif
#if HOST_CHARSET == HOST_CHARSET_ASCII
#define SOURCE_CHARSET "UTF-8"
#define LAST_POSSIBLY_BASIC_SOURCE_CHAR 0x7e
#elif HOST_CHARSET == HOST_CHARSET_EBCDIC
#define SOURCE_CHARSET "UTF-EBCDIC"
#define LAST_POSSIBLY_BASIC_SOURCE_CHAR 0xFF
#else
#error "Unrecognized basic host character set"
#endif
#ifndef EILSEQ
#define EILSEQ EINVAL
#endif
/* This structure is used for a resizable string buffer throughout. */
/* Don't call it strbuf, as that conflicts with unistd.h on systems
such as DYNIX/ptx where unistd.h includes stropts.h. */
struct _cpp_strbuf
{
uchar *text;
size_t asize;
size_t len;
};
/* This is enough to hold any string that fits on a single 80-column
line, even if iconv quadruples its size (e.g. conversion from
ASCII to UTF-32) rounded up to a power of two. */
#define OUTBUF_BLOCK_SIZE 256
/* Conversions between UTF-8 and UTF-16/32 are implemented by custom
logic. This is because a depressing number of systems lack iconv,
or have have iconv libraries that do not do these conversions, so
we need a fallback implementation for them. To ensure the fallback
doesn't break due to neglect, it is used on all systems.
UTF-32 encoding is nice and simple: a four-byte binary number,
constrained to the range 00000000-7FFFFFFF to avoid questions of
signedness. We do have to cope with big- and little-endian
variants.
UTF-16 encoding uses two-byte binary numbers, again in big- and
little-endian variants, for all values in the 00000000-0000FFFF
range. Values in the 00010000-0010FFFF range are encoded as pairs
of two-byte numbers, called "surrogate pairs": given a number S in
this range, it is mapped to a pair (H, L) as follows:
H = (S - 0x10000) / 0x400 + 0xD800
L = (S - 0x10000) % 0x400 + 0xDC00
Two-byte values in the D800...DFFF range are ill-formed except as a
component of a surrogate pair. Even if the encoding within a
two-byte value is little-endian, the H member of the surrogate pair
comes first.
There is no way to encode values in the 00110000-7FFFFFFF range,
which is not currently a problem as there are no assigned code
points in that range; however, the author expects that it will
eventually become necessary to abandon UTF-16 due to this
limitation. Note also that, because of these pairs, UTF-16 does
not meet the requirements of the C standard for a wide character
encoding (see 3.7.3 and 6.4.4.4p11).
UTF-8 encoding looks like this:
value range encoded as
00000000-0000007F 0xxxxxxx
00000080-000007FF 110xxxxx 10xxxxxx
00000800-0000FFFF 1110xxxx 10xxxxxx 10xxxxxx
00010000-001FFFFF 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
00200000-03FFFFFF 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
04000000-7FFFFFFF 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
Values in the 0000D800 ... 0000DFFF range (surrogates) are invalid,
which means that three-byte sequences ED xx yy, with A0 <= xx <= BF,
never occur. Note also that any value that can be encoded by a
given row of the table can also be encoded by all successive rows,
but this is not done; only the shortest possible encoding for any
given value is valid. For instance, the character 07C0 could be
encoded as any of DF 80, E0 9F 80, F0 80 9F 80, F8 80 80 9F 80, or
FC 80 80 80 9F 80. Only the first is valid.
An implementation note: the transformation from UTF-16 to UTF-8, or
vice versa, is easiest done by using UTF-32 as an intermediary. */
/* Internal primitives which go from an UTF-8 byte stream to native-endian
UTF-32 in a cppchar_t, or vice versa; this avoids an extra marshal/unmarshal
operation in several places below. */
static inline int
one_utf8_to_cppchar (const uchar **inbufp, size_t *inbytesleftp,
cppchar_t *cp)
{
static const uchar masks[6] = { 0x7F, 0x1F, 0x0F, 0x07, 0x02, 0x01 };
static const uchar patns[6] = { 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
cppchar_t c;
const uchar *inbuf = *inbufp;
size_t nbytes, i;
if (*inbytesleftp < 1)
return EINVAL;
c = *inbuf;
if (c < 0x80)
{
*cp = c;
*inbytesleftp -= 1;
*inbufp += 1;
return 0;
}
/* The number of leading 1-bits in the first byte indicates how many
bytes follow. */
for (nbytes = 2; nbytes < 7; nbytes++)
if ((c & ~masks[nbytes-1]) == patns[nbytes-1])
goto found;
return EILSEQ;
found:
if (*inbytesleftp < nbytes)
return EINVAL;
c = (c & masks[nbytes-1]);
inbuf++;
for (i = 1; i < nbytes; i++)
{
cppchar_t n = *inbuf++;
if ((n & 0xC0) != 0x80)
return EILSEQ;
c = ((c << 6) + (n & 0x3F));
}
/* Make sure the shortest possible encoding was used. */
if (c <= 0x7F && nbytes > 1) return EILSEQ;
if (c <= 0x7FF && nbytes > 2) return EILSEQ;
if (c <= 0xFFFF && nbytes > 3) return EILSEQ;
if (c <= 0x1FFFFF && nbytes > 4) return EILSEQ;
if (c <= 0x3FFFFFF && nbytes > 5) return EILSEQ;
/* Make sure the character is valid. */
if (c > 0x7FFFFFFF || (c >= 0xD800 && c <= 0xDFFF)) return EILSEQ;
*cp = c;
*inbufp = inbuf;
*inbytesleftp -= nbytes;
return 0;
}
static inline int
one_cppchar_to_utf8 (cppchar_t c, uchar **outbufp, size_t *outbytesleftp)
{
static const uchar masks[6] = { 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
static const uchar limits[6] = { 0x80, 0xE0, 0xF0, 0xF8, 0xFC, 0xFE };
size_t nbytes;
uchar buf[6], *p = &buf[6];
uchar *outbuf = *outbufp;
nbytes = 1;
if (c < 0x80)
*--p = c;
else
{
do
{
*--p = ((c & 0x3F) | 0x80);
c >>= 6;
nbytes++;
}
while (c >= 0x3F || (c & limits[nbytes-1]));
*--p = (c | masks[nbytes-1]);
}
if (*outbytesleftp < nbytes)
return E2BIG;
while (p < &buf[6])
*outbuf++ = *p++;
*outbytesleftp -= nbytes;
*outbufp = outbuf;
return 0;
}
/* The following four functions transform one character between the two
encodings named in the function name. All have the signature
int (*)(iconv_t bigend, const uchar **inbufp, size_t *inbytesleftp,
uchar **outbufp, size_t *outbytesleftp)
BIGEND must have the value 0 or 1, coerced to (iconv_t); it is
interpreted as a boolean indicating whether big-endian or
little-endian encoding is to be used for the member of the pair
that is not UTF-8.
INBUFP, INBYTESLEFTP, OUTBUFP, OUTBYTESLEFTP work exactly as they
do for iconv.
The return value is either 0 for success, or an errno value for
failure, which may be E2BIG (need more space), EILSEQ (ill-formed
input sequence), ir EINVAL (incomplete input sequence). */
static inline int
one_utf8_to_utf32 (iconv_t bigend, const uchar **inbufp, size_t *inbytesleftp,
uchar **outbufp, size_t *outbytesleftp)
{
uchar *outbuf;
cppchar_t s = 0;
int rval;
/* Check for space first, since we know exactly how much we need. */
if (*outbytesleftp < 4)
return E2BIG;
rval = one_utf8_to_cppchar (inbufp, inbytesleftp, &s);
if (rval)
return rval;
outbuf = *outbufp;
outbuf[bigend ? 3 : 0] = (s & 0x000000FF);
outbuf[bigend ? 2 : 1] = (s & 0x0000FF00) >> 8;
outbuf[bigend ? 1 : 2] = (s & 0x00FF0000) >> 16;
outbuf[bigend ? 0 : 3] = (s & 0xFF000000) >> 24;
*outbufp += 4;
*outbytesleftp -= 4;
return 0;
}
static inline int
one_utf32_to_utf8 (iconv_t bigend, const uchar **inbufp, size_t *inbytesleftp,
uchar **outbufp, size_t *outbytesleftp)
{
cppchar_t s;
int rval;
const uchar *inbuf;
if (*inbytesleftp < 4)
return EINVAL;
inbuf = *inbufp;
s = inbuf[bigend ? 0 : 3] << 24;
s += inbuf[bigend ? 1 : 2] << 16;
s += inbuf[bigend ? 2 : 1] << 8;
s += inbuf[bigend ? 3 : 0];
if (s >= 0x7FFFFFFF || (s >= 0xD800 && s <= 0xDFFF))
return EILSEQ;
rval = one_cppchar_to_utf8 (s, outbufp, outbytesleftp);
if (rval)
return rval;
*inbufp += 4;
*inbytesleftp -= 4;
return 0;
}
static inline int
one_utf8_to_utf16 (iconv_t bigend, const uchar **inbufp, size_t *inbytesleftp,
uchar **outbufp, size_t *outbytesleftp)
{
int rval;
cppchar_t s = 0;
const uchar *save_inbuf = *inbufp;
size_t save_inbytesleft = *inbytesleftp;
uchar *outbuf = *outbufp;
rval = one_utf8_to_cppchar (inbufp, inbytesleftp, &s);
if (rval)
return rval;
if (s > 0x0010FFFF)
{
*inbufp = save_inbuf;
*inbytesleftp = save_inbytesleft;
return EILSEQ;
}
if (s < 0xFFFF)
{
if (*outbytesleftp < 2)
{
*inbufp = save_inbuf;
*inbytesleftp = save_inbytesleft;
return E2BIG;
}
outbuf[bigend ? 1 : 0] = (s & 0x00FF);
outbuf[bigend ? 0 : 1] = (s & 0xFF00) >> 8;
*outbufp += 2;
*outbytesleftp -= 2;
return 0;
}
else
{
cppchar_t hi, lo;
if (*outbytesleftp < 4)
{
*inbufp = save_inbuf;
*inbytesleftp = save_inbytesleft;
return E2BIG;
}
hi = (s - 0x10000) / 0x400 + 0xD800;
lo = (s - 0x10000) % 0x400 + 0xDC00;
/* Even if we are little-endian, put the high surrogate first.
??? Matches practice? */
outbuf[bigend ? 1 : 0] = (hi & 0x00FF);
outbuf[bigend ? 0 : 1] = (hi & 0xFF00) >> 8;
outbuf[bigend ? 3 : 2] = (lo & 0x00FF);
outbuf[bigend ? 2 : 3] = (lo & 0xFF00) >> 8;
*outbufp += 4;
*outbytesleftp -= 4;
return 0;
}
}
static inline int
one_utf16_to_utf8 (iconv_t bigend, const uchar **inbufp, size_t *inbytesleftp,
uchar **outbufp, size_t *outbytesleftp)
{
cppchar_t s;
const uchar *inbuf = *inbufp;
int rval;
if (*inbytesleftp < 2)
return EINVAL;
s = inbuf[bigend ? 0 : 1] << 8;
s += inbuf[bigend ? 1 : 0];
/* Low surrogate without immediately preceding high surrogate is invalid. */
if (s >= 0xDC00 && s <= 0xDFFF)
return EILSEQ;
/* High surrogate must have a following low surrogate. */
else if (s >= 0xD800 && s <= 0xDBFF)
{
cppchar_t hi = s, lo;
if (*inbytesleftp < 4)
return EINVAL;
lo = inbuf[bigend ? 2 : 3] << 8;
lo += inbuf[bigend ? 3 : 2];
if (lo < 0xDC00 || lo > 0xDFFF)
return EILSEQ;
s = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000;
}
rval = one_cppchar_to_utf8 (s, outbufp, outbytesleftp);
if (rval)
return rval;
/* Success - update the input pointers (one_cppchar_to_utf8 has done
the output pointers for us). */
if (s <= 0xFFFF)
{
*inbufp += 2;
*inbytesleftp -= 2;
}
else
{
*inbufp += 4;
*inbytesleftp -= 4;
}
return 0;
}
/* Helper routine for the next few functions. The 'const' on
one_conversion means that we promise not to modify what function is
pointed to, which lets the inliner see through it. */
static inline bool
conversion_loop (int (*const one_conversion)(iconv_t, const uchar **, size_t *,
uchar **, size_t *),
iconv_t cd, const uchar *from, size_t flen, struct _cpp_strbuf *to)
{
const uchar *inbuf;
uchar *outbuf;
size_t inbytesleft, outbytesleft;
int rval;
inbuf = from;
inbytesleft = flen;
outbuf = to->text + to->len;
outbytesleft = to->asize - to->len;
for (;;)
{
do
rval = one_conversion (cd, &inbuf, &inbytesleft,
&outbuf, &outbytesleft);
while (inbytesleft && !rval);
if (__builtin_expect (inbytesleft == 0, 1))
{
to->len = to->asize - outbytesleft;
return true;
}
if (rval != E2BIG)
{
errno = rval;
return false;
}
outbytesleft += OUTBUF_BLOCK_SIZE;
to->asize += OUTBUF_BLOCK_SIZE;
to->text = XRESIZEVEC (uchar, to->text, to->asize);
outbuf = to->text + to->asize - outbytesleft;
}
}
/* These functions convert entire strings between character sets.
They all have the signature
bool (*)(iconv_t cd, const uchar *from, size_t flen, struct _cpp_strbuf *to);
The input string FROM is converted as specified by the function
name plus the iconv descriptor CD (which may be fake), and the
result appended to TO. On any error, false is returned, otherwise true. */
/* These four use the custom conversion code above. */
static bool
convert_utf8_utf16 (iconv_t cd, const uchar *from, size_t flen,
struct _cpp_strbuf *to)
{
return conversion_loop (one_utf8_to_utf16, cd, from, flen, to);
}
static bool
convert_utf8_utf32 (iconv_t cd, const uchar *from, size_t flen,
struct _cpp_strbuf *to)
{
return conversion_loop (one_utf8_to_utf32, cd, from, flen, to);
}
static bool
convert_utf16_utf8 (iconv_t cd, const uchar *from, size_t flen,
struct _cpp_strbuf *to)
{
return conversion_loop (one_utf16_to_utf8, cd, from, flen, to);
}
static bool
convert_utf32_utf8 (iconv_t cd, const uchar *from, size_t flen,
struct _cpp_strbuf *to)
{
return conversion_loop (one_utf32_to_utf8, cd, from, flen, to);
}
/* Identity conversion, used when we have no alternative. */
static bool
convert_no_conversion (iconv_t cd ATTRIBUTE_UNUSED,
const uchar *from, size_t flen, struct _cpp_strbuf *to)
{
if (to->len + flen > to->asize)
{
to->asize = to->len + flen;
to->text = XRESIZEVEC (uchar, to->text, to->asize);
}
memcpy (to->text + to->len, from, flen);
to->len += flen;
return true;
}
/* And this one uses the system iconv primitive. It's a little
different, since iconv's interface is a little different. */
#if HAVE_ICONV
static bool
convert_using_iconv (iconv_t cd, const uchar *from, size_t flen,
struct _cpp_strbuf *to)
{
ICONV_CONST char *inbuf;
char *outbuf;
size_t inbytesleft, outbytesleft;
/* Reset conversion descriptor and check that it is valid. */
if (iconv (cd, 0, 0, 0, 0) == (size_t)-1)
return false;
inbuf = (ICONV_CONST char *)from;
inbytesleft = flen;
outbuf = (char *)to->text + to->len;
outbytesleft = to->asize - to->len;
for (;;)
{
iconv (cd, &inbuf, &inbytesleft, &outbuf, &outbytesleft);
if (__builtin_expect (inbytesleft == 0, 1))
{
to->len = to->asize - outbytesleft;
return true;
}
if (errno != E2BIG)
return false;
outbytesleft += OUTBUF_BLOCK_SIZE;
to->asize += OUTBUF_BLOCK_SIZE;
to->text = XRESIZEVEC (uchar, to->text, to->asize);
outbuf = (char *)to->text + to->asize - outbytesleft;
}
}
#else
#define convert_using_iconv 0 /* prevent undefined symbol error below */
#endif
/* Arrange for the above custom conversion logic to be used automatically
when conversion between a suitable pair of character sets is requested. */
#define APPLY_CONVERSION(CONVERTER, FROM, FLEN, TO) \
CONVERTER.func (CONVERTER.cd, FROM, FLEN, TO)
struct conversion
{
const char *pair;
convert_f func;
iconv_t fake_cd;
};
static const struct conversion conversion_tab[] = {
{ "UTF-8/UTF-32LE", convert_utf8_utf32, (iconv_t)0 },
{ "UTF-8/UTF-32BE", convert_utf8_utf32, (iconv_t)1 },
{ "UTF-8/UTF-16LE", convert_utf8_utf16, (iconv_t)0 },
{ "UTF-8/UTF-16BE", convert_utf8_utf16, (iconv_t)1 },
{ "UTF-32LE/UTF-8", convert_utf32_utf8, (iconv_t)0 },
{ "UTF-32BE/UTF-8", convert_utf32_utf8, (iconv_t)1 },
{ "UTF-16LE/UTF-8", convert_utf16_utf8, (iconv_t)0 },
{ "UTF-16BE/UTF-8", convert_utf16_utf8, (iconv_t)1 },
};
/* Subroutine of cpp_init_iconv: initialize and return a
cset_converter structure for conversion from FROM to TO. If
iconv_open() fails, issue an error and return an identity
converter. Silently return an identity converter if FROM and TO
are identical. */
static struct cset_converter
init_iconv_desc (cpp_reader *pfile, const char *to, const char *from)
{
struct cset_converter ret;
char *pair;
size_t i;
if (!strcasecmp (to, from))
{
ret.func = convert_no_conversion;
ret.cd = (iconv_t) -1;
return ret;
}
pair = (char *) alloca(strlen(to) + strlen(from) + 2);
strcpy(pair, from);
strcat(pair, "/");
strcat(pair, to);
for (i = 0; i < ARRAY_SIZE (conversion_tab); i++)
if (!strcasecmp (pair, conversion_tab[i].pair))
{
ret.func = conversion_tab[i].func;
ret.cd = conversion_tab[i].fake_cd;
return ret;
}
/* No custom converter - try iconv. */
if (HAVE_ICONV)
{
ret.func = convert_using_iconv;
ret.cd = iconv_open (to, from);
if (ret.cd == (iconv_t) -1)
{
if (errno == EINVAL)
cpp_error (pfile, CPP_DL_ERROR, /* FIXME should be DL_SORRY */
"conversion from %s to %s not supported by iconv",
from, to);
else
cpp_errno (pfile, CPP_DL_ERROR, "iconv_open");
ret.func = convert_no_conversion;
}
}
else
{
cpp_error (pfile, CPP_DL_ERROR, /* FIXME: should be DL_SORRY */
"no iconv implementation, cannot convert from %s to %s",
from, to);
ret.func = convert_no_conversion;
ret.cd = (iconv_t) -1;
}
return ret;
}
/* If charset conversion is requested, initialize iconv(3) descriptors
for conversion from the source character set to the execution
character sets. If iconv is not present in the C library, and
conversion is requested, issue an error. */
void
cpp_init_iconv (cpp_reader *pfile)
{
const char *ncset = CPP_OPTION (pfile, narrow_charset);
const char *wcset = CPP_OPTION (pfile, wide_charset);
const char *default_wcset;
bool be = CPP_OPTION (pfile, bytes_big_endian);
if (CPP_OPTION (pfile, wchar_precision) >= 32)
default_wcset = be ? "UTF-32BE" : "UTF-32LE";
else if (CPP_OPTION (pfile, wchar_precision) >= 16)
default_wcset = be ? "UTF-16BE" : "UTF-16LE";
else
/* This effectively means that wide strings are not supported,
so don't do any conversion at all. */
default_wcset = SOURCE_CHARSET;
if (!ncset)
ncset = SOURCE_CHARSET;
if (!wcset)
wcset = default_wcset;
pfile->narrow_cset_desc = init_iconv_desc (pfile, ncset, SOURCE_CHARSET);
pfile->wide_cset_desc = init_iconv_desc (pfile, wcset, SOURCE_CHARSET);
}
/* Destroy iconv(3) descriptors set up by cpp_init_iconv, if necessary. */
void
_cpp_destroy_iconv (cpp_reader *pfile)
{
if (HAVE_ICONV)
{
if (pfile->narrow_cset_desc.func == convert_using_iconv)
iconv_close (pfile->narrow_cset_desc.cd);
if (pfile->wide_cset_desc.func == convert_using_iconv)
iconv_close (pfile->wide_cset_desc.cd);
}
}
/* Utility routine for use by a full compiler. C is a character taken
from the *basic* source character set, encoded in the host's
execution encoding. Convert it to (the target's) execution
encoding, and return that value.
Issues an internal error if C's representation in the narrow
execution character set fails to be a single-byte value (C99
5.2.1p3: "The representation of each member of the source and
execution character sets shall fit in a byte.") May also issue an
internal error if C fails to be a member of the basic source
character set (testing this exactly is too hard, especially when
the host character set is EBCDIC). */
cppchar_t
cpp_host_to_exec_charset (cpp_reader *pfile, cppchar_t c)
{
uchar sbuf[1];
struct _cpp_strbuf tbuf;
/* This test is merely an approximation, but it suffices to catch
the most important thing, which is that we don't get handed a
character outside the unibyte range of the host character set. */
if (c > LAST_POSSIBLY_BASIC_SOURCE_CHAR)
{
cpp_error (pfile, CPP_DL_ICE,
"character 0x%lx is not in the basic source character set\n",
(unsigned long)c);
return 0;
}
/* Being a character in the unibyte range of the host character set,
we can safely splat it into a one-byte buffer and trust that that
is a well-formed string. */
sbuf[0] = c;
/* This should never need to reallocate, but just in case... */
tbuf.asize = 1;
tbuf.text = XNEWVEC (uchar, tbuf.asize);
tbuf.len = 0;
if (!APPLY_CONVERSION (pfile->narrow_cset_desc, sbuf, 1, &tbuf))
{
cpp_errno (pfile, CPP_DL_ICE, "converting to execution character set");
return 0;
}
if (tbuf.len != 1)
{
cpp_error (pfile, CPP_DL_ICE,
"character 0x%lx is not unibyte in execution character set",
(unsigned long)c);
return 0;
}
c = tbuf.text[0];
free(tbuf.text);
return c;
}
/* Utility routine that computes a mask of the form 0000...111... with
WIDTH 1-bits. */
static inline size_t
width_to_mask (size_t width)
{
width = MIN (width, BITS_PER_CPPCHAR_T);
if (width >= CHAR_BIT * sizeof (size_t))
return ~(size_t) 0;
else
return ((size_t) 1 << width) - 1;
}
/* A large table of unicode character information. */
enum {
/* Valid in a C99 identifier? */
C99 = 1,
/* Valid in a C99 identifier, but not as the first character? */
DIG = 2,
/* Valid in a C++ identifier? */
CXX = 4,
/* NFC representation is not valid in an identifier? */
CID = 8,
/* Might be valid NFC form? */
NFC = 16,
/* Might be valid NFKC form? */
NKC = 32,
/* Certain preceding characters might make it not valid NFC/NKFC form? */
CTX = 64
};
static const struct {
/* Bitmap of flags above. */
unsigned char flags;
/* Combining class of the character. */
unsigned char combine;
/* Last character in the range described by this entry. */
unsigned short end;
} ucnranges[] = {
#include "ucnid.h"
};
/* Returns 1 if C is valid in an identifier, 2 if C is valid except at
the start of an identifier, and 0 if C is not valid in an
identifier. We assume C has already gone through the checks of
_cpp_valid_ucn. Also update NST for C if returning nonzero. The
algorithm is a simple binary search on the table defined in
ucnid.h. */
static int
ucn_valid_in_identifier (cpp_reader *pfile, cppchar_t c,
struct normalize_state *nst)
{
int mn, mx, md;
if (c > 0xFFFF)
return 0;
mn = 0;
mx = ARRAY_SIZE (ucnranges) - 1;
while (mx != mn)
{
md = (mn + mx) / 2;
if (c <= ucnranges[md].end)
mx = md;
else
mn = md + 1;
}
/* When -pedantic, we require the character to have been listed by
the standard for the current language. Otherwise, we accept the
union of the acceptable sets for C++98 and C99. */
if (! (ucnranges[mn].flags & (C99 | CXX)))
return 0;
if (CPP_PEDANTIC (pfile)
&& ((CPP_OPTION (pfile, c99) && !(ucnranges[mn].flags & C99))
|| (CPP_OPTION (pfile, cplusplus)
&& !(ucnranges[mn].flags & CXX))))
return 0;
/* Update NST. */
if (ucnranges[mn].combine != 0 && ucnranges[mn].combine < nst->prev_class)
nst->level = normalized_none;
else if (ucnranges[mn].flags & CTX)
{
bool safe;
cppchar_t p = nst->previous;
/* Easy cases from Bengali, Oriya, Tamil, Jannada, and Malayalam. */
if (c == 0x09BE)
safe = p != 0x09C7; /* Use 09CB instead of 09C7 09BE. */
else if (c == 0x0B3E)
safe = p != 0x0B47; /* Use 0B4B instead of 0B47 0B3E. */
else if (c == 0x0BBE)
safe = p != 0x0BC6 && p != 0x0BC7; /* Use 0BCA/0BCB instead. */
else if (c == 0x0CC2)
safe = p != 0x0CC6; /* Use 0CCA instead of 0CC6 0CC2. */
else if (c == 0x0D3E)
safe = p != 0x0D46 && p != 0x0D47; /* Use 0D4A/0D4B instead. */
/* For Hangul, characters in the range AC00-D7A3 are NFC/NFKC,
and are combined algorithmically from a sequence of the form
1100-1112 1161-1175 11A8-11C2
(if the third is not present, it is treated as 11A7, which is not
really a valid character).
Unfortunately, C99 allows (only) the NFC form, but C++ allows
only the combining characters. */
else if (c >= 0x1161 && c <= 0x1175)
safe = p < 0x1100 || p > 0x1112;
else if (c >= 0x11A8 && c <= 0x11C2)
safe = (p < 0xAC00 || p > 0xD7A3 || (p - 0xAC00) % 28 != 0);
else
{
/* Uh-oh, someone updated ucnid.h without updating this code. */
cpp_error (pfile, CPP_DL_ICE, "Character %x might not be NFKC", c);
safe = true;
}
if (!safe && c < 0x1161)
nst->level = normalized_none;
else if (!safe)
nst->level = MAX (nst->level, normalized_identifier_C);
}
else if (ucnranges[mn].flags & NKC)
;
else if (ucnranges[mn].flags & NFC)
nst->level = MAX (nst->level, normalized_C);
else if (ucnranges[mn].flags & CID)
nst->level = MAX (nst->level, normalized_identifier_C);
else
nst->level = normalized_none;
nst->previous = c;
nst->prev_class = ucnranges[mn].combine;
/* In C99, UCN digits may not begin identifiers. */
if (CPP_OPTION (pfile, c99) && (ucnranges[mn].flags & DIG))
return 2;
return 1;
}
/* [lex.charset]: The character designated by the universal character
name \UNNNNNNNN is that character whose character short name in
ISO/IEC 10646 is NNNNNNNN; the character designated by the
universal character name \uNNNN is that character whose character
short name in ISO/IEC 10646 is 0000NNNN. If the hexadecimal value
for a universal character name is less than 0x20 or in the range
0x7F-0x9F (inclusive), or if the universal character name
designates a character in the basic source character set, then the
program is ill-formed.
*PSTR must be preceded by "\u" or "\U"; it is assumed that the
buffer end is delimited by a non-hex digit. Returns zero if the
UCN has not been consumed.
Otherwise the nonzero value of the UCN, whether valid or invalid,
is returned. Diagnostics are emitted for invalid values. PSTR
is updated to point one beyond the UCN, or to the syntactically
invalid character.
IDENTIFIER_POS is 0 when not in an identifier, 1 for the start of
an identifier, or 2 otherwise. */
cppchar_t
_cpp_valid_ucn (cpp_reader *pfile, const uchar **pstr,
const uchar *limit, int identifier_pos,
struct normalize_state *nst)
{
cppchar_t result, c;
unsigned int length;
const uchar *str = *pstr;
const uchar *base = str - 2;
if (!CPP_OPTION (pfile, cplusplus) && !CPP_OPTION (pfile, c99))
cpp_error (pfile, CPP_DL_WARNING,
"universal character names are only valid in C++ and C99");
else if (CPP_WTRADITIONAL (pfile) && identifier_pos == 0)
cpp_error (pfile, CPP_DL_WARNING,
"the meaning of '\\%c' is different in traditional C",
(int) str[-1]);
if (str[-1] == 'u')
length = 4;
else if (str[-1] == 'U')
length = 8;
else
{
cpp_error (pfile, CPP_DL_ICE, "In _cpp_valid_ucn but not a UCN");
length = 4;
}
result = 0;
do
{
c = *str;
if (!ISXDIGIT (c))
break;
str++;
result = (result << 4) + hex_value (c);
}
while (--length && str < limit);
/* Partial UCNs are not valid in strings, but decompose into
multiple tokens in identifiers, so we can't give a helpful
error message in that case. */
if (length && identifier_pos)
return 0;
*pstr = str;
if (length)
{
cpp_error (pfile, CPP_DL_ERROR,
"incomplete universal character name %.*s",
(int) (str - base), base);
result = 1;
}
/* The standard permits $, @ and ` to be specified as UCNs. We use
hex escapes so that this also works with EBCDIC hosts. */
else if ((result < 0xa0
&& (result != 0x24 && result != 0x40 && result != 0x60))
|| (result & 0x80000000)
|| (result >= 0xD800 && result <= 0xDFFF))
{
cpp_error (pfile, CPP_DL_ERROR,
"%.*s is not a valid universal character",
(int) (str - base), base);
result = 1;