OpenCores
URL https://opencores.org/ocsvn/openrisc_me/openrisc_me/trunk

Subversion Repositories openrisc_me

[/] [openrisc/] [trunk/] [gnu-src/] [gdb-7.1/] [gdb/] [charset.c] - Blame information for rev 227

Details | Compare with Previous | View Log

Line No. Rev Author Line
1 227 jeremybenn
/* Character set conversion support for GDB.
2
 
3
   Copyright (C) 2001, 2003, 2007, 2008, 2009, 2010
4
   Free Software Foundation, Inc.
5
 
6
   This file is part of GDB.
7
 
8
   This program is free software; you can redistribute it and/or modify
9
   it under the terms of the GNU General Public License as published by
10
   the Free Software Foundation; either version 3 of the License, or
11
   (at your option) any later version.
12
 
13
   This program is distributed in the hope that it will be useful,
14
   but WITHOUT ANY WARRANTY; without even the implied warranty of
15
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
16
   GNU General Public License for more details.
17
 
18
   You should have received a copy of the GNU General Public License
19
   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
20
 
21
#include "defs.h"
22
#include "charset.h"
23
#include "gdbcmd.h"
24
#include "gdb_assert.h"
25
#include "gdb_obstack.h"
26
#include "gdb_wait.h"
27
#include "charset-list.h"
28
#include "vec.h"
29
#include "environ.h"
30
 
31
#include <stddef.h>
32
#include "gdb_string.h"
33
#include <ctype.h>
34
 
35
 
36
/* How GDB's character set support works
37
 
38
   GDB has three global settings:
39
 
40
   - The `current host character set' is the character set GDB should
41
     use in talking to the user, and which (hopefully) the user's
42
     terminal knows how to display properly.  Most users should not
43
     change this.
44
 
45
   - The `current target character set' is the character set the
46
     program being debugged uses.
47
 
48
   - The `current target wide character set' is the wide character set
49
     the program being debugged uses, that is, the encoding used for
50
     wchar_t.
51
 
52
   There are commands to set each of these, and mechanisms for
53
   choosing reasonable default values.  GDB has a global list of
54
   character sets that it can use as its host or target character
55
   sets.
56
 
57
   The header file `charset.h' declares various functions that
58
   different pieces of GDB need to perform tasks like:
59
 
60
   - printing target strings and characters to the user's terminal
61
     (mostly target->host conversions),
62
 
63
   - building target-appropriate representations of strings and
64
     characters the user enters in expressions (mostly host->target
65
     conversions),
66
 
67
     and so on.
68
 
69
   To avoid excessive code duplication and maintenance efforts,
70
   GDB simply requires a capable iconv function.  Users on platforms
71
   without a suitable iconv can use the GNU iconv library.  */
72
 
73
 
74
#ifdef PHONY_ICONV
75
 
76
/* Provide a phony iconv that does as little as possible.  Also,
77
   arrange for there to be a single available character set.  */
78
 
79
#undef GDB_DEFAULT_HOST_CHARSET
80
#define GDB_DEFAULT_HOST_CHARSET "ISO-8859-1"
81
#define GDB_DEFAULT_TARGET_CHARSET "ISO-8859-1"
82
#define GDB_DEFAULT_TARGET_WIDE_CHARSET "ISO-8859-1"
83
#undef DEFAULT_CHARSET_NAMES
84
#define DEFAULT_CHARSET_NAMES GDB_DEFAULT_HOST_CHARSET ,
85
 
86
#undef iconv_t
87
#define iconv_t int
88
#undef iconv_open
89
#undef iconv
90
#undef iconv_close
91
 
92
#undef ICONV_CONST
93
#define ICONV_CONST const
94
 
95
/* Some systems don't have EILSEQ, so we define it here, but not as
96
   EINVAL, because callers of `iconv' want to distinguish EINVAL and
97
   EILSEQ.  This is what iconv.h from libiconv does as well.  Note
98
   that wchar.h may also define EILSEQ, so this needs to be after we
99
   include wchar.h, which happens in defs.h through gdb_wchar.h.  */
100
#ifndef EILSEQ
101
#define EILSEQ ENOENT
102
#endif
103
 
104
iconv_t
105
iconv_open (const char *to, const char *from)
106
{
107
  /* We allow conversions from UTF-32BE, wchar_t, and the host charset.
108
     We allow conversions to wchar_t and the host charset.  */
109
  if (strcmp (from, "UTF-32BE") && strcmp (from, "wchar_t")
110
      && strcmp (from, GDB_DEFAULT_HOST_CHARSET))
111
    return -1;
112
  if (strcmp (to, "wchar_t") && strcmp (to, GDB_DEFAULT_HOST_CHARSET))
113
    return -1;
114
 
115
  /* Return 1 if we are converting from UTF-32BE, 0 otherwise.  This is
116
     used as a flag in calls to iconv.  */
117
  return !strcmp (from, "UTF-32BE");
118
}
119
 
120
int
121
iconv_close (iconv_t arg)
122
{
123
  return 0;
124
}
125
 
126
size_t
127
iconv (iconv_t utf_flag, const char **inbuf, size_t *inbytesleft,
128
       char **outbuf, size_t *outbytesleft)
129
{
130
  if (utf_flag)
131
    {
132
      while (*inbytesleft >= 4)
133
        {
134
          size_t j;
135
          unsigned long c = 0;
136
 
137
          for (j = 0; j < 4; ++j)
138
            {
139
              c <<= 8;
140
              c += (*inbuf)[j] & 0xff;
141
            }
142
 
143
          if (c >= 256)
144
            {
145
              errno = EILSEQ;
146
              return -1;
147
            }
148
          **outbuf = c & 0xff;
149
          ++*outbuf;
150
          --*outbytesleft;
151
 
152
          ++*inbuf;
153
          *inbytesleft -= 4;
154
        }
155
      if (*inbytesleft < 4)
156
        {
157
          errno = EINVAL;
158
          return -1;
159
        }
160
    }
161
  else
162
    {
163
      /* In all other cases we simply copy input bytes to the
164
         output.  */
165
      size_t amt = *inbytesleft;
166
      if (amt > *outbytesleft)
167
        amt = *outbytesleft;
168
      memcpy (*outbuf, *inbuf, amt);
169
      *inbuf += amt;
170
      *outbuf += amt;
171
      *inbytesleft -= amt;
172
      *outbytesleft -= amt;
173
    }
174
 
175
  if (*inbytesleft)
176
    {
177
      errno = E2BIG;
178
      return -1;
179
    }
180
 
181
  /* The number of non-reversible conversions -- but they were all
182
     reversible.  */
183
  return 0;
184
}
185
 
186
#endif
187
 
188
 
189
 
190
/* The global lists of character sets and translations.  */
191
 
192
 
193
#ifndef GDB_DEFAULT_TARGET_CHARSET
194
#define GDB_DEFAULT_TARGET_CHARSET "ISO-8859-1"
195
#endif
196
 
197
#ifndef GDB_DEFAULT_TARGET_WIDE_CHARSET
198
#define GDB_DEFAULT_TARGET_WIDE_CHARSET "UTF-32"
199
#endif
200
 
201
static const char *auto_host_charset_name = GDB_DEFAULT_HOST_CHARSET;
202
static const char *host_charset_name = "auto";
203
static void
204
show_host_charset_name (struct ui_file *file, int from_tty,
205
                        struct cmd_list_element *c,
206
                        const char *value)
207
{
208
  if (!strcmp (value, "auto"))
209
    fprintf_filtered (file,
210
                      _("The host character set is \"auto; currently %s\".\n"),
211
                      auto_host_charset_name);
212
  else
213
    fprintf_filtered (file, _("The host character set is \"%s\".\n"), value);
214
}
215
 
216
static const char *target_charset_name = GDB_DEFAULT_TARGET_CHARSET;
217
static void
218
show_target_charset_name (struct ui_file *file, int from_tty,
219
                          struct cmd_list_element *c, const char *value)
220
{
221
  fprintf_filtered (file, _("The target character set is \"%s\".\n"),
222
                    value);
223
}
224
 
225
static const char *target_wide_charset_name = GDB_DEFAULT_TARGET_WIDE_CHARSET;
226
static void
227
show_target_wide_charset_name (struct ui_file *file, int from_tty,
228
                               struct cmd_list_element *c, const char *value)
229
{
230
  fprintf_filtered (file, _("The target wide character set is \"%s\".\n"),
231
                    value);
232
}
233
 
234
static const char *default_charset_names[] =
235
{
236
  DEFAULT_CHARSET_NAMES
237
 
238
};
239
 
240
static const char **charset_enum;
241
 
242
 
243
/* If the target wide character set has big- or little-endian
244
   variants, these are the corresponding names.  */
245
static const char *target_wide_charset_be_name;
246
static const char *target_wide_charset_le_name;
247
 
248
/* A helper function for validate which sets the target wide big- and
249
   little-endian character set names, if possible.  */
250
 
251
static void
252
set_be_le_names (void)
253
{
254
  int i, len;
255
 
256
  target_wide_charset_le_name = NULL;
257
  target_wide_charset_be_name = NULL;
258
 
259
  len = strlen (target_wide_charset_name);
260
  for (i = 0; charset_enum[i]; ++i)
261
    {
262
      if (strncmp (target_wide_charset_name, charset_enum[i], len))
263
        continue;
264
      if ((charset_enum[i][len] == 'B'
265
           || charset_enum[i][len] == 'L')
266
          && charset_enum[i][len + 1] == 'E'
267
          && charset_enum[i][len + 2] == '\0')
268
        {
269
          if (charset_enum[i][len] == 'B')
270
            target_wide_charset_be_name = charset_enum[i];
271
          else
272
            target_wide_charset_le_name = charset_enum[i];
273
        }
274
    }
275
}
276
 
277
/* 'Set charset', 'set host-charset', 'set target-charset', 'set
278
   target-wide-charset', 'set charset' sfunc's.  */
279
 
280
static void
281
validate (void)
282
{
283
  iconv_t desc;
284
  const char *host_cset = host_charset ();
285
 
286
  desc = iconv_open (target_wide_charset_name, host_cset);
287
  if (desc == (iconv_t) -1)
288
    error ("Cannot convert between character sets `%s' and `%s'",
289
           target_wide_charset_name, host_cset);
290
  iconv_close (desc);
291
 
292
  desc = iconv_open (target_charset_name, host_cset);
293
  if (desc == (iconv_t) -1)
294
    error ("Cannot convert between character sets `%s' and `%s'",
295
           target_charset_name, host_cset);
296
  iconv_close (desc);
297
 
298
  set_be_le_names ();
299
}
300
 
301
/* This is the sfunc for the 'set charset' command.  */
302
static void
303
set_charset_sfunc (char *charset, int from_tty, struct cmd_list_element *c)
304
{
305
  /* CAREFUL: set the target charset here as well. */
306
  target_charset_name = host_charset_name;
307
  validate ();
308
}
309
 
310
/* 'set host-charset' command sfunc.  We need a wrapper here because
311
   the function needs to have a specific signature.  */
312
static void
313
set_host_charset_sfunc (char *charset, int from_tty,
314
                        struct cmd_list_element *c)
315
{
316
  validate ();
317
}
318
 
319
/* Wrapper for the 'set target-charset' command.  */
320
static void
321
set_target_charset_sfunc (char *charset, int from_tty,
322
                          struct cmd_list_element *c)
323
{
324
  validate ();
325
}
326
 
327
/* Wrapper for the 'set target-wide-charset' command.  */
328
static void
329
set_target_wide_charset_sfunc (char *charset, int from_tty,
330
                               struct cmd_list_element *c)
331
{
332
  validate ();
333
}
334
 
335
/* sfunc for the 'show charset' command.  */
336
static void
337
show_charset (struct ui_file *file, int from_tty, struct cmd_list_element *c,
338
              const char *name)
339
{
340
  show_host_charset_name (file, from_tty, c, host_charset_name);
341
  show_target_charset_name (file, from_tty, c, target_charset_name);
342
  show_target_wide_charset_name (file, from_tty, c, target_wide_charset_name);
343
}
344
 
345
 
346
/* Accessor functions.  */
347
 
348
const char *
349
host_charset (void)
350
{
351
  if (!strcmp (host_charset_name, "auto"))
352
    return auto_host_charset_name;
353
  return host_charset_name;
354
}
355
 
356
const char *
357
target_charset (void)
358
{
359
  return target_charset_name;
360
}
361
 
362
const char *
363
target_wide_charset (enum bfd_endian byte_order)
364
{
365
  if (byte_order == BFD_ENDIAN_BIG)
366
    {
367
      if (target_wide_charset_be_name)
368
        return target_wide_charset_be_name;
369
    }
370
  else
371
    {
372
      if (target_wide_charset_le_name)
373
        return target_wide_charset_le_name;
374
    }
375
 
376
  return target_wide_charset_name;
377
}
378
 
379
 
380
/* Host character set management.  For the time being, we assume that
381
   the host character set is some superset of ASCII.  */
382
 
383
char
384
host_letter_to_control_character (char c)
385
{
386
  if (c == '?')
387
    return 0177;
388
  return c & 0237;
389
}
390
 
391
/* Convert a host character, C, to its hex value.  C must already have
392
   been validated using isxdigit.  */
393
 
394
int
395
host_hex_value (char c)
396
{
397
  if (isdigit (c))
398
    return c - '0';
399
  if (c >= 'a' && c <= 'f')
400
    return 10 + c - 'a';
401
  gdb_assert (c >= 'A' && c <= 'F');
402
  return 10 + c - 'A';
403
}
404
 
405
 
406
/* Public character management functions.  */
407
 
408
/* A cleanup function which is run to close an iconv descriptor.  */
409
 
410
static void
411
cleanup_iconv (void *p)
412
{
413
  iconv_t *descp = p;
414
  iconv_close (*descp);
415
}
416
 
417
void
418
convert_between_encodings (const char *from, const char *to,
419
                           const gdb_byte *bytes, unsigned int num_bytes,
420
                           int width, struct obstack *output,
421
                           enum transliterations translit)
422
{
423
  iconv_t desc;
424
  struct cleanup *cleanups;
425
  size_t inleft;
426
  char *inp;
427
  unsigned int space_request;
428
 
429
  /* Often, the host and target charsets will be the same.  */
430
  if (!strcmp (from, to))
431
    {
432
      obstack_grow (output, bytes, num_bytes);
433
      return;
434
    }
435
 
436
  desc = iconv_open (to, from);
437
  if (desc == (iconv_t) -1)
438
    perror_with_name ("Converting character sets");
439
  cleanups = make_cleanup (cleanup_iconv, &desc);
440
 
441
  inleft = num_bytes;
442
  inp = (char *) bytes;
443
 
444
  space_request = num_bytes;
445
 
446
  while (inleft > 0)
447
    {
448
      char *outp;
449
      size_t outleft, r;
450
      int old_size;
451
 
452
      old_size = obstack_object_size (output);
453
      obstack_blank (output, space_request);
454
 
455
      outp = obstack_base (output) + old_size;
456
      outleft = space_request;
457
 
458
      r = iconv (desc, (ICONV_CONST char **) &inp, &inleft, &outp, &outleft);
459
 
460
      /* Now make sure that the object on the obstack only includes
461
         bytes we have converted.  */
462
      obstack_blank (output, - (int) outleft);
463
 
464
      if (r == (size_t) -1)
465
        {
466
          switch (errno)
467
            {
468
            case EILSEQ:
469
              {
470
                int i;
471
 
472
                /* Invalid input sequence.  */
473
                if (translit == translit_none)
474
                  error (_("Could not convert character to `%s' character set"),
475
                         to);
476
 
477
                /* We emit escape sequence for the bytes, skip them,
478
                   and try again.  */
479
                for (i = 0; i < width; ++i)
480
                  {
481
                    char octal[5];
482
 
483
                    sprintf (octal, "\\%.3o", *inp & 0xff);
484
                    obstack_grow_str (output, octal);
485
 
486
                    ++inp;
487
                    --inleft;
488
                  }
489
              }
490
              break;
491
 
492
            case E2BIG:
493
              /* We ran out of space in the output buffer.  Make it
494
                 bigger next time around.  */
495
              space_request *= 2;
496
              break;
497
 
498
            case EINVAL:
499
              /* Incomplete input sequence.  FIXME: ought to report this
500
                 to the caller somehow.  */
501
              inleft = 0;
502
              break;
503
 
504
            default:
505
              perror_with_name ("Internal error while converting character sets");
506
            }
507
        }
508
    }
509
 
510
  do_cleanups (cleanups);
511
}
512
 
513
 
514
 
515
/* An iterator that returns host wchar_t's from a target string.  */
516
struct wchar_iterator
517
{
518
  /* The underlying iconv descriptor.  */
519
  iconv_t desc;
520
 
521
  /* The input string.  This is updated as convert characters.  */
522
  char *input;
523
  /* The number of bytes remaining in the input.  */
524
  size_t bytes;
525
 
526
  /* The width of an input character.  */
527
  size_t width;
528
 
529
  /* The output buffer and its size.  */
530
  gdb_wchar_t *out;
531
  size_t out_size;
532
};
533
 
534
/* Create a new iterator.  */
535
struct wchar_iterator *
536
make_wchar_iterator (const gdb_byte *input, size_t bytes, const char *charset,
537
                     size_t width)
538
{
539
  struct wchar_iterator *result;
540
  iconv_t desc;
541
 
542
  desc = iconv_open (INTERMEDIATE_ENCODING, charset);
543
  if (desc == (iconv_t) -1)
544
    perror_with_name ("Converting character sets");
545
 
546
  result = XNEW (struct wchar_iterator);
547
  result->desc = desc;
548
  result->input = (char *) input;
549
  result->bytes = bytes;
550
  result->width = width;
551
 
552
  result->out = XNEW (gdb_wchar_t);
553
  result->out_size = 1;
554
 
555
  return result;
556
}
557
 
558
static void
559
do_cleanup_iterator (void *p)
560
{
561
  struct wchar_iterator *iter = p;
562
 
563
  iconv_close (iter->desc);
564
  xfree (iter->out);
565
  xfree (iter);
566
}
567
 
568
struct cleanup *
569
make_cleanup_wchar_iterator (struct wchar_iterator *iter)
570
{
571
  return make_cleanup (do_cleanup_iterator, iter);
572
}
573
 
574
int
575
wchar_iterate (struct wchar_iterator *iter,
576
               enum wchar_iterate_result *out_result,
577
               gdb_wchar_t **out_chars,
578
               const gdb_byte **ptr,
579
               size_t *len)
580
{
581
  size_t out_request;
582
 
583
  /* Try to convert some characters.  At first we try to convert just
584
     a single character.  The reason for this is that iconv does not
585
     necessarily update its outgoing arguments when it encounters an
586
     invalid input sequence -- but we want to reliably report this to
587
     our caller so it can emit an escape sequence.  */
588
  out_request = 1;
589
  while (iter->bytes > 0)
590
    {
591
      char *outptr = (char *) &iter->out[0];
592
      char *orig_inptr = iter->input;
593
      size_t orig_in = iter->bytes;
594
      size_t out_avail = out_request * sizeof (gdb_wchar_t);
595
      size_t num;
596
      gdb_wchar_t result;
597
 
598
      size_t r = iconv (iter->desc,
599
                        (ICONV_CONST char **) &iter->input, &iter->bytes,
600
                        &outptr, &out_avail);
601
      if (r == (size_t) -1)
602
        {
603
          switch (errno)
604
            {
605
            case EILSEQ:
606
              /* Invalid input sequence.  Skip it, and let the caller
607
                 know about it.  */
608
              *out_result = wchar_iterate_invalid;
609
              *ptr = iter->input;
610
              *len = iter->width;
611
              iter->input += iter->width;
612
              iter->bytes -= iter->width;
613
              return 0;
614
 
615
            case E2BIG:
616
              /* We ran out of space.  We still might have converted a
617
                 character; if so, return it.  Otherwise, grow the
618
                 buffer and try again.  */
619
              if (out_avail < out_request * sizeof (gdb_wchar_t))
620
                break;
621
 
622
              ++out_request;
623
              if (out_request > iter->out_size)
624
                {
625
                  iter->out_size = out_request;
626
                  iter->out = xrealloc (iter->out,
627
                                        out_request * sizeof (gdb_wchar_t));
628
                }
629
              continue;
630
 
631
            case EINVAL:
632
              /* Incomplete input sequence.  Let the caller know, and
633
                 arrange for future calls to see EOF.  */
634
              *out_result = wchar_iterate_incomplete;
635
              *ptr = iter->input;
636
              *len = iter->bytes;
637
              iter->bytes = 0;
638
              return 0;
639
 
640
            default:
641
              perror_with_name ("Internal error while converting character sets");
642
            }
643
        }
644
 
645
      /* We converted something.  */
646
      num = out_request - out_avail / sizeof (gdb_wchar_t);
647
      *out_result = wchar_iterate_ok;
648
      *out_chars = iter->out;
649
      *ptr = orig_inptr;
650
      *len = orig_in - iter->bytes;
651
      return num;
652
    }
653
 
654
  /* Really done.  */
655
  *out_result = wchar_iterate_eof;
656
  return -1;
657
}
658
 
659
 
660
/* The charset.c module initialization function.  */
661
 
662
extern initialize_file_ftype _initialize_charset; /* -Wmissing-prototype */
663
 
664
typedef char *char_ptr;
665
DEF_VEC_P (char_ptr);
666
 
667
static VEC (char_ptr) *charsets;
668
 
669
#ifdef PHONY_ICONV
670
 
671
static void
672
find_charset_names (void)
673
{
674
  VEC_safe_push (char_ptr, charsets, GDB_DEFAULT_HOST_CHARSET);
675
  VEC_safe_push (char_ptr, charsets, NULL);
676
}
677
 
678
#else /* PHONY_ICONV */
679
 
680
/* Sometimes, libiconv redefines iconvlist as libiconvlist -- but
681
   provides different symbols in the static and dynamic libraries.
682
   So, configure may see libiconvlist but not iconvlist.  But, calling
683
   iconvlist is the right thing to do and will work.  Hence we do a
684
   check here but unconditionally call iconvlist below.  */
685
#if defined (HAVE_ICONVLIST) || defined (HAVE_LIBICONVLIST)
686
 
687
/* A helper function that adds some character sets to the vector of
688
   all character sets.  This is a callback function for iconvlist.  */
689
 
690
static int
691
add_one (unsigned int count, const char *const *names, void *data)
692
{
693
  unsigned int i;
694
 
695
  for (i = 0; i < count; ++i)
696
    VEC_safe_push (char_ptr, charsets, xstrdup (names[i]));
697
 
698
  return 0;
699
}
700
 
701
static void
702
find_charset_names (void)
703
{
704
  iconvlist (add_one, NULL);
705
  VEC_safe_push (char_ptr, charsets, NULL);
706
}
707
 
708
#else
709
 
710
/* Return non-zero if LINE (output from iconv) should be ignored.
711
   Older iconv programs (e.g. 2.2.2) include the human readable
712
   introduction even when stdout is not a tty.  Newer versions omit
713
   the intro if stdout is not a tty.  */
714
 
715
static int
716
ignore_line_p (const char *line)
717
{
718
  /* This table is used to filter the output.  If this text appears
719
     anywhere in the line, it is ignored (strstr is used).  */
720
  static const char * const ignore_lines[] =
721
    {
722
      "The following",
723
      "not necessarily",
724
      "the FROM and TO",
725
      "listed with several",
726
      NULL
727
    };
728
  int i;
729
 
730
  for (i = 0; ignore_lines[i] != NULL; ++i)
731
    {
732
      if (strstr (line, ignore_lines[i]) != NULL)
733
        return 1;
734
    }
735
 
736
  return 0;
737
}
738
 
739
static void
740
find_charset_names (void)
741
{
742
  struct pex_obj *child;
743
  char *args[3];
744
  int err, status;
745
  int fail = 1;
746
  struct gdb_environ *iconv_env;
747
 
748
  /* Older iconvs, e.g. 2.2.2, don't omit the intro text if stdout is not
749
     a tty.  We need to recognize it and ignore it.  This text is subject
750
     to translation, so force LANGUAGE=C.  */
751
  iconv_env = make_environ ();
752
  init_environ (iconv_env);
753
  set_in_environ (iconv_env, "LANGUAGE", "C");
754
  set_in_environ (iconv_env, "LC_ALL", "C");
755
 
756
  child = pex_init (0, "iconv", NULL);
757
 
758
  args[0] = "iconv";
759
  args[1] = "-l";
760
  args[2] = NULL;
761
  /* Note that we simply ignore errors here.  */
762
  if (!pex_run_in_environment (child, PEX_SEARCH | PEX_STDERR_TO_STDOUT,
763
                               "iconv", args, environ_vector (iconv_env),
764
                               NULL, NULL, &err))
765
    {
766
      FILE *in = pex_read_output (child, 0);
767
 
768
      /* POSIX says that iconv -l uses an unspecified format.  We
769
         parse the glibc and libiconv formats; feel free to add others
770
         as needed.  */
771
 
772
      while (!feof (in))
773
        {
774
          /* The size of buf is chosen arbitrarily.  */
775
          char buf[1024];
776
          char *start, *r;
777
          int len, keep_going;
778
 
779
          r = fgets (buf, sizeof (buf), in);
780
          if (!r)
781
            break;
782
          len = strlen (r);
783
          if (len <= 3)
784
            continue;
785
          if (ignore_line_p (r))
786
            continue;
787
 
788
          /* Strip off the newline.  */
789
          --len;
790
          /* Strip off one or two '/'s.  glibc will print lines like
791
             "8859_7//", but also "10646-1:1993/UCS4/".  */
792
          if (buf[len - 1] == '/')
793
            --len;
794
          if (buf[len - 1] == '/')
795
            --len;
796
          buf[len] = '\0';
797
 
798
          /* libiconv will print multiple entries per line, separated
799
             by spaces.  Older iconvs will print multiple entries per line,
800
             indented by two spaces, and separated by ", "
801
             (i.e. the human readable form).  */
802
          start = buf;
803
          while (1)
804
            {
805
              int keep_going;
806
              char *p;
807
 
808
              /* Skip leading blanks.  */
809
              for (p = start; *p && *p == ' '; ++p)
810
                ;
811
              start = p;
812
              /* Find the next space, comma, or end-of-line.  */
813
              for ( ; *p && *p != ' ' && *p != ','; ++p)
814
                ;
815
              /* Ignore an empty result.  */
816
              if (p == start)
817
                break;
818
              keep_going = *p;
819
              *p = '\0';
820
              VEC_safe_push (char_ptr, charsets, xstrdup (start));
821
              if (!keep_going)
822
                break;
823
              /* Skip any extra spaces.  */
824
              for (start = p + 1; *start && *start == ' '; ++start)
825
                ;
826
            }
827
        }
828
 
829
      if (pex_get_status (child, 1, &status)
830
          && WIFEXITED (status) && !WEXITSTATUS (status))
831
        fail = 0;
832
 
833
    }
834
 
835
  pex_free (child);
836
  free_environ (iconv_env);
837
 
838
  if (fail)
839
    {
840
      /* Some error occurred, so drop the vector.  */
841
      int ix;
842
      char *elt;
843
      for (ix = 0; VEC_iterate (char_ptr, charsets, ix, elt); ++ix)
844
        xfree (elt);
845
      VEC_truncate (char_ptr, charsets, 0);
846
    }
847
  else
848
    VEC_safe_push (char_ptr, charsets, NULL);
849
}
850
 
851
#endif /* HAVE_ICONVLIST || HAVE_LIBICONVLIST */
852
#endif /* PHONY_ICONV */
853
 
854
void
855
_initialize_charset (void)
856
{
857
  struct cmd_list_element *new_cmd;
858
 
859
  /* The first element is always "auto"; then we skip it for the
860
     commands where it is not allowed.  */
861
  VEC_safe_push (char_ptr, charsets, xstrdup ("auto"));
862
  find_charset_names ();
863
 
864
  if (VEC_length (char_ptr, charsets) > 1)
865
    charset_enum = (const char **) VEC_address (char_ptr, charsets);
866
  else
867
    charset_enum = default_charset_names;
868
 
869
#ifndef PHONY_ICONV
870
#ifdef HAVE_LANGINFO_CODESET
871
  auto_host_charset_name = nl_langinfo (CODESET);
872
  /* Solaris will return `646' here -- but the Solaris iconv then
873
     does not accept this.  Darwin (and maybe FreeBSD) may return "" here,
874
     which GNU libiconv doesn't like (infinite loop).  */
875
  if (!strcmp (auto_host_charset_name, "646") || !*auto_host_charset_name)
876
    auto_host_charset_name = "ASCII";
877
  target_charset_name = auto_host_charset_name;
878
 
879
  set_be_le_names ();
880
#endif
881
#endif
882
 
883
  add_setshow_enum_cmd ("charset", class_support,
884
                        &charset_enum[1], &host_charset_name, _("\
885
Set the host and target character sets."), _("\
886
Show the host and target character sets."), _("\
887
The `host character set' is the one used by the system GDB is running on.\n\
888
The `target character set' is the one used by the program being debugged.\n\
889
You may only use supersets of ASCII for your host character set; GDB does\n\
890
not support any others.\n\
891
To see a list of the character sets GDB supports, type `set charset <TAB>'."),
892
                        /* Note that the sfunc below needs to set
893
                           target_charset_name, because the 'set
894
                           charset' command sets two variables.  */
895
                        set_charset_sfunc,
896
                        show_charset,
897
                        &setlist, &showlist);
898
 
899
  add_setshow_enum_cmd ("host-charset", class_support,
900
                        charset_enum, &host_charset_name, _("\
901
Set the host character set."), _("\
902
Show the host character set."), _("\
903
The `host character set' is the one used by the system GDB is running on.\n\
904
You may only use supersets of ASCII for your host character set; GDB does\n\
905
not support any others.\n\
906
To see a list of the character sets GDB supports, type `set host-charset <TAB>'."),
907
                        set_host_charset_sfunc,
908
                        show_host_charset_name,
909
                        &setlist, &showlist);
910
 
911
  add_setshow_enum_cmd ("target-charset", class_support,
912
                        &charset_enum[1], &target_charset_name, _("\
913
Set the target character set."), _("\
914
Show the target character set."), _("\
915
The `target character set' is the one used by the program being debugged.\n\
916
GDB translates characters and strings between the host and target\n\
917
character sets as needed.\n\
918
To see a list of the character sets GDB supports, type `set target-charset'<TAB>"),
919
                        set_target_charset_sfunc,
920
                        show_target_charset_name,
921
                        &setlist, &showlist);
922
 
923
  add_setshow_enum_cmd ("target-wide-charset", class_support,
924
                        &charset_enum[1], &target_wide_charset_name,
925
                        _("\
926
Set the target wide character set."), _("\
927
Show the target wide character set."), _("\
928
The `target wide character set' is the one used by the program being debugged.\n\
929
In particular it is the encoding used by `wchar_t'.\n\
930
GDB translates characters and strings between the host and target\n\
931
character sets as needed.\n\
932
To see a list of the character sets GDB supports, type\n\
933
`set target-wide-charset'<TAB>"),
934
                        set_target_wide_charset_sfunc,
935
                        show_target_wide_charset_name,
936
                        &setlist, &showlist);
937
}

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.