Commit | Line | Data |
---|---|---|
234b45d4 | 1 | /* Character set conversion support for GDB. |
1bac305b | 2 | |
0fb0cc75 | 3 | Copyright (C) 2001, 2003, 2007, 2008, 2009 Free Software Foundation, Inc. |
234b45d4 KB |
4 | |
5 | This file is part of GDB. | |
6 | ||
7 | This program is free software; you can redistribute it and/or modify | |
8 | it under the terms of the GNU General Public License as published by | |
a9762ec7 | 9 | the Free Software Foundation; either version 3 of the License, or |
234b45d4 KB |
10 | (at your option) any later version. |
11 | ||
12 | This program is distributed in the hope that it will be useful, | |
13 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 | GNU General Public License for more details. | |
16 | ||
17 | You should have received a copy of the GNU General Public License | |
a9762ec7 | 18 | along with this program. If not, see <http://www.gnu.org/licenses/>. */ |
234b45d4 KB |
19 | |
20 | #include "defs.h" | |
21 | #include "charset.h" | |
22 | #include "gdbcmd.h" | |
23 | #include "gdb_assert.h" | |
6c7a06a3 TT |
24 | #include "gdb_obstack.h" |
25 | #include "charset-list.h" | |
26 | #include "vec.h" | |
234b45d4 KB |
27 | |
28 | #include <stddef.h> | |
4ef3f3be | 29 | #include "gdb_string.h" |
234b45d4 KB |
30 | #include <ctype.h> |
31 | ||
234b45d4 KB |
32 | \f |
33 | /* How GDB's character set support works | |
34 | ||
6c7a06a3 | 35 | GDB has three global settings: |
234b45d4 KB |
36 | |
37 | - The `current host character set' is the character set GDB should | |
38 | use in talking to the user, and which (hopefully) the user's | |
6c7a06a3 TT |
39 | terminal knows how to display properly. Most users should not |
40 | change this. | |
234b45d4 KB |
41 | |
42 | - The `current target character set' is the character set the | |
43 | program being debugged uses. | |
44 | ||
6c7a06a3 TT |
45 | - The `current target wide character set' is the wide character set |
46 | the program being debugged uses, that is, the encoding used for | |
47 | wchar_t. | |
48 | ||
234b45d4 KB |
49 | There are commands to set each of these, and mechanisms for |
50 | choosing reasonable default values. GDB has a global list of | |
51 | character sets that it can use as its host or target character | |
52 | sets. | |
53 | ||
54 | The header file `charset.h' declares various functions that | |
55 | different pieces of GDB need to perform tasks like: | |
56 | ||
57 | - printing target strings and characters to the user's terminal | |
58 | (mostly target->host conversions), | |
59 | ||
60 | - building target-appropriate representations of strings and | |
61 | characters the user enters in expressions (mostly host->target | |
62 | conversions), | |
63 | ||
6c7a06a3 TT |
64 | and so on. |
65 | ||
66 | To avoid excessive code duplication and maintenance efforts, | |
67 | GDB simply requires a capable iconv function. Users on platforms | |
68 | without a suitable iconv can use the GNU iconv library. */ | |
234b45d4 KB |
69 | |
70 | \f | |
6c7a06a3 | 71 | #ifdef PHONY_ICONV |
234b45d4 | 72 | |
6c7a06a3 TT |
73 | /* Provide a phony iconv that does as little as possible. Also, |
74 | arrange for there to be a single available character set. */ | |
234b45d4 | 75 | |
6c7a06a3 TT |
76 | #undef GDB_DEFAULT_HOST_CHARSET |
77 | #define GDB_DEFAULT_HOST_CHARSET "ISO-8859-1" | |
78 | #define GDB_DEFAULT_TARGET_CHARSET "ISO-8859-1" | |
79 | #define GDB_DEFAULT_TARGET_WIDE_CHARSET "ISO-8859-1" | |
80 | #undef DEFAULT_CHARSET_NAMES | |
81 | #define DEFAULT_CHARSET_NAMES GDB_DEFAULT_HOST_CHARSET , | |
82 | ||
83 | #undef iconv_t | |
84 | #define iconv_t int | |
85 | #undef iconv_open | |
86 | #undef iconv | |
87 | #undef iconv_close | |
88 | ||
0dd7fb99 TT |
89 | #undef ICONV_CONST |
90 | #define ICONV_CONST const | |
91 | ||
6c7a06a3 TT |
92 | iconv_t |
93 | iconv_open (const char *to, const char *from) | |
94 | { | |
95 | /* We allow conversions from UCS-4BE, wchar_t, and the host charset. | |
96 | We allow conversions to wchar_t and the host charset. */ | |
97 | if (strcmp (from, "UCS-4BE") && strcmp (from, "wchar_t") | |
98 | && strcmp (from, GDB_DEFAULT_HOST_CHARSET)) | |
99 | return -1; | |
100 | if (strcmp (to, "wchar_t") && strcmp (to, GDB_DEFAULT_HOST_CHARSET)) | |
101 | return -1; | |
234b45d4 | 102 | |
6c7a06a3 TT |
103 | /* Return 1 if we are converting from UCS-4BE, 0 otherwise. This is |
104 | used as a flag in calls to iconv. */ | |
105 | return !strcmp (from, "UCS-4BE"); | |
106 | } | |
234b45d4 | 107 | |
6c7a06a3 TT |
108 | int |
109 | iconv_close (iconv_t arg) | |
110 | { | |
111 | return 0; | |
112 | } | |
234b45d4 | 113 | |
6c7a06a3 | 114 | size_t |
0dd7fb99 | 115 | iconv (iconv_t ucs_flag, const char **inbuf, size_t *inbytesleft, |
6c7a06a3 TT |
116 | char **outbuf, size_t *outbytesleft) |
117 | { | |
118 | if (ucs_flag) | |
119 | { | |
120 | while (*inbytesleft >= 4) | |
121 | { | |
122 | size_t j; | |
123 | unsigned long c = 0; | |
124 | ||
125 | for (j = 0; j < 4; ++j) | |
126 | { | |
127 | c <<= 8; | |
128 | c += (*inbuf)[j] & 0xff; | |
129 | } | |
130 | ||
131 | if (c >= 256) | |
132 | { | |
133 | errno = EILSEQ; | |
134 | return -1; | |
135 | } | |
136 | **outbuf = c & 0xff; | |
137 | ++*outbuf; | |
138 | --*outbytesleft; | |
139 | ||
140 | ++*inbuf; | |
141 | *inbytesleft -= 4; | |
142 | } | |
143 | if (*inbytesleft < 4) | |
144 | { | |
145 | errno = EINVAL; | |
146 | return -1; | |
147 | } | |
148 | } | |
149 | else | |
150 | { | |
151 | /* In all other cases we simply copy input bytes to the | |
152 | output. */ | |
153 | size_t amt = *inbytesleft; | |
154 | if (amt > *outbytesleft) | |
155 | amt = *outbytesleft; | |
156 | memcpy (*outbuf, *inbuf, amt); | |
157 | *inbuf += amt; | |
158 | *outbuf += amt; | |
159 | *inbytesleft -= amt; | |
160 | *outbytesleft -= amt; | |
161 | } | |
234b45d4 | 162 | |
6c7a06a3 TT |
163 | if (*inbytesleft) |
164 | { | |
165 | errno = E2BIG; | |
166 | return -1; | |
167 | } | |
234b45d4 | 168 | |
6c7a06a3 TT |
169 | /* The number of non-reversible conversions -- but they were all |
170 | reversible. */ | |
171 | return 0; | |
172 | } | |
234b45d4 | 173 | |
6c7a06a3 | 174 | #endif |
234b45d4 KB |
175 | |
176 | ||
177 | \f | |
178 | /* The global lists of character sets and translations. */ | |
179 | ||
180 | ||
e33d66ec EZ |
181 | #ifndef GDB_DEFAULT_TARGET_CHARSET |
182 | #define GDB_DEFAULT_TARGET_CHARSET "ISO-8859-1" | |
183 | #endif | |
184 | ||
6c7a06a3 TT |
185 | #ifndef GDB_DEFAULT_TARGET_WIDE_CHARSET |
186 | #define GDB_DEFAULT_TARGET_WIDE_CHARSET "UCS-4" | |
187 | #endif | |
188 | ||
189 | static const char *auto_host_charset_name = GDB_DEFAULT_HOST_CHARSET; | |
190 | static const char *host_charset_name = "auto"; | |
920d2a44 AC |
191 | static void |
192 | show_host_charset_name (struct ui_file *file, int from_tty, | |
193 | struct cmd_list_element *c, | |
194 | const char *value) | |
195 | { | |
6c7a06a3 TT |
196 | if (!strcmp (value, "auto")) |
197 | fprintf_filtered (file, | |
198 | _("The host character set is \"auto; currently %s\".\n"), | |
199 | auto_host_charset_name); | |
200 | else | |
201 | fprintf_filtered (file, _("The host character set is \"%s\".\n"), value); | |
920d2a44 AC |
202 | } |
203 | ||
e33d66ec | 204 | static const char *target_charset_name = GDB_DEFAULT_TARGET_CHARSET; |
920d2a44 AC |
205 | static void |
206 | show_target_charset_name (struct ui_file *file, int from_tty, | |
207 | struct cmd_list_element *c, const char *value) | |
208 | { | |
209 | fprintf_filtered (file, _("The target character set is \"%s\".\n"), | |
210 | value); | |
211 | } | |
212 | ||
6c7a06a3 TT |
213 | static const char *target_wide_charset_name = GDB_DEFAULT_TARGET_WIDE_CHARSET; |
214 | static void | |
215 | show_target_wide_charset_name (struct ui_file *file, int from_tty, | |
216 | struct cmd_list_element *c, const char *value) | |
e33d66ec | 217 | { |
6c7a06a3 TT |
218 | fprintf_filtered (file, _("The target wide character set is \"%s\".\n"), |
219 | value); | |
220 | } | |
e33d66ec | 221 | |
6c7a06a3 | 222 | static const char *default_charset_names[] = |
e33d66ec | 223 | { |
6c7a06a3 | 224 | DEFAULT_CHARSET_NAMES |
e33d66ec EZ |
225 | 0 |
226 | }; | |
234b45d4 | 227 | |
6c7a06a3 | 228 | static const char **charset_enum; |
234b45d4 | 229 | |
6c7a06a3 TT |
230 | \f |
231 | /* If the target wide character set has big- or little-endian | |
232 | variants, these are the corresponding names. */ | |
233 | static const char *target_wide_charset_be_name; | |
234 | static const char *target_wide_charset_le_name; | |
234b45d4 | 235 | |
6c7a06a3 TT |
236 | /* A helper function for validate which sets the target wide big- and |
237 | little-endian character set names, if possible. */ | |
234b45d4 | 238 | |
6c7a06a3 TT |
239 | static void |
240 | set_be_le_names (void) | |
234b45d4 | 241 | { |
6c7a06a3 | 242 | int i, len; |
234b45d4 | 243 | |
6c7a06a3 TT |
244 | target_wide_charset_le_name = NULL; |
245 | target_wide_charset_be_name = NULL; | |
234b45d4 | 246 | |
6c7a06a3 TT |
247 | len = strlen (target_wide_charset_name); |
248 | for (i = 0; charset_enum[i]; ++i) | |
249 | { | |
250 | if (strncmp (target_wide_charset_name, charset_enum[i], len)) | |
251 | continue; | |
252 | if ((charset_enum[i][len] == 'B' | |
253 | || charset_enum[i][len] == 'L') | |
254 | && charset_enum[i][len + 1] == 'E' | |
255 | && charset_enum[i][len + 2] == '\0') | |
256 | { | |
257 | if (charset_enum[i][len] == 'B') | |
258 | target_wide_charset_be_name = charset_enum[i]; | |
259 | else | |
260 | target_wide_charset_le_name = charset_enum[i]; | |
261 | } | |
262 | } | |
234b45d4 KB |
263 | } |
264 | ||
6c7a06a3 TT |
265 | /* 'Set charset', 'set host-charset', 'set target-charset', 'set |
266 | target-wide-charset', 'set charset' sfunc's. */ | |
234b45d4 KB |
267 | |
268 | static void | |
6c7a06a3 | 269 | validate (void) |
234b45d4 | 270 | { |
6c7a06a3 TT |
271 | iconv_t desc; |
272 | const char *host_cset = host_charset (); | |
234b45d4 | 273 | |
6c7a06a3 TT |
274 | desc = iconv_open (target_wide_charset_name, host_cset); |
275 | if (desc == (iconv_t) -1) | |
276 | error ("Cannot convert between character sets `%s' and `%s'", | |
277 | target_wide_charset_name, host_cset); | |
278 | iconv_close (desc); | |
234b45d4 | 279 | |
6c7a06a3 TT |
280 | desc = iconv_open (target_charset_name, host_cset); |
281 | if (desc == (iconv_t) -1) | |
282 | error ("Cannot convert between character sets `%s' and `%s'", | |
283 | target_charset_name, host_cset); | |
284 | iconv_close (desc); | |
234b45d4 | 285 | |
6c7a06a3 | 286 | set_be_le_names (); |
234b45d4 KB |
287 | } |
288 | ||
6c7a06a3 TT |
289 | /* This is the sfunc for the 'set charset' command. */ |
290 | static void | |
291 | set_charset_sfunc (char *charset, int from_tty, struct cmd_list_element *c) | |
234b45d4 | 292 | { |
6c7a06a3 TT |
293 | /* CAREFUL: set the target charset here as well. */ |
294 | target_charset_name = host_charset_name; | |
295 | validate (); | |
234b45d4 KB |
296 | } |
297 | ||
6c7a06a3 TT |
298 | /* 'set host-charset' command sfunc. We need a wrapper here because |
299 | the function needs to have a specific signature. */ | |
300 | static void | |
301 | set_host_charset_sfunc (char *charset, int from_tty, | |
302 | struct cmd_list_element *c) | |
234b45d4 | 303 | { |
6c7a06a3 | 304 | validate (); |
234b45d4 KB |
305 | } |
306 | ||
6c7a06a3 TT |
307 | /* Wrapper for the 'set target-charset' command. */ |
308 | static void | |
309 | set_target_charset_sfunc (char *charset, int from_tty, | |
310 | struct cmd_list_element *c) | |
234b45d4 | 311 | { |
6c7a06a3 | 312 | validate (); |
234b45d4 KB |
313 | } |
314 | ||
6c7a06a3 TT |
315 | /* Wrapper for the 'set target-wide-charset' command. */ |
316 | static void | |
317 | set_target_wide_charset_sfunc (char *charset, int from_tty, | |
318 | struct cmd_list_element *c) | |
234b45d4 | 319 | { |
6c7a06a3 | 320 | validate (); |
234b45d4 KB |
321 | } |
322 | ||
6c7a06a3 TT |
323 | /* sfunc for the 'show charset' command. */ |
324 | static void | |
325 | show_charset (struct ui_file *file, int from_tty, struct cmd_list_element *c, | |
326 | const char *name) | |
234b45d4 | 327 | { |
6c7a06a3 TT |
328 | show_host_charset_name (file, from_tty, c, host_charset_name); |
329 | show_target_charset_name (file, from_tty, c, target_charset_name); | |
330 | show_target_wide_charset_name (file, from_tty, c, target_wide_charset_name); | |
234b45d4 KB |
331 | } |
332 | ||
234b45d4 | 333 | \f |
6c7a06a3 | 334 | /* Accessor functions. */ |
234b45d4 | 335 | |
6c7a06a3 TT |
336 | const char * |
337 | host_charset (void) | |
234b45d4 | 338 | { |
6c7a06a3 TT |
339 | if (!strcmp (host_charset_name, "auto")) |
340 | return auto_host_charset_name; | |
341 | return host_charset_name; | |
234b45d4 KB |
342 | } |
343 | ||
6c7a06a3 TT |
344 | const char * |
345 | target_charset (void) | |
234b45d4 | 346 | { |
6c7a06a3 | 347 | return target_charset_name; |
234b45d4 | 348 | } |
234b45d4 | 349 | |
6c7a06a3 TT |
350 | const char * |
351 | target_wide_charset (void) | |
234b45d4 | 352 | { |
6c7a06a3 | 353 | if (gdbarch_byte_order (current_gdbarch) == BFD_ENDIAN_BIG) |
234b45d4 | 354 | { |
6c7a06a3 TT |
355 | if (target_wide_charset_be_name) |
356 | return target_wide_charset_be_name; | |
234b45d4 | 357 | } |
6c7a06a3 | 358 | else |
234b45d4 | 359 | { |
6c7a06a3 TT |
360 | if (target_wide_charset_le_name) |
361 | return target_wide_charset_le_name; | |
234b45d4 KB |
362 | } |
363 | ||
6c7a06a3 | 364 | return target_wide_charset_name; |
234b45d4 KB |
365 | } |
366 | ||
234b45d4 | 367 | \f |
6c7a06a3 TT |
368 | /* Host character set management. For the time being, we assume that |
369 | the host character set is some superset of ASCII. */ | |
234b45d4 | 370 | |
6c7a06a3 TT |
371 | char |
372 | host_letter_to_control_character (char c) | |
234b45d4 | 373 | { |
6c7a06a3 TT |
374 | if (c == '?') |
375 | return 0177; | |
376 | return c & 0237; | |
234b45d4 KB |
377 | } |
378 | ||
6c7a06a3 TT |
379 | /* Convert a host character, C, to its hex value. C must already have |
380 | been validated using isxdigit. */ | |
234b45d4 | 381 | |
6c7a06a3 TT |
382 | int |
383 | host_hex_value (char c) | |
234b45d4 | 384 | { |
6c7a06a3 TT |
385 | if (isdigit (c)) |
386 | return c - '0'; | |
387 | if (c >= 'a' && c <= 'f') | |
388 | return 10 + c - 'a'; | |
389 | gdb_assert (c >= 'A' && c <= 'F'); | |
390 | return 10 + c - 'A'; | |
234b45d4 KB |
391 | } |
392 | ||
234b45d4 | 393 | \f |
6c7a06a3 | 394 | /* Public character management functions. */ |
234b45d4 | 395 | |
6c7a06a3 | 396 | /* A cleanup function which is run to close an iconv descriptor. */ |
234b45d4 | 397 | |
6c7a06a3 TT |
398 | static void |
399 | cleanup_iconv (void *p) | |
234b45d4 | 400 | { |
6c7a06a3 TT |
401 | iconv_t *descp = p; |
402 | iconv_close (*descp); | |
234b45d4 KB |
403 | } |
404 | ||
6c7a06a3 TT |
405 | void |
406 | convert_between_encodings (const char *from, const char *to, | |
407 | const gdb_byte *bytes, unsigned int num_bytes, | |
408 | int width, struct obstack *output, | |
409 | enum transliterations translit) | |
410 | { | |
411 | iconv_t desc; | |
412 | struct cleanup *cleanups; | |
413 | size_t inleft; | |
414 | char *inp; | |
415 | unsigned int space_request; | |
416 | ||
417 | /* Often, the host and target charsets will be the same. */ | |
418 | if (!strcmp (from, to)) | |
419 | { | |
420 | obstack_grow (output, bytes, num_bytes); | |
421 | return; | |
422 | } | |
234b45d4 | 423 | |
6c7a06a3 TT |
424 | desc = iconv_open (to, from); |
425 | if (desc == (iconv_t) -1) | |
426 | perror_with_name ("Converting character sets"); | |
427 | cleanups = make_cleanup (cleanup_iconv, &desc); | |
234b45d4 | 428 | |
6c7a06a3 TT |
429 | inleft = num_bytes; |
430 | inp = (char *) bytes; | |
234b45d4 | 431 | |
6c7a06a3 | 432 | space_request = num_bytes; |
234b45d4 | 433 | |
6c7a06a3 | 434 | while (inleft > 0) |
234b45d4 | 435 | { |
6c7a06a3 TT |
436 | char *outp; |
437 | size_t outleft, r; | |
438 | int old_size; | |
439 | ||
440 | old_size = obstack_object_size (output); | |
441 | obstack_blank (output, space_request); | |
442 | ||
443 | outp = obstack_base (output) + old_size; | |
444 | outleft = space_request; | |
445 | ||
0dd7fb99 | 446 | r = iconv (desc, (ICONV_CONST char **) &inp, &inleft, &outp, &outleft); |
6c7a06a3 TT |
447 | |
448 | /* Now make sure that the object on the obstack only includes | |
449 | bytes we have converted. */ | |
450 | obstack_blank (output, - (int) outleft); | |
451 | ||
452 | if (r == (size_t) -1) | |
453 | { | |
454 | switch (errno) | |
455 | { | |
456 | case EILSEQ: | |
457 | { | |
458 | int i; | |
459 | ||
460 | /* Invalid input sequence. */ | |
461 | if (translit == translit_none) | |
462 | error (_("Could not convert character to `%s' character set"), | |
463 | to); | |
464 | ||
465 | /* We emit escape sequence for the bytes, skip them, | |
466 | and try again. */ | |
467 | for (i = 0; i < width; ++i) | |
468 | { | |
469 | char octal[5]; | |
470 | ||
471 | sprintf (octal, "\\%.3o", *inp & 0xff); | |
472 | obstack_grow_str (output, octal); | |
473 | ||
474 | ++inp; | |
475 | --inleft; | |
476 | } | |
477 | } | |
478 | break; | |
479 | ||
480 | case E2BIG: | |
481 | /* We ran out of space in the output buffer. Make it | |
482 | bigger next time around. */ | |
483 | space_request *= 2; | |
484 | break; | |
485 | ||
486 | case EINVAL: | |
487 | /* Incomplete input sequence. FIXME: ought to report this | |
488 | to the caller somehow. */ | |
489 | inleft = 0; | |
490 | break; | |
491 | ||
492 | default: | |
493 | perror_with_name ("Internal error while converting character sets"); | |
494 | } | |
495 | } | |
234b45d4 | 496 | } |
234b45d4 | 497 | |
6c7a06a3 | 498 | do_cleanups (cleanups); |
234b45d4 KB |
499 | } |
500 | ||
e33d66ec | 501 | \f |
e33d66ec | 502 | |
6c7a06a3 TT |
503 | /* An iterator that returns host wchar_t's from a target string. */ |
504 | struct wchar_iterator | |
e33d66ec | 505 | { |
6c7a06a3 TT |
506 | /* The underlying iconv descriptor. */ |
507 | iconv_t desc; | |
e33d66ec | 508 | |
6c7a06a3 TT |
509 | /* The input string. This is updated as convert characters. */ |
510 | char *input; | |
511 | /* The number of bytes remaining in the input. */ | |
512 | size_t bytes; | |
e33d66ec | 513 | |
6c7a06a3 TT |
514 | /* The width of an input character. */ |
515 | size_t width; | |
e33d66ec | 516 | |
6c7a06a3 TT |
517 | /* The output buffer and its size. */ |
518 | gdb_wchar_t *out; | |
519 | size_t out_size; | |
520 | }; | |
234b45d4 | 521 | |
6c7a06a3 TT |
522 | /* Create a new iterator. */ |
523 | struct wchar_iterator * | |
524 | make_wchar_iterator (const gdb_byte *input, size_t bytes, const char *charset, | |
525 | size_t width) | |
234b45d4 | 526 | { |
6c7a06a3 TT |
527 | struct wchar_iterator *result; |
528 | iconv_t desc; | |
234b45d4 | 529 | |
6c7a06a3 TT |
530 | desc = iconv_open ("wchar_t", charset); |
531 | if (desc == (iconv_t) -1) | |
532 | perror_with_name ("Converting character sets"); | |
234b45d4 | 533 | |
6c7a06a3 TT |
534 | result = XNEW (struct wchar_iterator); |
535 | result->desc = desc; | |
536 | result->input = (char *) input; | |
537 | result->bytes = bytes; | |
538 | result->width = width; | |
234b45d4 | 539 | |
6c7a06a3 TT |
540 | result->out = XNEW (gdb_wchar_t); |
541 | result->out_size = 1; | |
234b45d4 | 542 | |
6c7a06a3 | 543 | return result; |
e33d66ec | 544 | } |
234b45d4 | 545 | |
e33d66ec | 546 | static void |
6c7a06a3 | 547 | do_cleanup_iterator (void *p) |
e33d66ec | 548 | { |
6c7a06a3 | 549 | struct wchar_iterator *iter = p; |
234b45d4 | 550 | |
6c7a06a3 TT |
551 | iconv_close (iter->desc); |
552 | xfree (iter->out); | |
553 | xfree (iter); | |
234b45d4 KB |
554 | } |
555 | ||
6c7a06a3 TT |
556 | struct cleanup * |
557 | make_cleanup_wchar_iterator (struct wchar_iterator *iter) | |
e33d66ec | 558 | { |
6c7a06a3 | 559 | return make_cleanup (do_cleanup_iterator, iter); |
e33d66ec | 560 | } |
234b45d4 | 561 | |
6c7a06a3 TT |
562 | int |
563 | wchar_iterate (struct wchar_iterator *iter, | |
564 | enum wchar_iterate_result *out_result, | |
565 | gdb_wchar_t **out_chars, | |
566 | const gdb_byte **ptr, | |
567 | size_t *len) | |
568 | { | |
569 | size_t out_request; | |
570 | ||
571 | /* Try to convert some characters. At first we try to convert just | |
572 | a single character. The reason for this is that iconv does not | |
573 | necessarily update its outgoing arguments when it encounters an | |
574 | invalid input sequence -- but we want to reliably report this to | |
575 | our caller so it can emit an escape sequence. */ | |
576 | out_request = 1; | |
577 | while (iter->bytes > 0) | |
e33d66ec | 578 | { |
6c7a06a3 TT |
579 | char *outptr = (char *) &iter->out[0]; |
580 | char *orig_inptr = iter->input; | |
581 | size_t orig_in = iter->bytes; | |
582 | size_t out_avail = out_request * sizeof (gdb_wchar_t); | |
583 | size_t num; | |
584 | gdb_wchar_t result; | |
585 | ||
0dd7fb99 TT |
586 | size_t r = iconv (iter->desc, |
587 | (ICONV_CONST char **) &iter->input, &iter->bytes, | |
6c7a06a3 TT |
588 | &outptr, &out_avail); |
589 | if (r == (size_t) -1) | |
590 | { | |
591 | switch (errno) | |
592 | { | |
593 | case EILSEQ: | |
594 | /* Invalid input sequence. Skip it, and let the caller | |
595 | know about it. */ | |
596 | *out_result = wchar_iterate_invalid; | |
597 | *ptr = iter->input; | |
598 | *len = iter->width; | |
599 | iter->input += iter->width; | |
600 | iter->bytes -= iter->width; | |
601 | return 0; | |
602 | ||
603 | case E2BIG: | |
604 | /* We ran out of space. We still might have converted a | |
605 | character; if so, return it. Otherwise, grow the | |
606 | buffer and try again. */ | |
607 | if (out_avail < out_request * sizeof (gdb_wchar_t)) | |
608 | break; | |
609 | ||
610 | ++out_request; | |
611 | if (out_request > iter->out_size) | |
612 | { | |
613 | iter->out_size = out_request; | |
614 | iter->out = xrealloc (iter->out, | |
615 | out_request * sizeof (gdb_wchar_t)); | |
616 | } | |
617 | continue; | |
618 | ||
619 | case EINVAL: | |
620 | /* Incomplete input sequence. Let the caller know, and | |
621 | arrange for future calls to see EOF. */ | |
622 | *out_result = wchar_iterate_incomplete; | |
623 | *ptr = iter->input; | |
624 | *len = iter->bytes; | |
625 | iter->bytes = 0; | |
626 | return 0; | |
627 | ||
628 | default: | |
629 | perror_with_name ("Internal error while converting character sets"); | |
630 | } | |
631 | } | |
632 | ||
633 | /* We converted something. */ | |
634 | num = out_request - out_avail / sizeof (gdb_wchar_t); | |
635 | *out_result = wchar_iterate_ok; | |
636 | *out_chars = iter->out; | |
637 | *ptr = orig_inptr; | |
638 | *len = orig_in - iter->bytes; | |
639 | return num; | |
e33d66ec | 640 | } |
6c7a06a3 TT |
641 | |
642 | /* Really done. */ | |
643 | *out_result = wchar_iterate_eof; | |
644 | return -1; | |
234b45d4 KB |
645 | } |
646 | ||
e33d66ec | 647 | \f |
6c7a06a3 | 648 | /* The charset.c module initialization function. */ |
234b45d4 | 649 | |
6c7a06a3 | 650 | extern initialize_file_ftype _initialize_charset; /* -Wmissing-prototype */ |
234b45d4 | 651 | |
6c7a06a3 TT |
652 | typedef char *char_ptr; |
653 | DEF_VEC_P (char_ptr); | |
234b45d4 | 654 | |
6c7a06a3 | 655 | static VEC (char_ptr) *charsets; |
234b45d4 | 656 | |
6c7a06a3 | 657 | #ifdef PHONY_ICONV |
234b45d4 | 658 | |
6c7a06a3 TT |
659 | static void |
660 | find_charset_names (void) | |
234b45d4 | 661 | { |
6c7a06a3 TT |
662 | VEC_safe_push (char_ptr, charsets, GDB_DEFAULT_HOST_CHARSET); |
663 | VEC_safe_push (char_ptr, charsets, NULL); | |
234b45d4 KB |
664 | } |
665 | ||
6c7a06a3 | 666 | #else /* PHONY_ICONV */ |
fc3b640d TT |
667 | |
668 | /* Sometimes, libiconv redefines iconvlist as libiconvlist -- but | |
669 | provides different symbols in the static and dynamic libraries. | |
670 | So, configure may see libiconvlist but not iconvlist. But, calling | |
671 | iconvlist is the right thing to do and will work. Hence we do a | |
672 | check here but unconditionally call iconvlist below. */ | |
673 | #if defined (HAVE_ICONVLIST) || defined (HAVE_LIBICONVLIST) | |
234b45d4 | 674 | |
6c7a06a3 TT |
675 | /* A helper function that adds some character sets to the vector of |
676 | all character sets. This is a callback function for iconvlist. */ | |
677 | ||
678 | static int | |
679 | add_one (unsigned int count, const char *const *names, void *data) | |
234b45d4 | 680 | { |
6c7a06a3 | 681 | unsigned int i; |
234b45d4 | 682 | |
6c7a06a3 TT |
683 | for (i = 0; i < count; ++i) |
684 | VEC_safe_push (char_ptr, charsets, xstrdup (names[i])); | |
234b45d4 | 685 | |
6c7a06a3 | 686 | return 0; |
234b45d4 KB |
687 | } |
688 | ||
6c7a06a3 TT |
689 | static void |
690 | find_charset_names (void) | |
234b45d4 | 691 | { |
6c7a06a3 TT |
692 | iconvlist (add_one, NULL); |
693 | VEC_safe_push (char_ptr, charsets, NULL); | |
234b45d4 KB |
694 | } |
695 | ||
6c7a06a3 | 696 | #else |
234b45d4 | 697 | |
6c7a06a3 TT |
698 | static void |
699 | find_charset_names (void) | |
234b45d4 | 700 | { |
6c7a06a3 | 701 | FILE *in; |
234b45d4 | 702 | |
6c7a06a3 TT |
703 | in = popen ("iconv -l", "r"); |
704 | /* It is ok to ignore errors; we'll fall back on a default. */ | |
705 | if (!in) | |
706 | return; | |
234b45d4 | 707 | |
6c7a06a3 TT |
708 | /* POSIX says that iconv -l uses an unspecified format. We parse |
709 | the glibc format; feel free to add others as needed. */ | |
710 | while (!feof (in)) | |
711 | { | |
712 | /* The size of buf is chosen arbitrarily. A character set name | |
713 | longer than this would not be very nice. */ | |
714 | char buf[80]; | |
715 | int len; | |
716 | char *r = fgets (buf, sizeof (buf), in); | |
717 | if (!r) | |
718 | break; | |
719 | len = strlen (r); | |
720 | if (len <= 3) | |
721 | continue; | |
722 | if (buf[len - 2] == '/' && buf[len - 3] == '/') | |
723 | buf[len - 3] = '\0'; | |
724 | VEC_safe_push (char_ptr, charsets, xstrdup (buf)); | |
725 | } | |
234b45d4 | 726 | |
6c7a06a3 | 727 | pclose (in); |
234b45d4 | 728 | |
6c7a06a3 TT |
729 | VEC_safe_push (char_ptr, charsets, NULL); |
730 | } | |
234b45d4 | 731 | |
fc3b640d | 732 | #endif /* HAVE_ICONVLIST || HAVE_LIBICONVLIST */ |
6c7a06a3 | 733 | #endif /* PHONY_ICONV */ |
234b45d4 KB |
734 | |
735 | void | |
736 | _initialize_charset (void) | |
737 | { | |
e33d66ec EZ |
738 | struct cmd_list_element *new_cmd; |
739 | ||
6c7a06a3 TT |
740 | /* The first element is always "auto"; then we skip it for the |
741 | commands where it is not allowed. */ | |
742 | VEC_safe_push (char_ptr, charsets, "auto"); | |
743 | find_charset_names (); | |
744 | ||
745 | if (VEC_length (char_ptr, charsets) > 1) | |
746 | charset_enum = (const char **) VEC_address (char_ptr, charsets); | |
747 | else | |
748 | charset_enum = default_charset_names; | |
749 | ||
750 | #ifndef PHONY_ICONV | |
751 | #ifdef HAVE_LANGINFO_CODESET | |
752 | auto_host_charset_name = nl_langinfo (CODESET); | |
753 | target_charset_name = auto_host_charset_name; | |
754 | ||
755 | set_be_le_names (); | |
756 | #endif | |
757 | #endif | |
e33d66ec | 758 | |
7ab04401 | 759 | add_setshow_enum_cmd ("charset", class_support, |
6c7a06a3 | 760 | &charset_enum[1], &host_charset_name, _("\ |
7ab04401 AC |
761 | Set the host and target character sets."), _("\ |
762 | Show the host and target character sets."), _("\ | |
3d263c1d BI |
763 | The `host character set' is the one used by the system GDB is running on.\n\ |
764 | The `target character set' is the one used by the program being debugged.\n\ | |
765 | You may only use supersets of ASCII for your host character set; GDB does\n\ | |
766 | not support any others.\n\ | |
767 | To see a list of the character sets GDB supports, type `set charset <TAB>'."), | |
7ab04401 AC |
768 | /* Note that the sfunc below needs to set |
769 | target_charset_name, because the 'set | |
770 | charset' command sets two variables. */ | |
771 | set_charset_sfunc, | |
772 | show_charset, | |
773 | &setlist, &showlist); | |
774 | ||
775 | add_setshow_enum_cmd ("host-charset", class_support, | |
6c7a06a3 | 776 | charset_enum, &host_charset_name, _("\ |
7ab04401 AC |
777 | Set the host character set."), _("\ |
778 | Show the host character set."), _("\ | |
3d263c1d BI |
779 | The `host character set' is the one used by the system GDB is running on.\n\ |
780 | You may only use supersets of ASCII for your host character set; GDB does\n\ | |
781 | not support any others.\n\ | |
782 | To see a list of the character sets GDB supports, type `set host-charset <TAB>'."), | |
7ab04401 | 783 | set_host_charset_sfunc, |
920d2a44 | 784 | show_host_charset_name, |
7ab04401 AC |
785 | &setlist, &showlist); |
786 | ||
787 | add_setshow_enum_cmd ("target-charset", class_support, | |
6c7a06a3 | 788 | &charset_enum[1], &target_charset_name, _("\ |
7ab04401 AC |
789 | Set the target character set."), _("\ |
790 | Show the target character set."), _("\ | |
3d263c1d BI |
791 | The `target character set' is the one used by the program being debugged.\n\ |
792 | GDB translates characters and strings between the host and target\n\ | |
793 | character sets as needed.\n\ | |
794 | To see a list of the character sets GDB supports, type `set target-charset'<TAB>"), | |
7ab04401 | 795 | set_target_charset_sfunc, |
920d2a44 | 796 | show_target_charset_name, |
7ab04401 | 797 | &setlist, &showlist); |
6c7a06a3 TT |
798 | |
799 | add_setshow_enum_cmd ("target-wide-charset", class_support, | |
800 | &charset_enum[1], &target_wide_charset_name, | |
801 | _("\ | |
802 | Set the target wide character set."), _("\ | |
803 | Show the target wide character set."), _("\ | |
804 | The `target wide character set' is the one used by the program being debugged.\n\ | |
805 | In particular it is the encoding used by `wchar_t'.\n\ | |
806 | GDB translates characters and strings between the host and target\n\ | |
807 | character sets as needed.\n\ | |
808 | To see a list of the character sets GDB supports, type\n\ | |
809 | `set target-wide-charset'<TAB>"), | |
810 | set_target_wide_charset_sfunc, | |
811 | show_target_wide_charset_name, | |
812 | &setlist, &showlist); | |
234b45d4 | 813 | } |