]>
Commit | Line | Data |
---|---|---|
b9305c66 | 1 | /* Mainly the interface between cpplib and the C front ends. |
517cbe13 | 2 | Copyright (C) 1987, 1988, 1989, 1992, 1994, 1995, 1996, 1997 |
5793b276 | 3 | 1998, 1999, 2000, 2001, 2002, 2003 Free Software Foundation, Inc. |
e8bbfc4e | 4 | |
1322177d | 5 | This file is part of GCC. |
e8bbfc4e | 6 | |
1322177d LB |
7 | GCC is free software; you can redistribute it and/or modify it under |
8 | the terms of the GNU General Public License as published by the Free | |
9 | Software Foundation; either version 2, or (at your option) any later | |
10 | version. | |
e8bbfc4e | 11 | |
1322177d LB |
12 | GCC is distributed in the hope that it will be useful, but WITHOUT ANY |
13 | WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
14 | FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License | |
15 | for more details. | |
e8bbfc4e RK |
16 | |
17 | You should have received a copy of the GNU General Public License | |
1322177d LB |
18 | along with GCC; see the file COPYING. If not, write to the Free |
19 | Software Foundation, 59 Temple Place - Suite 330, Boston, MA | |
20 | 02111-1307, USA. */ | |
e8bbfc4e | 21 | |
e9a25f70 | 22 | #include "config.h" |
670ee920 | 23 | #include "system.h" |
4977bab6 ZW |
24 | #include "coretypes.h" |
25 | #include "tm.h" | |
e8bbfc4e | 26 | |
11ad4784 | 27 | #include "real.h" |
e8bbfc4e RK |
28 | #include "rtl.h" |
29 | #include "tree.h" | |
eb3aaa5b | 30 | #include "expr.h" |
e8bbfc4e | 31 | #include "input.h" |
d6f4ec51 | 32 | #include "output.h" |
e8bbfc4e | 33 | #include "c-tree.h" |
52dabb6c | 34 | #include "c-common.h" |
e8bbfc4e | 35 | #include "flags.h" |
0e5921e8 | 36 | #include "timevar.h" |
8b97c5f8 | 37 | #include "cpplib.h" |
3d6f7931 | 38 | #include "c-pragma.h" |
5f6da302 | 39 | #include "toplev.h" |
ab87f8c8 | 40 | #include "intl.h" |
7bdb32b9 | 41 | #include "tm_p.h" |
0e5921e8 | 42 | #include "splay-tree.h" |
7f905405 | 43 | #include "debug.h" |
ab87f8c8 | 44 | |
e8bbfc4e | 45 | #ifdef MULTIBYTE_CHARS |
56f48ce9 | 46 | #include "mbchar.h" |
e8bbfc4e | 47 | #include <locale.h> |
56f48ce9 | 48 | #endif /* MULTIBYTE_CHARS */ |
e8bbfc4e | 49 | |
67821e3a | 50 | /* The current line map. */ |
47d89cf3 | 51 | static const struct line_map *map; |
67821e3a | 52 | |
97293897 NB |
53 | /* The line used to refresh the lineno global variable after each token. */ |
54 | static unsigned int src_lineno; | |
55 | ||
0e5921e8 ZW |
56 | /* We may keep statistics about how long which files took to compile. */ |
57 | static int header_time, body_time; | |
58 | static splay_tree file_info_tree; | |
3ab6dd7c | 59 | |
e8bbfc4e RK |
60 | /* File used for outputting assembler code. */ |
61 | extern FILE *asm_out_file; | |
62 | ||
12a39b12 JM |
63 | #undef WCHAR_TYPE_SIZE |
64 | #define WCHAR_TYPE_SIZE TYPE_PRECISION (wchar_type_node) | |
e8bbfc4e RK |
65 | |
66 | /* Number of bytes in a wide character. */ | |
67 | #define WCHAR_BYTES (WCHAR_TYPE_SIZE / BITS_PER_UNIT) | |
68 | ||
0e5921e8 ZW |
69 | int pending_lang_change; /* If we need to switch languages - C++ only */ |
70 | int c_header_level; /* depth in C headers - C++ only */ | |
fbb18613 JM |
71 | |
72 | /* Nonzero tells yylex to ignore \ in string constants. */ | |
73 | static int ignore_escape_flag; | |
e9a25f70 | 74 | |
ceeedfc1 NB |
75 | static tree interpret_integer PARAMS ((const cpp_token *, unsigned int)); |
76 | static tree interpret_float PARAMS ((const cpp_token *, unsigned int)); | |
77 | static enum integer_type_kind | |
78 | narrowest_unsigned_type PARAMS ((tree, unsigned int)); | |
79 | static enum integer_type_kind | |
80 | narrowest_signed_type PARAMS ((tree, unsigned int)); | |
62ae2529 JJ |
81 | static tree lex_string PARAMS ((const unsigned char *, unsigned int, |
82 | int)); | |
c8a96070 | 83 | static tree lex_charconst PARAMS ((const cpp_token *)); |
0e5921e8 ZW |
84 | static void update_header_times PARAMS ((const char *)); |
85 | static int dump_one_header PARAMS ((splay_tree_node, void *)); | |
97293897 | 86 | static void cb_line_change PARAMS ((cpp_reader *, const cpp_token *, int)); |
8bbbef34 NB |
87 | static void cb_ident PARAMS ((cpp_reader *, unsigned int, |
88 | const cpp_string *)); | |
47d89cf3 | 89 | static void cb_file_change PARAMS ((cpp_reader *, const struct line_map *)); |
8bbbef34 NB |
90 | static void cb_def_pragma PARAMS ((cpp_reader *, unsigned int)); |
91 | static void cb_define PARAMS ((cpp_reader *, unsigned int, | |
92 | cpp_hashnode *)); | |
93 | static void cb_undef PARAMS ((cpp_reader *, unsigned int, | |
94 | cpp_hashnode *)); | |
e31c7eec | 95 | \f |
63973df3 NB |
96 | void |
97 | init_c_lex () | |
e3d1fd32 | 98 | { |
b61c5ed0 | 99 | struct cpp_callbacks *cb; |
0e5921e8 ZW |
100 | struct c_fileinfo *toplevel; |
101 | ||
f5e99456 | 102 | /* Set up filename timing. Must happen before cpp_read_main_file. */ |
0e5921e8 ZW |
103 | file_info_tree = splay_tree_new ((splay_tree_compare_fn)strcmp, |
104 | 0, | |
105 | (splay_tree_delete_value_fn)free); | |
a8a05998 | 106 | toplevel = get_fileinfo ("<top level>"); |
0e5921e8 ZW |
107 | if (flag_detailed_statistics) |
108 | { | |
109 | header_time = 0; | |
110 | body_time = get_run_time (); | |
111 | toplevel->time = body_time; | |
112 | } | |
113 | ||
b61c5ed0 NB |
114 | cb = cpp_get_callbacks (parse_in); |
115 | ||
9d10c9a9 | 116 | cb->register_builtins = cb_register_builtins; |
97293897 | 117 | cb->line_change = cb_line_change; |
b61c5ed0 NB |
118 | cb->ident = cb_ident; |
119 | cb->file_change = cb_file_change; | |
120 | cb->def_pragma = cb_def_pragma; | |
17211ab5 GK |
121 | cb->valid_pch = c_common_valid_pch; |
122 | cb->read_pch = c_common_read_pch; | |
0e5921e8 | 123 | |
65289a3a NB |
124 | /* Set the debug callbacks if we can use them. */ |
125 | if (debug_info_level == DINFO_LEVEL_VERBOSE | |
7a0c8d71 DR |
126 | && (write_symbols == DWARF_DEBUG || write_symbols == DWARF2_DEBUG |
127 | || write_symbols == VMS_AND_DWARF2_DEBUG)) | |
65289a3a | 128 | { |
b61c5ed0 NB |
129 | cb->define = cb_define; |
130 | cb->undef = cb_undef; | |
65289a3a | 131 | } |
e3d1fd32 PB |
132 | } |
133 | ||
3aac38d7 | 134 | /* A thin wrapper around the real parser that initializes the |
bebc7e8b ZW |
135 | integrated preprocessor after debug output has been initialized. |
136 | Also, make sure the start_source_file debug hook gets called for | |
137 | the primary source file. */ | |
3aac38d7 | 138 | |
52dabb6c | 139 | void |
ff45c01e NB |
140 | c_common_parse_file (set_yydebug) |
141 | int set_yydebug ATTRIBUTE_UNUSED; | |
3aac38d7 | 142 | { |
ff45c01e NB |
143 | #if YYDEBUG != 0 |
144 | yydebug = set_yydebug; | |
145 | #else | |
146 | warning ("YYDEBUG not defined"); | |
147 | #endif | |
148 | ||
bebc7e8b | 149 | (*debug_hooks->start_source_file) (lineno, input_filename); |
f5e99456 | 150 | cpp_finish_options (parse_in); |
3aac38d7 | 151 | |
17211ab5 GK |
152 | pch_init(); |
153 | ||
52dabb6c | 154 | yyparse (); |
94a50397 | 155 | free_parser_stacks (); |
3aac38d7 RH |
156 | } |
157 | ||
0e5921e8 ZW |
158 | struct c_fileinfo * |
159 | get_fileinfo (name) | |
160 | const char *name; | |
e3d1fd32 | 161 | { |
0e5921e8 ZW |
162 | splay_tree_node n; |
163 | struct c_fileinfo *fi; | |
164 | ||
165 | n = splay_tree_lookup (file_info_tree, (splay_tree_key) name); | |
166 | if (n) | |
167 | return (struct c_fileinfo *) n->value; | |
168 | ||
169 | fi = (struct c_fileinfo *) xmalloc (sizeof (struct c_fileinfo)); | |
170 | fi->time = 0; | |
171 | fi->interface_only = 0; | |
172 | fi->interface_unknown = 1; | |
173 | splay_tree_insert (file_info_tree, (splay_tree_key) name, | |
174 | (splay_tree_value) fi); | |
175 | return fi; | |
e56e519d | 176 | } |
e3d1fd32 | 177 | |
0e5921e8 ZW |
178 | static void |
179 | update_header_times (name) | |
180 | const char *name; | |
e8bbfc4e | 181 | { |
0e5921e8 ZW |
182 | /* Changing files again. This means currently collected time |
183 | is charged against header time, and body time starts back at 0. */ | |
184 | if (flag_detailed_statistics) | |
e8bbfc4e | 185 | { |
0e5921e8 ZW |
186 | int this_time = get_run_time (); |
187 | struct c_fileinfo *file = get_fileinfo (name); | |
188 | header_time += this_time - body_time; | |
189 | file->time += this_time - body_time; | |
190 | body_time = this_time; | |
e8bbfc4e RK |
191 | } |
192 | } | |
193 | ||
0e5921e8 ZW |
194 | static int |
195 | dump_one_header (n, dummy) | |
196 | splay_tree_node n; | |
197 | void *dummy ATTRIBUTE_UNUSED; | |
e8bbfc4e | 198 | { |
0e5921e8 ZW |
199 | print_time ((const char *) n->key, |
200 | ((struct c_fileinfo *) n->value)->time); | |
201 | return 0; | |
e8bbfc4e | 202 | } |
e8bbfc4e RK |
203 | |
204 | void | |
0e5921e8 | 205 | dump_time_statistics () |
e8bbfc4e | 206 | { |
0e5921e8 ZW |
207 | struct c_fileinfo *file = get_fileinfo (input_filename); |
208 | int this_time = get_run_time (); | |
209 | file->time += this_time - body_time; | |
210 | ||
211 | fprintf (stderr, "\n******\n"); | |
212 | print_time ("header files (total)", header_time); | |
213 | print_time ("main file (total)", this_time - body_time); | |
214 | fprintf (stderr, "ratio = %g : 1\n", | |
215 | (double)header_time / (double)(this_time - body_time)); | |
216 | fprintf (stderr, "\n******\n"); | |
217 | ||
218 | splay_tree_foreach (file_info_tree, dump_one_header, 0); | |
e8bbfc4e | 219 | } |
a6124a42 | 220 | |
0e5921e8 | 221 | static void |
8bbbef34 | 222 | cb_ident (pfile, line, str) |
27e2564a | 223 | cpp_reader *pfile ATTRIBUTE_UNUSED; |
8bbbef34 | 224 | unsigned int line ATTRIBUTE_UNUSED; |
f5720527 | 225 | const cpp_string *str ATTRIBUTE_UNUSED; |
0e5921e8 | 226 | { |
0e5921e8 | 227 | #ifdef ASM_OUTPUT_IDENT |
27e2564a | 228 | if (! flag_no_ident) |
0e5921e8 | 229 | { |
27e2564a | 230 | /* Convert escapes in the string. */ |
4977bab6 | 231 | tree value ATTRIBUTE_UNUSED = lex_string (str->text, str->len, 0); |
27e2564a | 232 | ASM_OUTPUT_IDENT (asm_out_file, TREE_STRING_POINTER (value)); |
0e5921e8 ZW |
233 | } |
234 | #endif | |
27e2564a NB |
235 | } |
236 | ||
97293897 NB |
237 | /* Called at the start of every non-empty line. TOKEN is the first |
238 | lexed token on the line. Used for diagnostic line numbers. */ | |
239 | static void | |
240 | cb_line_change (pfile, token, parsing_args) | |
241 | cpp_reader *pfile ATTRIBUTE_UNUSED; | |
242 | const cpp_token *token; | |
243 | int parsing_args ATTRIBUTE_UNUSED; | |
244 | { | |
245 | src_lineno = SOURCE_LINE (map, token->line); | |
246 | } | |
247 | ||
27e2564a | 248 | static void |
47d89cf3 | 249 | cb_file_change (pfile, new_map) |
27e2564a | 250 | cpp_reader *pfile ATTRIBUTE_UNUSED; |
47d89cf3 | 251 | const struct line_map *new_map; |
27e2564a | 252 | { |
47d89cf3 | 253 | unsigned int to_line = SOURCE_LINE (new_map, new_map->to_line); |
d82fc108 | 254 | |
47d89cf3 | 255 | if (new_map->reason == LC_ENTER) |
fbb18613 | 256 | { |
5bea1ccf JM |
257 | /* Don't stack the main buffer on the input stack; |
258 | we already did in compile_file. */ | |
47d89cf3 NB |
259 | if (map == NULL) |
260 | main_input_filename = new_map->to_file; | |
d82fc108 | 261 | else |
0e5921e8 | 262 | { |
f78ce0b7 JB |
263 | int included_at = SOURCE_LINE (new_map - 1, new_map->from_line - 1); |
264 | ||
265 | lineno = included_at; | |
47d89cf3 | 266 | push_srcloc (new_map->to_file, 1); |
f78ce0b7 | 267 | (*debug_hooks->start_source_file) (included_at, new_map->to_file); |
27e2564a NB |
268 | #ifndef NO_IMPLICIT_EXTERN_C |
269 | if (c_header_level) | |
270 | ++c_header_level; | |
47d89cf3 | 271 | else if (new_map->sysp == 2) |
27e2564a NB |
272 | { |
273 | c_header_level = 1; | |
274 | ++pending_lang_change; | |
275 | } | |
0e5921e8 | 276 | #endif |
27e2564a | 277 | } |
fbb18613 | 278 | } |
47d89cf3 | 279 | else if (new_map->reason == LC_LEAVE) |
fbb18613 | 280 | { |
0e5921e8 | 281 | #ifndef NO_IMPLICIT_EXTERN_C |
47d89cf3 NB |
282 | if (c_header_level && --c_header_level == 0) |
283 | { | |
284 | if (new_map->sysp == 2) | |
285 | warning ("badly nested C headers from preprocessor"); | |
286 | --pending_lang_change; | |
287 | } | |
47d89cf3 NB |
288 | #endif |
289 | pop_srcloc (); | |
290 | ||
291 | (*debug_hooks->end_source_file) (to_line); | |
e8bbfc4e | 292 | } |
fbb18613 | 293 | |
47d89cf3 NB |
294 | update_header_times (new_map->to_file); |
295 | in_system_header = new_map->sysp != 0; | |
296 | input_filename = new_map->to_file; | |
297 | lineno = to_line; | |
298 | map = new_map; | |
0e5921e8 | 299 | |
0e5921e8 ZW |
300 | /* Hook for C++. */ |
301 | extract_interface_info (); | |
302 | } | |
8b97c5f8 ZW |
303 | |
304 | static void | |
8bbbef34 | 305 | cb_def_pragma (pfile, line) |
8b97c5f8 | 306 | cpp_reader *pfile; |
67821e3a | 307 | unsigned int line; |
8b97c5f8 ZW |
308 | { |
309 | /* Issue a warning message if we have been asked to do so. Ignore | |
310 | unknown pragmas in system headers unless an explicit | |
ec5c56db | 311 | -Wunknown-pragmas has been given. */ |
8b97c5f8 ZW |
312 | if (warn_unknown_pragmas > in_system_header) |
313 | { | |
06470238 | 314 | const unsigned char *space, *name; |
4ed5bcfb | 315 | const cpp_token *s; |
23356f93 | 316 | |
06470238 | 317 | space = name = (const unsigned char *) ""; |
4ed5bcfb | 318 | s = cpp_get_token (pfile); |
06470238 NB |
319 | if (s->type != CPP_EOF) |
320 | { | |
321 | space = cpp_token_as_text (pfile, s); | |
322 | s = cpp_get_token (pfile); | |
323 | if (s->type == CPP_NAME) | |
324 | name = cpp_token_as_text (pfile, s); | |
325 | } | |
8b97c5f8 | 326 | |
67821e3a | 327 | lineno = SOURCE_LINE (map, line); |
06470238 | 328 | warning ("ignoring #pragma %s %s", space, name); |
8b97c5f8 ZW |
329 | } |
330 | } | |
0e5921e8 | 331 | |
65289a3a NB |
332 | /* #define callback for DWARF and DWARF2 debug info. */ |
333 | static void | |
8bbbef34 | 334 | cb_define (pfile, line, node) |
65289a3a | 335 | cpp_reader *pfile; |
67821e3a | 336 | unsigned int line; |
65289a3a NB |
337 | cpp_hashnode *node; |
338 | { | |
67821e3a | 339 | (*debug_hooks->define) (SOURCE_LINE (map, line), |
7f905405 | 340 | (const char *) cpp_macro_definition (pfile, node)); |
65289a3a NB |
341 | } |
342 | ||
343 | /* #undef callback for DWARF and DWARF2 debug info. */ | |
344 | static void | |
8bbbef34 | 345 | cb_undef (pfile, line, node) |
67821e3a NB |
346 | cpp_reader *pfile ATTRIBUTE_UNUSED; |
347 | unsigned int line; | |
65289a3a NB |
348 | cpp_hashnode *node; |
349 | { | |
67821e3a | 350 | (*debug_hooks->undef) (SOURCE_LINE (map, line), |
7f905405 | 351 | (const char *) NODE_NAME (node)); |
65289a3a NB |
352 | } |
353 | ||
0e5921e8 ZW |
354 | #if 0 /* not yet */ |
355 | /* Returns nonzero if C is a universal-character-name. Give an error if it | |
356 | is not one which may appear in an identifier, as per [extendid]. | |
357 | ||
358 | Note that extended character support in identifiers has not yet been | |
359 | implemented. It is my personal opinion that this is not a desirable | |
360 | feature. Portable code cannot count on support for more than the basic | |
361 | identifier character set. */ | |
362 | ||
363 | static inline int | |
364 | is_extended_char (c) | |
365 | int c; | |
366 | { | |
367 | #ifdef TARGET_EBCDIC | |
368 | return 0; | |
369 | #else | |
370 | /* ASCII. */ | |
371 | if (c < 0x7f) | |
372 | return 0; | |
373 | ||
374 | /* None of the valid chars are outside the Basic Multilingual Plane (the | |
375 | low 16 bits). */ | |
376 | if (c > 0xffff) | |
377 | { | |
378 | error ("universal-character-name '\\U%08x' not valid in identifier", c); | |
379 | return 1; | |
380 | } | |
381 | ||
382 | /* Latin */ | |
383 | if ((c >= 0x00c0 && c <= 0x00d6) | |
384 | || (c >= 0x00d8 && c <= 0x00f6) | |
385 | || (c >= 0x00f8 && c <= 0x01f5) | |
386 | || (c >= 0x01fa && c <= 0x0217) | |
387 | || (c >= 0x0250 && c <= 0x02a8) | |
388 | || (c >= 0x1e00 && c <= 0x1e9a) | |
389 | || (c >= 0x1ea0 && c <= 0x1ef9)) | |
390 | return 1; | |
391 | ||
392 | /* Greek */ | |
393 | if ((c == 0x0384) | |
394 | || (c >= 0x0388 && c <= 0x038a) | |
395 | || (c == 0x038c) | |
396 | || (c >= 0x038e && c <= 0x03a1) | |
397 | || (c >= 0x03a3 && c <= 0x03ce) | |
398 | || (c >= 0x03d0 && c <= 0x03d6) | |
399 | || (c == 0x03da) | |
400 | || (c == 0x03dc) | |
401 | || (c == 0x03de) | |
402 | || (c == 0x03e0) | |
403 | || (c >= 0x03e2 && c <= 0x03f3) | |
404 | || (c >= 0x1f00 && c <= 0x1f15) | |
405 | || (c >= 0x1f18 && c <= 0x1f1d) | |
406 | || (c >= 0x1f20 && c <= 0x1f45) | |
407 | || (c >= 0x1f48 && c <= 0x1f4d) | |
408 | || (c >= 0x1f50 && c <= 0x1f57) | |
409 | || (c == 0x1f59) | |
410 | || (c == 0x1f5b) | |
411 | || (c == 0x1f5d) | |
412 | || (c >= 0x1f5f && c <= 0x1f7d) | |
413 | || (c >= 0x1f80 && c <= 0x1fb4) | |
414 | || (c >= 0x1fb6 && c <= 0x1fbc) | |
415 | || (c >= 0x1fc2 && c <= 0x1fc4) | |
416 | || (c >= 0x1fc6 && c <= 0x1fcc) | |
417 | || (c >= 0x1fd0 && c <= 0x1fd3) | |
418 | || (c >= 0x1fd6 && c <= 0x1fdb) | |
419 | || (c >= 0x1fe0 && c <= 0x1fec) | |
420 | || (c >= 0x1ff2 && c <= 0x1ff4) | |
421 | || (c >= 0x1ff6 && c <= 0x1ffc)) | |
422 | return 1; | |
423 | ||
424 | /* Cyrillic */ | |
425 | if ((c >= 0x0401 && c <= 0x040d) | |
426 | || (c >= 0x040f && c <= 0x044f) | |
427 | || (c >= 0x0451 && c <= 0x045c) | |
428 | || (c >= 0x045e && c <= 0x0481) | |
429 | || (c >= 0x0490 && c <= 0x04c4) | |
430 | || (c >= 0x04c7 && c <= 0x04c8) | |
431 | || (c >= 0x04cb && c <= 0x04cc) | |
432 | || (c >= 0x04d0 && c <= 0x04eb) | |
433 | || (c >= 0x04ee && c <= 0x04f5) | |
434 | || (c >= 0x04f8 && c <= 0x04f9)) | |
435 | return 1; | |
436 | ||
437 | /* Armenian */ | |
438 | if ((c >= 0x0531 && c <= 0x0556) | |
439 | || (c >= 0x0561 && c <= 0x0587)) | |
440 | return 1; | |
441 | ||
442 | /* Hebrew */ | |
443 | if ((c >= 0x05d0 && c <= 0x05ea) | |
444 | || (c >= 0x05f0 && c <= 0x05f4)) | |
445 | return 1; | |
446 | ||
447 | /* Arabic */ | |
448 | if ((c >= 0x0621 && c <= 0x063a) | |
449 | || (c >= 0x0640 && c <= 0x0652) | |
450 | || (c >= 0x0670 && c <= 0x06b7) | |
451 | || (c >= 0x06ba && c <= 0x06be) | |
452 | || (c >= 0x06c0 && c <= 0x06ce) | |
453 | || (c >= 0x06e5 && c <= 0x06e7)) | |
454 | return 1; | |
455 | ||
456 | /* Devanagari */ | |
457 | if ((c >= 0x0905 && c <= 0x0939) | |
458 | || (c >= 0x0958 && c <= 0x0962)) | |
459 | return 1; | |
460 | ||
461 | /* Bengali */ | |
462 | if ((c >= 0x0985 && c <= 0x098c) | |
463 | || (c >= 0x098f && c <= 0x0990) | |
464 | || (c >= 0x0993 && c <= 0x09a8) | |
465 | || (c >= 0x09aa && c <= 0x09b0) | |
466 | || (c == 0x09b2) | |
467 | || (c >= 0x09b6 && c <= 0x09b9) | |
468 | || (c >= 0x09dc && c <= 0x09dd) | |
469 | || (c >= 0x09df && c <= 0x09e1) | |
470 | || (c >= 0x09f0 && c <= 0x09f1)) | |
471 | return 1; | |
472 | ||
473 | /* Gurmukhi */ | |
474 | if ((c >= 0x0a05 && c <= 0x0a0a) | |
475 | || (c >= 0x0a0f && c <= 0x0a10) | |
476 | || (c >= 0x0a13 && c <= 0x0a28) | |
477 | || (c >= 0x0a2a && c <= 0x0a30) | |
478 | || (c >= 0x0a32 && c <= 0x0a33) | |
479 | || (c >= 0x0a35 && c <= 0x0a36) | |
480 | || (c >= 0x0a38 && c <= 0x0a39) | |
481 | || (c >= 0x0a59 && c <= 0x0a5c) | |
482 | || (c == 0x0a5e)) | |
483 | return 1; | |
484 | ||
485 | /* Gujarati */ | |
486 | if ((c >= 0x0a85 && c <= 0x0a8b) | |
487 | || (c == 0x0a8d) | |
488 | || (c >= 0x0a8f && c <= 0x0a91) | |
489 | || (c >= 0x0a93 && c <= 0x0aa8) | |
490 | || (c >= 0x0aaa && c <= 0x0ab0) | |
491 | || (c >= 0x0ab2 && c <= 0x0ab3) | |
492 | || (c >= 0x0ab5 && c <= 0x0ab9) | |
493 | || (c == 0x0ae0)) | |
494 | return 1; | |
495 | ||
496 | /* Oriya */ | |
497 | if ((c >= 0x0b05 && c <= 0x0b0c) | |
498 | || (c >= 0x0b0f && c <= 0x0b10) | |
499 | || (c >= 0x0b13 && c <= 0x0b28) | |
500 | || (c >= 0x0b2a && c <= 0x0b30) | |
501 | || (c >= 0x0b32 && c <= 0x0b33) | |
502 | || (c >= 0x0b36 && c <= 0x0b39) | |
503 | || (c >= 0x0b5c && c <= 0x0b5d) | |
504 | || (c >= 0x0b5f && c <= 0x0b61)) | |
505 | return 1; | |
506 | ||
507 | /* Tamil */ | |
508 | if ((c >= 0x0b85 && c <= 0x0b8a) | |
509 | || (c >= 0x0b8e && c <= 0x0b90) | |
510 | || (c >= 0x0b92 && c <= 0x0b95) | |
511 | || (c >= 0x0b99 && c <= 0x0b9a) | |
512 | || (c == 0x0b9c) | |
513 | || (c >= 0x0b9e && c <= 0x0b9f) | |
514 | || (c >= 0x0ba3 && c <= 0x0ba4) | |
515 | || (c >= 0x0ba8 && c <= 0x0baa) | |
516 | || (c >= 0x0bae && c <= 0x0bb5) | |
517 | || (c >= 0x0bb7 && c <= 0x0bb9)) | |
518 | return 1; | |
519 | ||
520 | /* Telugu */ | |
521 | if ((c >= 0x0c05 && c <= 0x0c0c) | |
522 | || (c >= 0x0c0e && c <= 0x0c10) | |
523 | || (c >= 0x0c12 && c <= 0x0c28) | |
524 | || (c >= 0x0c2a && c <= 0x0c33) | |
525 | || (c >= 0x0c35 && c <= 0x0c39) | |
526 | || (c >= 0x0c60 && c <= 0x0c61)) | |
527 | return 1; | |
528 | ||
529 | /* Kannada */ | |
530 | if ((c >= 0x0c85 && c <= 0x0c8c) | |
531 | || (c >= 0x0c8e && c <= 0x0c90) | |
532 | || (c >= 0x0c92 && c <= 0x0ca8) | |
533 | || (c >= 0x0caa && c <= 0x0cb3) | |
534 | || (c >= 0x0cb5 && c <= 0x0cb9) | |
535 | || (c >= 0x0ce0 && c <= 0x0ce1)) | |
536 | return 1; | |
537 | ||
538 | /* Malayalam */ | |
539 | if ((c >= 0x0d05 && c <= 0x0d0c) | |
540 | || (c >= 0x0d0e && c <= 0x0d10) | |
541 | || (c >= 0x0d12 && c <= 0x0d28) | |
542 | || (c >= 0x0d2a && c <= 0x0d39) | |
543 | || (c >= 0x0d60 && c <= 0x0d61)) | |
544 | return 1; | |
545 | ||
546 | /* Thai */ | |
547 | if ((c >= 0x0e01 && c <= 0x0e30) | |
548 | || (c >= 0x0e32 && c <= 0x0e33) | |
549 | || (c >= 0x0e40 && c <= 0x0e46) | |
550 | || (c >= 0x0e4f && c <= 0x0e5b)) | |
551 | return 1; | |
552 | ||
553 | /* Lao */ | |
554 | if ((c >= 0x0e81 && c <= 0x0e82) | |
555 | || (c == 0x0e84) | |
556 | || (c == 0x0e87) | |
557 | || (c == 0x0e88) | |
558 | || (c == 0x0e8a) | |
559 | || (c == 0x0e0d) | |
560 | || (c >= 0x0e94 && c <= 0x0e97) | |
561 | || (c >= 0x0e99 && c <= 0x0e9f) | |
562 | || (c >= 0x0ea1 && c <= 0x0ea3) | |
563 | || (c == 0x0ea5) | |
564 | || (c == 0x0ea7) | |
565 | || (c == 0x0eaa) | |
566 | || (c == 0x0eab) | |
567 | || (c >= 0x0ead && c <= 0x0eb0) | |
568 | || (c == 0x0eb2) | |
569 | || (c == 0x0eb3) | |
570 | || (c == 0x0ebd) | |
571 | || (c >= 0x0ec0 && c <= 0x0ec4) | |
572 | || (c == 0x0ec6)) | |
573 | return 1; | |
574 | ||
575 | /* Georgian */ | |
576 | if ((c >= 0x10a0 && c <= 0x10c5) | |
577 | || (c >= 0x10d0 && c <= 0x10f6)) | |
578 | return 1; | |
579 | ||
580 | /* Hiragana */ | |
581 | if ((c >= 0x3041 && c <= 0x3094) | |
582 | || (c >= 0x309b && c <= 0x309e)) | |
583 | return 1; | |
584 | ||
585 | /* Katakana */ | |
586 | if ((c >= 0x30a1 && c <= 0x30fe)) | |
587 | return 1; | |
588 | ||
589 | /* Bopmofo */ | |
590 | if ((c >= 0x3105 && c <= 0x312c)) | |
591 | return 1; | |
592 | ||
593 | /* Hangul */ | |
594 | if ((c >= 0x1100 && c <= 0x1159) | |
595 | || (c >= 0x1161 && c <= 0x11a2) | |
596 | || (c >= 0x11a8 && c <= 0x11f9)) | |
597 | return 1; | |
598 | ||
599 | /* CJK Unified Ideographs */ | |
600 | if ((c >= 0xf900 && c <= 0xfa2d) | |
601 | || (c >= 0xfb1f && c <= 0xfb36) | |
602 | || (c >= 0xfb38 && c <= 0xfb3c) | |
603 | || (c == 0xfb3e) | |
604 | || (c >= 0xfb40 && c <= 0xfb41) | |
605 | || (c >= 0xfb42 && c <= 0xfb44) | |
606 | || (c >= 0xfb46 && c <= 0xfbb1) | |
607 | || (c >= 0xfbd3 && c <= 0xfd3f) | |
608 | || (c >= 0xfd50 && c <= 0xfd8f) | |
609 | || (c >= 0xfd92 && c <= 0xfdc7) | |
610 | || (c >= 0xfdf0 && c <= 0xfdfb) | |
611 | || (c >= 0xfe70 && c <= 0xfe72) | |
612 | || (c == 0xfe74) | |
613 | || (c >= 0xfe76 && c <= 0xfefc) | |
614 | || (c >= 0xff21 && c <= 0xff3a) | |
615 | || (c >= 0xff41 && c <= 0xff5a) | |
616 | || (c >= 0xff66 && c <= 0xffbe) | |
617 | || (c >= 0xffc2 && c <= 0xffc7) | |
618 | || (c >= 0xffca && c <= 0xffcf) | |
619 | || (c >= 0xffd2 && c <= 0xffd7) | |
620 | || (c >= 0xffda && c <= 0xffdc) | |
621 | || (c >= 0x4e00 && c <= 0x9fa5)) | |
622 | return 1; | |
623 | ||
624 | error ("universal-character-name '\\u%04x' not valid in identifier", c); | |
625 | return 1; | |
626 | #endif | |
627 | } | |
628 | ||
629 | /* Add the UTF-8 representation of C to the token_buffer. */ | |
630 | ||
631 | static void | |
632 | utf8_extend_token (c) | |
633 | int c; | |
e8bbfc4e | 634 | { |
0e5921e8 ZW |
635 | int shift, mask; |
636 | ||
637 | if (c <= 0x0000007f) | |
638 | { | |
639 | extend_token (c); | |
640 | return; | |
641 | } | |
642 | else if (c <= 0x000007ff) | |
643 | shift = 6, mask = 0xc0; | |
644 | else if (c <= 0x0000ffff) | |
645 | shift = 12, mask = 0xe0; | |
646 | else if (c <= 0x001fffff) | |
647 | shift = 18, mask = 0xf0; | |
648 | else if (c <= 0x03ffffff) | |
649 | shift = 24, mask = 0xf8; | |
e8bbfc4e | 650 | else |
0e5921e8 ZW |
651 | shift = 30, mask = 0xfc; |
652 | ||
653 | extend_token (mask | (c >> shift)); | |
654 | do | |
655 | { | |
656 | shift -= 6; | |
657 | extend_token ((unsigned char) (0x80 | (c >> shift))); | |
658 | } | |
659 | while (shift); | |
e8bbfc4e | 660 | } |
0e5921e8 | 661 | #endif |
e8bbfc4e | 662 | \f |
0e5921e8 ZW |
663 | int |
664 | c_lex (value) | |
665 | tree *value; | |
fbb18613 | 666 | { |
4ed5bcfb | 667 | const cpp_token *tok; |
0e5921e8 ZW |
668 | |
669 | retry: | |
670 | timevar_push (TV_CPP); | |
4ed5bcfb NB |
671 | do |
672 | tok = cpp_get_token (parse_in); | |
673 | while (tok->type == CPP_PADDING); | |
0e5921e8 ZW |
674 | timevar_pop (TV_CPP); |
675 | ||
676 | /* The C++ front end does horrible things with the current line | |
677 | number. To ensure an accurate line number, we must reset it | |
23356f93 | 678 | every time we return a token. */ |
97293897 | 679 | lineno = src_lineno; |
0e5921e8 ZW |
680 | |
681 | *value = NULL_TREE; | |
a23c9413 | 682 | switch (tok->type) |
0e5921e8 | 683 | { |
4ed5bcfb | 684 | /* Issue this error here, where we can get at tok->val.c. */ |
0e5921e8 | 685 | case CPP_OTHER: |
4ed5bcfb NB |
686 | if (ISGRAPH (tok->val.c)) |
687 | error ("stray '%c' in program", tok->val.c); | |
0e5921e8 | 688 | else |
4ed5bcfb | 689 | error ("stray '\\%o' in program", tok->val.c); |
0e5921e8 ZW |
690 | goto retry; |
691 | ||
0e5921e8 | 692 | case CPP_NAME: |
4ed5bcfb | 693 | *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node)); |
0e5921e8 | 694 | break; |
fbb18613 | 695 | |
0e5921e8 | 696 | case CPP_NUMBER: |
ceeedfc1 NB |
697 | { |
698 | unsigned int flags = cpp_classify_number (parse_in, tok); | |
699 | ||
700 | switch (flags & CPP_N_CATEGORY) | |
701 | { | |
702 | case CPP_N_INVALID: | |
703 | /* cpplib has issued an error. */ | |
e8f2b18d | 704 | *value = error_mark_node; |
ceeedfc1 NB |
705 | break; |
706 | ||
707 | case CPP_N_INTEGER: | |
708 | *value = interpret_integer (tok, flags); | |
709 | break; | |
710 | ||
711 | case CPP_N_FLOATING: | |
712 | *value = interpret_float (tok, flags); | |
713 | break; | |
714 | ||
715 | default: | |
716 | abort (); | |
717 | } | |
718 | } | |
0e5921e8 | 719 | break; |
93868d11 | 720 | |
0e5921e8 ZW |
721 | case CPP_CHAR: |
722 | case CPP_WCHAR: | |
4ed5bcfb | 723 | *value = lex_charconst (tok); |
0e5921e8 | 724 | break; |
fbb18613 | 725 | |
0e5921e8 ZW |
726 | case CPP_STRING: |
727 | case CPP_WSTRING: | |
62ae2529 JJ |
728 | *value = lex_string (tok->val.str.text, tok->val.str.len, |
729 | tok->type == CPP_WSTRING); | |
0e5921e8 | 730 | break; |
fbb18613 | 731 | |
0e5921e8 ZW |
732 | /* These tokens should not be visible outside cpplib. */ |
733 | case CPP_HEADER_NAME: | |
734 | case CPP_COMMENT: | |
735 | case CPP_MACRO_ARG: | |
0e5921e8 ZW |
736 | abort (); |
737 | ||
738 | default: break; | |
739 | } | |
740 | ||
a23c9413 | 741 | return tok->type; |
0e5921e8 | 742 | } |
8d9bfdc5 | 743 | |
ceeedfc1 NB |
744 | /* Returns the narrowest C-visible unsigned type, starting with the |
745 | minimum specified by FLAGS, that can fit VALUE, or itk_none if | |
746 | there isn't one. */ | |
747 | static enum integer_type_kind | |
748 | narrowest_unsigned_type (value, flags) | |
749 | tree value; | |
750 | unsigned int flags; | |
0e5921e8 | 751 | { |
ceeedfc1 | 752 | enum integer_type_kind itk; |
56f48ce9 | 753 | |
ceeedfc1 NB |
754 | if ((flags & CPP_N_WIDTH) == CPP_N_SMALL) |
755 | itk = itk_unsigned_int; | |
756 | else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM) | |
757 | itk = itk_unsigned_long; | |
758 | else | |
759 | itk = itk_unsigned_long_long; | |
e8bbfc4e | 760 | |
ceeedfc1 NB |
761 | /* int_fits_type_p must think the type of its first argument is |
762 | wider than its second argument, or it won't do the proper check. */ | |
763 | TREE_TYPE (value) = widest_unsigned_literal_type_node; | |
e8bbfc4e | 764 | |
ceeedfc1 NB |
765 | for (; itk < itk_none; itk += 2 /* skip unsigned types */) |
766 | if (int_fits_type_p (value, integer_types[itk])) | |
767 | return itk; | |
56f48ce9 | 768 | |
ceeedfc1 NB |
769 | return itk_none; |
770 | } | |
e8bbfc4e | 771 | |
ceeedfc1 NB |
772 | /* Ditto, but narrowest signed type. */ |
773 | static enum integer_type_kind | |
774 | narrowest_signed_type (value, flags) | |
775 | tree value; | |
776 | unsigned int flags; | |
777 | { | |
778 | enum integer_type_kind itk; | |
e8bbfc4e | 779 | |
ceeedfc1 NB |
780 | if ((flags & CPP_N_WIDTH) == CPP_N_SMALL) |
781 | itk = itk_int; | |
782 | else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM) | |
783 | itk = itk_long; | |
784 | else | |
785 | itk = itk_long_long; | |
e8bbfc4e | 786 | |
ceeedfc1 NB |
787 | /* int_fits_type_p must think the type of its first argument is |
788 | wider than its second argument, or it won't do the proper check. */ | |
789 | TREE_TYPE (value) = widest_unsigned_literal_type_node; | |
e8bbfc4e | 790 | |
ceeedfc1 NB |
791 | for (; itk < itk_none; itk += 2 /* skip signed types */) |
792 | if (int_fits_type_p (value, integer_types[itk])) | |
793 | return itk; | |
15e5ad76 | 794 | |
ceeedfc1 NB |
795 | return itk_none; |
796 | } | |
15e5ad76 | 797 | |
ceeedfc1 NB |
798 | /* Interpret TOKEN, an integer with FLAGS as classified by cpplib. */ |
799 | static tree | |
800 | interpret_integer (token, flags) | |
801 | const cpp_token *token; | |
802 | unsigned int flags; | |
803 | { | |
804 | tree value, type; | |
805 | enum integer_type_kind itk; | |
806 | cpp_num integer; | |
807 | cpp_options *options = cpp_get_options (parse_in); | |
808 | ||
809 | integer = cpp_interpret_integer (parse_in, token, flags); | |
810 | integer = cpp_num_sign_extend (integer, options->precision); | |
811 | value = build_int_2_wide (integer.low, integer.high); | |
812 | ||
813 | /* The type of a constant with a U suffix is straightforward. */ | |
814 | if (flags & CPP_N_UNSIGNED) | |
815 | itk = narrowest_unsigned_type (value, flags); | |
0e5921e8 ZW |
816 | else |
817 | { | |
ceeedfc1 NB |
818 | /* The type of a potentially-signed integer constant varies |
819 | depending on the base it's in, the standard in use, and the | |
820 | length suffixes. */ | |
821 | enum integer_type_kind itk_u = narrowest_unsigned_type (value, flags); | |
822 | enum integer_type_kind itk_s = narrowest_signed_type (value, flags); | |
823 | ||
824 | /* In both C89 and C99, octal and hex constants may be signed or | |
825 | unsigned, whichever fits tighter. We do not warn about this | |
826 | choice differing from the traditional choice, as the constant | |
827 | is probably a bit pattern and either way will work. */ | |
828 | if ((flags & CPP_N_RADIX) != CPP_N_DECIMAL) | |
829 | itk = MIN (itk_u, itk_s); | |
830 | else | |
0e5921e8 | 831 | { |
ceeedfc1 NB |
832 | /* In C99, decimal constants are always signed. |
833 | In C89, decimal constants that don't fit in long have | |
8d9afc4e | 834 | undefined behavior; we try to make them unsigned long. |
ceeedfc1 NB |
835 | In GCC's extended C89, that last is true of decimal |
836 | constants that don't fit in long long, too. */ | |
837 | ||
838 | itk = itk_s; | |
839 | if (itk_s > itk_u && itk_s > itk_long) | |
0e5921e8 | 840 | { |
ceeedfc1 | 841 | if (!flag_isoc99) |
0e5921e8 | 842 | { |
ceeedfc1 NB |
843 | if (itk_u < itk_unsigned_long) |
844 | itk_u = itk_unsigned_long; | |
845 | itk = itk_u; | |
56508306 | 846 | warning ("this decimal constant is unsigned only in ISO C90"); |
0e5921e8 | 847 | } |
ceeedfc1 | 848 | else if (warn_traditional) |
56508306 | 849 | warning ("this decimal constant would be unsigned in ISO C90"); |
0e5921e8 ZW |
850 | } |
851 | } | |
ceeedfc1 | 852 | } |
56f48ce9 | 853 | |
ceeedfc1 NB |
854 | if (itk == itk_none) |
855 | /* cpplib has already issued a warning for overflow. */ | |
856 | type = ((flags & CPP_N_UNSIGNED) | |
857 | ? widest_unsigned_literal_type_node | |
858 | : widest_integer_literal_type_node); | |
859 | else | |
860 | type = integer_types[itk]; | |
e8bbfc4e | 861 | |
ceeedfc1 NB |
862 | if (itk > itk_unsigned_long |
863 | && (flags & CPP_N_WIDTH) != CPP_N_LARGE | |
864 | && ! in_system_header && ! flag_isoc99) | |
865 | pedwarn ("integer constant is too large for \"%s\" type", | |
866 | (flags & CPP_N_UNSIGNED) ? "unsigned long" : "long"); | |
0468bc75 | 867 | |
ceeedfc1 | 868 | TREE_TYPE (value) = type; |
e8bbfc4e | 869 | |
ceeedfc1 NB |
870 | /* Convert imaginary to a complex type. */ |
871 | if (flags & CPP_N_IMAGINARY) | |
872 | value = build_complex (NULL_TREE, convert (type, integer_zero_node), value); | |
e8bbfc4e | 873 | |
ceeedfc1 NB |
874 | return value; |
875 | } | |
e8bbfc4e | 876 | |
ceeedfc1 NB |
877 | /* Interpret TOKEN, a floating point number with FLAGS as classified |
878 | by cpplib. */ | |
879 | static tree | |
880 | interpret_float (token, flags) | |
881 | const cpp_token *token; | |
882 | unsigned int flags; | |
883 | { | |
884 | tree type; | |
885 | tree value; | |
886 | REAL_VALUE_TYPE real; | |
887 | char *copy; | |
888 | size_t copylen; | |
889 | const char *typename; | |
e8bbfc4e | 890 | |
ceeedfc1 NB |
891 | /* FIXME: make %T work in error/warning, then we don't need typename. */ |
892 | if ((flags & CPP_N_WIDTH) == CPP_N_LARGE) | |
893 | { | |
894 | type = long_double_type_node; | |
895 | typename = "long double"; | |
896 | } | |
897 | else if ((flags & CPP_N_WIDTH) == CPP_N_SMALL | |
898 | || flag_single_precision_constant) | |
899 | { | |
900 | type = float_type_node; | |
901 | typename = "float"; | |
902 | } | |
903 | else | |
904 | { | |
905 | type = double_type_node; | |
906 | typename = "double"; | |
907 | } | |
e8bbfc4e | 908 | |
ceeedfc1 NB |
909 | /* Copy the constant to a nul-terminated buffer. If the constant |
910 | has any suffixes, cut them off; REAL_VALUE_ATOF/ REAL_VALUE_HTOF | |
911 | can't handle them. */ | |
912 | copylen = token->val.str.len; | |
913 | if ((flags & CPP_N_WIDTH) != CPP_N_MEDIUM) | |
914 | /* Must be an F or L suffix. */ | |
915 | copylen--; | |
916 | if (flags & CPP_N_IMAGINARY) | |
917 | /* I or J suffix. */ | |
918 | copylen--; | |
919 | ||
920 | copy = alloca (copylen + 1); | |
921 | memcpy (copy, token->val.str.text, copylen); | |
922 | copy[copylen] = '\0'; | |
923 | ||
efdc7e19 RH |
924 | real_from_string (&real, copy); |
925 | real_convert (&real, TYPE_MODE (type), &real); | |
fbb18613 | 926 | |
ceeedfc1 NB |
927 | /* A diagnostic is required for "soft" overflow by some ISO C |
928 | testsuites. This is not pedwarn, because some people don't want | |
929 | an error for this. | |
930 | ??? That's a dubious reason... is this a mandatory diagnostic or | |
931 | isn't it? -- zw, 2001-08-21. */ | |
932 | if (REAL_VALUE_ISINF (real) && pedantic) | |
933 | warning ("floating constant exceeds range of \"%s\"", typename); | |
fbb18613 | 934 | |
ceeedfc1 NB |
935 | /* Create a node with determined type and value. */ |
936 | value = build_real (type, real); | |
937 | if (flags & CPP_N_IMAGINARY) | |
938 | value = build_complex (NULL_TREE, convert (type, integer_zero_node), value); | |
e8bbfc4e | 939 | |
0e5921e8 | 940 | return value; |
0e5921e8 | 941 | } |
e8bbfc4e | 942 | |
0e5921e8 ZW |
943 | static tree |
944 | lex_string (str, len, wide) | |
62ae2529 | 945 | const unsigned char *str; |
0e5921e8 ZW |
946 | unsigned int len; |
947 | int wide; | |
948 | { | |
949 | tree value; | |
950 | char *buf = alloca ((len + 1) * (wide ? WCHAR_BYTES : 1)); | |
951 | char *q = buf; | |
62ae2529 | 952 | const unsigned char *p = str, *limit = str + len; |
4268e8bb | 953 | cppchar_t c; |
e9a25f70 | 954 | |
0e5921e8 ZW |
955 | #ifdef MULTIBYTE_CHARS |
956 | /* Reset multibyte conversion state. */ | |
9714cf43 | 957 | (void) local_mbtowc (NULL, NULL, 0); |
0e5921e8 | 958 | #endif |
e9a25f70 | 959 | |
0e5921e8 ZW |
960 | while (p < limit) |
961 | { | |
962 | #ifdef MULTIBYTE_CHARS | |
963 | wchar_t wc; | |
964 | int char_len; | |
965 | ||
62ae2529 | 966 | char_len = local_mbtowc (&wc, (const char *) p, limit - p); |
0e5921e8 ZW |
967 | if (char_len == -1) |
968 | { | |
53fcdc76 | 969 | warning ("ignoring invalid multibyte character"); |
0e5921e8 ZW |
970 | char_len = 1; |
971 | c = *p++; | |
972 | } | |
973 | else | |
974 | { | |
975 | p += char_len; | |
976 | c = wc; | |
977 | } | |
978 | #else | |
979 | c = *p++; | |
980 | #endif | |
981 | ||
982 | if (c == '\\' && !ignore_escape_flag) | |
4268e8bb | 983 | c = cpp_parse_escape (parse_in, &p, limit, wide); |
0e5921e8 | 984 | |
64cdc383 MH |
985 | /* Add this single character into the buffer either as a wchar_t, |
986 | a multibyte sequence, or as a single byte. */ | |
0e5921e8 ZW |
987 | if (wide) |
988 | { | |
989 | unsigned charwidth = TYPE_PRECISION (char_type_node); | |
5c80f6e6 | 990 | unsigned bytemask = (1 << charwidth) - 1; |
0e5921e8 ZW |
991 | int byte; |
992 | ||
993 | for (byte = 0; byte < WCHAR_BYTES; ++byte) | |
994 | { | |
995 | int n; | |
996 | if (byte >= (int) sizeof (c)) | |
997 | n = 0; | |
998 | else | |
999 | n = (c >> (byte * charwidth)) & bytemask; | |
1000 | if (BYTES_BIG_ENDIAN) | |
1001 | q[WCHAR_BYTES - byte - 1] = n; | |
1002 | else | |
1003 | q[byte] = n; | |
1004 | } | |
1005 | q += WCHAR_BYTES; | |
1006 | } | |
64cdc383 MH |
1007 | #ifdef MULTIBYTE_CHARS |
1008 | else if (char_len > 1) | |
1009 | { | |
4b7e68e7 | 1010 | /* We're dealing with a multibyte character. */ |
64cdc383 MH |
1011 | for ( ; char_len >0; --char_len) |
1012 | { | |
1013 | *q++ = *(p - char_len); | |
1014 | } | |
1015 | } | |
1016 | #endif | |
0e5921e8 ZW |
1017 | else |
1018 | { | |
1019 | *q++ = c; | |
1020 | } | |
e8bbfc4e RK |
1021 | } |
1022 | ||
0e5921e8 ZW |
1023 | /* Terminate the string value, either with a single byte zero |
1024 | or with a wide zero. */ | |
e8bbfc4e | 1025 | |
0e5921e8 ZW |
1026 | if (wide) |
1027 | { | |
1028 | memset (q, 0, WCHAR_BYTES); | |
1029 | q += WCHAR_BYTES; | |
1030 | } | |
1031 | else | |
1032 | { | |
1033 | *q++ = '\0'; | |
1034 | } | |
1035 | ||
1036 | value = build_string (q - buf, buf); | |
1037 | ||
1038 | if (wide) | |
1039 | TREE_TYPE (value) = wchar_array_type_node; | |
1040 | else | |
1041 | TREE_TYPE (value) = char_array_type_node; | |
e8bbfc4e RK |
1042 | return value; |
1043 | } | |
1044 | ||
c8a96070 | 1045 | /* Converts a (possibly wide) character constant token into a tree. */ |
0e5921e8 | 1046 | static tree |
c8a96070 NB |
1047 | lex_charconst (token) |
1048 | const cpp_token *token; | |
e8bbfc4e | 1049 | { |
4268e8bb | 1050 | cppchar_t result; |
9340544b | 1051 | tree type, value; |
c8a96070 | 1052 | unsigned int chars_seen; |
4268e8bb | 1053 | int unsignedp; |
ceeedfc1 | 1054 | |
a5a49440 | 1055 | result = cpp_interpret_charconst (parse_in, token, |
4268e8bb | 1056 | &chars_seen, &unsignedp); |
9340544b | 1057 | |
4268e8bb NB |
1058 | /* Cast to cppchar_signed_t to get correct sign-extension of RESULT |
1059 | before possibly widening to HOST_WIDE_INT for build_int_2. */ | |
1060 | if (unsignedp || (cppchar_signed_t) result >= 0) | |
1061 | value = build_int_2 (result, 0); | |
1062 | else | |
1063 | value = build_int_2 ((cppchar_signed_t) result, -1); | |
9340544b | 1064 | |
4268e8bb NB |
1065 | if (token->type == CPP_WCHAR) |
1066 | type = wchar_type_node; | |
1067 | /* In C, a character constant has type 'int'. | |
1068 | In C++ 'char', but multi-char charconsts have type 'int'. */ | |
0f7866e7 | 1069 | else if ((c_language == clk_c) || chars_seen > 1) |
4268e8bb NB |
1070 | type = integer_type_node; |
1071 | else | |
1072 | type = char_type_node; | |
9340544b | 1073 | |
4268e8bb | 1074 | TREE_TYPE (value) = type; |
0e5921e8 | 1075 | return value; |
e8bbfc4e | 1076 | } |