PulkoMandy | 17fc759 | 2022-07-28 18:27:54 +0200 | [diff] [blame] | 1 | /* |
| 2 | * C and T preprocessor, and integrated lexer |
| 3 | * (c) Thomas Pornin 1999, 2000 |
| 4 | * |
| 5 | * Redistribution and use in source and binary forms, with or without |
| 6 | * modification, are permitted provided that the following conditions |
| 7 | * are met: |
| 8 | * 1. Redistributions of source code must retain the above copyright |
| 9 | * notice, this list of conditions and the following disclaimer. |
| 10 | * 2. Redistributions in binary form must reproduce the above copyright |
| 11 | * notice, this list of conditions and the following disclaimer in the |
| 12 | * documentation and/or other materials provided with the distribution. |
| 13 | * 4. The name of the authors may not be used to endorse or promote |
| 14 | * products derived from this software without specific prior written |
| 15 | * permission. |
| 16 | * |
| 17 | * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR |
| 18 | * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED |
| 19 | * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
| 20 | * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE |
| 21 | * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR |
| 22 | * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT |
| 23 | * OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR |
| 24 | * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, |
| 25 | * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE |
| 26 | * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, |
| 27 | * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 28 | * |
| 29 | */ |
| 30 | |
| 31 | /*vb*/ |
| 32 | #ifdef HAVE_MISRA |
| 33 | void misra(int,...); |
| 34 | #endif |
| 35 | extern void handle_deps(char *,int); |
| 36 | |
| 37 | #define VERS_MAJ 1 |
| 38 | #define VERS_MIN 0 |
| 39 | /* uncomment the following if you cannot set it with a compiler flag */ |
| 40 | /* #define STAND_ALONE */ |
| 41 | |
| 42 | #include "tune.h" |
| 43 | #ifdef UCPP_MMAP |
| 44 | #ifndef _POSIX_SOURCE |
| 45 | #define _POSIX_SOURCE 1 |
| 46 | #endif |
| 47 | #endif |
| 48 | #include <stdio.h> |
| 49 | #include <string.h> |
| 50 | #include <stdarg.h> |
| 51 | #include <setjmp.h> |
| 52 | #include <stddef.h> |
| 53 | #include <limits.h> |
| 54 | #include <time.h> |
| 55 | #include "ucppi.h" |
| 56 | #include "mem.h" |
| 57 | #include "hash.h" |
| 58 | #ifdef UCPP_MMAP |
| 59 | #include <unistd.h> |
| 60 | #include <sys/types.h> |
| 61 | #include <sys/mman.h> |
| 62 | #include <fcntl.h> |
| 63 | #endif |
| 64 | |
| 65 | /* |
| 66 | * The standard path where includes are looked for. |
| 67 | */ |
| 68 | #ifdef STAND_ALONE |
| 69 | static char *include_path_std[] = { STD_INCLUDE_PATH, 0 }; |
| 70 | #endif |
| 71 | static char **include_path; |
| 72 | static size_t include_path_nb = 0; |
| 73 | |
| 74 | int no_special_macros = 0; |
| 75 | int emit_dependencies = 0, emit_defines = 0, emit_assertions = 0; |
| 76 | FILE *emit_output; |
| 77 | |
| 78 | #ifdef STAND_ALONE |
| 79 | static char *system_macros_def[] = { STD_MACROS, 0 }; |
| 80 | static char *system_assertions_def[] = { STD_ASSERT, 0 }; |
| 81 | #endif |
| 82 | |
| 83 | char *current_filename, *current_long_filename = 0; |
| 84 | static int current_incdir = -1; |
| 85 | |
| 86 | #ifndef NO_UCPP_ERROR_FUNCTIONS |
| 87 | /* |
| 88 | * "ouch" is the name for an internal ucpp error. If AUDIT is not defined, |
| 89 | * no code calling this function will be generated; a "ouch" may still be |
| 90 | * emitted by getmem() (in mem.c) if MEM_CHECK is defined, but this "ouch" |
| 91 | * does not use this function. |
| 92 | */ |
| 93 | #endif /* NO_UCPP_ERROR_FUNCTIONS */ |
| 94 | |
| 95 | /* |
| 96 | * Some memory allocations are manually garbage-collected; essentially, |
| 97 | * strings duplicated in the process of macro replacement. Each such |
| 98 | * string is referenced in the garbage_fifo, which is cleared when all |
| 99 | * nested macros have been resolved. |
| 100 | */ |
| 101 | |
| 102 | struct garbage_fifo { |
| 103 | char **garbage; |
| 104 | size_t ngarb, memgarb; |
| 105 | }; |
| 106 | |
| 107 | /* |
| 108 | * throw_away() marks a string to be collected later |
| 109 | */ |
| 110 | void throw_away(struct garbage_fifo *gf, char *n) |
| 111 | { |
| 112 | wan(gf->garbage, gf->ngarb, n, gf->memgarb); |
| 113 | } |
| 114 | |
| 115 | /* |
| 116 | * free marked strings |
| 117 | */ |
| 118 | void garbage_collect(struct garbage_fifo *gf) |
| 119 | { |
| 120 | size_t i; |
| 121 | |
| 122 | for (i = 0; i < gf->ngarb; i ++) freemem(gf->garbage[i]); |
| 123 | gf->ngarb = 0; |
| 124 | } |
| 125 | |
| 126 | static void init_garbage_fifo(struct garbage_fifo *gf) |
| 127 | { |
| 128 | gf->garbage = getmem((gf->memgarb = GARBAGE_LIST_MEMG) |
| 129 | * sizeof(char *)); |
| 130 | gf->ngarb = 0; |
| 131 | } |
| 132 | |
| 133 | /* |
| 134 | * order is important: it must match the token-constants declared as an |
| 135 | * enum in the header file. |
| 136 | */ |
| 137 | char *operators_name[] = { |
| 138 | " ", "\n", " ", |
| 139 | "0000", "name", "bunch", "pragma", "context", |
| 140 | "\"dummy string\"", "'dummy char'", |
| 141 | "/", "/=", "-", "--", "-=", "->", "+", "++", "+=", "<", "<=", "<<", |
| 142 | "<<=", ">", ">=", ">>", ">>=", "=", "==", |
| 143 | #ifdef CAST_OP |
| 144 | "=>", |
| 145 | #endif |
| 146 | "~", "!=", "&", "&&", "&=", "|", "||", "|=", "%", "%=", "*", "*=", |
| 147 | "^", "^=", "!", "~=", "{", "}", "[", "]", "(", ")", ",", "?", ";", |
| 148 | ":", ".", "...", "#", "##", " ", "ouch", "<:", ":>", "<%", "%>", |
| 149 | "%:", "%:%:" |
| 150 | }; |
| 151 | |
| 152 | /* the ascii representation of a token */ |
| 153 | #ifdef SEMPER_FIDELIS |
| 154 | #define tname(x) (ttWHI((x).type) ? " " : S_TOKEN((x).type) \ |
| 155 | ? (x).name : operators_name[(x).type]) |
| 156 | #else |
| 157 | #define tname(x) (S_TOKEN((x).type) ? (x).name \ |
| 158 | : operators_name[(x).type]) |
| 159 | #endif |
| 160 | |
| 161 | char *token_name(struct token *t) |
| 162 | { |
| 163 | return tname(*t); |
| 164 | } |
| 165 | |
| 166 | /* |
| 167 | * To speed up deeply nested and repeated inclusions, we: |
| 168 | * -- use a hash table to remember where we found each file |
| 169 | * -- remember when the file is protected by a #ifndef/#define/#endif |
| 170 | * construction; we can then avoid including several times a file |
| 171 | * when this is not necessary. |
| 172 | * -- remember in which directory, in the include path, the file was found. |
| 173 | */ |
| 174 | struct found_file { |
| 175 | char *name; /* first field */ |
| 176 | char *long_name; |
| 177 | char *protect; |
| 178 | int incdir; |
| 179 | }; |
| 180 | |
| 181 | static struct HT *found_files = 0, *found_files_loc = 0; |
| 182 | |
| 183 | static struct found_file *new_found_file(void) |
| 184 | { |
| 185 | struct found_file *ff = getmem(sizeof(struct found_file)); |
| 186 | |
| 187 | ff->protect = 0; |
| 188 | ff->incdir = -1; |
| 189 | return ff; |
| 190 | } |
| 191 | |
| 192 | static void del_found_file(void *m) |
| 193 | { |
| 194 | struct found_file *ff = (struct found_file *)m; |
| 195 | |
| 196 | freemem(ff->name); |
| 197 | freemem(ff->long_name); |
| 198 | if (ff->protect) freemem(ff->protect); |
| 199 | } |
| 200 | |
| 201 | /* |
| 202 | * To keep up with the #ifndef/#define/#endif protection mechanism |
| 203 | * detection. |
| 204 | */ |
| 205 | struct protect protect_detect; |
| 206 | static struct protect *protect_detect_stack = 0; |
| 207 | |
| 208 | /* |
| 209 | * "current_filename" is used for error reporting, and the __FILE__ macro |
| 210 | */ |
| 211 | static void set_current_filename(char *x) |
| 212 | { |
| 213 | current_filename = x; |
| 214 | } |
| 215 | |
| 216 | void set_init_filename(char *x, int real_file) |
| 217 | { |
| 218 | set_current_filename(sdup(x)); |
| 219 | current_long_filename = 0; |
| 220 | current_incdir = -1; |
| 221 | if (real_file) { |
| 222 | protect_detect.macro = 0; |
| 223 | protect_detect.state = 1; |
| 224 | protect_detect.ff = new_found_file(); |
| 225 | protect_detect.ff->name = sdup(x); |
| 226 | protect_detect.ff->long_name = sdup(x); |
| 227 | putHT(found_files_loc, protect_detect.ff); |
| 228 | } else { |
| 229 | protect_detect.state = 0; |
| 230 | } |
| 231 | } |
| 232 | |
| 233 | static void init_found_files(void) |
| 234 | { |
| 235 | if (found_files) killHT(found_files); |
| 236 | found_files = newHT(128, cmp_struct, hash_struct, del_found_file); |
| 237 | if (found_files_loc) killHT(found_files_loc); |
| 238 | found_files_loc = newHT(128, cmp_struct, hash_struct, del_found_file); |
| 239 | } |
| 240 | |
| 241 | /* |
| 242 | * convert a Unix-style path for the host architecture |
| 243 | */ |
| 244 | #if defined(MSDOS) || defined(ATARI) || defined(_WIN32) |
| 245 | static char *convert_path(char *path) |
| 246 | { |
| 247 | /* on msdos systems, replace all / by \ */ |
| 248 | char *c, *newp = sdup(path); |
| 249 | |
| 250 | for (c = newp; *c; c ++) if (*c == '/') *c = '\\'; |
| 251 | return newp; |
| 252 | } |
| 253 | |
| 254 | #elif defined AMIGA |
| 255 | static char *convert_path(char *path) |
| 256 | { |
| 257 | char *newp = getmem(1 + strlen(path)); |
| 258 | char *s = path, *d = newp; |
| 259 | |
| 260 | if (*s == '.') { /* convert "." to "" */ |
| 261 | if (*(s+1) == '\0') |
| 262 | s++; |
| 263 | } |
| 264 | while (*s) { |
| 265 | if (*s == '.') { |
| 266 | if (*(s+1) == '/') { |
| 267 | s += 2; /* "./" is discarded */ |
| 268 | continue; |
| 269 | } |
| 270 | else if (*(s+1) == '.') { |
| 271 | if (*(s+2) == '/') |
| 272 | s += 2; /* "../" is converted to "/" */ |
| 273 | } |
| 274 | } |
| 275 | *d++ = *s++; |
| 276 | } |
| 277 | *d = '\0'; |
| 278 | |
| 279 | return newp; |
| 280 | } |
| 281 | |
| 282 | #else /* no conversion */ |
| 283 | static char *convert_path(char *path) |
| 284 | { |
| 285 | return sdup(path); |
| 286 | } |
| 287 | #endif |
| 288 | |
| 289 | /* |
| 290 | * Set the lexer state at the beginning of a file. |
| 291 | */ |
| 292 | static void reinit_lexer_state(struct lexer_state *ls, int wb) |
| 293 | { |
| 294 | #ifndef NO_UCPP_BUF |
| 295 | ls->input_buf = wb ? getmem(INPUT_BUF_MEMG) : 0; |
| 296 | #ifdef UCPP_MMAP |
| 297 | ls->from_mmap = 0; |
| 298 | #endif |
| 299 | #endif |
| 300 | ls->ebuf = ls->pbuf = 0; |
| 301 | ls->nlka = 0; |
| 302 | ls->macfile = 0; |
| 303 | ls->discard = 1; |
| 304 | ls->last = 0; /* we suppose '\n' is not 0 */ |
| 305 | ls->line = 1; |
| 306 | ls->ltwnl = 1; |
| 307 | ls->oline = 1; |
| 308 | ls->pending_token = 0; |
| 309 | ls->cli = 0; |
| 310 | ls->copy_line[COPY_LINE_LENGTH - 1] = 0; |
| 311 | ls->ifnest = 0; |
| 312 | ls->condf[0] = ls->condf[1] = 0; |
| 313 | } |
| 314 | |
| 315 | /* |
| 316 | * Initialize the struct lexer_state, with optional input and output buffers. |
| 317 | */ |
| 318 | void init_buf_lexer_state(struct lexer_state *ls, int wb) |
| 319 | { |
| 320 | reinit_lexer_state(ls, wb); |
| 321 | #ifndef NO_UCPP_BUF |
| 322 | ls->output_buf = wb ? getmem(OUTPUT_BUF_MEMG) : 0; |
| 323 | #endif |
| 324 | ls->sbuf = 0; |
| 325 | |
| 326 | ls->ctok = getmem(sizeof(struct token)); |
| 327 | ls->ctok->name = getmem(ls->tknl = TOKEN_NAME_MEMG); |
| 328 | ls->pending_token = 0; |
| 329 | |
| 330 | ls->flags = 0; |
| 331 | ls->count_trigraphs = 0; |
| 332 | ls->gf = getmem(sizeof(struct garbage_fifo)); |
| 333 | init_garbage_fifo(ls->gf); |
| 334 | ls->condcomp = 1; |
| 335 | ls->condnest = 0; |
| 336 | #ifdef INMACRO_FLAG |
| 337 | ls->inmacro = 0; |
| 338 | ls->macro_count = 0; |
| 339 | #endif |
| 340 | } |
| 341 | |
| 342 | /* |
| 343 | * Initialize the (complex) struct lexer_state. |
| 344 | */ |
| 345 | void init_lexer_state(struct lexer_state *ls) |
| 346 | { |
| 347 | init_buf_lexer_state(ls, 1); |
| 348 | } |
| 349 | |
| 350 | /* |
| 351 | * Restore what is needed from a lexer_state. This is used for #include. |
| 352 | */ |
| 353 | static void restore_lexer_state(struct lexer_state *ls, |
| 354 | struct lexer_state *lsbak) |
| 355 | { |
| 356 | #ifndef NO_UCPP_BUF |
| 357 | freemem(ls->input_buf); |
| 358 | ls->input_buf = lsbak->input_buf; |
| 359 | #ifdef UCPP_MMAP |
| 360 | ls->from_mmap = lsbak->from_mmap; |
| 361 | ls->input_buf_sav = lsbak->input_buf_sav; |
| 362 | #endif |
| 363 | #endif |
| 364 | ls->input = lsbak->input; |
| 365 | ls->ebuf = lsbak->ebuf; |
| 366 | ls->pbuf = lsbak->pbuf; |
| 367 | ls->nlka = lsbak->nlka; |
| 368 | ls->macfile = lsbak->macfile; |
| 369 | ls->discard = lsbak->discard; |
| 370 | ls->line = lsbak->line; |
| 371 | ls->oline = lsbak->oline; |
| 372 | ls->ifnest = lsbak->ifnest; |
| 373 | ls->condf[0] = lsbak->condf[0]; |
| 374 | ls->condf[1] = lsbak->condf[1]; |
| 375 | } |
| 376 | |
| 377 | /* |
| 378 | * file_context (and the two functions push_ and pop_) are used to save |
| 379 | * all that is needed when including a file. |
| 380 | */ |
| 381 | static struct file_context { |
| 382 | struct lexer_state ls; |
| 383 | char *name, *long_name; |
| 384 | int incdir; |
| 385 | } *ls_stack; |
| 386 | static size_t ls_depth = 0; |
| 387 | |
| 388 | static void push_file_context(struct lexer_state *ls) |
| 389 | { |
| 390 | struct file_context fc; |
| 391 | |
| 392 | if (ls_depth > 1023) |
| 393 | error(ls->line - 1, "too many nested includes"); |
| 394 | fc.name = current_filename; |
| 395 | fc.long_name = current_long_filename; |
| 396 | fc.incdir = current_incdir; |
| 397 | mmv(&(fc.ls), ls, sizeof(struct lexer_state)); |
| 398 | aol(ls_stack, ls_depth, fc, LS_STACK_MEMG); |
| 399 | ls_depth --; |
| 400 | aol(protect_detect_stack, ls_depth, protect_detect, LS_STACK_MEMG); |
| 401 | protect_detect.macro = 0; |
| 402 | } |
| 403 | |
| 404 | static void pop_file_context(struct lexer_state *ls) |
| 405 | { |
| 406 | #ifdef AUDIT |
| 407 | if (ls_depth <= 0) ouch("prepare to meet thy creator"); |
| 408 | #endif |
| 409 | restore_lexer_state(ls, &(ls_stack[-- ls_depth].ls)); |
| 410 | if (protect_detect.macro) freemem(protect_detect.macro); |
| 411 | protect_detect = protect_detect_stack[ls_depth]; |
| 412 | set_current_filename(ls_stack[ls_depth].name); |
| 413 | current_long_filename = ls_stack[ls_depth].long_name; |
| 414 | current_incdir = ls_stack[ls_depth].incdir; |
| 415 | } |
| 416 | |
| 417 | /* |
| 418 | * report_context() returns the list of successive includers of the |
| 419 | * current file, ending with a dummy entry with a negative line number. |
| 420 | * The caller is responsible for freeing the returned pointer. |
| 421 | */ |
| 422 | struct stack_context *report_context(void) |
| 423 | { |
| 424 | struct stack_context *sc; |
| 425 | size_t i; |
| 426 | |
| 427 | sc = getmem((ls_depth + 1) * sizeof(struct stack_context)); |
| 428 | for (i = 0; i < ls_depth; i ++) { |
| 429 | sc[i].name = ls_stack[ls_depth - i - 1].name; |
| 430 | sc[i].long_name = ls_stack[ls_depth - i - 1].long_name; |
| 431 | sc[i].line = ls_stack[ls_depth - i - 1].ls.line - 1; |
| 432 | } |
| 433 | sc[ls_depth].line = -1; |
| 434 | return sc; |
| 435 | } |
| 436 | |
| 437 | /* |
| 438 | * init_lexer_mode() is used to end initialization of a struct lexer_state |
| 439 | * if it must be used for a lexer |
| 440 | */ |
| 441 | void init_lexer_mode(struct lexer_state *ls) |
| 442 | { |
| 443 | ls->flags = DEFAULT_LEXER_FLAGS; |
| 444 | ls->output_fifo = getmem(sizeof(struct token_fifo)); |
| 445 | ls->output_fifo->art = ls->output_fifo->nt = 0; |
| 446 | ls->toplevel_of = ls->output_fifo; |
| 447 | ls->save_ctok = ls->ctok; |
| 448 | } |
| 449 | |
| 450 | /* |
| 451 | * release memory used by a struct lexer_state |
| 452 | */ |
| 453 | void free_lexer_state(struct lexer_state *ls) |
| 454 | { |
| 455 | #ifndef NO_UCPP_BUF |
| 456 | if (ls->input_buf) freemem(ls->input_buf); |
| 457 | if (ls->output_buf) freemem(ls->output_buf); |
| 458 | #endif |
| 459 | freemem(ls->ctok->name); |
| 460 | freemem(ls->ctok); |
| 461 | if (!(ls->flags & LEXER)) garbage_collect(ls->gf); |
| 462 | freemem(ls->gf); |
| 463 | } |
| 464 | |
| 465 | /* |
| 466 | * Print line information. |
| 467 | */ |
| 468 | static void print_line_info(struct lexer_state *ls, unsigned long flags) |
| 469 | { |
| 470 | char *fn = current_long_filename ? |
| 471 | current_long_filename : current_filename; |
| 472 | char *b, *d; |
| 473 | |
| 474 | b = getmem(50 + strlen(fn)); |
| 475 | if (flags & GCC_LINE_NUM) { |
| 476 | sprintf(b, "# %ld \"%s\"\n", ls->line, fn); |
| 477 | } else { |
| 478 | sprintf(b, "#line %ld \"%s\"\n", ls->line, fn); |
| 479 | } |
| 480 | for (d = b; *d; d ++) put_char(ls, (unsigned char)(*d)); |
| 481 | freemem(b); |
| 482 | } |
| 483 | |
| 484 | /* |
| 485 | * Enter a file; this implies the possible emission of a #line directive. |
| 486 | * The flags used are passed as second parameter instead of being |
| 487 | * extracted from the struct lexer_state. |
| 488 | * |
| 489 | * As a command-line option, gcc-like directives (with only a '#', |
| 490 | * without 'line') may be produced. |
| 491 | */ |
| 492 | void enter_file(struct lexer_state *ls, unsigned long flags) |
| 493 | { |
| 494 | char *fn = current_long_filename ? |
| 495 | current_long_filename : current_filename; |
| 496 | |
| 497 | if (!(flags & LINE_NUM)) return; |
| 498 | if ((flags & LEXER) && !(flags & TEXT_OUTPUT)) { |
| 499 | struct token t; |
| 500 | |
| 501 | t.type = CONTEXT; |
| 502 | t.line = ls->line; |
| 503 | t.name = fn; |
| 504 | print_token(ls, &t, 0); |
| 505 | return; |
| 506 | } |
| 507 | print_line_info(ls, flags); |
| 508 | ls->oline --; /* emitted #line troubled oline */ |
| 509 | } |
| 510 | |
| 511 | #ifdef UCPP_MMAP |
| 512 | /* |
| 513 | * We open() the file, then fdopen() it and fseek() to its end. If the |
| 514 | * fseek() worked, we try to mmap() the file, up to the point where we |
| 515 | * arrived. |
| 516 | * On an architecture where end-of-lines are multibytes and translated |
| 517 | * into single '\n', bad things could happen. We strongly hope that, if |
| 518 | * we could fseek() to the end but could not mmap(), then we can get back. |
| 519 | */ |
| 520 | static void *find_file_map; |
| 521 | static size_t map_length; |
| 522 | |
| 523 | FILE *fopen_mmap_file(char *name) |
| 524 | { |
| 525 | FILE *f; |
| 526 | int fd; |
| 527 | long l; |
| 528 | |
| 529 | find_file_map = 0; |
| 530 | fd = open(name, O_RDONLY, 0); |
| 531 | if (fd < 0) return 0; |
| 532 | l = lseek(fd, 0, SEEK_END); |
| 533 | f = fdopen(fd, "r"); |
| 534 | if (!f) { |
| 535 | close(fd); |
| 536 | return 0; |
| 537 | } |
| 538 | if (l < 0) return f; /* not seekable */ |
| 539 | map_length = l; |
| 540 | if ((find_file_map = mmap(0, map_length, PROT_READ, |
| 541 | MAP_PRIVATE, fd, 0)) == MAP_FAILED) { |
| 542 | /* we could not mmap() the file; get back */ |
| 543 | find_file_map = 0; |
| 544 | if (fseek(f, 0, SEEK_SET)) { |
| 545 | /* bwaah... can't get back. This file is cursed. */ |
| 546 | fclose(f); |
| 547 | return 0; |
| 548 | } |
| 549 | } |
| 550 | return f; |
| 551 | } |
| 552 | |
| 553 | void set_input_file(struct lexer_state *ls, FILE *f) |
| 554 | { |
| 555 | ls->input = f; |
| 556 | if (find_file_map) { |
| 557 | ls->from_mmap = 1; |
| 558 | ls->input_buf_sav = ls->input_buf; |
| 559 | ls->input_buf = find_file_map; |
| 560 | ls->pbuf = 0; |
| 561 | ls->ebuf = map_length; |
| 562 | } else { |
| 563 | ls->from_mmap = 0; |
| 564 | } |
| 565 | } |
| 566 | #endif |
| 567 | |
| 568 | /* |
| 569 | * Find a file by looking through the include path. |
| 570 | * return value: a FILE * on the file, opened in "r" mode, or 0. |
| 571 | * |
| 572 | * find_file_error will contain: |
| 573 | * FF_ERROR on error (file not found or impossible to read) |
| 574 | * FF_PROTECT file is protected and therefore useless to read |
| 575 | * FF_KNOWN file is already known |
| 576 | * FF_UNKNOWN file was not already known |
| 577 | */ |
| 578 | static int find_file_error; |
| 579 | |
| 580 | enum { FF_ERROR, FF_PROTECT, FF_KNOWN, FF_UNKNOWN }; |
| 581 | |
| 582 | static FILE *find_file(char *name, int localdir) |
| 583 | { |
| 584 | FILE *f; |
| 585 | int i, incdir = -1; |
| 586 | size_t nl = strlen(name); |
| 587 | char *s = 0; |
| 588 | struct found_file *ff = 0; |
| 589 | int lf = 0; |
| 590 | |
| 591 | find_file_error = FF_ERROR; |
| 592 | protect_detect.state = -1; |
| 593 | protect_detect.macro = 0; |
| 594 | if (localdir) { |
| 595 | int i; |
| 596 | char *rfn = current_long_filename ? current_long_filename |
| 597 | : current_filename; |
| 598 | |
| 599 | for (i = strlen(rfn) - 1; i >= 0; i --) |
| 600 | #if defined(MSDOS) || defined(ATARI) || defined(_WIN32) |
| 601 | if (rfn[i] == '\\') break; |
| 602 | #elif defined AMIGA |
| 603 | if (rfn[i] == '/' || rfn[i] == ':') break; |
| 604 | #else |
| 605 | if (rfn[i] == '/') break; |
| 606 | #endif |
| 607 | |
| 608 | #if defined(MSDOS) || defined(ATARI) || defined(_WIN32) |
| 609 | if (i >= 0 && *name != '\\' && (nl < 2 || name[1] != ':')) |
| 610 | #elif defined AMIGA |
| 611 | if (i >= 0 && strchr(name,':') == 0) |
| 612 | #else |
| 613 | if (i >= 0 && *name != '/') |
| 614 | #endif |
| 615 | { |
| 616 | /* |
| 617 | * current file is somewhere else, and the provided |
| 618 | * file name is not absolute, so we must adjust the |
| 619 | * base for looking for the file; besides, |
| 620 | * found_files and found_files_loc are irrelevant |
| 621 | * for this search. |
| 622 | */ |
| 623 | s = getmem(i + 2 + nl); |
| 624 | mmv(s, rfn, i + 1); |
| 625 | mmv(s + i + 1, name, nl); |
| 626 | s[i + 1 + nl] = 0; |
| 627 | ff = getHT(found_files_loc, &s); |
| 628 | } else ff = getHT(found_files_loc, &name); |
| 629 | |
| 630 | if (!ff) { |
| 631 | #ifdef UCPP_MMAP |
| 632 | if (f = fopen_mmap_file(s ? s : name)) { |
| 633 | #else |
| 634 | if (f = fopen(s ? s : name, "r")) { |
| 635 | #endif |
| 636 | lf = 1; |
| 637 | protect_detect.ff = new_found_file(); |
| 638 | goto found_file; |
| 639 | } |
| 640 | } |
| 641 | } |
| 642 | |
| 643 | if (!ff) ff = getHT(found_files, s ? &s : &name); |
| 644 | if (ff) { |
| 645 | if (ff->protect && get_macro(ff->protect)) { |
| 646 | /* file is protected, do not include it */ |
| 647 | find_file_error = FF_PROTECT; |
| 648 | return 0; |
| 649 | } |
| 650 | protect_detect.ff = ff; |
| 651 | #ifdef UCPP_MMAP |
| 652 | f = fopen_mmap_file(ff->long_name); |
| 653 | #else |
| 654 | f = fopen(ff->long_name, "r"); |
| 655 | #endif |
| 656 | if (!f) return 0; |
| 657 | goto found_file_2; |
| 658 | } else { |
| 659 | protect_detect.ff = new_found_file(); |
| 660 | } |
| 661 | /* |
| 662 | * This is the first time we find the file, or it was not protected. |
| 663 | */ |
| 664 | if (s) freemem(s); |
| 665 | for (i = 0; i < include_path_nb; i ++) { |
| 666 | size_t ni = strlen(include_path[i]); |
| 667 | |
| 668 | s = getmem(ni + nl + 2); |
| 669 | mmv(s, include_path[i], ni); |
| 670 | #ifdef AMIGA |
| 671 | /* contributed by Volker Barthelmann */ |
| 672 | if (ni > 0 && s[ni - 1] != ':' && s[ni - 1] != '/') { |
| 673 | s[ni] = '/'; |
| 674 | mmv(s + ni + 1, name, nl + 1); |
| 675 | } else { |
| 676 | mmv(s + ni, name, nl + 1); |
| 677 | } |
| 678 | #else |
| 679 | s[ni] = '/'; |
| 680 | mmv(s + ni + 1, name, nl + 1); |
| 681 | #endif |
| 682 | incdir = i; |
| 683 | #ifdef UCPP_MMAP |
| 684 | f = fopen_mmap_file(s); |
| 685 | #else |
| 686 | f = fopen(s, "r"); |
| 687 | #endif |
| 688 | if (f) goto found_file; |
| 689 | freemem(s); |
| 690 | } |
| 691 | return 0; |
| 692 | found_file: |
| 693 | if (f && ((emit_dependencies == 1 && lf) |
| 694 | || emit_dependencies == 2)) { |
| 695 | fprintf(emit_output, " %s", s ? s : name); |
| 696 | } |
| 697 | if (!ff) { |
| 698 | struct found_file *nff = protect_detect.ff; |
| 699 | |
| 700 | nff->name = sdup(name); |
| 701 | nff->long_name = s ? s : sdup(name); |
| 702 | nff->incdir = incdir; |
| 703 | #if 0 |
| 704 | putHT(lf ? found_files_loc : found_files, nff); |
| 705 | #endif |
| 706 | s = 0; |
| 707 | find_file_error = FF_UNKNOWN; |
| 708 | ff = nff; |
| 709 | } else find_file_error = FF_KNOWN; |
| 710 | found_file_2: |
| 711 | if (s) freemem(s); |
| 712 | current_long_filename = ff->long_name; |
| 713 | #ifdef NO_LIBC_BUF |
| 714 | setbuf(f, 0); |
| 715 | #endif |
| 716 | current_incdir = ff->incdir; |
| 717 | return f; |
| 718 | } |
| 719 | |
| 720 | /* |
| 721 | * Find the named file by looking through the end of the include path. |
| 722 | * This one is not as optimized as find_file() since it is rarely used. |
| 723 | */ |
| 724 | static FILE *find_file_next(char *name) |
| 725 | { |
| 726 | int i; |
| 727 | size_t nl = strlen(name); |
| 728 | FILE *f; |
| 729 | |
| 730 | for (i = current_incdir + 1; i < include_path_nb; i ++) { |
| 731 | char *s; |
| 732 | size_t ni = strlen(include_path[i]); |
| 733 | |
| 734 | s = getmem(ni + nl + 2); |
| 735 | mmv(s, include_path[i], ni); |
| 736 | #ifdef AMIGA |
| 737 | /* contributed by Volker Barthelmann */ |
| 738 | if (ni > 0 && s[ni - 1] != ':' && s[ni - 1] != '/') { |
| 739 | s[ni] = '/'; |
| 740 | mmv(s + ni + 1, name, nl + 1); |
| 741 | } else { |
| 742 | mmv(s + ni, name, nl + 1); |
| 743 | } |
| 744 | #else |
| 745 | s[ni] = '/'; |
| 746 | mmv(s + ni + 1, name, nl + 1); |
| 747 | #endif |
| 748 | |
| 749 | #ifdef UCPP_MMAP |
| 750 | f = fopen_mmap_file(s); |
| 751 | #else |
| 752 | f = fopen(s, "r"); |
| 753 | #endif |
| 754 | if (f) { |
| 755 | if (emit_dependencies == 2) { |
| 756 | fprintf(emit_output, " %s", s ? s : name); |
| 757 | } |
| 758 | current_long_filename = s; |
| 759 | current_incdir = i; |
| 760 | return f; |
| 761 | } |
| 762 | freemem(s); |
| 763 | } |
| 764 | return 0; |
| 765 | } |
| 766 | |
| 767 | /* |
| 768 | * The #if directive. This function parse the expression, performs macro |
| 769 | * expansion (and handles the "defined" operator), and call eval_expr. |
| 770 | * return value: 1 if the expression is true, 0 if it is false, -1 on error. |
| 771 | */ |
| 772 | static int handle_if(struct lexer_state *ls) |
| 773 | { |
| 774 | struct token_fifo tf, tf1, tf2, tf3, *save_tf; |
| 775 | long l = ls->line; |
| 776 | unsigned long z; |
| 777 | int ret = 0, ltww = 1; |
| 778 | |
| 779 | /* first, get the whole line */ |
| 780 | tf.art = tf.nt = 0; |
| 781 | while (!next_token(ls) && ls->ctok->type != NEWLINE) { |
| 782 | struct token t; |
| 783 | |
| 784 | if (ltww && ttMWS(ls->ctok->type)) continue; |
| 785 | ltww = ttMWS(ls->ctok->type); |
| 786 | t.type = ls->ctok->type; |
| 787 | t.line = l; |
| 788 | if (S_TOKEN(ls->ctok->type)) { |
| 789 | t.name = sdup(ls->ctok->name); |
| 790 | throw_away(ls->gf, t.name); |
| 791 | } |
| 792 | aol(tf.t, tf.nt, t, TOKEN_LIST_MEMG); |
| 793 | } |
| 794 | if (ltww && tf.nt) if ((-- tf.nt) == 0) freemem(tf.t); |
| 795 | if (tf.nt == 0) { |
| 796 | error(l, "void condition for a #if/#elif"); |
| 797 | return -1; |
| 798 | } |
| 799 | /* handle the "defined" operator */ |
| 800 | tf1.art = tf1.nt = 0; |
| 801 | while (tf.art < tf.nt) { |
| 802 | struct token *ct, rt; |
| 803 | struct macro *m; |
| 804 | size_t nidx, eidx; |
| 805 | |
| 806 | ct = tf.t + (tf.art ++); |
| 807 | if (ct->type == NAME && !strcmp(ct->name, "defined")) { |
| 808 | if (tf.art >= tf.nt) goto store_token; |
| 809 | nidx = tf.art; |
| 810 | if (ttMWS(tf.t[nidx].type)) |
| 811 | if (++ nidx >= tf.nt) goto store_token; |
| 812 | if (tf.t[nidx].type == NAME) { |
| 813 | eidx = nidx; |
| 814 | goto check_macro; |
| 815 | } |
| 816 | if (tf.t[nidx].type != LPAR) { |
| 817 | #ifdef HAVE_MISRA |
| 818 | misra_neu(100,19,14,0); |
| 819 | #endif |
| 820 | goto store_token;} |
| 821 | if (++ nidx >= tf.nt) goto store_token; |
| 822 | if (ttMWS(tf.t[nidx].type)) |
| 823 | if (++ nidx >= tf.nt) goto store_token; |
| 824 | if (tf.t[nidx].type != NAME) { |
| 825 | #ifdef HAVE_MISRA |
| 826 | misra_neu(100,19,14,0); |
| 827 | #endif |
| 828 | goto store_token;} |
| 829 | eidx = nidx + 1; |
| 830 | if (eidx >= tf.nt) goto store_token; |
| 831 | if (ttMWS(tf.t[eidx].type)) |
| 832 | if (++ eidx >= tf.nt) goto store_token; |
| 833 | if (tf.t[eidx].type != RPAR) goto store_token; |
| 834 | goto check_macro; |
| 835 | } |
| 836 | store_token: |
| 837 | aol(tf1.t, tf1.nt, *ct, TOKEN_LIST_MEMG); |
| 838 | continue; |
| 839 | |
| 840 | check_macro: |
| 841 | m = get_macro(tf.t[nidx].name); |
| 842 | rt.type = NUMBER; |
| 843 | rt.name = m ? "1L" : "0L"; |
| 844 | aol(tf1.t, tf1.nt, rt, TOKEN_LIST_MEMG); |
| 845 | tf.art = eidx + 1; |
| 846 | } |
| 847 | freemem(tf.t); |
| 848 | if (tf1.nt == 0) { |
| 849 | error(l, "void condition (after expansion) for a #if/#elif"); |
| 850 | return -1; |
| 851 | } |
| 852 | |
| 853 | /* perform all macro substitutions */ |
| 854 | tf2.art = tf2.nt = 0; |
| 855 | save_tf = ls->output_fifo; |
| 856 | ls->output_fifo = &tf2; |
| 857 | while (tf1.art < tf1.nt) { |
| 858 | struct token *ct; |
| 859 | |
| 860 | ct = tf1.t + (tf1.art ++); |
| 861 | if (ct->type == NAME) { |
| 862 | struct macro *m = get_macro(ct->name); |
| 863 | |
| 864 | if (m) { |
| 865 | if (substitute_macro(ls, m, &tf1, 0, |
| 866 | #ifdef NO_PRAGMA_IN_DIRECTIVE |
| 867 | 1, |
| 868 | #else |
| 869 | 0, |
| 870 | #endif |
| 871 | ct->line)) { |
| 872 | ls->output_fifo = save_tf; |
| 873 | return -1; |
| 874 | } |
| 875 | continue; |
| 876 | } |
| 877 | } else if ((ct->type == SHARP || ct->type == DIG_SHARP) |
| 878 | && (ls->flags & HANDLE_ASSERTIONS)) { |
| 879 | /* we have an assertion; parse it */ |
| 880 | int nnp, ltww = 1; |
| 881 | size_t i = tf1.art; |
| 882 | struct token_fifo atl; |
| 883 | char *aname; |
| 884 | struct assert *a; |
| 885 | int av = 0; |
| 886 | struct token rt; |
| 887 | |
| 888 | atl.art = atl.nt = 0; |
| 889 | while (i < tf1.nt && ttMWS(tf1.t[i].type)) i ++; |
| 890 | if (i >= tf1.nt) goto assert_error; |
| 891 | if (tf1.t[i].type != NAME) goto assert_error; |
| 892 | aname = tf1.t[i ++].name; |
| 893 | while (i < tf1.nt && ttMWS(tf1.t[i].type)) i ++; |
| 894 | if (i >= tf1.nt) goto assert_generic; |
| 895 | if (tf1.t[i].type != LPAR) goto assert_generic; |
| 896 | i ++; |
| 897 | for (nnp = 1; nnp && i < tf1.nt; i ++) { |
| 898 | if (ltww && ttMWS(tf1.t[i].type)) continue; |
| 899 | if (tf1.t[i].type == LPAR) nnp ++; |
| 900 | else if (tf1.t[i].type == RPAR |
| 901 | && (-- nnp) == 0) { |
| 902 | tf1.art = i + 1; |
| 903 | break; |
| 904 | } |
| 905 | ltww = ttMWS(tf1.t[i].type); |
| 906 | aol(atl.t, atl.nt, tf1.t[i], TOKEN_LIST_MEMG); |
| 907 | } |
| 908 | if (nnp) goto assert_error; |
| 909 | if (ltww && atl.nt && (-- atl.nt) == 0) freemem(atl.t); |
| 910 | if (atl.nt == 0) goto assert_error; |
| 911 | |
| 912 | /* the assertion is in aname and atl; check it */ |
| 913 | a = get_assertion(aname); |
| 914 | if (a) for (i = 0; i < a->nbval; i ++) |
| 915 | if (!cmp_token_list(&atl, a->val + i)) { |
| 916 | av = 1; |
| 917 | break; |
| 918 | } |
| 919 | rt.type = NUMBER; |
| 920 | rt.name = av ? "1" : "0"; |
| 921 | aol(tf2.t, tf2.nt, rt, TOKEN_LIST_MEMG); |
| 922 | continue; |
| 923 | |
| 924 | assert_generic: |
| 925 | tf1.art = i; |
| 926 | rt.type = NUMBER; |
| 927 | rt.name = get_assertion(aname) ? "1" : "0"; |
| 928 | aol(tf2.t, tf2.nt, rt, TOKEN_LIST_MEMG); |
| 929 | continue; |
| 930 | |
| 931 | assert_error: |
| 932 | error(l, "syntax error for assertion in #if"); |
| 933 | ls->output_fifo = save_tf; |
| 934 | return -1; |
| 935 | } |
| 936 | aol(tf2.t, tf2.nt, *ct, TOKEN_LIST_MEMG); |
| 937 | } |
| 938 | ls->output_fifo = save_tf; |
| 939 | freemem(tf1.t); |
| 940 | if (tf2.nt == 0) { |
| 941 | error(l, "void condition (after expansion) for a #if/#elif"); |
| 942 | return -1; |
| 943 | } |
| 944 | |
| 945 | /* |
| 946 | * suppress whitespace and replace rogue identifiers by 0 |
| 947 | */ |
| 948 | tf3.art = tf3.nt = 0; |
| 949 | while (tf2.art < tf2.nt) { |
| 950 | struct token *ct = tf2.t + (tf2.art ++); |
| 951 | |
| 952 | if (ttMWS(ct->type)) continue; |
| 953 | if (ct->type == NAME) { |
| 954 | /* |
| 955 | * a rogue identifier; we replace it with "0". |
| 956 | */ |
| 957 | struct token rt; |
| 958 | #ifdef HAVE_MISRA |
| 959 | misra_neu(97,19,11,0); |
| 960 | #endif |
| 961 | |
| 962 | rt.type = NUMBER; |
| 963 | rt.name = "0"; |
| 964 | aol(tf3.t, tf3.nt, rt, TOKEN_LIST_MEMG); |
| 965 | continue; |
| 966 | } |
| 967 | aol(tf3.t, tf3.nt, *ct, TOKEN_LIST_MEMG); |
| 968 | } |
| 969 | freemem(tf2.t); |
| 970 | |
| 971 | if (tf3.nt == 0) { |
| 972 | error(l, "void condition (after expansion) for a #if/#elif"); |
| 973 | return -1; |
| 974 | } |
| 975 | eval_line = l; |
| 976 | z = eval_expr(&tf3, &ret, (ls->flags & WARN_STANDARD) != 0); |
| 977 | freemem(tf3.t); |
| 978 | if (ret) return -1; |
| 979 | return (z != 0); |
| 980 | } |
| 981 | |
| 982 | /* |
| 983 | * A #include was found; parse the end of line, replace macros if |
| 984 | * necessary. |
| 985 | * |
| 986 | * If nex is set to non-zero, the directive is considered as a #include_next |
| 987 | * (extension to C99) |
| 988 | */ |
| 989 | static int handle_include(struct lexer_state *ls, unsigned long flags, int nex) |
| 990 | { |
| 991 | int c, string_fname = 0; |
| 992 | char *fname; |
| 993 | size_t fname_ptr = 0; |
| 994 | long l = ls->line; |
| 995 | int x, y; |
| 996 | FILE *f; |
| 997 | struct token_fifo tf, tf2, *save_tf; |
| 998 | size_t nl; |
| 999 | int tgd; |
| 1000 | |
| 1001 | #define left_angle(t) ((t) == LT || (t) == LEQ || (t) == LSH \ |
| 1002 | || (t) == ASLSH || (t) == DIG_LBRK || (t) == LBRA) |
| 1003 | #define right_angle(t) ((t) == GT || (t) == RSH || (t) == ARROW \ |
| 1004 | || (t) == DIG_RBRK || (t) == DIG_RBRA) |
| 1005 | |
| 1006 | #ifdef HAVE_MISRA |
| 1007 | extern int misracheck,misratok; |
| 1008 | |
| 1009 | if(misracheck&&misratok) misra_neu(87,19,1,0); |
| 1010 | #endif |
| 1011 | tf.t=0; /*vb*/ |
| 1012 | while ((c = grap_char(ls)) >= 0 && c != '\n') { |
| 1013 | if (space_char(c)) { |
| 1014 | discard_char(ls); |
| 1015 | continue; |
| 1016 | } |
| 1017 | if (c == '<') { |
| 1018 | discard_char(ls); |
| 1019 | while ((c = grap_char(ls)) >= 0) { |
| 1020 | discard_char(ls); |
| 1021 | if (c == '\n') goto include_error; |
| 1022 | if (c == '>') break; |
| 1023 | aol(fname, fname_ptr, (char)c, FNAME_MEMG); |
| 1024 | } |
| 1025 | aol(fname, fname_ptr, (char)0, FNAME_MEMG); |
| 1026 | string_fname = 0; |
| 1027 | goto do_include; |
| 1028 | } else if (c == '"') { |
| 1029 | discard_char(ls); |
| 1030 | while ((c = grap_char(ls)) >= 0) { |
| 1031 | discard_char(ls); |
| 1032 | if (c == '\n') goto include_error; |
| 1033 | if (c == '"') break; |
| 1034 | aol(fname, fname_ptr, (char)c, FNAME_MEMG); |
| 1035 | } |
| 1036 | aol(fname, fname_ptr, (char)0, FNAME_MEMG); |
| 1037 | string_fname = 1; |
| 1038 | goto do_include; |
| 1039 | } |
| 1040 | goto include_macro; |
| 1041 | } |
| 1042 | |
| 1043 | include_error: |
| 1044 | error(l, "invalid '#include'"); |
| 1045 | return 1; |
| 1046 | |
| 1047 | include_macro: |
| 1048 | #ifdef HAVE_MISRA |
| 1049 | misra(90); |
| 1050 | #endif |
| 1051 | tf.art = tf.nt = 0; |
| 1052 | while (!next_token(ls) && ls->ctok->type != NEWLINE) { |
| 1053 | if (!ttMWS(ls->ctok->type)) { |
| 1054 | struct token t; |
| 1055 | |
| 1056 | t.type = ls->ctok->type; |
| 1057 | t.line = l; |
| 1058 | if (S_TOKEN(ls->ctok->type)) { |
| 1059 | t.name = sdup(ls->ctok->name); |
| 1060 | throw_away(ls->gf, t.name); |
| 1061 | } |
| 1062 | aol(tf.t, tf.nt, t, TOKEN_LIST_MEMG); |
| 1063 | } |
| 1064 | } |
| 1065 | tf2.art = tf2.nt = 0; |
| 1066 | save_tf = ls->output_fifo; |
| 1067 | ls->output_fifo = &tf2; |
| 1068 | while (tf.art < tf.nt) { |
| 1069 | struct token *ct; |
| 1070 | |
| 1071 | ct = tf.t + (tf.art ++); |
| 1072 | if (ct->type == NAME) { |
| 1073 | struct macro *m = get_macro(ct->name); |
| 1074 | if (m) { |
| 1075 | if (substitute_macro(ls, m, &tf, 0, |
| 1076 | #ifdef NO_PRAGMA_IN_DIRECTIVE |
| 1077 | 1, |
| 1078 | #else |
| 1079 | 0, |
| 1080 | #endif |
| 1081 | ct->line)) { |
| 1082 | ls->output_fifo = save_tf; |
| 1083 | return -1; |
| 1084 | } |
| 1085 | continue; |
| 1086 | } |
| 1087 | } |
| 1088 | aol(tf2.t, tf2.nt, *ct, TOKEN_LIST_MEMG); |
| 1089 | } |
| 1090 | freemem(tf.t); |
| 1091 | ls->output_fifo = save_tf; |
| 1092 | for (x = 0; x < tf2.nt && ttWHI(tf2.t[x].type); x ++); |
| 1093 | for (y = tf2.nt - 1; y >= 0 && ttWHI(tf2.t[y].type); y --); |
| 1094 | if (x >= tf2.nt) goto include_macro_err; |
| 1095 | if (tf2.t[x].type == STRING) { |
| 1096 | if (y != x) goto include_macro_err; |
| 1097 | if (tf2.t[x].name[0] == 'L') { |
| 1098 | if (ls->flags & WARN_STANDARD) |
| 1099 | warning(l, "wide string for #include"); |
| 1100 | fname = sdup(tf2.t[x].name); |
| 1101 | nl = strlen(fname); |
| 1102 | *(fname + nl - 1) = 0; |
| 1103 | mmvwo(fname, fname + 2, nl - 2); |
| 1104 | } else { |
| 1105 | fname = sdup(tf2.t[x].name); |
| 1106 | nl = strlen(fname); |
| 1107 | *(fname + nl - 1) = 0; |
| 1108 | mmvwo(fname, fname + 1, nl - 1); |
| 1109 | } |
| 1110 | string_fname = 1; |
| 1111 | } else if (left_angle(tf2.t[x].type) && right_angle(tf2.t[y].type)) { |
| 1112 | int i, j; |
| 1113 | |
| 1114 | if (ls->flags & WARN_ANNOYING) warning(l, "reconstruction " |
| 1115 | "of <foo> in #include"); |
| 1116 | for (j = 0, i = x; i <= y; i ++) if (!ttWHI(tf2.t[i].type)) |
| 1117 | j += strlen(tname(tf2.t[i])); |
| 1118 | fname = getmem(j + 1); |
| 1119 | for (j = 0, i = x; i <= y; i ++) { |
| 1120 | if (ttWHI(tf2.t[i].type)) continue; |
| 1121 | strcpy(fname + j, tname(tf2.t[i])); |
| 1122 | j += strlen(tname(tf2.t[i])); |
| 1123 | } |
| 1124 | *(fname + j - 1) = 0; |
| 1125 | mmvwo(fname, fname + 1, j); |
| 1126 | string_fname = 0; |
| 1127 | } else goto include_macro_err; |
| 1128 | freemem(tf2.t); |
| 1129 | goto do_include_next; |
| 1130 | |
| 1131 | include_macro_err: |
| 1132 | error(l, "macro expansion did not produce a valid filename " |
| 1133 | "for #include"); |
| 1134 | if (tf2.nt) freemem(tf2.t); |
| 1135 | return 1; |
| 1136 | |
| 1137 | do_include: |
| 1138 | tgd = 1; |
| 1139 | while (!next_token(ls)) { |
| 1140 | if (tgd && !ttWHI(ls->ctok->type) |
| 1141 | && (ls->flags & WARN_STANDARD)) { |
| 1142 | warning(l, "trailing garbage in #include"); |
| 1143 | tgd = 0; |
| 1144 | } |
| 1145 | if (ls->ctok->type == NEWLINE) break; |
| 1146 | } |
| 1147 | |
| 1148 | /* the increment of ls->line is intended so that the line |
| 1149 | numbering is reported correctly in report_context() even if |
| 1150 | the #include is at the end of the line with no trailing newline */ |
| 1151 | if (ls->ctok->type != NEWLINE) ls->line ++; |
| 1152 | do_include_next: |
| 1153 | if (!(ls->flags & LEXER) && (ls->flags & KEEP_OUTPUT)) |
| 1154 | put_char(ls, '\n'); |
| 1155 | push_file_context(ls); |
| 1156 | reinit_lexer_state(ls, 1); |
| 1157 | #ifdef HAVE_MISRA |
| 1158 | if(misracheck){ |
| 1159 | char *p; |
| 1160 | for(p=fname;*p;p++) |
| 1161 | if(*p=='\''||*p=='\\'||*p=='\"'||(*p=='/'&&p[1]=='*')) |
| 1162 | misra_neu(88,19,2,-1); |
| 1163 | } |
| 1164 | #endif |
| 1165 | { |
| 1166 | char *conv_fname = convert_path(fname); |
| 1167 | freemem(fname); |
| 1168 | fname = conv_fname; |
| 1169 | } |
| 1170 | f = nex ? find_file_next(fname) : find_file(fname, string_fname); |
| 1171 | if (!f) { |
| 1172 | pop_file_context(ls); |
| 1173 | if (find_file_error == FF_ERROR) { |
| 1174 | error(l, "file '%s' not found", fname); |
| 1175 | return 1; |
| 1176 | } |
| 1177 | /* file was found, but it is useless to include it again */ |
| 1178 | return 0; |
| 1179 | } |
| 1180 | #ifdef UCPP_MMAP |
| 1181 | set_input_file(ls, f); |
| 1182 | #else |
| 1183 | ls->input = f; |
| 1184 | #endif |
| 1185 | #ifdef HAVE_MISRA |
| 1186 | if (!strcmp("signal.h",fname)) misra_neu(123,20,8,-1); |
| 1187 | if (!strcmp("stdio.h",fname)) misra_neu(124,20,9,-1); |
| 1188 | if (!strcmp("time.h",fname)) misra(127,20,12,-1); |
| 1189 | #endif |
| 1190 | set_current_filename(fname); |
| 1191 | handle_deps(current_long_filename,string_fname); |
| 1192 | enter_file(ls, flags); |
| 1193 | return 0; |
| 1194 | |
| 1195 | #undef left_angle |
| 1196 | #undef right_angle |
| 1197 | } |
| 1198 | |
| 1199 | /* |
| 1200 | * for #line directives |
| 1201 | */ |
| 1202 | static int handle_line(struct lexer_state *ls, unsigned long flags) |
| 1203 | { |
| 1204 | char *fname; |
| 1205 | long l = ls->line; |
| 1206 | struct token_fifo tf, tf2, *save_tf; |
| 1207 | size_t nl, j; |
| 1208 | unsigned long z; |
| 1209 | |
| 1210 | tf.art = tf.nt = 0; |
| 1211 | while (!next_token(ls) && ls->ctok->type != NEWLINE) { |
| 1212 | if (!ttMWS(ls->ctok->type)) { |
| 1213 | struct token t; |
| 1214 | |
| 1215 | t.type = ls->ctok->type; |
| 1216 | t.line = l; |
| 1217 | if (S_TOKEN(ls->ctok->type)) { |
| 1218 | t.name = sdup(ls->ctok->name); |
| 1219 | throw_away(ls->gf, t.name); |
| 1220 | } |
| 1221 | aol(tf.t, tf.nt, t, TOKEN_LIST_MEMG); |
| 1222 | } |
| 1223 | } |
| 1224 | tf2.art = tf2.nt = 0; |
| 1225 | save_tf = ls->output_fifo; |
| 1226 | ls->output_fifo = &tf2; |
| 1227 | while (tf.art < tf.nt) { |
| 1228 | struct token *ct; |
| 1229 | |
| 1230 | ct = tf.t + (tf.art ++); |
| 1231 | if (ct->type == NAME) { |
| 1232 | struct macro *m = get_macro(ct->name); |
| 1233 | if (m) { |
| 1234 | if (substitute_macro(ls, m, &tf, 0, |
| 1235 | #ifdef NO_PRAGMA_IN_DIRECTIVE |
| 1236 | 1, |
| 1237 | #else |
| 1238 | 0, |
| 1239 | #endif |
| 1240 | ct->line)) { |
| 1241 | ls->output_fifo = save_tf; |
| 1242 | return -1; |
| 1243 | } |
| 1244 | continue; |
| 1245 | } |
| 1246 | } |
| 1247 | aol(tf2.t, tf2.nt, *ct, TOKEN_LIST_MEMG); |
| 1248 | } |
| 1249 | freemem(tf.t); |
| 1250 | for (tf2.art = 0; tf2.art < tf2.nt && ttWHI(tf2.t[tf2.art].type); |
| 1251 | tf2.art ++); |
| 1252 | ls->output_fifo = save_tf; |
| 1253 | if (tf2.art == tf2.nt || (tf2.t[tf2.art].type != NUMBER |
| 1254 | && tf2.t[tf2.art].type != CHAR)) { |
| 1255 | error(l, "not a valid number for #line"); |
| 1256 | goto line_macro_err; |
| 1257 | } |
| 1258 | for (j = 0; tf2.t[tf2.art].name[j]; j ++) |
| 1259 | if (tf2.t[tf2.art].name[j] < '0' |
| 1260 | || tf2.t[tf2.art].name[j] > '9') |
| 1261 | if (ls->flags & WARN_STANDARD) |
| 1262 | warning(l, "non-standard line number in #line"); |
| 1263 | if (catch(eval_exception)) goto line_macro_err; |
| 1264 | z = strtoconst(tf2.t[tf2.art].name); |
| 1265 | if (j > 10 || z > 2147483647U) { |
| 1266 | error(l, "out-of-bound line number for #line"); |
| 1267 | goto line_macro_err; |
| 1268 | } |
| 1269 | ls->oline = ls->line = z; |
| 1270 | if ((++ tf2.art) < tf2.nt) { |
| 1271 | int i; |
| 1272 | |
| 1273 | for (i = tf2.art; i < tf2.nt && ttMWS(tf2.t[i].type); i ++); |
| 1274 | if (i < tf2.nt) { |
| 1275 | if (tf2.t[i].type != STRING) { |
| 1276 | error(l, "not a valid filename for #line"); |
| 1277 | goto line_macro_err; |
| 1278 | } |
| 1279 | if (tf2.t[i].name[0] == 'L') { |
| 1280 | if (ls->flags & WARN_STANDARD) { |
| 1281 | warning(l, "wide string for #line"); |
| 1282 | } |
| 1283 | fname = sdup(tf2.t[i].name); |
| 1284 | nl = strlen(fname); |
| 1285 | *(fname + nl - 1) = 0; |
| 1286 | mmvwo(fname, fname + 2, nl - 2); |
| 1287 | } else { |
| 1288 | fname = sdup(tf2.t[i].name); |
| 1289 | nl = strlen(fname); |
| 1290 | *(fname + nl - 1) = 0; |
| 1291 | mmvwo(fname, fname + 1, nl - 1); |
| 1292 | } |
| 1293 | set_current_filename(fname); |
| 1294 | } |
| 1295 | for (i ++; i < tf2.nt && ttMWS(tf2.t[i].type); i ++); |
| 1296 | if (i < tf2.nt && (ls->flags & WARN_STANDARD)) { |
| 1297 | warning(l, "trailing garbage in #line"); |
| 1298 | } |
| 1299 | } |
| 1300 | freemem(tf2.t); |
| 1301 | enter_file(ls, flags); |
| 1302 | return 0; |
| 1303 | |
| 1304 | line_macro_err: |
| 1305 | if (tf2.nt) freemem(tf2.t); |
| 1306 | return 1; |
| 1307 | } |
| 1308 | |
| 1309 | /* |
| 1310 | * a #error or #warning directive: we emit the message without any |
| 1311 | * modification (except the usual backslash+newline and trigraphs) |
| 1312 | */ |
| 1313 | static void handle_error(struct lexer_state *ls,int warnflag) |
| 1314 | { |
| 1315 | int c; |
| 1316 | size_t p = 0, lp = 128; |
| 1317 | long l = ls->line; |
| 1318 | unsigned char *buf = getmem(lp); |
| 1319 | |
| 1320 | while ((c = grap_char(ls)) >= 0 && c != '\n') { |
| 1321 | discard_char(ls); |
| 1322 | wan(buf, p, (unsigned char)c, lp); |
| 1323 | } |
| 1324 | wan(buf, p, 0, lp); |
| 1325 | if (warnflag) |
| 1326 | warning(l, "#warning%s", buf); |
| 1327 | else |
| 1328 | error(l, "#error%s", buf); |
| 1329 | freemem(buf); |
| 1330 | } |
| 1331 | |
| 1332 | /* |
| 1333 | * convert digraph tokens to their standard equivalent. |
| 1334 | */ |
| 1335 | static int undig(int type) |
| 1336 | { |
| 1337 | static int ud[6] = { LBRK, RBRK, LBRA, RBRA, SHARP, DSHARP }; |
| 1338 | |
| 1339 | return ud[type - DIG_LBRK]; |
| 1340 | } |
| 1341 | |
| 1342 | #ifdef PRAGMA_TOKENIZE |
| 1343 | /* |
| 1344 | * Make a compressed representation of a token list; the contents of |
| 1345 | * the token_fifo are freed. Values equal to 0 are replaced by |
| 1346 | * PRAGMA_TOKEN_END (by default, (unsigned char)'\n') and the compressed |
| 1347 | * string is padded by a 0 (so that it may be * handled like a string). |
| 1348 | * Digraph tokens are replaced by their non-digraph equivalents. |
| 1349 | */ |
| 1350 | struct comp_token_fifo compress_token_list(struct token_fifo *tf) |
| 1351 | { |
| 1352 | struct comp_token_fifo ct; |
| 1353 | size_t l; |
| 1354 | |
| 1355 | for (l = 0, tf->art = 0; tf->art < tf->nt; tf->art ++) { |
| 1356 | l ++; |
| 1357 | if (S_TOKEN(tf->t[tf->art].type)) |
| 1358 | l += strlen(tf->t[tf->art].name) + 1; |
| 1359 | } |
| 1360 | ct.t = getmem((ct.length = l) + 1); |
| 1361 | for (l = 0, tf->art = 0; tf->art < tf->nt; tf->art ++) { |
| 1362 | int tt = tf->t[tf->art].type; |
| 1363 | |
| 1364 | if (tt == 0) tt = PRAGMA_TOKEN_END; |
| 1365 | if (tt > DIGRAPH_TOKENS && tt < DIGRAPH_TOKENS_END) |
| 1366 | tt = undig(tt); |
| 1367 | ct.t[l ++] = tt; |
| 1368 | if (S_TOKEN(tt)) { |
| 1369 | char *tn = tf->t[tf->art].name; |
| 1370 | size_t sl = strlen(tn); |
| 1371 | |
| 1372 | mmv(ct.t + l, tn, sl); |
| 1373 | l += sl; |
| 1374 | ct.t[l ++] = PRAGMA_TOKEN_END; |
| 1375 | freemem(tn); |
| 1376 | } |
| 1377 | } |
| 1378 | ct.t[l] = 0; |
| 1379 | if (tf->nt) freemem(tf->t); |
| 1380 | ct.rp = 0; |
| 1381 | return ct; |
| 1382 | } |
| 1383 | #endif |
| 1384 | |
| 1385 | /* |
| 1386 | * A #pragma directive: we make a PRAGMA token containing the rest of |
| 1387 | * the line. |
| 1388 | * |
| 1389 | * We strongly hope that we are called only in LEXER mode. |
| 1390 | */ |
| 1391 | static void handle_pragma(struct lexer_state *ls) |
| 1392 | { |
| 1393 | unsigned char *buf; |
| 1394 | struct token t; |
| 1395 | long l = ls->line; |
| 1396 | |
| 1397 | #ifdef PRAGMA_TOKENIZE |
| 1398 | struct token_fifo tf; |
| 1399 | |
| 1400 | tf.art = tf.nt = 0; |
| 1401 | while (!next_token(ls) && ls->ctok->type != NEWLINE) |
| 1402 | if (!ttMWS(ls->ctok->type)) break; |
| 1403 | if (ls->ctok->type != NEWLINE) { |
| 1404 | do { |
| 1405 | struct token t; |
| 1406 | |
| 1407 | t.type = ls->ctok->type; |
| 1408 | if (ttMWS(t.type)) continue; |
| 1409 | if (S_TOKEN(t.type)) t.name = sdup(ls->ctok->name); |
| 1410 | aol(tf.t, tf.nt, t, TOKEN_LIST_MEMG); |
| 1411 | } while (!next_token(ls) && ls->ctok->type != NEWLINE); |
| 1412 | } |
| 1413 | if (tf.nt == 0) { |
| 1414 | /* void pragma are silently ignored */ |
| 1415 | return; |
| 1416 | } |
| 1417 | buf = (compress_token_list(&tf)).t; |
| 1418 | #else |
| 1419 | int c, x = 1, y = 32; |
| 1420 | |
| 1421 | while ((c = grap_char(ls)) >= 0 && c != '\n') { |
| 1422 | discard_char(ls); |
| 1423 | if (!space_char(c)) break; |
| 1424 | } |
| 1425 | /* void #pragma are ignored */ |
| 1426 | if (c == '\n') return; |
| 1427 | buf = getmem(y); |
| 1428 | buf[0] = c; |
| 1429 | while ((c = grap_char(ls)) >= 0 && c != '\n') { |
| 1430 | discard_char(ls); |
| 1431 | wan(buf, x, c, y); |
| 1432 | } |
| 1433 | for (x --; x >= 0 && space_char(buf[x]); x --); |
| 1434 | x ++; |
| 1435 | wan(buf, x, 0, y); |
| 1436 | #endif |
| 1437 | t.type = PRAGMA; |
| 1438 | t.line = l; |
| 1439 | t.name = (char *)buf; |
| 1440 | aol(ls->output_fifo->t, ls->output_fifo->nt, t, TOKEN_LIST_MEMG); |
| 1441 | throw_away(ls->gf, (char *)buf); |
| 1442 | } |
| 1443 | |
| 1444 | /* |
| 1445 | * We saw a # at the beginning of a line (or preceeded only by whitespace). |
| 1446 | * We check the directive name and act accordingly. |
| 1447 | */ |
| 1448 | static int handle_cpp(struct lexer_state *ls, int sharp_type) |
| 1449 | { |
| 1450 | #define condfset(x) do { \ |
| 1451 | ls->condf[(x) / 32] |= 1UL << ((x) % 32); \ |
| 1452 | } while (0) |
| 1453 | #define condfclr(x) do { \ |
| 1454 | ls->condf[(x) / 32] &= ~(1UL << ((x) % 32)); \ |
| 1455 | } while (0) |
| 1456 | #define condfval(x) ((ls->condf[(x) / 32] & (1UL << ((x) % 32))) != 0) |
| 1457 | |
| 1458 | long l = ls->line; |
| 1459 | unsigned long save_flags = ls->flags; |
| 1460 | int ret = 0; |
| 1461 | |
| 1462 | save_flags = ls->flags; |
| 1463 | ls->flags |= LEXER; |
| 1464 | while (!next_token(ls)) { |
| 1465 | int t = ls->ctok->type; |
| 1466 | |
| 1467 | switch (t) { |
| 1468 | case COMMENT: |
| 1469 | if (ls->flags & WARN_ANNOYING) { |
| 1470 | warning(l, "comment in the middle of " |
| 1471 | "a cpp directive"); |
| 1472 | } |
| 1473 | /* fall through */ |
| 1474 | case NONE: |
| 1475 | continue; |
| 1476 | case NEWLINE: |
| 1477 | /* null directive */ |
| 1478 | if (ls->flags & WARN_ANNOYING) { |
| 1479 | /* truly an annoying warning; null directives |
| 1480 | are rare but may increase readability of |
| 1481 | some source files, and they are legal */ |
| 1482 | warning(l, "null cpp directive"); |
| 1483 | } |
| 1484 | if (!(ls->flags & LEXER)) put_char(ls, '\n'); |
| 1485 | goto handle_exit2; |
| 1486 | case NAME: |
| 1487 | break; |
| 1488 | default: |
| 1489 | if (ls->flags & FAIL_SHARP) { |
| 1490 | if (ls->condcomp) { |
| 1491 | error(l, "rogue '#'"); |
| 1492 | ret = 1; |
| 1493 | } else { |
| 1494 | if (ls->flags & WARN_STANDARD) { |
| 1495 | warning(l, "rogue '#' in code " |
| 1496 | "compiled out"); |
| 1497 | ret = 0; |
| 1498 | } |
| 1499 | } |
| 1500 | ls->flags = save_flags; |
| 1501 | goto handle_warp_ign; |
| 1502 | } else { |
| 1503 | struct token u; |
| 1504 | |
| 1505 | u.type = sharp_type; |
| 1506 | u.line = l; |
| 1507 | ls->flags = save_flags; |
| 1508 | print_token(ls, &u, 0); |
| 1509 | print_token(ls, ls->ctok, 0); |
| 1510 | if (ls->flags & WARN_ANNOYING) { |
| 1511 | warning(l, "rogue '#' dumped"); |
| 1512 | } |
| 1513 | goto handle_exit3; |
| 1514 | } |
| 1515 | } |
| 1516 | if (ls->condcomp) { |
| 1517 | if (!strcmp(ls->ctok->name, "define")) { |
| 1518 | ret = handle_define(ls); |
| 1519 | goto handle_exit; |
| 1520 | } else if (!strcmp(ls->ctok->name, "undef")) { |
| 1521 | ret = handle_undef(ls); |
| 1522 | goto handle_exit; |
| 1523 | } else if (!strcmp(ls->ctok->name, "if")) { |
| 1524 | if ((++ ls->ifnest) > 63) goto too_many_if; |
| 1525 | condfclr(ls->ifnest - 1); |
| 1526 | ret = handle_if(ls); |
| 1527 | if (ret > 0) ret = 0; |
| 1528 | else if (ret == 0) { |
| 1529 | ls->condcomp = 0; |
| 1530 | ls->condmet = 0; |
| 1531 | ls->condnest = ls->ifnest - 1; |
| 1532 | } |
| 1533 | else ret = 1; |
| 1534 | goto handle_exit; |
| 1535 | } else if (!strcmp(ls->ctok->name, "ifdef")) { |
| 1536 | if ((++ ls->ifnest) > 63) goto too_many_if; |
| 1537 | condfclr(ls->ifnest - 1); |
| 1538 | |
| 1539 | ret = handle_ifdef(ls); |
| 1540 | if (ret > 0) ret = 0; |
| 1541 | else if (ret == 0) { |
| 1542 | ls->condcomp = 0; |
| 1543 | ls->condmet = 0; |
| 1544 | ls->condnest = ls->ifnest - 1; |
| 1545 | } |
| 1546 | else ret = 1; |
| 1547 | goto handle_exit; |
| 1548 | } else if (!strcmp(ls->ctok->name, "ifndef")) { |
| 1549 | if ((++ ls->ifnest) > 63) goto too_many_if; |
| 1550 | condfclr(ls->ifnest - 1); |
| 1551 | ret = handle_ifndef(ls); |
| 1552 | if (ret > 0) ret = 0; |
| 1553 | else if (ret == 0) { |
| 1554 | ls->condcomp = 0; |
| 1555 | ls->condmet = 0; |
| 1556 | ls->condnest = ls->ifnest - 1; |
| 1557 | } |
| 1558 | else ret = 1; |
| 1559 | goto handle_exit; |
| 1560 | } else if (!strcmp(ls->ctok->name, "else")) { |
| 1561 | if (ls->ifnest == 0 |
| 1562 | || condfval(ls->ifnest - 1)) { |
| 1563 | error(l, "rogue #else"); |
| 1564 | ret = 1; |
| 1565 | goto handle_warp; |
| 1566 | } |
| 1567 | condfset(ls->ifnest - 1); |
| 1568 | if (ls->ifnest == 1) protect_detect.state = 0; |
| 1569 | ls->condcomp = 0; |
| 1570 | ls->condmet = 1; |
| 1571 | ls->condnest = ls->ifnest - 1; |
| 1572 | goto handle_warp; |
| 1573 | } else if (!strcmp(ls->ctok->name, "elif")) { |
| 1574 | if (ls->ifnest == 0 |
| 1575 | || condfval(ls->ifnest - 1)) { |
| 1576 | error(l, "rogue #elif"); |
| 1577 | ret = 1; |
| 1578 | goto handle_warp_ign; |
| 1579 | } |
| 1580 | if (ls->ifnest == 1) protect_detect.state = 0; |
| 1581 | ls->condcomp = 0; |
| 1582 | ls->condmet = 1; |
| 1583 | ls->condnest = ls->ifnest - 1; |
| 1584 | goto handle_warp_ign; |
| 1585 | } else if (!strcmp(ls->ctok->name, "endif")) { |
| 1586 | if (ls->ifnest == 0) { |
| 1587 | error(l, "unmatched #endif"); |
| 1588 | ret = 1; |
| 1589 | goto handle_warp; |
| 1590 | } |
| 1591 | if ((-- ls->ifnest) == 0 |
| 1592 | && protect_detect.state == 2) { |
| 1593 | protect_detect.state = 3; |
| 1594 | } |
| 1595 | goto handle_warp; |
| 1596 | } else if (!strcmp(ls->ctok->name, "include")) { |
| 1597 | ret = handle_include(ls, save_flags, 0); |
| 1598 | goto handle_exit3; |
| 1599 | } else if (!strcmp(ls->ctok->name, "include_next")) { |
| 1600 | ret = handle_include(ls, save_flags, 1); |
| 1601 | goto handle_exit3; |
| 1602 | } else if (!strcmp(ls->ctok->name, "pragma")) { |
| 1603 | if (!(save_flags & LEXER)) { |
| 1604 | #ifdef PRAGMA_DUMP |
| 1605 | /* dump #pragma in output */ |
| 1606 | struct token u; |
| 1607 | |
| 1608 | u.type = sharp_type; |
| 1609 | u.line = l; |
| 1610 | ls->flags = save_flags; |
| 1611 | print_token(ls, &u, 0); |
| 1612 | print_token(ls, ls->ctok, 0); |
| 1613 | while (ls->flags |= LEXER, |
| 1614 | !next_token(ls)) { |
| 1615 | long save_line; |
| 1616 | |
| 1617 | ls->flags &= ~LEXER; |
| 1618 | save_line = ls->line; |
| 1619 | ls->line = l; |
| 1620 | print_token(ls, ls->ctok, 0); |
| 1621 | ls->line = save_line; |
| 1622 | if (ls->ctok->type == NEWLINE) |
| 1623 | break; |
| 1624 | } |
| 1625 | goto handle_exit3; |
| 1626 | #else |
| 1627 | if (ls->flags & WARN_PRAGMA) |
| 1628 | warning(l, "#pragma ignored " |
| 1629 | "and not dumped"); |
| 1630 | goto handle_warp_ign; |
| 1631 | #endif |
| 1632 | } |
| 1633 | if (!(ls->flags & HANDLE_PRAGMA)) |
| 1634 | goto handle_warp_ign; |
| 1635 | handle_pragma(ls); |
| 1636 | goto handle_exit; |
| 1637 | } else if (!strcmp(ls->ctok->name, "error")) { |
| 1638 | ret = 1; |
| 1639 | handle_error(ls,0); |
| 1640 | goto handle_exit; |
| 1641 | } else if (!strcmp(ls->ctok->name, "warning")) { |
| 1642 | ret = 0; |
| 1643 | handle_error(ls,1); |
| 1644 | goto handle_exit; |
| 1645 | } else if (!strcmp(ls->ctok->name, "line")) { |
| 1646 | ret = handle_line(ls, save_flags); |
| 1647 | goto handle_exit; |
| 1648 | } else if ((ls->flags & HANDLE_ASSERTIONS) |
| 1649 | && !strcmp(ls->ctok->name, "assert")) { |
| 1650 | ret = handle_assert(ls); |
| 1651 | goto handle_exit; |
| 1652 | } else if ((ls->flags & HANDLE_ASSERTIONS) |
| 1653 | && !strcmp(ls->ctok->name, "unassert")) { |
| 1654 | ret = handle_unassert(ls); |
| 1655 | goto handle_exit; |
| 1656 | } |
| 1657 | } else { |
| 1658 | if (!strcmp(ls->ctok->name, "else")) { |
| 1659 | if (condfval(ls->ifnest - 1) |
| 1660 | && (ls->flags & WARN_STANDARD)) { |
| 1661 | warning(l, "rogue #else in code " |
| 1662 | "compiled out"); |
| 1663 | } |
| 1664 | if (ls->condnest == ls->ifnest - 1) { |
| 1665 | if (!ls->condmet) ls->condcomp = 1; |
| 1666 | } |
| 1667 | condfset(ls->ifnest - 1); |
| 1668 | if (ls->ifnest == 1) protect_detect.state = 0; |
| 1669 | goto handle_warp; |
| 1670 | } else if (!strcmp(ls->ctok->name, "elif")) { |
| 1671 | if (condfval(ls->ifnest - 1) |
| 1672 | && (ls->flags & WARN_STANDARD)) { |
| 1673 | warning(l, "rogue #elif in code " |
| 1674 | "compiled out"); |
| 1675 | } |
| 1676 | if (ls->condnest != ls->ifnest - 1 |
| 1677 | || ls->condmet) |
| 1678 | goto handle_warp_ign; |
| 1679 | if (ls->ifnest == 1) protect_detect.state = 0; |
| 1680 | ret = handle_if(ls); |
| 1681 | if (ret > 0) { |
| 1682 | ls->condcomp = 1; |
| 1683 | ls->condmet = 1; |
| 1684 | ret = 0; |
| 1685 | } else if (ret < 0) ret = 1; |
| 1686 | goto handle_exit; |
| 1687 | } else if (!strcmp(ls->ctok->name, "endif")) { |
| 1688 | if ((-- ls->ifnest) == ls->condnest) { |
| 1689 | if (ls->ifnest == 0 && |
| 1690 | protect_detect.state == 2) |
| 1691 | protect_detect.state = 3; |
| 1692 | ls->condcomp = 1; |
| 1693 | } |
| 1694 | goto handle_warp; |
| 1695 | } else if (!strcmp(ls->ctok->name, "if") |
| 1696 | || !strcmp(ls->ctok->name, "ifdef") |
| 1697 | || !strcmp(ls->ctok->name, "ifndef")) { |
| 1698 | if ((++ ls->ifnest) > 63) goto too_many_if; |
| 1699 | condfclr(ls->ifnest - 1); |
| 1700 | } |
| 1701 | goto handle_warp_ign; |
| 1702 | } |
| 1703 | /* |
| 1704 | * Unrecognized directive. We emit either an error or |
| 1705 | * an annoying warning, depending on a command-line switch. |
| 1706 | */ |
| 1707 | if (ls->flags & FAIL_SHARP) { |
| 1708 | error(l, "unknown cpp directive '#%s'", |
| 1709 | ls->ctok->name); |
| 1710 | goto handle_warp_ign; |
| 1711 | } else { |
| 1712 | struct token u; |
| 1713 | |
| 1714 | u.type = sharp_type; |
| 1715 | u.line = l; |
| 1716 | ls->flags = save_flags; |
| 1717 | print_token(ls, &u, 0); |
| 1718 | print_token(ls, ls->ctok, 0); |
| 1719 | if (ls->flags & WARN_ANNOYING) { |
| 1720 | warning(l, "rogue '#' dumped"); |
| 1721 | } |
| 1722 | } |
| 1723 | } |
| 1724 | return 1; |
| 1725 | |
| 1726 | handle_warp_ign: |
| 1727 | while (!next_token(ls)) if (ls->ctok->type == NEWLINE) break; |
| 1728 | goto handle_exit; |
| 1729 | handle_warp: |
| 1730 | while (!next_token(ls)) { |
| 1731 | if (!ttWHI(ls->ctok->type) && (ls->flags & WARN_STANDARD)) { |
| 1732 | warning(l, "trailing garbage in " |
| 1733 | "preprocessing directive"); |
| 1734 | } |
| 1735 | if (ls->ctok->type == NEWLINE) break; |
| 1736 | } |
| 1737 | handle_exit: |
| 1738 | if (!(ls->flags & LEXER)) put_char(ls, '\n'); |
| 1739 | handle_exit3: |
| 1740 | if (protect_detect.state == 1) { |
| 1741 | protect_detect.state = 0; |
| 1742 | } else if (protect_detect.state == -1) { |
| 1743 | /* just after the #include */ |
| 1744 | protect_detect.state = 1; |
| 1745 | } |
| 1746 | handle_exit2: |
| 1747 | ls->flags = save_flags; |
| 1748 | return ret; |
| 1749 | too_many_if: |
| 1750 | error(l, "too many levels of conditional inclusion (max 63)"); |
| 1751 | ret = 1; |
| 1752 | goto handle_warp; |
| 1753 | #undef condfset |
| 1754 | #undef condfclr |
| 1755 | #undef condfval |
| 1756 | } |
| 1757 | |
| 1758 | /* |
| 1759 | * This is the main entry function. It maintains count of #, and call the |
| 1760 | * appropriate functions when it encounters a cpp directive or a macro |
| 1761 | * name. |
| 1762 | * return value: positive on error; CPPERR_EOF means "end of input reached" |
| 1763 | */ |
| 1764 | int cpp(struct lexer_state *ls) |
| 1765 | { |
| 1766 | int r = 0; |
| 1767 | |
| 1768 | while (next_token(ls)) { |
| 1769 | if (protect_detect.state == 3) { |
| 1770 | /* Cool ! A protection mechanism has been detected */ |
| 1771 | protect_detect.ff->protect = protect_detect.macro; |
| 1772 | protect_detect.macro = 0; |
| 1773 | } |
| 1774 | if (ls->ifnest) { |
| 1775 | error(ls->line, "unterminated #if construction " |
| 1776 | "(depth %ld)", (long)ls->ifnest); |
| 1777 | r = CPPERR_NEST; |
| 1778 | } |
| 1779 | if (ls_depth == 0) return CPPERR_EOF; |
| 1780 | fclose(ls->input); |
| 1781 | freemem(current_filename); |
| 1782 | pop_file_context(ls); |
| 1783 | if (!(ls->flags & LEXER) && !ls->ltwnl) put_char(ls, '\n'); |
| 1784 | enter_file(ls, ls->flags); |
| 1785 | } |
| 1786 | if (!(ls->ltwnl && (ls->ctok->type == SHARP |
| 1787 | || ls->ctok->type == DIG_SHARP)) |
| 1788 | && protect_detect.state == 1 && !ttWHI(ls->ctok->type)) { |
| 1789 | /* the first non-whitespace token encountered is not |
| 1790 | a sharp introducing a cpp directive */ |
| 1791 | protect_detect.state = 0; |
| 1792 | } |
| 1793 | if (protect_detect.state == 3 && !ttWHI(ls->ctok->type)) { |
| 1794 | /* a non-whitespace token encountered after the #endif */ |
| 1795 | protect_detect.state = 0; |
| 1796 | } |
| 1797 | if (ls->condcomp) { |
| 1798 | if (ls->ltwnl && (ls->ctok->type == SHARP |
| 1799 | || ls->ctok->type == DIG_SHARP)) { |
| 1800 | int x = handle_cpp(ls, ls->ctok->type); |
| 1801 | |
| 1802 | ls->ltwnl = 1; |
| 1803 | return r ? r : x; |
| 1804 | } |
| 1805 | if (ls->ctok->type == NAME) { |
| 1806 | struct macro *m; |
| 1807 | |
| 1808 | if ((m = get_macro(ls->ctok->name)) != 0) { |
| 1809 | int x; |
| 1810 | |
| 1811 | x = substitute_macro(ls, m, 0, 1, 0, |
| 1812 | ls->ctok->line); |
| 1813 | if (!(ls->flags & LEXER)) |
| 1814 | garbage_collect(ls->gf); |
| 1815 | return r ? r : x; |
| 1816 | } |
| 1817 | if (!(ls->flags & LEXER)) |
| 1818 | print_token(ls, ls->ctok, 0); |
| 1819 | } |
| 1820 | } else { |
| 1821 | if (ls->ltwnl && (ls->ctok->type == SHARP |
| 1822 | || ls->ctok->type == DIG_SHARP)) { |
| 1823 | int x = handle_cpp(ls, ls->ctok->type); |
| 1824 | |
| 1825 | ls->ltwnl = 1; |
| 1826 | return r ? r : x; |
| 1827 | } |
| 1828 | } |
| 1829 | if (ls->ctok->type == NEWLINE) ls->ltwnl = 1; |
| 1830 | else if (!ttWHI(ls->ctok->type)) ls->ltwnl = 0; |
| 1831 | return r ? r : -1; |
| 1832 | } |
| 1833 | |
| 1834 | #ifndef STAND_ALONE |
| 1835 | /* |
| 1836 | * llex() and lex() are the lexing functions, when the preprocessor is |
| 1837 | * linked to another code. llex() should be called only by lex(). |
| 1838 | */ |
| 1839 | static int llex(struct lexer_state *ls) |
| 1840 | { |
| 1841 | struct token_fifo *tf = ls->output_fifo; |
| 1842 | int r; |
| 1843 | |
| 1844 | if (tf->nt != 0) { |
| 1845 | if (tf->art < tf->nt) { |
| 1846 | #ifdef INMACRO_FLAG |
| 1847 | if (!ls->inmacro) { |
| 1848 | ls->inmacro = 1; |
| 1849 | ls->macro_count ++; |
| 1850 | } |
| 1851 | #endif |
| 1852 | ls->ctok = tf->t + (tf->art ++); |
| 1853 | if (ls->ctok->type > DIGRAPH_TOKENS |
| 1854 | && ls->ctok->type < DIGRAPH_TOKENS_END) { |
| 1855 | ls->ctok->type = undig(ls->ctok->type); |
| 1856 | } |
| 1857 | return 0; |
| 1858 | } else { |
| 1859 | #ifdef INMACRO_FLAG |
| 1860 | ls->inmacro = 0; |
| 1861 | #endif |
| 1862 | if (tf->nt) freemem(tf->t); |
| 1863 | tf->art = tf->nt = 0; |
| 1864 | garbage_collect(ls->gf); |
| 1865 | ls->ctok = ls->save_ctok; |
| 1866 | } |
| 1867 | |
| 1868 | } |
| 1869 | r = cpp(ls); |
| 1870 | if (ls->ctok->type > DIGRAPH_TOKENS |
| 1871 | && ls->ctok->type < LAST_MEANINGFUL_TOKEN) { |
| 1872 | ls->ctok->type = undig(ls->ctok->type); |
| 1873 | } |
| 1874 | if (r > 0) return r; |
| 1875 | if (r < 0) return 0; |
| 1876 | return llex(ls); |
| 1877 | } |
| 1878 | |
| 1879 | /* |
| 1880 | * lex() reads the next token from the processed stream and stores it |
| 1881 | * into ls->ctok. |
| 1882 | * return value: non zero on error (including CPPERR_EOF, which is not |
| 1883 | * quite an error) |
| 1884 | */ |
| 1885 | int lex(struct lexer_state *ls) |
| 1886 | { |
| 1887 | int r; |
| 1888 | |
| 1889 | do { |
| 1890 | r = llex(ls); |
| 1891 | #ifdef SEMPER_FIDELIS |
| 1892 | } while (!r && !ls->condcomp); |
| 1893 | #else |
| 1894 | } while (!r && (!ls->condcomp || (ttWHI(ls->ctok->type) && |
| 1895 | (!(ls->flags & LINE_NUM) || ls->ctok->type != NEWLINE)))); |
| 1896 | #endif |
| 1897 | return r; |
| 1898 | } |
| 1899 | #endif |
| 1900 | |
| 1901 | /* |
| 1902 | * check_cpp_errors() must be called when the end of input is reached; |
| 1903 | * it checks pending errors due to truncated constructs (actually none, |
| 1904 | * this is reserved for future evolutions). |
| 1905 | */ |
| 1906 | int check_cpp_errors(struct lexer_state *ls) |
| 1907 | { |
| 1908 | return 0; |
| 1909 | } |
| 1910 | |
| 1911 | /* |
| 1912 | * init_cpp() initializes everything that should be initialized. |
| 1913 | * If standard_inc is non-zero, the standard include directories are |
| 1914 | * added to the search path. |
| 1915 | */ |
| 1916 | void init_cpp(void) |
| 1917 | { |
| 1918 | init_cppm(); |
| 1919 | init_buf_lexer_state(&dsharp_lexer, 0); |
| 1920 | #ifdef PRAGMA_TOKENIZE |
| 1921 | init_buf_lexer_state(&tokenize_lexer, 0); |
| 1922 | #endif |
| 1923 | } |
| 1924 | |
| 1925 | /* |
| 1926 | * (re)init the global tables. |
| 1927 | * If standard_assertions is non 0, init the assertions table. |
| 1928 | */ |
| 1929 | void init_tables(int with_assertions) |
| 1930 | { |
| 1931 | time_t t; |
| 1932 | struct tm *ct; |
| 1933 | |
| 1934 | time(&t); |
| 1935 | ct = localtime(&t); |
| 1936 | #ifdef MSDOS |
| 1937 | /* no function strftime() in TurboC 2.01 */ |
| 1938 | { |
| 1939 | char *c = asctime(ct); |
| 1940 | |
| 1941 | compile_time[0] = '"'; |
| 1942 | mmv(compile_time + 1, c + 11, 8); |
| 1943 | compile_time[9] = '"'; |
| 1944 | compile_time[10] = 0; |
| 1945 | compile_date[0] = '"'; |
| 1946 | mmv(compile_date + 1, c + 4, 7); |
| 1947 | mmv(compile_date + 8, c + 20, 4); |
| 1948 | compile_date[12] = '"'; |
| 1949 | compile_date[13] = 0; |
| 1950 | } |
| 1951 | #else |
| 1952 | strftime(compile_time, 12, "\"%H:%M:%S\"", ct); |
| 1953 | strftime(compile_date, 24, "\"%b %d %Y\"", ct); |
| 1954 | #endif |
| 1955 | sprintf(compile_adate,"\"(%02d.%02d.%02d)\"", |
| 1956 | ct->tm_mday,ct->tm_mon+1,ct->tm_year%100); |
| 1957 | init_macros(); |
| 1958 | if (with_assertions) init_assertions(); |
| 1959 | init_found_files(); |
| 1960 | } |
| 1961 | |
| 1962 | /* |
| 1963 | * Resets the include path. |
| 1964 | */ |
| 1965 | void init_include_path(char *incpath[]) |
| 1966 | { |
| 1967 | if (include_path_nb) { |
| 1968 | int i; |
| 1969 | |
| 1970 | for (i = 0; i < include_path_nb; i ++) |
| 1971 | freemem(include_path[i]); |
| 1972 | freemem(include_path); |
| 1973 | include_path_nb = 0; |
| 1974 | } |
| 1975 | if (incpath) { |
| 1976 | int i; |
| 1977 | |
| 1978 | for (i = 0; incpath[i]; i ++) |
| 1979 | aol(include_path, include_path_nb, convert_path(incpath[i]), INCPATH_MEMG); |
| 1980 | } |
| 1981 | } |
| 1982 | |
| 1983 | /* |
| 1984 | * add_incpath() adds "path" to the standard include path. |
| 1985 | */ |
| 1986 | void add_incpath(char *path) |
| 1987 | { |
| 1988 | aol(include_path, include_path_nb, convert_path(path), INCPATH_MEMG); |
| 1989 | } |
| 1990 | |
| 1991 | #ifdef STAND_ALONE |
| 1992 | /* |
| 1993 | * print some help |
| 1994 | */ |
| 1995 | static void usage(char *command_name) |
| 1996 | { |
| 1997 | fprintf(stderr, |
| 1998 | "Usage: %s [options] [file]\n" |
| 1999 | "language options:\n" |
| 2000 | " -C keep comments in output\n" |
| 2001 | " -s keep '#' when no cpp directive is recognized\n" |
| 2002 | " -l do not emit line numbers\n" |
| 2003 | " -lg emit gcc-like line numbers\n" |
| 2004 | " -CC disable C++-like comments\n" |
| 2005 | " -a, -na, -a0 handle (or not) assertions\n" |
| 2006 | " -V disable macros with extra arguments\n" |
| 2007 | " -u understand UTF-8 in source\n" |
| 2008 | " -X enable -a, -u and -Y\n" |
| 2009 | " -c90 mimic C90 behaviour\n" |
| 2010 | " -t disable trigraph support\n" |
| 2011 | "warning options:\n" |
| 2012 | " -wt emit a final warning when trigaphs are encountered\n" |
| 2013 | " -wtt emit warnings for each trigaph encountered\n" |
| 2014 | " -wa emit warnings that are usually useless\n" |
| 2015 | " -w0 disable standard warnings\n" |
| 2016 | "directory options:\n" |
| 2017 | " -I directory add 'directory' before the standard include path\n" |
| 2018 | " -J directory add 'directory' after the standard include path\n" |
| 2019 | " -zI do not use the standard include path\n" |
| 2020 | " -M emit Makefile-like dependencies instead of normal " |
| 2021 | "output\n" |
| 2022 | " -Ma emit also dependancies for system files\n" |
| 2023 | " -o file store output in file\n" |
| 2024 | "macro and assertion options:\n" |
| 2025 | " -Dmacro predefine 'macro'\n" |
| 2026 | " -Dmacro=def predefine 'macro' with 'def' content\n" |
| 2027 | " -Umacro undefine 'macro'\n" |
| 2028 | " -Afoo(bar) assert foo(bar)\n" |
| 2029 | " -Bfoo(bar) unassert foo(bar)\n" |
| 2030 | " -Y predefine system-dependant macros\n" |
| 2031 | " -Z do not predefine special macros\n" |
| 2032 | " -d emit defined macros\n" |
| 2033 | " -e emit assertions\n" |
| 2034 | "misc options:\n" |
| 2035 | " -v print version number and settings\n" |
| 2036 | " -h show this help\n", |
| 2037 | command_name); |
| 2038 | } |
| 2039 | |
| 2040 | /* |
| 2041 | * print version and compile-time settings |
| 2042 | */ |
| 2043 | static void version(void) |
| 2044 | { |
| 2045 | size_t i; |
| 2046 | |
| 2047 | fprintf(stderr, "ucpp version %d.%d\n", VERS_MAJ, VERS_MIN); |
| 2048 | fprintf(stderr, "search path:\n"); |
| 2049 | for (i = 0; i < include_path_nb; i ++) |
| 2050 | fprintf(stderr, " %s\n", include_path[i]); |
| 2051 | } |
| 2052 | |
| 2053 | /* |
| 2054 | * parse_opt() initializes many things according to the command-line |
| 2055 | * options. |
| 2056 | * Return values: |
| 2057 | * 0 on success |
| 2058 | * 1 on semantic error (redefinition of a special macro, for instance) |
| 2059 | * 2 on syntaxic error (unknown options for instance) |
| 2060 | */ |
| 2061 | static int parse_opt(int argc, char *argv[], struct lexer_state *ls) |
| 2062 | { |
| 2063 | int i, ret = 0; |
| 2064 | char *filename = 0; |
| 2065 | int with_std_incpath = 1; |
| 2066 | int print_version = 0, print_defs = 0, print_asserts = 0; |
| 2067 | int system_macros = 0, standard_assertions = 1; |
| 2068 | |
| 2069 | init_lexer_state(ls); |
| 2070 | ls->flags = DEFAULT_CPP_FLAGS; |
| 2071 | emit_output = ls->output = stdout; |
| 2072 | for (i = 1; i < argc; i ++) if (argv[i][0] == '-') { |
| 2073 | if (!strcmp(argv[i], "-h")) { |
| 2074 | return 2; |
| 2075 | } else if (!strcmp(argv[i], "-C")) { |
| 2076 | ls->flags &= ~DISCARD_COMMENTS; |
| 2077 | } else if (!strcmp(argv[i], "-CC")) { |
| 2078 | ls->flags &= ~CPLUSPLUS_COMMENTS; |
| 2079 | } else if (!strcmp(argv[i], "-a")) { |
| 2080 | ls->flags |= HANDLE_ASSERTIONS; |
| 2081 | } else if (!strcmp(argv[i], "-na")) { |
| 2082 | ls->flags |= HANDLE_ASSERTIONS; |
| 2083 | standard_assertions = 0; |
| 2084 | } else if (!strcmp(argv[i], "-a0")) { |
| 2085 | ls->flags &= ~HANDLE_ASSERTIONS; |
| 2086 | } else if (!strcmp(argv[i], "-V")) { |
| 2087 | ls->flags &= ~MACRO_VAARG; |
| 2088 | } else if (!strcmp(argv[i], "-u")) { |
| 2089 | ls->flags |= UTF8_SOURCE; |
| 2090 | } else if (!strcmp(argv[i], "-X")) { |
| 2091 | ls->flags |= HANDLE_ASSERTIONS; |
| 2092 | ls->flags |= UTF8_SOURCE; |
| 2093 | system_macros = 1; |
| 2094 | } else if (!strcmp(argv[i], "-c90")) { |
| 2095 | ls->flags &= ~MACRO_VAARG; |
| 2096 | ls->flags &= ~CPLUSPLUS_COMMENTS; |
| 2097 | c99_compliant = 0; |
| 2098 | c99_hosted = -1; |
| 2099 | } else if (!strcmp(argv[i], "-t")) { |
| 2100 | ls->flags &= ~HANDLE_TRIGRAPHS; |
| 2101 | } else if (!strcmp(argv[i], "-wt")) { |
| 2102 | ls->flags |= WARN_TRIGRAPHS; |
| 2103 | } else if (!strcmp(argv[i], "-wtt")) { |
| 2104 | ls->flags |= WARN_TRIGRAPHS_MORE; |
| 2105 | } else if (!strcmp(argv[i], "-wa")) { |
| 2106 | ls->flags |= WARN_ANNOYING; |
| 2107 | } else if (!strcmp(argv[i], "-w0")) { |
| 2108 | ls->flags &= ~WARN_STANDARD; |
| 2109 | ls->flags &= ~WARN_PRAGMA; |
| 2110 | } else if (!strcmp(argv[i], "-s")) { |
| 2111 | ls->flags &= ~FAIL_SHARP; |
| 2112 | } else if (!strcmp(argv[i], "-l")) { |
| 2113 | ls->flags &= ~LINE_NUM; |
| 2114 | } else if (!strcmp(argv[i], "-lg")) { |
| 2115 | ls->flags |= GCC_LINE_NUM; |
| 2116 | } else if (!strcmp(argv[i], "-M")) { |
| 2117 | ls->flags &= ~KEEP_OUTPUT; |
| 2118 | emit_dependencies = 1; |
| 2119 | } else if (!strcmp(argv[i], "-Ma")) { |
| 2120 | ls->flags &= ~KEEP_OUTPUT; |
| 2121 | emit_dependencies = 2; |
| 2122 | } else if (!strcmp(argv[i], "-Y")) { |
| 2123 | system_macros = 1; |
| 2124 | } else if (!strcmp(argv[i], "-Z")) { |
| 2125 | no_special_macros = 1; |
| 2126 | } else if (!strcmp(argv[i], "-d")) { |
| 2127 | ls->flags &= ~KEEP_OUTPUT; |
| 2128 | print_defs = 1; |
| 2129 | } else if (!strcmp(argv[i], "-e")) { |
| 2130 | ls->flags &= ~KEEP_OUTPUT; |
| 2131 | print_asserts = 1; |
| 2132 | } else if (!strcmp(argv[i], "-zI")) { |
| 2133 | with_std_incpath = 0; |
| 2134 | } else if (!strcmp(argv[i], "-I") || !strcmp(argv[i], "-J")) { |
| 2135 | i ++; |
| 2136 | } else if (!strcmp(argv[i], "-o")) { |
| 2137 | if ((++ i) >= argc) { |
| 2138 | error(-1, "missing filename after -o"); |
| 2139 | return 2; |
| 2140 | } |
| 2141 | if (argv[i][0] == '-' && argv[i][1] == 0) { |
| 2142 | emit_output = ls->output = stdout; |
| 2143 | } else { |
| 2144 | ls->output = fopen(argv[i], "w"); |
| 2145 | if (!ls->output) { |
| 2146 | error(-1, "failed to open for " |
| 2147 | "writing: %s", argv[i]); |
| 2148 | return 2; |
| 2149 | } |
| 2150 | emit_output = ls->output; |
| 2151 | } |
| 2152 | } else if (!strcmp(argv[i], "-v")) { |
| 2153 | print_version = 1; |
| 2154 | } else if (argv[i][1] != 'I' && argv[i][1] != 'J' |
| 2155 | && argv[i][1] != 'D' && argv[i][1] != 'U' |
| 2156 | && argv[i][1] != 'A' && argv[i][1] != 'B') |
| 2157 | warning(-1, "unknown option '%s'", argv[i]); |
| 2158 | } else { |
| 2159 | if (filename != 0) { |
| 2160 | error(-1, "spurious filename '%s'", argv[i]); |
| 2161 | return 2; |
| 2162 | } |
| 2163 | filename = argv[i]; |
| 2164 | } |
| 2165 | init_tables(ls->flags & HANDLE_ASSERTIONS); |
| 2166 | init_include_path(0); |
| 2167 | if (filename) { |
| 2168 | #ifdef UCPP_MMAP |
| 2169 | FILE *f = fopen_mmap_file(filename); |
| 2170 | |
| 2171 | ls->input = 0; |
| 2172 | if (f) set_input_file(ls, f); |
| 2173 | #else |
| 2174 | ls->input = fopen(filename, "r"); |
| 2175 | #endif |
| 2176 | if (!ls->input) { |
| 2177 | error(-1, "file '%s' not found", filename); |
| 2178 | return 1; |
| 2179 | } |
| 2180 | #ifdef NO_LIBC_BUF |
| 2181 | setbuf(ls->input, 0); |
| 2182 | #endif |
| 2183 | set_init_filename(filename, 1); |
| 2184 | } else { |
| 2185 | ls->input = stdin; |
| 2186 | set_init_filename("<stdin>", 0); |
| 2187 | } |
| 2188 | for (i = 1; i < argc; i ++) |
| 2189 | if (argv[i][0] == '-' && argv[i][1] == 'I') |
| 2190 | add_incpath(argv[i][2] ? argv[i] + 2 : argv[i + 1]); |
| 2191 | if (system_macros) for (i = 0; system_macros_def[i]; i ++) |
| 2192 | ret = ret || define_macro(ls, system_macros_def[i]); |
| 2193 | for (i = 1; i < argc; i ++) |
| 2194 | if (argv[i][0] == '-' && argv[i][1] == 'D') |
| 2195 | ret = ret || define_macro(ls, argv[i] + 2); |
| 2196 | for (i = 1; i < argc; i ++) |
| 2197 | if (argv[i][0] == '-' && argv[i][1] == 'U') |
| 2198 | ret = ret || undef_macro(ls, argv[i] + 2); |
| 2199 | if (ls->flags & HANDLE_ASSERTIONS) { |
| 2200 | if (standard_assertions) |
| 2201 | for (i = 0; system_assertions_def[i]; i ++) |
| 2202 | make_assertion(system_assertions_def[i]); |
| 2203 | for (i = 1; i < argc; i ++) |
| 2204 | if (argv[i][0] == '-' && argv[i][1] == 'A') |
| 2205 | ret = ret || make_assertion(argv[i] + 2); |
| 2206 | for (i = 1; i < argc; i ++) |
| 2207 | if (argv[i][0] == '-' && argv[i][1] == 'B') |
| 2208 | ret = ret || destroy_assertion(argv[i] + 2); |
| 2209 | } else { |
| 2210 | for (i = 1; i < argc; i ++) |
| 2211 | if (argv[i][0] == '-' |
| 2212 | && (argv[i][1] == 'A' || argv[i][1] == 'B')) |
| 2213 | warning(-1, "assertions disabled"); |
| 2214 | } |
| 2215 | if (with_std_incpath) { |
| 2216 | for (i = 0; include_path_std[i]; i ++) |
| 2217 | add_incpath(include_path_std[i]); |
| 2218 | } |
| 2219 | for (i = 1; i < argc; i ++) |
| 2220 | if (argv[i][0] == '-' && argv[i][1] == 'J') |
| 2221 | add_incpath(argv[i][2] ? argv[i] + 2 : argv[i + 1]); |
| 2222 | |
| 2223 | if (print_version) { |
| 2224 | version(); |
| 2225 | return 1; |
| 2226 | } |
| 2227 | if (print_defs) { |
| 2228 | print_defines(); |
| 2229 | emit_defines = 1; |
| 2230 | } |
| 2231 | if (print_asserts && (ls->flags & HANDLE_ASSERTIONS)) { |
| 2232 | print_assertions(); |
| 2233 | emit_assertions = 1; |
| 2234 | } |
| 2235 | return ret; |
| 2236 | } |
| 2237 | |
| 2238 | int main(int argc, char *argv[]) |
| 2239 | { |
| 2240 | struct lexer_state ls; |
| 2241 | int r, fr = 0; |
| 2242 | |
| 2243 | init_cpp(); |
| 2244 | if ((r = parse_opt(argc, argv, &ls)) != 0) { |
| 2245 | if (r == 2) usage(argv[0]); |
| 2246 | return EXIT_FAILURE; |
| 2247 | } |
| 2248 | enter_file(&ls, ls.flags); |
| 2249 | while ((r = cpp(&ls)) < CPPERR_EOF) fr = fr || (r > 0); |
| 2250 | fr = fr || check_cpp_errors(&ls); |
| 2251 | if (ls.flags & KEEP_OUTPUT) { |
| 2252 | /* while (ls.line > ls.oline) put_char(&ls, '\n'); */ |
| 2253 | put_char(&ls, '\n'); |
| 2254 | } |
| 2255 | if (emit_dependencies) fputc('\n', emit_output); |
| 2256 | #ifndef NO_UCPP_BUF |
| 2257 | if (!(ls.flags & LEXER)) { |
| 2258 | flush_output(&ls); |
| 2259 | } |
| 2260 | #endif |
| 2261 | if (ls.flags & WARN_TRIGRAPHS && ls.count_trigraphs) |
| 2262 | warning(0, "%ld trigraphs encountered", ls.count_trigraphs); |
| 2263 | return fr ? EXIT_FAILURE : EXIT_SUCCESS; |
| 2264 | } |
| 2265 | #endif |