gcc/cppinit.c

   1 /* CPP Library.
   2    Copyright (C) 1986, 1987, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
   3    1999, 2000, 2001, 2002, 2003 Free Software Foundation, Inc.
   4    Contributed by Per Bothner, 1994-95.
   5    Based on CCCP program by Paul Rubin, June 1986
   6    Adapted to ANSI C, Richard Stallman, Jan 1987
   7
   8 This program is free software; you can redistribute it and/or modify it
   9 under the terms of the GNU General Public License as published by the
  10 Free Software Foundation; either version 2, or (at your option) any
  11 later version.
  12
  13 This program is distributed in the hope that it will be useful,
  14 but WITHOUT ANY WARRANTY; without even the implied warranty of
  15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16 GNU General Public License for more details.
  17
  18 You should have received a copy of the GNU General Public License
  19 along with this program; if not, write to the Free Software
  20 Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  21
  22 #include "config.h"
  23 #include "system.h"
  24 #include "coretypes.h"
  25 #include "tm.h"
  26 #include "cpplib.h"
  27 #include "cpphash.h"
  28 #include "mkdeps.h"
  29
  30 static void init_library (void);
  31 static void mark_named_operators (cpp_reader *);
  32 static void read_original_filename (cpp_reader *);
  33 static void post_options (cpp_reader *);
  34
  35 /* If we have designated initializers (GCC >2.7) these tables can be
  36    initialized, constant data.  Otherwise, they have to be filled in at
  37    runtime.  */
  38 #if HAVE_DESIGNATED_INITIALIZERS
  39
  40 #define init_trigraph_map()  /* Nothing.  */
  41 #define TRIGRAPH_MAP \
  42 __extension__ const uchar _cpp_trigraph_map[UCHAR_MAX + 1] = {
  43
  44 #define END };
  45 #define s(p, v) [p] = v,
  46
  47 #else
  48
  49 #define TRIGRAPH_MAP uchar _cpp_trigraph_map[UCHAR_MAX + 1] = { 0 }; \
  50  static void init_trigraph_map (void) { \
  51  unsigned char *x = _cpp_trigraph_map;
  52
  53 #define END }
  54 #define s(p, v) x[p] = v;
  55
  56 #endif
  57
  58 TRIGRAPH_MAP
  59   s('=', '#')   s(')', ']')     s('!', '|')
  60   s('(', '[')   s('\'', '^')    s('>', '}')
  61   s('/', '\\')  s('<', '{')     s('-', '~')
  62 END
  63
  64 #undef s
  65 #undef END
  66 #undef TRIGRAPH_MAP
  67
  68 /* A set of booleans indicating what CPP features each source language
  69    requires.  */
  70 struct lang_flags
  71 {
  72   char c99;
  73   char cplusplus;
  74   char extended_numbers;
  75   char std;
  76   char cplusplus_comments;
  77   char digraphs;
  78 };
  79
  80 static const struct lang_flags lang_defaults[] =
  81 { /*              c99 c++ xnum std  //   digr  */
  82   /* GNUC89 */  { 0,  0,  1,   0,   1,   1     },
  83   /* GNUC99 */  { 1,  0,  1,   0,   1,   1     },
  84   /* STDC89 */  { 0,  0,  0,   1,   0,   0     },
  85   /* STDC94 */  { 0,  0,  0,   1,   0,   1     },
  86   /* STDC99 */  { 1,  0,  1,   1,   1,   1     },
  87   /* GNUCXX */  { 0,  1,  1,   0,   1,   1     },
  88   /* CXX98  */  { 0,  1,  1,   1,   1,   1     },
  89   /* ASM    */  { 0,  0,  1,   0,   1,   0     }
  90 };
  91
  92 /* Sets internal flags correctly for a given language.  */
  93 void
  94 cpp_set_lang (cpp_reader *pfile, enum c_lang lang)
  95 {
  96   const struct lang_flags *l = &lang_defaults[(int) lang];
  97
  98   CPP_OPTION (pfile, lang) = lang;
  99
 100   CPP_OPTION (pfile, c99)                = l->c99;
 101   CPP_OPTION (pfile, cplusplus)          = l->cplusplus;
 102   CPP_OPTION (pfile, extended_numbers)   = l->extended_numbers;
 103   CPP_OPTION (pfile, std)                = l->std;
 104   CPP_OPTION (pfile, trigraphs)          = l->std;
 105   CPP_OPTION (pfile, cplusplus_comments) = l->cplusplus_comments;
 106   CPP_OPTION (pfile, digraphs)           = l->digraphs;
 107 }
 108
 109 /* Initialize library global state.  */
 110 static void
 111 init_library (void)
 112 {
 113   static int initialized = 0;
 114
 115   if (! initialized)
 116     {
 117       initialized = 1;
 118
 119       /* Set up the trigraph map.  This doesn't need to do anything if
 120          we were compiled with a compiler that supports C99 designated
 121          initializers.  */
 122       init_trigraph_map ();
 123     }
 124 }
 125
 126 /* Initialize a cpp_reader structure.  */
 127 cpp_reader *
 128 cpp_create_reader (enum c_lang lang, hash_table *table)
 129 {
 130   cpp_reader *pfile;
 131
 132   /* Initialize this instance of the library if it hasn't been already.  */
 133   init_library ();
 134
 135   pfile = (cpp_reader *) xcalloc (1, sizeof (cpp_reader));
 136
 137   cpp_set_lang (pfile, lang);
 138   CPP_OPTION (pfile, warn_import) = 1;
 139   CPP_OPTION (pfile, warn_multichar) = 1;
 140   CPP_OPTION (pfile, discard_comments) = 1;
 141   CPP_OPTION (pfile, discard_comments_in_macro_exp) = 1;
 142   CPP_OPTION (pfile, show_column) = 1;
 143   CPP_OPTION (pfile, tabstop) = 8;
 144   CPP_OPTION (pfile, operator_names) = 1;
 145   CPP_OPTION (pfile, warn_trigraphs) = 2;
 146   CPP_OPTION (pfile, warn_endif_labels) = 1;
 147   CPP_OPTION (pfile, warn_deprecated) = 1;
 148   CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
 149   CPP_OPTION (pfile, dollars_in_ident) = 1;
 150   CPP_OPTION (pfile, warn_dollars) = 1;
 151
 152   /* Default CPP arithmetic to something sensible for the host for the
 153      benefit of dumb users like fix-header.  */
 154   CPP_OPTION (pfile, precision) = CHAR_BIT * sizeof (long);
 155   CPP_OPTION (pfile, char_precision) = CHAR_BIT;
 156   CPP_OPTION (pfile, wchar_precision) = CHAR_BIT * sizeof (int);
 157   CPP_OPTION (pfile, int_precision) = CHAR_BIT * sizeof (int);
 158   CPP_OPTION (pfile, unsigned_char) = 0;
 159   CPP_OPTION (pfile, unsigned_wchar) = 1;
 160
 161   /* Initialize the line map.  Start at logical line 1, so we can use
 162      a line number of zero for special states.  */
 163   init_line_maps (&pfile->line_maps);
 164
 165   /* Initialize lexer state.  */
 166   pfile->state.save_comments = ! CPP_OPTION (pfile, discard_comments);
 167
 168   /* Set up static tokens.  */
 169   pfile->avoid_paste.type = CPP_PADDING;
 170   pfile->avoid_paste.val.source = NULL;
 171   pfile->eof.type = CPP_EOF;
 172   pfile->eof.flags = 0;
 173
 174   /* Create a token buffer for the lexer.  */
 175   _cpp_init_tokenrun (&pfile->base_run, 250);
 176   pfile->cur_run = &pfile->base_run;
 177   pfile->cur_token = pfile->base_run.base;
 178
 179   /* Initialize the base context.  */
 180   pfile->context = &pfile->base_context;
 181   pfile->base_context.macro = 0;
 182   pfile->base_context.prev = pfile->base_context.next = 0;
 183
 184   /* Aligned and unaligned storage.  */
 185   pfile->a_buff = _cpp_get_buff (pfile, 0);
 186   pfile->u_buff = _cpp_get_buff (pfile, 0);
 187
 188   /* The expression parser stack.  */
 189   _cpp_expand_op_stack (pfile);
 190
 191   /* Initialize the buffer obstack.  */
 192   gcc_obstack_init (&pfile->buffer_ob);
 193
 194   _cpp_init_includes (pfile);
 195
 196   _cpp_init_hashtable (pfile, table);
 197
 198   return pfile;
 199 }
 200
 201 /* Free resources used by PFILE.  Accessing PFILE after this function
 202    returns leads to undefined behavior.  Returns the error count.  */
 203 void
 204 cpp_destroy (cpp_reader *pfile)
 205 {
 206   cpp_context *context, *contextn;
 207   tokenrun *run, *runn;
 208
 209   free (pfile->op_stack);
 210
 211   while (CPP_BUFFER (pfile) != NULL)
 212     _cpp_pop_buffer (pfile);
 213
 214   if (pfile->out.base)
 215     free (pfile->out.base);
 216
 217   if (pfile->macro_buffer)
 218     {
 219       free (pfile->macro_buffer);
 220       pfile->macro_buffer = NULL;
 221       pfile->macro_buffer_len = 0;
 222     }
 223
 224   if (pfile->deps)
 225     deps_free (pfile->deps);
 226   obstack_free (&pfile->buffer_ob, 0);
 227
 228   _cpp_destroy_hashtable (pfile);
 229   _cpp_cleanup_includes (pfile);
 230
 231   _cpp_free_buff (pfile->a_buff);
 232   _cpp_free_buff (pfile->u_buff);
 233   _cpp_free_buff (pfile->free_buffs);
 234
 235   for (run = &pfile->base_run; run; run = runn)
 236     {
 237       runn = run->next;
 238       free (run->base);
 239       if (run != &pfile->base_run)
 240         free (run);
 241     }
 242
 243   for (context = pfile->base_context.next; context; context = contextn)
 244     {
 245       contextn = context->next;
 246       free (context);
 247     }
 248
 249   free_line_maps (&pfile->line_maps);
 250   free (pfile);
 251 }
 252
 253 /* This structure defines one built-in identifier.  A node will be
 254    entered in the hash table under the name NAME, with value VALUE.
 255
 256    There are two tables of these.  builtin_array holds all the
 257    "builtin" macros: these are handled by builtin_macro() in
 258    cppmacro.c.  Builtin is somewhat of a misnomer -- the property of
 259    interest is that these macros require special code to compute their
 260    expansions.  The value is a "builtin_type" enumerator.
 261
 262    operator_array holds the C++ named operators.  These are keywords
 263    which act as aliases for punctuators.  In C++, they cannot be
 264    altered through #define, and #if recognizes them as operators.  In
 265    C, these are not entered into the hash table at all (but see
 266    <iso646.h>).  The value is a token-type enumerator.  */
 267 struct builtin
 268 {
 269   const uchar *name;
 270   unsigned short len;
 271   unsigned short value;
 272 };
 273
 274 #define B(n, t)    { DSC(n), t }
 275 static const struct builtin builtin_array[] =
 276 {
 277   B("__TIME__",          BT_TIME),
 278   B("__DATE__",          BT_DATE),
 279   B("__FILE__",          BT_FILE),
 280   B("__BASE_FILE__",     BT_BASE_FILE),
 281   B("__LINE__",          BT_SPECLINE),
 282   B("__INCLUDE_LEVEL__", BT_INCLUDE_LEVEL),
 283   /* Keep builtins not used for -traditional-cpp at the end, and
 284      update init_builtins() if any more are added.  */
 285   B("_Pragma",           BT_PRAGMA),
 286   B("__STDC__",          BT_STDC),
 287 };
 288
 289 static const struct builtin operator_array[] =
 290 {
 291   B("and",      CPP_AND_AND),
 292   B("and_eq",   CPP_AND_EQ),
 293   B("bitand",   CPP_AND),
 294   B("bitor",    CPP_OR),
 295   B("compl",    CPP_COMPL),
 296   B("not",      CPP_NOT),
 297   B("not_eq",   CPP_NOT_EQ),
 298   B("or",       CPP_OR_OR),
 299   B("or_eq",    CPP_OR_EQ),
 300   B("xor",      CPP_XOR),
 301   B("xor_eq",   CPP_XOR_EQ)
 302 };
 303 #undef B
 304
 305 /* Mark the C++ named operators in the hash table.  */
 306 static void
 307 mark_named_operators (cpp_reader *pfile)
 308 {
 309   const struct builtin *b;
 310
 311   for (b = operator_array;
 312        b < (operator_array + ARRAY_SIZE (operator_array));
 313        b++)
 314     {
 315       cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
 316       hp->flags |= NODE_OPERATOR;
 317       hp->is_directive = 0;
 318       hp->directive_index = b->value;
 319     }
 320 }
 321
 322 /* Read the builtins table above and enter them, and language-specific
 323    macros, into the hash table.  HOSTED is true if this is a hosted
 324    environment.  */
 325 void
 326 cpp_init_builtins (cpp_reader *pfile, int hosted)
 327 {
 328   const struct builtin *b;
 329   size_t n = ARRAY_SIZE (builtin_array);
 330
 331   if (CPP_OPTION (pfile, traditional))
 332     n -= 2;
 333
 334   for(b = builtin_array; b < builtin_array + n; b++)
 335     {
 336       cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
 337       hp->type = NT_MACRO;
 338       hp->flags |= NODE_BUILTIN | NODE_WARN;
 339       hp->value.builtin = b->value;
 340     }
 341
 342   if (CPP_OPTION (pfile, cplusplus))
 343     _cpp_define_builtin (pfile, "__cplusplus 1");
 344   else if (CPP_OPTION (pfile, lang) == CLK_ASM)
 345     _cpp_define_builtin (pfile, "__ASSEMBLER__ 1");
 346   else if (CPP_OPTION (pfile, lang) == CLK_STDC94)
 347     _cpp_define_builtin (pfile, "__STDC_VERSION__ 199409L");
 348   else if (CPP_OPTION (pfile, c99))
 349     _cpp_define_builtin (pfile, "__STDC_VERSION__ 199901L");
 350
 351   if (hosted)
 352     _cpp_define_builtin (pfile, "__STDC_HOSTED__ 1");
 353   else
 354     _cpp_define_builtin (pfile, "__STDC_HOSTED__ 0");
 355
 356   if (CPP_OPTION (pfile, objc))
 357     _cpp_define_builtin (pfile, "__OBJC__ 1");
 358 }
 359
 360 /* Sanity-checks are dependent on command-line options, so it is
 361    called as a subroutine of cpp_read_main_file ().  */
 362 #if ENABLE_CHECKING
 363 static void sanity_checks (cpp_reader *);
 364 static void sanity_checks (cpp_reader *pfile)
 365 {
 366   cppchar_t test = 0;
 367   size_t max_precision = 2 * CHAR_BIT * sizeof (cpp_num_part);
 368
 369   /* Sanity checks for assumptions about CPP arithmetic and target
 370      type precisions made by cpplib.  */
 371   test--;
 372   if (test < 1)
 373     cpp_error (pfile, DL_ICE, "cppchar_t must be an unsigned type");
 374
 375   if (CPP_OPTION (pfile, precision) > max_precision)
 376     cpp_error (pfile, DL_ICE,
 377                "preprocessor arithmetic has maximum precision of %lu bits;"
 378                " target requires %lu bits",
 379                (unsigned long) max_precision,
 380                (unsigned long) CPP_OPTION (pfile, precision));
 381
 382   if (CPP_OPTION (pfile, precision) < CPP_OPTION (pfile, int_precision))
 383     cpp_error (pfile, DL_ICE,
 384                "CPP arithmetic must be at least as precise as a target int");
 385
 386   if (CPP_OPTION (pfile, char_precision) < 8)
 387     cpp_error (pfile, DL_ICE, "target char is less than 8 bits wide");
 388
 389   if (CPP_OPTION (pfile, wchar_precision) < CPP_OPTION (pfile, char_precision))
 390     cpp_error (pfile, DL_ICE,
 391                "target wchar_t is narrower than target char");
 392
 393   if (CPP_OPTION (pfile, int_precision) < CPP_OPTION (pfile, char_precision))
 394     cpp_error (pfile, DL_ICE,
 395                "target int is narrower than target char");
 396
 397   /* This is assumed in eval_token() and could be fixed if necessary.  */
 398   if (sizeof (cppchar_t) > sizeof (cpp_num_part))
 399     cpp_error (pfile, DL_ICE, "CPP half-integer narrower than CPP character");
 400
 401   if (CPP_OPTION (pfile, wchar_precision) > BITS_PER_CPPCHAR_T)
 402     cpp_error (pfile, DL_ICE,
 403                "CPP on this host cannot handle wide character constants over"
 404                " %lu bits, but the target requires %lu bits",
 405                (unsigned long) BITS_PER_CPPCHAR_T,
 406                (unsigned long) CPP_OPTION (pfile, wchar_precision));
 407 }
 408 #else
 409 # define sanity_checks(PFILE)
 410 #endif
 411
 412 /* Add a dependency target.  Can be called any number of times before
 413    cpp_read_main_file().  If no targets have been added before
 414    cpp_read_main_file(), then the default target is used.  */
 415 void
 416 cpp_add_dependency_target (cpp_reader *pfile, const char *target, int quote)
 417 {
 418   if (!pfile->deps)
 419     pfile->deps = deps_init ();
 420
 421   deps_add_target (pfile->deps, target, quote);
 422 }
 423
 424 /* This is called after options have been parsed, and partially
 425    processed.  Setup for processing input from the file named FNAME,
 426    or stdin if it is the empty string.  Return the original filename
 427    on success (e.g. foo.i->foo.c), or NULL on failure.  */
 428 const char *
 429 cpp_read_main_file (cpp_reader *pfile, const char *fname)
 430 {
 431   sanity_checks (pfile);
 432
 433   post_options (pfile);
 434
 435   /* Mark named operators before handling command line macros.  */
 436   if (CPP_OPTION (pfile, cplusplus) && CPP_OPTION (pfile, operator_names))
 437     mark_named_operators (pfile);
 438
 439   if (CPP_OPTION (pfile, deps.style) != DEPS_NONE)
 440     {
 441       if (!pfile->deps)
 442         pfile->deps = deps_init ();
 443
 444       /* Set the default target (if there is none already).  */
 445       deps_add_default_target (pfile->deps, fname);
 446     }
 447
 448   /* Open the main input file.  */
 449   pfile->line = 1;
 450   if (!_cpp_read_file (pfile, fname))
 451     return NULL;
 452
 453   /* Set this here so the client can change the option if it wishes,
 454      and after stacking the main file so we don't trace the main
 455      file.  */
 456   pfile->line_maps.trace_includes = CPP_OPTION (pfile, print_include_names);
 457
 458   /* For foo.i, read the original filename foo.c now, for the benefit
 459      of the front ends.  */
 460   if (CPP_OPTION (pfile, preprocessed))
 461     read_original_filename (pfile);
 462
 463   return pfile->map->to_file;
 464 }
 465
 466 /* For preprocessed files, if the first tokens are of the form # NUM.
 467    handle the directive so we know the original file name.  This will
 468    generate file_change callbacks, which the front ends must handle
 469    appropriately given their state of initialization.  */
 470 static void
 471 read_original_filename (cpp_reader *pfile)
 472 {
 473   const cpp_token *token, *token1;
 474
 475   /* Lex ahead; if the first tokens are of the form # NUM, then
 476      process the directive, otherwise back up.  */
 477   token = _cpp_lex_direct (pfile);
 478   if (token->type == CPP_HASH)
 479     {
 480       token1 = _cpp_lex_direct (pfile);
 481       _cpp_backup_tokens (pfile, 1);
 482
 483       /* If it's a #line directive, handle it.  */
 484       if (token1->type == CPP_NUMBER)
 485         {
 486           _cpp_handle_directive (pfile, token->flags & PREV_WHITE);
 487           return;
 488         }
 489     }
 490
 491   /* Backup as if nothing happened.  */
 492   _cpp_backup_tokens (pfile, 1);
 493 }
 494
 495 /* This is called at the end of preprocessing.  It pops the last
 496    buffer and writes dependency output, and returns the number of
 497    errors.
 498
 499    Maybe it should also reset state, such that you could call
 500    cpp_start_read with a new filename to restart processing.  */
 501 int
 502 cpp_finish (cpp_reader *pfile, FILE *deps_stream)
 503 {
 504   /* Warn about unused macros before popping the final buffer.  */
 505   if (CPP_OPTION (pfile, warn_unused_macros))
 506     cpp_forall_identifiers (pfile, _cpp_warn_if_unused_macro, NULL);
 507
 508   /* cpplex.c leaves the final buffer on the stack.  This it so that
 509      it returns an unending stream of CPP_EOFs to the client.  If we
 510      popped the buffer, we'd dereference a NULL buffer pointer and
 511      segfault.  It's nice to allow the client to do worry-free excess
 512      cpp_get_token calls.  */
 513   while (pfile->buffer)
 514     _cpp_pop_buffer (pfile);
 515
 516   /* Don't write the deps file if there are errors.  */
 517   if (CPP_OPTION (pfile, deps.style) != DEPS_NONE
 518       && deps_stream && pfile->errors == 0)
 519     {
 520       deps_write (pfile->deps, deps_stream, 72);
 521
 522       if (CPP_OPTION (pfile, deps.phony_targets))
 523         deps_phony_targets (pfile->deps, deps_stream);
 524     }
 525
 526   /* Report on headers that could use multiple include guards.  */
 527   if (CPP_OPTION (pfile, print_include_names))
 528     _cpp_report_missing_guards (pfile);
 529
 530   return pfile->errors;
 531 }
 532
 533 static void
 534 post_options (cpp_reader *pfile)
 535 {
 536   /* -Wtraditional is not useful in C++ mode.  */
 537   if (CPP_OPTION (pfile, cplusplus))
 538     CPP_OPTION (pfile, warn_traditional) = 0;
 539
 540   /* Permanently disable macro expansion if we are rescanning
 541      preprocessed text.  Read preprocesed source in ISO mode.  */
 542   if (CPP_OPTION (pfile, preprocessed))
 543     {
 544       pfile->state.prevent_expansion = 1;
 545       CPP_OPTION (pfile, traditional) = 0;
 546     }
 547
 548   if (CPP_OPTION (pfile, warn_trigraphs) == 2)
 549     CPP_OPTION (pfile, warn_trigraphs) = !CPP_OPTION (pfile, trigraphs);
 550
 551   if (CPP_OPTION (pfile, traditional))
 552     {
 553       /* Traditional CPP does not accurately track column information.  */
 554       CPP_OPTION (pfile, show_column) = 0;
 555       CPP_OPTION (pfile, trigraphs) = 0;
 556       CPP_OPTION (pfile, warn_trigraphs) = 0;
 557     }
 558 }