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                PARAMS ((void));
  31 static void mark_named_operators        PARAMS ((cpp_reader *));
  32 static void read_original_filename      PARAMS ((cpp_reader *));
  33 static void post_options                PARAMS ((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 PARAMS ((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 (pfile, lang)
  95      cpp_reader *pfile;
  96      enum c_lang lang;
  97 {
  98   const struct lang_flags *l = &lang_defaults[(int) lang];
  99
 100   CPP_OPTION (pfile, lang) = lang;
 101
 102   CPP_OPTION (pfile, c99)                = l->c99;
 103   CPP_OPTION (pfile, cplusplus)          = l->cplusplus;
 104   CPP_OPTION (pfile, extended_numbers)   = l->extended_numbers;
 105   CPP_OPTION (pfile, std)                = l->std;
 106   CPP_OPTION (pfile, trigraphs)          = l->std;
 107   CPP_OPTION (pfile, cplusplus_comments) = l->cplusplus_comments;
 108   CPP_OPTION (pfile, digraphs)           = l->digraphs;
 109 }
 110
 111 /* Initialize library global state.  */
 112 static void
 113 init_library ()
 114 {
 115   static int initialized = 0;
 116
 117   if (! initialized)
 118     {
 119       initialized = 1;
 120
 121       /* Set up the trigraph map.  This doesn't need to do anything if
 122          we were compiled with a compiler that supports C99 designated
 123          initializers.  */
 124       init_trigraph_map ();
 125     }
 126 }
 127
 128 /* Initialize a cpp_reader structure.  */
 129 cpp_reader *
 130 cpp_create_reader (lang, table)
 131      enum c_lang lang;
 132      hash_table *table;
 133 {
 134   cpp_reader *pfile;
 135
 136   /* Initialize this instance of the library if it hasn't been already.  */
 137   init_library ();
 138
 139   pfile = (cpp_reader *) xcalloc (1, sizeof (cpp_reader));
 140
 141   cpp_set_lang (pfile, lang);
 142   CPP_OPTION (pfile, warn_import) = 1;
 143   CPP_OPTION (pfile, warn_multichar) = 1;
 144   CPP_OPTION (pfile, discard_comments) = 1;
 145   CPP_OPTION (pfile, discard_comments_in_macro_exp) = 1;
 146   CPP_OPTION (pfile, show_column) = 1;
 147   CPP_OPTION (pfile, tabstop) = 8;
 148   CPP_OPTION (pfile, operator_names) = 1;
 149   CPP_OPTION (pfile, warn_trigraphs) = 2;
 150   CPP_OPTION (pfile, warn_endif_labels) = 1;
 151   CPP_OPTION (pfile, warn_deprecated) = 1;
 152   CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
 153   CPP_OPTION (pfile, dollars_in_ident) = 1;
 154
 155   /* Default CPP arithmetic to something sensible for the host for the
 156      benefit of dumb users like fix-header.  */
 157   CPP_OPTION (pfile, precision) = CHAR_BIT * sizeof (long);
 158   CPP_OPTION (pfile, char_precision) = CHAR_BIT;
 159   CPP_OPTION (pfile, wchar_precision) = CHAR_BIT * sizeof (int);
 160   CPP_OPTION (pfile, int_precision) = CHAR_BIT * sizeof (int);
 161   CPP_OPTION (pfile, unsigned_char) = 0;
 162   CPP_OPTION (pfile, unsigned_wchar) = 1;
 163
 164   /* Initialize the line map.  Start at logical line 1, so we can use
 165      a line number of zero for special states.  */
 166   init_line_maps (&pfile->line_maps);
 167
 168   /* Initialize lexer state.  */
 169   pfile->state.save_comments = ! CPP_OPTION (pfile, discard_comments);
 170
 171   /* Set up static tokens.  */
 172   pfile->avoid_paste.type = CPP_PADDING;
 173   pfile->avoid_paste.val.source = NULL;
 174   pfile->eof.type = CPP_EOF;
 175   pfile->eof.flags = 0;
 176
 177   /* Create a token buffer for the lexer.  */
 178   _cpp_init_tokenrun (&pfile->base_run, 250);
 179   pfile->cur_run = &pfile->base_run;
 180   pfile->cur_token = pfile->base_run.base;
 181
 182   /* Initialize the base context.  */
 183   pfile->context = &pfile->base_context;
 184   pfile->base_context.macro = 0;
 185   pfile->base_context.prev = pfile->base_context.next = 0;
 186
 187   /* Aligned and unaligned storage.  */
 188   pfile->a_buff = _cpp_get_buff (pfile, 0);
 189   pfile->u_buff = _cpp_get_buff (pfile, 0);
 190
 191   /* The expression parser stack.  */
 192   _cpp_expand_op_stack (pfile);
 193
 194   /* Initialize the buffer obstack.  */
 195   gcc_obstack_init (&pfile->buffer_ob);
 196
 197   _cpp_init_includes (pfile);
 198
 199   _cpp_init_hashtable (pfile, table);
 200
 201   return pfile;
 202 }
 203
 204 /* Free resources used by PFILE.  Accessing PFILE after this function
 205    returns leads to undefined behavior.  Returns the error count.  */
 206 void
 207 cpp_destroy (pfile)
 208      cpp_reader *pfile;
 209 {
 210   cpp_context *context, *contextn;
 211   tokenrun *run, *runn;
 212
 213   free (pfile->op_stack);
 214
 215   while (CPP_BUFFER (pfile) != NULL)
 216     _cpp_pop_buffer (pfile);
 217
 218   if (pfile->out.base)
 219     free (pfile->out.base);
 220
 221   if (pfile->macro_buffer)
 222     {
 223       free ((PTR) pfile->macro_buffer);
 224       pfile->macro_buffer = NULL;
 225       pfile->macro_buffer_len = 0;
 226     }
 227
 228   if (pfile->deps)
 229     deps_free (pfile->deps);
 230   obstack_free (&pfile->buffer_ob, 0);
 231
 232   _cpp_destroy_hashtable (pfile);
 233   _cpp_cleanup_includes (pfile);
 234
 235   _cpp_free_buff (pfile->a_buff);
 236   _cpp_free_buff (pfile->u_buff);
 237   _cpp_free_buff (pfile->free_buffs);
 238
 239   for (run = &pfile->base_run; run; run = runn)
 240     {
 241       runn = run->next;
 242       free (run->base);
 243       if (run != &pfile->base_run)
 244         free (run);
 245     }
 246
 247   for (context = pfile->base_context.next; context; context = contextn)
 248     {
 249       contextn = context->next;
 250       free (context);
 251     }
 252
 253   free_line_maps (&pfile->line_maps);
 254   free (pfile);
 255 }
 256
 257 /* This structure defines one built-in identifier.  A node will be
 258    entered in the hash table under the name NAME, with value VALUE.
 259
 260    There are two tables of these.  builtin_array holds all the
 261    "builtin" macros: these are handled by builtin_macro() in
 262    cppmacro.c.  Builtin is somewhat of a misnomer -- the property of
 263    interest is that these macros require special code to compute their
 264    expansions.  The value is a "builtin_type" enumerator.
 265
 266    operator_array holds the C++ named operators.  These are keywords
 267    which act as aliases for punctuators.  In C++, they cannot be
 268    altered through #define, and #if recognizes them as operators.  In
 269    C, these are not entered into the hash table at all (but see
 270    <iso646.h>).  The value is a token-type enumerator.  */
 271 struct builtin
 272 {
 273   const uchar *name;
 274   unsigned short len;
 275   unsigned short value;
 276 };
 277
 278 #define B(n, t)    { DSC(n), t }
 279 static const struct builtin builtin_array[] =
 280 {
 281   B("__TIME__",          BT_TIME),
 282   B("__DATE__",          BT_DATE),
 283   B("__FILE__",          BT_FILE),
 284   B("__BASE_FILE__",     BT_BASE_FILE),
 285   B("__LINE__",          BT_SPECLINE),
 286   B("__INCLUDE_LEVEL__", BT_INCLUDE_LEVEL),
 287   /* Keep builtins not used for -traditional-cpp at the end, and
 288      update init_builtins() if any more are added.  */
 289   B("_Pragma",           BT_PRAGMA),
 290   B("__STDC__",          BT_STDC),
 291 };
 292
 293 static const struct builtin operator_array[] =
 294 {
 295   B("and",      CPP_AND_AND),
 296   B("and_eq",   CPP_AND_EQ),
 297   B("bitand",   CPP_AND),
 298   B("bitor",    CPP_OR),
 299   B("compl",    CPP_COMPL),
 300   B("not",      CPP_NOT),
 301   B("not_eq",   CPP_NOT_EQ),
 302   B("or",       CPP_OR_OR),
 303   B("or_eq",    CPP_OR_EQ),
 304   B("xor",      CPP_XOR),
 305   B("xor_eq",   CPP_XOR_EQ)
 306 };
 307 #undef B
 308
 309 /* Mark the C++ named operators in the hash table.  */
 310 static void
 311 mark_named_operators (pfile)
 312      cpp_reader *pfile;
 313 {
 314   const struct builtin *b;
 315
 316   for (b = operator_array;
 317        b < (operator_array + ARRAY_SIZE (operator_array));
 318        b++)
 319     {
 320       cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
 321       hp->flags |= NODE_OPERATOR;
 322       hp->is_directive = 0;
 323       hp->directive_index = b->value;
 324     }
 325 }
 326
 327 /* Read the builtins table above and enter them, and language-specific
 328    macros, into the hash table.  HOSTED is true if this is a hosted
 329    environment.  */
 330 void
 331 cpp_init_builtins (pfile, hosted)
 332      cpp_reader *pfile;
 333      int hosted;
 334 {
 335   const struct builtin *b;
 336   size_t n = ARRAY_SIZE (builtin_array);
 337
 338   if (CPP_OPTION (pfile, traditional))
 339     n -= 2;
 340
 341   for(b = builtin_array; b < builtin_array + n; b++)
 342     {
 343       cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
 344       hp->type = NT_MACRO;
 345       hp->flags |= NODE_BUILTIN | NODE_WARN;
 346       hp->value.builtin = b->value;
 347     }
 348
 349   if (CPP_OPTION (pfile, cplusplus))
 350     _cpp_define_builtin (pfile, "__cplusplus 1");
 351   else if (CPP_OPTION (pfile, lang) == CLK_ASM)
 352     _cpp_define_builtin (pfile, "__ASSEMBLER__ 1");
 353   else if (CPP_OPTION (pfile, lang) == CLK_STDC94)
 354     _cpp_define_builtin (pfile, "__STDC_VERSION__ 199409L");
 355   else if (CPP_OPTION (pfile, c99))
 356     _cpp_define_builtin (pfile, "__STDC_VERSION__ 199901L");
 357
 358   if (hosted)
 359     _cpp_define_builtin (pfile, "__STDC_HOSTED__ 1");
 360   else
 361     _cpp_define_builtin (pfile, "__STDC_HOSTED__ 0");
 362
 363   if (CPP_OPTION (pfile, objc))
 364     _cpp_define_builtin (pfile, "__OBJC__ 1");
 365 }
 366
 367 /* Sanity-checks are dependent on command-line options, so it is
 368    called as a subroutine of cpp_read_main_file ().  */
 369 #if ENABLE_CHECKING
 370 static void sanity_checks PARAMS ((cpp_reader *));
 371 static void sanity_checks (pfile)
 372      cpp_reader *pfile;
 373 {
 374   cppchar_t test = 0;
 375   size_t max_precision = 2 * CHAR_BIT * sizeof (cpp_num_part);
 376
 377   /* Sanity checks for assumptions about CPP arithmetic and target
 378      type precisions made by cpplib.  */
 379   test--;
 380   if (test < 1)
 381     cpp_error (pfile, DL_ICE, "cppchar_t must be an unsigned type");
 382
 383   if (CPP_OPTION (pfile, precision) > max_precision)
 384     cpp_error (pfile, DL_ICE,
 385                "preprocessor arithmetic has maximum precision of %lu bits; target requires %lu bits",
 386                (unsigned long) max_precision,
 387                (unsigned long) CPP_OPTION (pfile, precision));
 388
 389   if (CPP_OPTION (pfile, precision) < CPP_OPTION (pfile, int_precision))
 390     cpp_error (pfile, DL_ICE,
 391                "CPP arithmetic must be at least as precise as a target int");
 392
 393   if (CPP_OPTION (pfile, char_precision) < 8)
 394     cpp_error (pfile, DL_ICE, "target char is less than 8 bits wide");
 395
 396   if (CPP_OPTION (pfile, wchar_precision) < CPP_OPTION (pfile, char_precision))
 397     cpp_error (pfile, DL_ICE,
 398                "target wchar_t is narrower than target char");
 399
 400   if (CPP_OPTION (pfile, int_precision) < CPP_OPTION (pfile, char_precision))
 401     cpp_error (pfile, DL_ICE,
 402                "target int is narrower than target char");
 403
 404   /* This is assumed in eval_token() and could be fixed if necessary.  */
 405   if (sizeof (cppchar_t) > sizeof (cpp_num_part))
 406     cpp_error (pfile, DL_ICE, "CPP half-integer narrower than CPP character");
 407
 408   if (CPP_OPTION (pfile, wchar_precision) > BITS_PER_CPPCHAR_T)
 409     cpp_error (pfile, DL_ICE,
 410                "CPP on this host cannot handle wide character constants over %lu bits, but the target requires %lu bits",
 411                (unsigned long) BITS_PER_CPPCHAR_T,
 412                (unsigned long) CPP_OPTION (pfile, wchar_precision));
 413 }
 414 #else
 415 # define sanity_checks(PFILE)
 416 #endif
 417
 418 /* Add a dependency target.  Can be called any number of times before
 419    cpp_read_main_file().  If no targets have been added before
 420    cpp_read_main_file(), then the default target is used.  */
 421 void
 422 cpp_add_dependency_target (pfile, target, quote)
 423      cpp_reader *pfile;
 424      const char *target;
 425      int quote;
 426 {
 427   if (!pfile->deps)
 428     pfile->deps = deps_init ();
 429
 430   deps_add_target (pfile->deps, target, quote);
 431 }
 432
 433 /* This is called after options have been parsed, and partially
 434    processed.  Setup for processing input from the file named FNAME,
 435    or stdin if it is the empty string.  Return the original filename
 436    on success (e.g. foo.i->foo.c), or NULL on failure.  */
 437 const char *
 438 cpp_read_main_file (pfile, fname)
 439      cpp_reader *pfile;
 440      const char *fname;
 441 {
 442   sanity_checks (pfile);
 443
 444   post_options (pfile);
 445
 446   /* Mark named operators before handling command line macros.  */
 447   if (CPP_OPTION (pfile, cplusplus) && CPP_OPTION (pfile, operator_names))
 448     mark_named_operators (pfile);
 449
 450   if (CPP_OPTION (pfile, deps.style) != DEPS_NONE)
 451     {
 452       if (!pfile->deps)
 453         pfile->deps = deps_init ();
 454
 455       /* Set the default target (if there is none already).  */
 456       deps_add_default_target (pfile->deps, fname);
 457     }
 458
 459   /* Open the main input file.  */
 460   pfile->line = 1;
 461   if (!_cpp_read_file (pfile, fname))
 462     return NULL;
 463
 464   /* Set this here so the client can change the option if it wishes,
 465      and after stacking the main file so we don't trace the main
 466      file.  */
 467   pfile->line_maps.trace_includes = CPP_OPTION (pfile, print_include_names);
 468
 469   /* For foo.i, read the original filename foo.c now, for the benefit
 470      of the front ends.  */
 471   if (CPP_OPTION (pfile, preprocessed))
 472     read_original_filename (pfile);
 473
 474   return pfile->map->to_file;
 475 }
 476
 477 /* For preprocessed files, if the first tokens are of the form # NUM.
 478    handle the directive so we know the original file name.  This will
 479    generate file_change callbacks, which the front ends must handle
 480    appropriately given their state of initialization.  */
 481 static void
 482 read_original_filename (pfile)
 483      cpp_reader *pfile;
 484 {
 485   const cpp_token *token, *token1;
 486
 487   /* Lex ahead; if the first tokens are of the form # NUM, then
 488      process the directive, otherwise back up.  */
 489   token = _cpp_lex_direct (pfile);
 490   if (token->type == CPP_HASH)
 491     {
 492       token1 = _cpp_lex_direct (pfile);
 493       _cpp_backup_tokens (pfile, 1);
 494
 495       /* If it's a #line directive, handle it.  */
 496       if (token1->type == CPP_NUMBER)
 497         {
 498           _cpp_handle_directive (pfile, token->flags & PREV_WHITE);
 499           return;
 500         }
 501     }
 502
 503   /* Backup as if nothing happened.  */
 504   _cpp_backup_tokens (pfile, 1);
 505 }
 506
 507 /* This is called at the end of preprocessing.  It pops the last
 508    buffer and writes dependency output, and returns the number of
 509    errors.
 510
 511    Maybe it should also reset state, such that you could call
 512    cpp_start_read with a new filename to restart processing.  */
 513 int
 514 cpp_finish (pfile, deps_stream)
 515      cpp_reader *pfile;
 516      FILE *deps_stream;
 517 {
 518   /* Warn about unused macros before popping the final buffer.  */
 519   if (CPP_OPTION (pfile, warn_unused_macros))
 520     cpp_forall_identifiers (pfile, _cpp_warn_if_unused_macro, NULL);
 521
 522   /* cpplex.c leaves the final buffer on the stack.  This it so that
 523      it returns an unending stream of CPP_EOFs to the client.  If we
 524      popped the buffer, we'd dereference a NULL buffer pointer and
 525      segfault.  It's nice to allow the client to do worry-free excess
 526      cpp_get_token calls.  */
 527   while (pfile->buffer)
 528     _cpp_pop_buffer (pfile);
 529
 530   /* Don't write the deps file if there are errors.  */
 531   if (CPP_OPTION (pfile, deps.style) != DEPS_NONE
 532       && deps_stream && pfile->errors == 0)
 533     {
 534       deps_write (pfile->deps, deps_stream, 72);
 535
 536       if (CPP_OPTION (pfile, deps.phony_targets))
 537         deps_phony_targets (pfile->deps, deps_stream);
 538     }
 539
 540   /* Report on headers that could use multiple include guards.  */
 541   if (CPP_OPTION (pfile, print_include_names))
 542     _cpp_report_missing_guards (pfile);
 543
 544   return pfile->errors;
 545 }
 546
 547 static void
 548 post_options (pfile)
 549      cpp_reader *pfile;
 550 {
 551   /* -Wtraditional is not useful in C++ mode.  */
 552   if (CPP_OPTION (pfile, cplusplus))
 553     CPP_OPTION (pfile, warn_traditional) = 0;
 554
 555   /* Permanently disable macro expansion if we are rescanning
 556      preprocessed text.  Read preprocesed source in ISO mode.  */
 557   if (CPP_OPTION (pfile, preprocessed))
 558     {
 559       pfile->state.prevent_expansion = 1;
 560       CPP_OPTION (pfile, traditional) = 0;
 561     }
 562
 563   if (CPP_OPTION (pfile, warn_trigraphs) == 2)
 564     CPP_OPTION (pfile, warn_trigraphs) = !CPP_OPTION (pfile, trigraphs);
 565
 566   if (CPP_OPTION (pfile, traditional))
 567     {
 568       /* Traditional CPP does not accurately track column information.  */
 569       CPP_OPTION (pfile, show_column) = 0;
 570       CPP_OPTION (pfile, trigraphs) = 0;
 571       CPP_OPTION (pfile, warn_trigraphs) = 0;
 572     }
 573
 574   /* C99 permits implementation-defined characters in identifiers.
 575      The documented meaning of -std= is to turn off extensions that
 576      conflict with the specified standard, and since a strictly
 577      conforming program cannot contain a '$', we do not condition
 578      their acceptance on the -std= setting.  */
 579   pfile->warn_dollars = CPP_PEDANTIC (pfile) && !CPP_OPTION (pfile, c99);
 580 }