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 dollars_in_ident;
  77   char cplusplus_comments;
  78   char digraphs;
  79 };
  80
  81 /* ??? Enable $ in identifiers in assembly? */
  82 static const struct lang_flags lang_defaults[] =
  83 { /*              c99 c++ xnum std dollar c++comm digr  */
  84   /* GNUC89 */  { 0,  0,  1,   0,   1,     1,      1     },
  85   /* GNUC99 */  { 1,  0,  1,   0,   1,     1,      1     },
  86   /* STDC89 */  { 0,  0,  0,   1,   0,     0,      0     },
  87   /* STDC94 */  { 0,  0,  0,   1,   0,     0,      1     },
  88   /* STDC99 */  { 1,  0,  1,   1,   0,     1,      1     },
  89   /* GNUCXX */  { 0,  1,  1,   0,   1,     1,      1     },
  90   /* CXX98  */  { 0,  1,  1,   1,   0,     1,      1     },
  91   /* ASM    */  { 0,  0,  1,   0,   0,     1,      0     }
  92 };
  93
  94 /* Sets internal flags correctly for a given language.  */
  95 void
  96 cpp_set_lang (pfile, lang)
  97      cpp_reader *pfile;
  98      enum c_lang lang;
  99 {
 100   const struct lang_flags *l = &lang_defaults[(int) lang];
 101
 102   CPP_OPTION (pfile, lang) = lang;
 103
 104   CPP_OPTION (pfile, c99)                = l->c99;
 105   CPP_OPTION (pfile, cplusplus)          = l->cplusplus;
 106   CPP_OPTION (pfile, extended_numbers)   = l->extended_numbers;
 107   CPP_OPTION (pfile, std)                = l->std;
 108   CPP_OPTION (pfile, trigraphs)          = l->std;
 109   CPP_OPTION (pfile, dollars_in_ident)   = l->dollars_in_ident;
 110   CPP_OPTION (pfile, cplusplus_comments) = l->cplusplus_comments;
 111   CPP_OPTION (pfile, digraphs)           = l->digraphs;
 112 }
 113
 114 /* Initialize library global state.  */
 115 static void
 116 init_library ()
 117 {
 118   static int initialized = 0;
 119
 120   if (! initialized)
 121     {
 122       initialized = 1;
 123
 124       /* Set up the trigraph map.  This doesn't need to do anything if
 125          we were compiled with a compiler that supports C99 designated
 126          initializers.  */
 127       init_trigraph_map ();
 128     }
 129 }
 130
 131 /* Initialize a cpp_reader structure.  */
 132 cpp_reader *
 133 cpp_create_reader (lang, table)
 134      enum c_lang lang;
 135      hash_table *table;
 136 {
 137   cpp_reader *pfile;
 138
 139   /* Initialize this instance of the library if it hasn't been already.  */
 140   init_library ();
 141
 142   pfile = (cpp_reader *) xcalloc (1, sizeof (cpp_reader));
 143
 144   cpp_set_lang (pfile, lang);
 145   CPP_OPTION (pfile, warn_import) = 1;
 146   CPP_OPTION (pfile, warn_multichar) = 1;
 147   CPP_OPTION (pfile, discard_comments) = 1;
 148   CPP_OPTION (pfile, discard_comments_in_macro_exp) = 1;
 149   CPP_OPTION (pfile, show_column) = 1;
 150   CPP_OPTION (pfile, tabstop) = 8;
 151   CPP_OPTION (pfile, operator_names) = 1;
 152   CPP_OPTION (pfile, warn_trigraphs) = 2;
 153   CPP_OPTION (pfile, warn_endif_labels) = 1;
 154   CPP_OPTION (pfile, warn_deprecated) = 1;
 155   CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
 156
 157   /* Default CPP arithmetic to something sensible for the host for the
 158      benefit of dumb users like fix-header.  */
 159   CPP_OPTION (pfile, precision) = CHAR_BIT * sizeof (long);
 160   CPP_OPTION (pfile, char_precision) = CHAR_BIT;
 161   CPP_OPTION (pfile, wchar_precision) = CHAR_BIT * sizeof (int);
 162   CPP_OPTION (pfile, int_precision) = CHAR_BIT * sizeof (int);
 163   CPP_OPTION (pfile, unsigned_char) = 0;
 164   CPP_OPTION (pfile, unsigned_wchar) = 1;
 165
 166   /* Initialize the line map.  Start at logical line 1, so we can use
 167      a line number of zero for special states.  */
 168   init_line_maps (&pfile->line_maps);
 169
 170   /* Initialize lexer state.  */
 171   pfile->state.save_comments = ! CPP_OPTION (pfile, discard_comments);
 172
 173   /* Set up static tokens.  */
 174   pfile->avoid_paste.type = CPP_PADDING;
 175   pfile->avoid_paste.val.source = NULL;
 176   pfile->eof.type = CPP_EOF;
 177   pfile->eof.flags = 0;
 178
 179   /* Create a token buffer for the lexer.  */
 180   _cpp_init_tokenrun (&pfile->base_run, 250);
 181   pfile->cur_run = &pfile->base_run;
 182   pfile->cur_token = pfile->base_run.base;
 183
 184   /* Initialize the base context.  */
 185   pfile->context = &pfile->base_context;
 186   pfile->base_context.macro = 0;
 187   pfile->base_context.prev = pfile->base_context.next = 0;
 188
 189   /* Aligned and unaligned storage.  */
 190   pfile->a_buff = _cpp_get_buff (pfile, 0);
 191   pfile->u_buff = _cpp_get_buff (pfile, 0);
 192
 193   /* The expression parser stack.  */
 194   _cpp_expand_op_stack (pfile);
 195
 196   /* Initialize the buffer obstack.  */
 197   gcc_obstack_init (&pfile->buffer_ob);
 198
 199   _cpp_init_includes (pfile);
 200
 201   _cpp_init_hashtable (pfile, table);
 202
 203   return pfile;
 204 }
 205
 206 /* Free resources used by PFILE.  Accessing PFILE after this function
 207    returns leads to undefined behavior.  Returns the error count.  */
 208 void
 209 cpp_destroy (pfile)
 210      cpp_reader *pfile;
 211 {
 212   cpp_context *context, *contextn;
 213   tokenrun *run, *runn;
 214
 215   free (pfile->op_stack);
 216
 217   while (CPP_BUFFER (pfile) != NULL)
 218     _cpp_pop_buffer (pfile);
 219
 220   if (pfile->out.base)
 221     free (pfile->out.base);
 222
 223   if (pfile->macro_buffer)
 224     {
 225       free ((PTR) pfile->macro_buffer);
 226       pfile->macro_buffer = NULL;
 227       pfile->macro_buffer_len = 0;
 228     }
 229
 230   if (pfile->deps)
 231     deps_free (pfile->deps);
 232   obstack_free (&pfile->buffer_ob, 0);
 233
 234   _cpp_destroy_hashtable (pfile);
 235   _cpp_cleanup_includes (pfile);
 236
 237   _cpp_free_buff (pfile->a_buff);
 238   _cpp_free_buff (pfile->u_buff);
 239   _cpp_free_buff (pfile->free_buffs);
 240
 241   for (run = &pfile->base_run; run; run = runn)
 242     {
 243       runn = run->next;
 244       free (run->base);
 245       if (run != &pfile->base_run)
 246         free (run);
 247     }
 248
 249   for (context = pfile->base_context.next; context; context = contextn)
 250     {
 251       contextn = context->next;
 252       free (context);
 253     }
 254
 255   free_line_maps (&pfile->line_maps);
 256   free (pfile);
 257 }
 258
 259 /* This structure defines one built-in identifier.  A node will be
 260    entered in the hash table under the name NAME, with value VALUE.
 261
 262    There are two tables of these.  builtin_array holds all the
 263    "builtin" macros: these are handled by builtin_macro() in
 264    cppmacro.c.  Builtin is somewhat of a misnomer -- the property of
 265    interest is that these macros require special code to compute their
 266    expansions.  The value is a "builtin_type" enumerator.
 267
 268    operator_array holds the C++ named operators.  These are keywords
 269    which act as aliases for punctuators.  In C++, they cannot be
 270    altered through #define, and #if recognizes them as operators.  In
 271    C, these are not entered into the hash table at all (but see
 272    <iso646.h>).  The value is a token-type enumerator.  */
 273 struct builtin
 274 {
 275   const uchar *name;
 276   unsigned short len;
 277   unsigned short value;
 278 };
 279
 280 #define B(n, t)    { DSC(n), t }
 281 static const struct builtin builtin_array[] =
 282 {
 283   B("__TIME__",          BT_TIME),
 284   B("__DATE__",          BT_DATE),
 285   B("__FILE__",          BT_FILE),
 286   B("__BASE_FILE__",     BT_BASE_FILE),
 287   B("__LINE__",          BT_SPECLINE),
 288   B("__INCLUDE_LEVEL__", BT_INCLUDE_LEVEL),
 289   /* Keep builtins not used for -traditional-cpp at the end, and
 290      update init_builtins() if any more are added.  */
 291   B("_Pragma",           BT_PRAGMA),
 292   B("__STDC__",          BT_STDC),
 293 };
 294
 295 static const struct builtin operator_array[] =
 296 {
 297   B("and",      CPP_AND_AND),
 298   B("and_eq",   CPP_AND_EQ),
 299   B("bitand",   CPP_AND),
 300   B("bitor",    CPP_OR),
 301   B("compl",    CPP_COMPL),
 302   B("not",      CPP_NOT),
 303   B("not_eq",   CPP_NOT_EQ),
 304   B("or",       CPP_OR_OR),
 305   B("or_eq",    CPP_OR_EQ),
 306   B("xor",      CPP_XOR),
 307   B("xor_eq",   CPP_XOR_EQ)
 308 };
 309 #undef B
 310
 311 /* Mark the C++ named operators in the hash table.  */
 312 static void
 313 mark_named_operators (pfile)
 314      cpp_reader *pfile;
 315 {
 316   const struct builtin *b;
 317
 318   for (b = operator_array;
 319        b < (operator_array + ARRAY_SIZE (operator_array));
 320        b++)
 321     {
 322       cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
 323       hp->flags |= NODE_OPERATOR;
 324       hp->is_directive = 0;
 325       hp->directive_index = b->value;
 326     }
 327 }
 328
 329 /* Read the builtins table above and enter them, and language-specific
 330    macros, into the hash table.  HOSTED is true if this is a hosted
 331    environment.  */
 332 void
 333 cpp_init_builtins (pfile, hosted)
 334      cpp_reader *pfile;
 335      int hosted;
 336 {
 337   const struct builtin *b;
 338   size_t n = ARRAY_SIZE (builtin_array);
 339
 340   if (CPP_OPTION (pfile, traditional))
 341     n -= 2;
 342
 343   for(b = builtin_array; b < builtin_array + n; b++)
 344     {
 345       cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
 346       hp->type = NT_MACRO;
 347       hp->flags |= NODE_BUILTIN | NODE_WARN;
 348       hp->value.builtin = b->value;
 349     }
 350
 351   if (CPP_OPTION (pfile, cplusplus))
 352     _cpp_define_builtin (pfile, "__cplusplus 1");
 353   else if (CPP_OPTION (pfile, lang) == CLK_ASM)
 354     _cpp_define_builtin (pfile, "__ASSEMBLER__ 1");
 355   else if (CPP_OPTION (pfile, lang) == CLK_STDC94)
 356     _cpp_define_builtin (pfile, "__STDC_VERSION__ 199409L");
 357   else if (CPP_OPTION (pfile, c99))
 358     _cpp_define_builtin (pfile, "__STDC_VERSION__ 199901L");
 359
 360   if (hosted)
 361     _cpp_define_builtin (pfile, "__STDC_HOSTED__ 1");
 362   else
 363     _cpp_define_builtin (pfile, "__STDC_HOSTED__ 0");
 364
 365   if (CPP_OPTION (pfile, objc))
 366     _cpp_define_builtin (pfile, "__OBJC__ 1");
 367 }
 368
 369 /* Sanity-checks are dependent on command-line options, so it is
 370    called as a subroutine of cpp_read_main_file ().  */
 371 #if ENABLE_CHECKING
 372 static void sanity_checks PARAMS ((cpp_reader *));
 373 static void sanity_checks (pfile)
 374      cpp_reader *pfile;
 375 {
 376   cppchar_t test = 0;
 377   size_t max_precision = 2 * CHAR_BIT * sizeof (cpp_num_part);
 378
 379   /* Sanity checks for assumptions about CPP arithmetic and target
 380      type precisions made by cpplib.  */
 381   test--;
 382   if (test < 1)
 383     cpp_error (pfile, DL_ICE, "cppchar_t must be an unsigned type");
 384
 385   if (CPP_OPTION (pfile, precision) > max_precision)
 386     cpp_error (pfile, DL_ICE,
 387                "preprocessor arithmetic has maximum precision of %lu bits; target requires %lu bits",
 388                (unsigned long) max_precision,
 389                (unsigned long) CPP_OPTION (pfile, precision));
 390
 391   if (CPP_OPTION (pfile, precision) < CPP_OPTION (pfile, int_precision))
 392     cpp_error (pfile, DL_ICE,
 393                "CPP arithmetic must be at least as precise as a target int");
 394
 395   if (CPP_OPTION (pfile, char_precision) < 8)
 396     cpp_error (pfile, DL_ICE, "target char is less than 8 bits wide");
 397
 398   if (CPP_OPTION (pfile, wchar_precision) < CPP_OPTION (pfile, char_precision))
 399     cpp_error (pfile, DL_ICE,
 400                "target wchar_t is narrower than target char");
 401
 402   if (CPP_OPTION (pfile, int_precision) < CPP_OPTION (pfile, char_precision))
 403     cpp_error (pfile, DL_ICE,
 404                "target int is narrower than target char");
 405
 406   /* This is assumed in eval_token() and could be fixed if necessary.  */
 407   if (sizeof (cppchar_t) > sizeof (cpp_num_part))
 408     cpp_error (pfile, DL_ICE, "CPP half-integer narrower than CPP character");
 409
 410   if (CPP_OPTION (pfile, wchar_precision) > BITS_PER_CPPCHAR_T)
 411     cpp_error (pfile, DL_ICE,
 412                "CPP on this host cannot handle wide character constants over %lu bits, but the target requires %lu bits",
 413                (unsigned long) BITS_PER_CPPCHAR_T,
 414                (unsigned long) CPP_OPTION (pfile, wchar_precision));
 415 }
 416 #else
 417 # define sanity_checks(PFILE)
 418 #endif
 419
 420 /* Add a dependency target.  Can be called any number of times before
 421    cpp_read_main_file().  If no targets have been added before
 422    cpp_read_main_file(), then the default target is used.  */
 423 void
 424 cpp_add_dependency_target (pfile, target, quote)
 425      cpp_reader *pfile;
 426      const char *target;
 427      int quote;
 428 {
 429   if (!pfile->deps)
 430     pfile->deps = deps_init ();
 431
 432   deps_add_target (pfile->deps, target, quote);
 433 }
 434
 435 /* This is called after options have been parsed, and partially
 436    processed.  Setup for processing input from the file named FNAME,
 437    or stdin if it is the empty string.  Return the original filename
 438    on success (e.g. foo.i->foo.c), or NULL on failure.  */
 439 const char *
 440 cpp_read_main_file (pfile, fname)
 441      cpp_reader *pfile;
 442      const char *fname;
 443 {
 444   sanity_checks (pfile);
 445
 446   post_options (pfile);
 447
 448   /* Mark named operators before handling command line macros.  */
 449   if (CPP_OPTION (pfile, cplusplus) && CPP_OPTION (pfile, operator_names))
 450     mark_named_operators (pfile);
 451
 452   if (CPP_OPTION (pfile, deps.style) != DEPS_NONE)
 453     {
 454       if (!pfile->deps)
 455         pfile->deps = deps_init ();
 456
 457       /* Set the default target (if there is none already).  */
 458       deps_add_default_target (pfile->deps, fname);
 459     }
 460
 461   /* Open the main input file.  */
 462   pfile->line = 1;
 463   if (!_cpp_read_file (pfile, fname))
 464     return NULL;
 465
 466   /* Set this here so the client can change the option if it wishes,
 467      and after stacking the main file so we don't trace the main
 468      file.  */
 469   pfile->line_maps.trace_includes = CPP_OPTION (pfile, print_include_names);
 470
 471   /* For foo.i, read the original filename foo.c now, for the benefit
 472      of the front ends.  */
 473   if (CPP_OPTION (pfile, preprocessed))
 474     read_original_filename (pfile);
 475
 476   return pfile->map->to_file;
 477 }
 478
 479 /* For preprocessed files, if the first tokens are of the form # NUM.
 480    handle the directive so we know the original file name.  This will
 481    generate file_change callbacks, which the front ends must handle
 482    appropriately given their state of initialization.  */
 483 static void
 484 read_original_filename (pfile)
 485      cpp_reader *pfile;
 486 {
 487   const cpp_token *token, *token1;
 488
 489   /* Lex ahead; if the first tokens are of the form # NUM, then
 490      process the directive, otherwise back up.  */
 491   token = _cpp_lex_direct (pfile);
 492   if (token->type == CPP_HASH)
 493     {
 494       token1 = _cpp_lex_direct (pfile);
 495       _cpp_backup_tokens (pfile, 1);
 496
 497       /* If it's a #line directive, handle it.  */
 498       if (token1->type == CPP_NUMBER)
 499         {
 500           _cpp_handle_directive (pfile, token->flags & PREV_WHITE);
 501           return;
 502         }
 503     }
 504
 505   /* Backup as if nothing happened.  */
 506   _cpp_backup_tokens (pfile, 1);
 507 }
 508
 509 /* This is called at the end of preprocessing.  It pops the last
 510    buffer and writes dependency output, and returns the number of
 511    errors.
 512
 513    Maybe it should also reset state, such that you could call
 514    cpp_start_read with a new filename to restart processing.  */
 515 int
 516 cpp_finish (pfile, deps_stream)
 517      cpp_reader *pfile;
 518      FILE *deps_stream;
 519 {
 520   /* Warn about unused macros before popping the final buffer.  */
 521   if (CPP_OPTION (pfile, warn_unused_macros))
 522     cpp_forall_identifiers (pfile, _cpp_warn_if_unused_macro, NULL);
 523
 524   /* cpplex.c leaves the final buffer on the stack.  This it so that
 525      it returns an unending stream of CPP_EOFs to the client.  If we
 526      popped the buffer, we'd dereference a NULL buffer pointer and
 527      segfault.  It's nice to allow the client to do worry-free excess
 528      cpp_get_token calls.  */
 529   while (pfile->buffer)
 530     _cpp_pop_buffer (pfile);
 531
 532   /* Don't write the deps file if there are errors.  */
 533   if (CPP_OPTION (pfile, deps.style) != DEPS_NONE
 534       && deps_stream && pfile->errors == 0)
 535     {
 536       deps_write (pfile->deps, deps_stream, 72);
 537
 538       if (CPP_OPTION (pfile, deps.phony_targets))
 539         deps_phony_targets (pfile->deps, deps_stream);
 540     }
 541
 542   /* Report on headers that could use multiple include guards.  */
 543   if (CPP_OPTION (pfile, print_include_names))
 544     _cpp_report_missing_guards (pfile);
 545
 546   return pfile->errors;
 547 }
 548
 549 static void
 550 post_options (pfile)
 551      cpp_reader *pfile;
 552 {
 553   /* -Wtraditional is not useful in C++ mode.  */
 554   if (CPP_OPTION (pfile, cplusplus))
 555     CPP_OPTION (pfile, warn_traditional) = 0;
 556
 557   /* Permanently disable macro expansion if we are rescanning
 558      preprocessed text.  Read preprocesed source in ISO mode.  */
 559   if (CPP_OPTION (pfile, preprocessed))
 560     {
 561       pfile->state.prevent_expansion = 1;
 562       CPP_OPTION (pfile, traditional) = 0;
 563     }
 564
 565   if (CPP_OPTION (pfile, warn_trigraphs) == 2)
 566     CPP_OPTION (pfile, warn_trigraphs) = !CPP_OPTION (pfile, trigraphs);
 567
 568   if (CPP_OPTION (pfile, traditional))
 569     {
 570       /* Traditional CPP does not accurately track column information.  */
 571       CPP_OPTION (pfile, show_column) = 0;
 572       CPP_OPTION (pfile, trigraphs) = 0;
 573       CPP_OPTION (pfile, warn_trigraphs) = 0;
 574     }
 575 }