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 "cpplib.h"
  25 #include "cpphash.h"
  26 #include "mkdeps.h"
  27
  28 static void init_library (void);
  29 static void mark_named_operators (cpp_reader *);
  30 static void read_original_filename (cpp_reader *);
  31 static void post_options (cpp_reader *);
  32
  33 /* If we have designated initializers (GCC >2.7) these tables can be
  34    initialized, constant data.  Otherwise, they have to be filled in at
  35    runtime.  */
  36 #if HAVE_DESIGNATED_INITIALIZERS
  37
  38 #define init_trigraph_map()  /* Nothing.  */
  39 #define TRIGRAPH_MAP \
  40 __extension__ const uchar _cpp_trigraph_map[UCHAR_MAX + 1] = {
  41
  42 #define END };
  43 #define s(p, v) [p] = v,
  44
  45 #else
  46
  47 #define TRIGRAPH_MAP uchar _cpp_trigraph_map[UCHAR_MAX + 1] = { 0 }; \
  48  static void init_trigraph_map (void) { \
  49  unsigned char *x = _cpp_trigraph_map;
  50
  51 #define END }
  52 #define s(p, v) x[p] = v;
  53
  54 #endif
  55
  56 TRIGRAPH_MAP
  57   s('=', '#')   s(')', ']')     s('!', '|')
  58   s('(', '[')   s('\'', '^')    s('>', '}')
  59   s('/', '\\')  s('<', '{')     s('-', '~')
  60 END
  61
  62 #undef s
  63 #undef END
  64 #undef TRIGRAPH_MAP
  65
  66 /* A set of booleans indicating what CPP features each source language
  67    requires.  */
  68 struct lang_flags
  69 {
  70   char c99;
  71   char cplusplus;
  72   char extended_numbers;
  73   char std;
  74   char cplusplus_comments;
  75   char digraphs;
  76 };
  77
  78 static const struct lang_flags lang_defaults[] =
  79 { /*              c99 c++ xnum std  //   digr  */
  80   /* GNUC89 */  { 0,  0,  1,   0,   1,   1     },
  81   /* GNUC99 */  { 1,  0,  1,   0,   1,   1     },
  82   /* STDC89 */  { 0,  0,  0,   1,   0,   0     },
  83   /* STDC94 */  { 0,  0,  0,   1,   0,   1     },
  84   /* STDC99 */  { 1,  0,  1,   1,   1,   1     },
  85   /* GNUCXX */  { 0,  1,  1,   0,   1,   1     },
  86   /* CXX98  */  { 0,  1,  1,   1,   1,   1     },
  87   /* ASM    */  { 0,  0,  1,   0,   1,   0     }
  88 };
  89
  90 /* Sets internal flags correctly for a given language.  */
  91 void
  92 cpp_set_lang (cpp_reader *pfile, enum c_lang lang)
  93 {
  94   const struct lang_flags *l = &lang_defaults[(int) lang];
  95
  96   CPP_OPTION (pfile, lang) = lang;
  97
  98   CPP_OPTION (pfile, c99)                = l->c99;
  99   CPP_OPTION (pfile, cplusplus)          = l->cplusplus;
 100   CPP_OPTION (pfile, extended_numbers)   = l->extended_numbers;
 101   CPP_OPTION (pfile, std)                = l->std;
 102   CPP_OPTION (pfile, trigraphs)          = l->std;
 103   CPP_OPTION (pfile, cplusplus_comments) = l->cplusplus_comments;
 104   CPP_OPTION (pfile, digraphs)           = l->digraphs;
 105 }
 106
 107 /* Initialize library global state.  */
 108 static void
 109 init_library (void)
 110 {
 111   static int initialized = 0;
 112
 113   if (! initialized)
 114     {
 115       initialized = 1;
 116
 117       /* Set up the trigraph map.  This doesn't need to do anything if
 118          we were compiled with a compiler that supports C99 designated
 119          initializers.  */
 120       init_trigraph_map ();
 121     }
 122 }
 123
 124 /* Initialize a cpp_reader structure.  */
 125 cpp_reader *
 126 cpp_create_reader (enum c_lang lang, hash_table *table)
 127 {
 128   cpp_reader *pfile;
 129
 130   /* Initialize this instance of the library if it hasn't been already.  */
 131   init_library ();
 132
 133   pfile = (cpp_reader *) xcalloc (1, sizeof (cpp_reader));
 134
 135   cpp_set_lang (pfile, lang);
 136   CPP_OPTION (pfile, warn_import) = 1;
 137   CPP_OPTION (pfile, warn_multichar) = 1;
 138   CPP_OPTION (pfile, discard_comments) = 1;
 139   CPP_OPTION (pfile, discard_comments_in_macro_exp) = 1;
 140   CPP_OPTION (pfile, show_column) = 1;
 141   CPP_OPTION (pfile, tabstop) = 8;
 142   CPP_OPTION (pfile, operator_names) = 1;
 143   CPP_OPTION (pfile, warn_trigraphs) = 2;
 144   CPP_OPTION (pfile, warn_endif_labels) = 1;
 145   CPP_OPTION (pfile, warn_deprecated) = 1;
 146   CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
 147   CPP_OPTION (pfile, dollars_in_ident) = 1;
 148   CPP_OPTION (pfile, warn_dollars) = 1;
 149
 150   /* Default CPP arithmetic to something sensible for the host for the
 151      benefit of dumb users like fix-header.  */
 152   CPP_OPTION (pfile, precision) = CHAR_BIT * sizeof (long);
 153   CPP_OPTION (pfile, char_precision) = CHAR_BIT;
 154   CPP_OPTION (pfile, wchar_precision) = CHAR_BIT * sizeof (int);
 155   CPP_OPTION (pfile, int_precision) = CHAR_BIT * sizeof (int);
 156   CPP_OPTION (pfile, unsigned_char) = 0;
 157   CPP_OPTION (pfile, unsigned_wchar) = 1;
 158   CPP_OPTION (pfile, bytes_big_endian) = 1;  /* does not matter */
 159
 160   /* Default to no charset conversion.  */
 161   CPP_OPTION (pfile, narrow_charset) = 0;
 162   CPP_OPTION (pfile, wide_charset) = 0;
 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   _obstack_begin (&pfile->buffer_ob, 0, 0,
 196                   (void *(*) (long)) xmalloc,
 197                   (void (*) (void *)) free);
 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 (cpp_reader *pfile)
 210 {
 211   cpp_context *context, *contextn;
 212   tokenrun *run, *runn;
 213
 214   free (pfile->op_stack);
 215
 216   while (CPP_BUFFER (pfile) != NULL)
 217     _cpp_pop_buffer (pfile);
 218
 219   if (pfile->out.base)
 220     free (pfile->out.base);
 221
 222   if (pfile->macro_buffer)
 223     {
 224       free (pfile->macro_buffer);
 225       pfile->macro_buffer = NULL;
 226       pfile->macro_buffer_len = 0;
 227     }
 228
 229   if (pfile->deps)
 230     deps_free (pfile->deps);
 231   obstack_free (&pfile->buffer_ob, 0);
 232
 233   _cpp_destroy_hashtable (pfile);
 234   _cpp_cleanup_includes (pfile);
 235   _cpp_destroy_iconv (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 (cpp_reader *pfile)
 314 {
 315   const struct builtin *b;
 316
 317   for (b = operator_array;
 318        b < (operator_array + ARRAY_SIZE (operator_array));
 319        b++)
 320     {
 321       cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
 322       hp->flags |= NODE_OPERATOR;
 323       hp->is_directive = 0;
 324       hp->directive_index = b->value;
 325     }
 326 }
 327
 328 /* Read the builtins table above and enter them, and language-specific
 329    macros, into the hash table.  HOSTED is true if this is a hosted
 330    environment.  */
 331 void
 332 cpp_init_builtins (cpp_reader *pfile, int hosted)
 333 {
 334   const struct builtin *b;
 335   size_t n = ARRAY_SIZE (builtin_array);
 336
 337   if (CPP_OPTION (pfile, traditional))
 338     n -= 2;
 339
 340   for(b = builtin_array; b < builtin_array + n; b++)
 341     {
 342       cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
 343       hp->type = NT_MACRO;
 344       hp->flags |= NODE_BUILTIN | NODE_WARN;
 345       hp->value.builtin = b->value;
 346     }
 347
 348   if (CPP_OPTION (pfile, cplusplus))
 349     _cpp_define_builtin (pfile, "__cplusplus 1");
 350   else if (CPP_OPTION (pfile, lang) == CLK_ASM)
 351     _cpp_define_builtin (pfile, "__ASSEMBLER__ 1");
 352   else if (CPP_OPTION (pfile, lang) == CLK_STDC94)
 353     _cpp_define_builtin (pfile, "__STDC_VERSION__ 199409L");
 354   else if (CPP_OPTION (pfile, c99))
 355     _cpp_define_builtin (pfile, "__STDC_VERSION__ 199901L");
 356
 357   if (hosted)
 358     _cpp_define_builtin (pfile, "__STDC_HOSTED__ 1");
 359   else
 360     _cpp_define_builtin (pfile, "__STDC_HOSTED__ 0");
 361
 362   if (CPP_OPTION (pfile, objc))
 363     _cpp_define_builtin (pfile, "__OBJC__ 1");
 364 }
 365
 366 /* Sanity-checks are dependent on command-line options, so it is
 367    called as a subroutine of cpp_read_main_file ().  */
 368 #if ENABLE_CHECKING
 369 static void sanity_checks (cpp_reader *);
 370 static void sanity_checks (cpp_reader *pfile)
 371 {
 372   cppchar_t test = 0;
 373   size_t max_precision = 2 * CHAR_BIT * sizeof (cpp_num_part);
 374
 375   /* Sanity checks for assumptions about CPP arithmetic and target
 376      type precisions made by cpplib.  */
 377   test--;
 378   if (test < 1)
 379     cpp_error (pfile, DL_ICE, "cppchar_t must be an unsigned type");
 380
 381   if (CPP_OPTION (pfile, precision) > max_precision)
 382     cpp_error (pfile, DL_ICE,
 383                "preprocessor arithmetic has maximum precision of %lu bits;"
 384                " target requires %lu bits",
 385                (unsigned long) max_precision,
 386                (unsigned long) CPP_OPTION (pfile, precision));
 387
 388   if (CPP_OPTION (pfile, precision) < CPP_OPTION (pfile, int_precision))
 389     cpp_error (pfile, DL_ICE,
 390                "CPP arithmetic must be at least as precise as a target int");
 391
 392   if (CPP_OPTION (pfile, char_precision) < 8)
 393     cpp_error (pfile, DL_ICE, "target char is less than 8 bits wide");
 394
 395   if (CPP_OPTION (pfile, wchar_precision) < CPP_OPTION (pfile, char_precision))
 396     cpp_error (pfile, DL_ICE,
 397                "target wchar_t is narrower than target char");
 398
 399   if (CPP_OPTION (pfile, int_precision) < CPP_OPTION (pfile, char_precision))
 400     cpp_error (pfile, DL_ICE,
 401                "target int is narrower than target char");
 402
 403   /* This is assumed in eval_token() and could be fixed if necessary.  */
 404   if (sizeof (cppchar_t) > sizeof (cpp_num_part))
 405     cpp_error (pfile, DL_ICE, "CPP half-integer narrower than CPP character");
 406
 407   if (CPP_OPTION (pfile, wchar_precision) > BITS_PER_CPPCHAR_T)
 408     cpp_error (pfile, DL_ICE,
 409                "CPP on this host cannot handle wide character constants over"
 410                " %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 (cpp_reader *pfile, const char *target, int quote)
 423 {
 424   if (!pfile->deps)
 425     pfile->deps = deps_init ();
 426
 427   deps_add_target (pfile->deps, target, quote);
 428 }
 429
 430 /* This sets up for processing input from the file FNAME.
 431    It returns false on error.  */
 432 bool
 433 cpp_read_next_file (cpp_reader *pfile, const char *fname)
 434 {
 435   /* Open the main input file.  */
 436   return _cpp_read_file (pfile, fname);
 437 }
 438
 439 /* This is called after options have been parsed, and partially
 440    processed.  Setup for processing input from the file named FNAME,
 441    or stdin if it is the empty string.  Return the original filename
 442    on success (e.g. foo.i->foo.c), or NULL on failure.  */
 443 const char *
 444 cpp_read_main_file (cpp_reader *pfile, const char *fname)
 445 {
 446   sanity_checks (pfile);
 447
 448   post_options (pfile);
 449
 450   /* Mark named operators before handling command line macros.  */
 451   if (CPP_OPTION (pfile, cplusplus) && CPP_OPTION (pfile, operator_names))
 452     mark_named_operators (pfile);
 453
 454   if (CPP_OPTION (pfile, deps.style) != DEPS_NONE)
 455     {
 456       if (!pfile->deps)
 457         pfile->deps = deps_init ();
 458
 459       /* Set the default target (if there is none already).  */
 460       deps_add_default_target (pfile->deps, fname);
 461     }
 462
 463   pfile->line = 1;
 464   if (!cpp_read_next_file (pfile, fname))
 465     return NULL;
 466
 467   /* Set this here so the client can change the option if it wishes,
 468      and after stacking the main file so we don't trace the main
 469      file.  */
 470   pfile->line_maps.trace_includes = CPP_OPTION (pfile, print_include_names);
 471
 472   /* For foo.i, read the original filename foo.c now, for the benefit
 473      of the front ends.  */
 474   if (CPP_OPTION (pfile, preprocessed))
 475     read_original_filename (pfile);
 476
 477   return pfile->map->to_file;
 478 }
 479
 480 /* For preprocessed files, if the first tokens are of the form # NUM.
 481    handle the directive so we know the original file name.  This will
 482    generate file_change callbacks, which the front ends must handle
 483    appropriately given their state of initialization.  */
 484 static void
 485 read_original_filename (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 (cpp_reader *pfile, 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 (cpp_reader *pfile)
 549 {
 550   /* -Wtraditional is not useful in C++ mode.  */
 551   if (CPP_OPTION (pfile, cplusplus))
 552     CPP_OPTION (pfile, warn_traditional) = 0;
 553
 554   /* Permanently disable macro expansion if we are rescanning
 555      preprocessed text.  Read preprocesed source in ISO mode.  */
 556   if (CPP_OPTION (pfile, preprocessed))
 557     {
 558       pfile->state.prevent_expansion = 1;
 559       CPP_OPTION (pfile, traditional) = 0;
 560     }
 561
 562   if (CPP_OPTION (pfile, warn_trigraphs) == 2)
 563     CPP_OPTION (pfile, warn_trigraphs) = !CPP_OPTION (pfile, trigraphs);
 564
 565   if (CPP_OPTION (pfile, traditional))
 566     {
 567       /* Traditional CPP does not accurately track column information.  */
 568       CPP_OPTION (pfile, show_column) = 0;
 569       CPP_OPTION (pfile, trigraphs) = 0;
 570       CPP_OPTION (pfile, warn_trigraphs) = 0;
 571     }
 572 }