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