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_endif_labels) = 1;
 153   CPP_OPTION (pfile, warn_deprecated) = 1;
 154   CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
 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.  */
 330 void
 331 cpp_init_builtins (pfile)
 332      cpp_reader *pfile;
 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 (CPP_OPTION (pfile, objc))
 358     _cpp_define_builtin (pfile, "__OBJC__ 1");
 359 }
 360
 361 /* Sanity-checks are dependent on command-line options, so it is
 362    called as a subroutine of cpp_read_main_file ().  */
 363 #if ENABLE_CHECKING
 364 static void sanity_checks PARAMS ((cpp_reader *));
 365 static void sanity_checks (pfile)
 366      cpp_reader *pfile;
 367 {
 368   cppchar_t test = 0;
 369   size_t max_precision = 2 * CHAR_BIT * sizeof (cpp_num_part);
 370
 371   /* Sanity checks for assumptions about CPP arithmetic and target
 372      type precisions made by cpplib.  */
 373   test--;
 374   if (test < 1)
 375     cpp_error (pfile, DL_ICE, "cppchar_t must be an unsigned type");
 376
 377   if (CPP_OPTION (pfile, precision) > max_precision)
 378     cpp_error (pfile, DL_ICE,
 379                "preprocessor arithmetic has maximum precision of %lu bits; target requires %lu bits",
 380                (unsigned long) max_precision,
 381                (unsigned long) CPP_OPTION (pfile, precision));
 382
 383   if (CPP_OPTION (pfile, precision) < CPP_OPTION (pfile, int_precision))
 384     cpp_error (pfile, DL_ICE,
 385                "CPP arithmetic must be at least as precise as a target int");
 386
 387   if (CPP_OPTION (pfile, char_precision) < 8)
 388     cpp_error (pfile, DL_ICE, "target char is less than 8 bits wide");
 389
 390   if (CPP_OPTION (pfile, wchar_precision) < CPP_OPTION (pfile, char_precision))
 391     cpp_error (pfile, DL_ICE,
 392                "target wchar_t is narrower than target char");
 393
 394   if (CPP_OPTION (pfile, int_precision) < CPP_OPTION (pfile, char_precision))
 395     cpp_error (pfile, DL_ICE,
 396                "target int is narrower than target char");
 397
 398   /* This is assumed in eval_token() and could be fixed if necessary.  */
 399   if (sizeof (cppchar_t) > sizeof (cpp_num_part))
 400     cpp_error (pfile, DL_ICE, "CPP half-integer narrower than CPP character");
 401
 402   if (CPP_OPTION (pfile, wchar_precision) > BITS_PER_CPPCHAR_T)
 403     cpp_error (pfile, DL_ICE,
 404                "CPP on this host cannot handle wide character constants over %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 (pfile, target, quote)
 417      cpp_reader *pfile;
 418      const char *target;
 419      int quote;
 420 {
 421   if (!pfile->deps)
 422     pfile->deps = deps_init ();
 423
 424   deps_add_target (pfile->deps, target, quote);
 425 }
 426
 427 /* This is called after options have been parsed, and partially
 428    processed.  Setup for processing input from the file named FNAME,
 429    or stdin if it is the empty string.  Return the original filename
 430    on success (e.g. foo.i->foo.c), or NULL on failure.  */
 431 const char *
 432 cpp_read_main_file (pfile, fname)
 433      cpp_reader *pfile;
 434      const char *fname;
 435 {
 436   sanity_checks (pfile);
 437
 438   post_options (pfile);
 439
 440   /* Mark named operators before handling command line macros.  */
 441   if (CPP_OPTION (pfile, cplusplus) && CPP_OPTION (pfile, operator_names))
 442     mark_named_operators (pfile);
 443
 444   if (CPP_OPTION (pfile, deps.style) != DEPS_NONE)
 445     {
 446       if (!pfile->deps)
 447         pfile->deps = deps_init ();
 448
 449       /* Set the default target (if there is none already).  */
 450       deps_add_default_target (pfile->deps, fname);
 451     }
 452
 453   /* Open the main input file.  */
 454   pfile->line = 1;
 455   if (!_cpp_read_file (pfile, fname))
 456     return NULL;
 457
 458   /* Set this here so the client can change the option if it wishes,
 459      and after stacking the main file so we don't trace the main
 460      file.  */
 461   pfile->line_maps.trace_includes = CPP_OPTION (pfile, print_include_names);
 462
 463   /* For foo.i, read the original filename foo.c now, for the benefit
 464      of the front ends.  */
 465   if (CPP_OPTION (pfile, preprocessed))
 466     read_original_filename (pfile);
 467
 468   return pfile->map->to_file;
 469 }
 470
 471 /* For preprocessed files, if the first tokens are of the form # NUM.
 472    handle the directive so we know the original file name.  This will
 473    generate file_change callbacks, which the front ends must handle
 474    appropriately given their state of initialization.  */
 475 static void
 476 read_original_filename (pfile)
 477      cpp_reader *pfile;
 478 {
 479   const cpp_token *token, *token1;
 480
 481   /* Lex ahead; if the first tokens are of the form # NUM, then
 482      process the directive, otherwise back up.  */
 483   token = _cpp_lex_direct (pfile);
 484   if (token->type == CPP_HASH)
 485     {
 486       token1 = _cpp_lex_direct (pfile);
 487       _cpp_backup_tokens (pfile, 1);
 488
 489       /* If it's a #line directive, handle it.  */
 490       if (token1->type == CPP_NUMBER)
 491         {
 492           _cpp_handle_directive (pfile, token->flags & PREV_WHITE);
 493           return;
 494         }
 495     }
 496
 497   /* Backup as if nothing happened.  */
 498   _cpp_backup_tokens (pfile, 1);
 499 }
 500
 501 /* This is called at the end of preprocessing.  It pops the last
 502    buffer and writes dependency output, and returns the number of
 503    errors.
 504
 505    Maybe it should also reset state, such that you could call
 506    cpp_start_read with a new filename to restart processing.  */
 507 int
 508 cpp_finish (pfile, deps_stream)
 509      cpp_reader *pfile;
 510      FILE *deps_stream;
 511 {
 512   /* Warn about unused macros before popping the final buffer.  */
 513   if (CPP_OPTION (pfile, warn_unused_macros))
 514     cpp_forall_identifiers (pfile, _cpp_warn_if_unused_macro, NULL);
 515
 516   /* cpplex.c leaves the final buffer on the stack.  This it so that
 517      it returns an unending stream of CPP_EOFs to the client.  If we
 518      popped the buffer, we'd dereference a NULL buffer pointer and
 519      segfault.  It's nice to allow the client to do worry-free excess
 520      cpp_get_token calls.  */
 521   while (pfile->buffer)
 522     _cpp_pop_buffer (pfile);
 523
 524   /* Don't write the deps file if there are errors.  */
 525   if (CPP_OPTION (pfile, deps.style) != DEPS_NONE
 526       && deps_stream && pfile->errors == 0)
 527     {
 528       deps_write (pfile->deps, deps_stream, 72);
 529
 530       if (CPP_OPTION (pfile, deps.phony_targets))
 531         deps_phony_targets (pfile->deps, deps_stream);
 532     }
 533
 534   /* Report on headers that could use multiple include guards.  */
 535   if (CPP_OPTION (pfile, print_include_names))
 536     _cpp_report_missing_guards (pfile);
 537
 538   return pfile->errors;
 539 }
 540
 541 static void
 542 post_options (pfile)
 543      cpp_reader *pfile;
 544 {
 545   /* -Wtraditional is not useful in C++ mode.  */
 546   if (CPP_OPTION (pfile, cplusplus))
 547     CPP_OPTION (pfile, warn_traditional) = 0;
 548
 549   /* Permanently disable macro expansion if we are rescanning
 550      preprocessed text.  Read preprocesed source in ISO mode.  */
 551   if (CPP_OPTION (pfile, preprocessed))
 552     {
 553       pfile->state.prevent_expansion = 1;
 554       CPP_OPTION (pfile, traditional) = 0;
 555     }
 556
 557   if (CPP_OPTION (pfile, traditional))
 558     {
 559       /* Traditional CPP does not accurately track column information.  */
 560       CPP_OPTION (pfile, show_column) = 0;
 561       CPP_OPTION (pfile, trigraphs) = 0;
 562       CPP_OPTION (pfile, warn_trigraphs) = 0;
 563     }
 564 }