src/m4.h

   1 /* GNU m4 -- A simple macro processor
   2
   3    Copyright (C) 1989, 1990, 1991, 1992, 1993, 1994, 2004, 2005, 2006, 2007,
   4    2008 Free Software Foundation, Inc.
   5
   6    This file is part of GNU M4.
   7
   8    GNU M4 is free software: you can redistribute it and/or modify
   9    it under the terms of the GNU General Public License as published by
  10    the Free Software Foundation, either version 3 of the License, or
  11    (at your option) any later version.
  12
  13    GNU M4 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, see <http://www.gnu.org/licenses/>.
  20 */
  21
  22 /* We use <config.h> instead of "config.h" so that a compilation
  23    using -I. -I$srcdir will use ./config.h rather than $srcdir/config.h
  24    (which it would do because it found this file in $srcdir).  */
  25
  26 #include <config.h>
  27
  28 #include <assert.h>
  29 #include <ctype.h>
  30 #include <errno.h>
  31 #include <limits.h>
  32 #include <stdbool.h>
  33 #include <stdint.h>
  34 #include <string.h>
  35 #include <sys/types.h>
  36
  37 #include "binary-io.h"
  38 #include "clean-temp.h"
  39 #include "cloexec.h"
  40 #include "close-stream.h"
  41 #include "closein.h"
  42 #include "error.h"
  43 #include "exitfail.h"
  44 #include "intprops.h"
  45 #include "obstack.h"
  46 #include "stdio--.h"
  47 #include "stdlib--.h"
  48 #include "unistd--.h"
  49 #include "vasnprintf.h"
  50 #include "verror.h"
  51 #include "xalloc.h"
  52 #include "xprintf.h"
  53 #include "xvasprintf.h"
  54
  55 /* Canonicalize UNIX recognition macros.  */
  56 #if defined unix || defined __unix || defined __unix__ \
  57   || defined _POSIX_VERSION || defined _POSIX2_VERSION \
  58   || defined __NetBSD__ || defined __OpenBSD__ \
  59   || defined __APPLE__ || defined __APPLE_CC__
  60 # define UNIX 1
  61 #endif
  62
  63 /* Canonicalize Windows recognition macros.  */
  64 #if (defined _WIN32 || defined __WIN32__) && !defined __CYGWIN__
  65 # define W32_NATIVE 1
  66 #endif
  67
  68 /* Canonicalize OS/2 recognition macro.  */
  69 #ifdef __EMX__
  70 # define OS2 1
  71 #endif
  72
  73 /* Used for version mismatch, when -R detects a frozen file it can't parse.  */
  74 #define EXIT_MISMATCH 63
  75
  76 /* M4 1.4.x is not yet internationalized.  But when it is, this can be
  77    redefined as gettext().  */
  78 #define _(STRING) STRING
  79 \f
  80 /* Various declarations.  */
  81
  82 struct string
  83   {
  84     char *string;               /* characters of the string */
  85     size_t length;              /* length of the string */
  86   };
  87 typedef struct string STRING;
  88
  89 /* Memory allocation.  */
  90 #define obstack_chunk_alloc     xmalloc
  91 #define obstack_chunk_free      free
  92
  93 /* These must come first.  */
  94 typedef struct input_block input_block;
  95 typedef struct token_data token_data;
  96 typedef struct macro_arguments macro_arguments;
  97 typedef void builtin_func (struct obstack *, int, macro_arguments *);
  98
  99 /* Gnulib's stdbool doesn't work with bool bitfields.  For nicer
 100    debugging, use bool when we know it works, but use the more
 101    portable unsigned int elsewhere.  */
 102 #if __GNUC__ > 2
 103 typedef bool bool_bitfield;
 104 #else
 105 typedef unsigned int bool_bitfield;
 106 #endif /* !__GNUC__ */
 107
 108 /* Take advantage of GNU C compiler source level optimization hints,
 109    using portable macros.  */
 110 #if __GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ > 6)
 111 #  define M4_GNUC_ATTRIBUTE(args)       __attribute__ (args)
 112 #else
 113 #  define M4_GNUC_ATTRIBUTE(args)
 114 #endif  /* __GNUC__ */
 115
 116 #define M4_GNUC_UNUSED          M4_GNUC_ATTRIBUTE ((__unused__))
 117 #define M4_GNUC_PRINTF(fmt, arg)                        \
 118   M4_GNUC_ATTRIBUTE ((__format__ (__printf__, fmt, arg)))
 119 \f
 120 /* File: m4.c  --- global definitions.  */
 121
 122 /* Option flags.  */
 123 extern int sync_output;                 /* -s */
 124 extern int debug_level;                 /* -d */
 125 extern size_t hash_table_size;          /* -H */
 126 extern int no_gnu_extensions;           /* -G */
 127 extern int prefix_all_builtins;         /* -P */
 128 extern int max_debug_argument_length;   /* -l */
 129 extern int suppress_warnings;           /* -Q */
 130 extern int warning_status;              /* -E */
 131 extern int nesting_limit;               /* -L */
 132 #ifdef ENABLE_CHANGEWORD
 133 extern const char *user_word_regexp;    /* -W */
 134 #endif
 135
 136 /* Error handling.  */
 137 extern int retcode;
 138 extern const char *program_name;
 139
 140 void m4_error (int, int, const char *, const char *, ...)
 141   M4_GNUC_PRINTF (4, 5);
 142 void m4_error_at_line (int, int, const char *, int, const char *,
 143                        const char *, ...) M4_GNUC_PRINTF (6, 7);
 144 void m4_warn (int, const char *, const char *, ...) M4_GNUC_PRINTF (3, 4);
 145 void m4_warn_at_line (int, const char *, int, const char *,
 146                       const char *, ...) M4_GNUC_PRINTF (5, 6);
 147
 148 #ifdef USE_STACKOVF
 149 void setup_stackovf_trap (char *const *, char *const *,
 150                           void (*handler) (void));
 151 #endif
 152 \f
 153 /* File: debug.c  --- debugging and tracing function.  */
 154
 155 extern FILE *debug;
 156
 157 /* The value of debug_level is a bitmask of the following.  */
 158
 159 /* a: show arglist in trace output */
 160 #define DEBUG_TRACE_ARGS 1
 161 /* e: show expansion in trace output */
 162 #define DEBUG_TRACE_EXPANSION 2
 163 /* q: quote args and expansion in trace output */
 164 #define DEBUG_TRACE_QUOTE 4
 165 /* t: trace all macros -- overrides trace{on,off} */
 166 #define DEBUG_TRACE_ALL 8
 167 /* l: add line numbers to trace output */
 168 #define DEBUG_TRACE_LINE 16
 169 /* f: add file name to trace output */
 170 #define DEBUG_TRACE_FILE 32
 171 /* p: trace path search of include files */
 172 #define DEBUG_TRACE_PATH 64
 173 /* c: show macro call before args collection */
 174 #define DEBUG_TRACE_CALL 128
 175 /* i: trace changes of input files */
 176 #define DEBUG_TRACE_INPUT 256
 177 /* x: add call id to trace output */
 178 #define DEBUG_TRACE_CALLID 512
 179
 180 /* V: very verbose --  print everything */
 181 #define DEBUG_TRACE_VERBOSE 1023
 182 /* default flags -- equiv: aeq */
 183 #define DEBUG_TRACE_DEFAULT 7
 184
 185 #define DEBUG_PRINT1(Fmt, Arg1)                                 \
 186   do                                                            \
 187     {                                                           \
 188       if (debug != NULL)                                        \
 189         xfprintf (debug, Fmt, Arg1);                            \
 190     }                                                           \
 191   while (0)
 192
 193 #define DEBUG_PRINT3(Fmt, Arg1, Arg2, Arg3)                     \
 194   do                                                            \
 195     {                                                           \
 196       if (debug != NULL)                                        \
 197         xfprintf (debug, Fmt, Arg1, Arg2, Arg3);                \
 198     }                                                           \
 199   while (0)
 200
 201 #define DEBUG_MESSAGE(Fmt)                                      \
 202   do                                                            \
 203     {                                                           \
 204       if (debug != NULL)                                        \
 205         {                                                       \
 206           debug_message_prefix ();                              \
 207           xfprintf (debug, Fmt);                                \
 208           putc ('\n', debug);                                   \
 209         }                                                       \
 210     }                                                           \
 211   while (0)
 212
 213 #define DEBUG_MESSAGE1(Fmt, Arg1)                               \
 214   do                                                            \
 215     {                                                           \
 216       if (debug != NULL)                                        \
 217         {                                                       \
 218           debug_message_prefix ();                              \
 219           xfprintf (debug, Fmt, Arg1);                          \
 220           putc ('\n', debug);                                   \
 221         }                                                       \
 222     }                                                           \
 223   while (0)
 224
 225 #define DEBUG_MESSAGE2(Fmt, Arg1, Arg2)                         \
 226   do                                                            \
 227     {                                                           \
 228       if (debug != NULL)                                        \
 229         {                                                       \
 230           debug_message_prefix ();                              \
 231           xfprintf (debug, Fmt, Arg1, Arg2);                    \
 232           putc ('\n', debug);                                   \
 233         }                                                       \
 234     }                                                           \
 235   while (0)
 236
 237 void debug_init (void);
 238 int debug_decode (const char *);
 239 void debug_flush_files (void);
 240 bool debug_set_output (const char *, const char *);
 241 void debug_message_prefix (void);
 242
 243 void trace_prepre (const char *, int);
 244 void trace_pre (const char *, int, macro_arguments *);
 245 void trace_post (const char *, int, macro_arguments *,
 246                  const input_block *);
 247
 248 bool obstack_print (struct obstack *, const char *, size_t, int *);
 249 \f
 250 /* File: input.c  --- lexical definitions.  */
 251
 252 typedef struct token_chain token_chain;
 253
 254 /* Various different token types.  Avoid overlap with token_data_type,
 255    since the shared prefix of the enumerators is a bit confusing.  */
 256 enum token_type
 257 {
 258   TOKEN_EOF = 4,/* End of file, TOKEN_VOID.  */
 259   TOKEN_STRING, /* Quoted string or comment, TOKEN_TEXT or TOKEN_COMP.  */
 260   TOKEN_WORD,   /* An identifier, TOKEN_TEXT.  */
 261   TOKEN_OPEN,   /* Active character `(', TOKEN_TEXT.  */
 262   TOKEN_COMMA,  /* Active character `,', TOKEN_TEXT.  */
 263   TOKEN_CLOSE,  /* Active character `)', TOKEN_TEXT.  */
 264   TOKEN_SIMPLE, /* Any other single character, TOKEN_TEXT.  */
 265   TOKEN_MACDEF  /* A macro's definition (see "defn"), TOKEN_FUNC.  */
 266 };
 267
 268 /* The data for a token, a macro argument, and a macro definition.  */
 269 enum token_data_type
 270 {
 271   TOKEN_VOID,   /* Token still being constructed, u is invalid.  */
 272   TOKEN_TEXT,   /* Straight text, u.u_t is valid.  */
 273   TOKEN_FUNC,   /* Builtin function definition, u.func is valid.  */
 274   TOKEN_COMP    /* Composite argument, u.chain is valid.  */
 275 };
 276
 277 /* Composite tokens are built of a linked list of chains.  */
 278 struct token_chain
 279 {
 280   token_chain *next;    /* Pointer to next link of chain.  */
 281   const char *str;      /* NUL-terminated string if text, else NULL.  */
 282   size_t len;           /* Length of str, else 0.  */
 283   int level;            /* Expansion level of link content, or -1.  */
 284   macro_arguments *argv;/* Reference to earlier $@.  */
 285   unsigned int index;   /* Argument index within argv.  */
 286   bool flatten;         /* True to treat builtins as text.  */
 287 };
 288
 289 /* The content of a token or macro argument.  */
 290 struct token_data
 291 {
 292   enum token_data_type type;
 293   union
 294     {
 295       struct
 296         {
 297           /* We don't support NUL in text, yet.  So len is just a
 298              cache for now.  But it will be essential if we ever DO
 299              support NUL.  */
 300           size_t len;
 301           char *text; /* The contents of the token.  */
 302           /* The value of quote_age when this token was scanned.  If
 303              this token is later encountered in the context of
 304              scanning a quoted string, and quote_age has not changed,
 305              then rescanning this string is provably unnecessary.  If
 306              zero, then this string potentially contains content that
 307              might change the parse on rescan.  Ignored for 0 len.  */
 308           unsigned int quote_age;
 309 #ifdef ENABLE_CHANGEWORD
 310           /* If changeword is in effect, and contains a () group, then
 311              this contains the entire token, while text contains the
 312              portion that matched the () group to form a macro name.
 313              Otherwise, this field is unused.  */
 314           const char *original_text;
 315 #endif
 316         }
 317       u_t;
 318       builtin_func *func;
 319
 320       /* Composite text: a linked list of straight text and $@
 321          placeholders.  */
 322       token_chain *chain;
 323     }
 324   u;
 325 };
 326
 327 #define TOKEN_DATA_TYPE(Td)             ((Td)->type)
 328 #define TOKEN_DATA_LEN(Td)              ((Td)->u.u_t.len)
 329 #define TOKEN_DATA_TEXT(Td)             ((Td)->u.u_t.text)
 330 #define TOKEN_DATA_QUOTE_AGE(Td)        ((Td)->u.u_t.quote_age)
 331 #ifdef ENABLE_CHANGEWORD
 332 # define TOKEN_DATA_ORIG_TEXT(Td)       ((Td)->u.u_t.original_text)
 333 #endif
 334 #define TOKEN_DATA_FUNC(Td)             ((Td)->u.func)
 335
 336 typedef enum token_type token_type;
 337 typedef enum token_data_type token_data_type;
 338
 339 void input_init (void);
 340 token_type peek_token (void);
 341 token_type next_token (token_data *, int *, struct obstack *, const char *);
 342 void skip_line (const char *);
 343
 344 /* push back input */
 345 void push_file (FILE *, const char *, bool);
 346 void push_macro (builtin_func *);
 347 struct obstack *push_string_init (void);
 348 bool push_token (token_data *, int);
 349 const input_block *push_string_finish (void);
 350 void push_wrapup (const char *);
 351 bool pop_wrapup (void);
 352 void input_print (struct obstack *, const input_block *);
 353
 354 /* current input file, and line */
 355 extern const char *current_file;
 356 extern int current_line;
 357
 358 /* left and right quote, begin and end comment */
 359 extern STRING bcomm, ecomm;
 360 extern STRING lquote, rquote;
 361
 362 #define DEF_LQUOTE "`"
 363 #define DEF_RQUOTE "\'"
 364 #define DEF_BCOMM "#"
 365 #define DEF_ECOMM "\n"
 366
 367 void set_quotes (const char *, const char *);
 368 void set_comment (const char *, const char *);
 369 #ifdef ENABLE_CHANGEWORD
 370 void set_word_regexp (const char *, const char *);
 371 #endif
 372 unsigned int quote_age (void);
 373 bool safe_quotes (void);
 374 \f
 375 /* File: output.c --- output functions.  */
 376 extern int current_diversion;
 377 extern int output_current_line;
 378
 379 void output_init (void);
 380 void output_exit (void);
 381 void output_text (const char *, int);
 382 void shipout_text (struct obstack *, const char *, int, int);
 383 void make_diversion (int);
 384 void insert_diversion (int);
 385 void insert_file (FILE *);
 386 void freeze_diversions (FILE *);
 387 \f
 388 /* File symtab.c  --- symbol table definitions.  */
 389
 390 /* Operation modes for lookup_symbol ().  */
 391 enum symbol_lookup
 392 {
 393   SYMBOL_LOOKUP,
 394   SYMBOL_INSERT,
 395   SYMBOL_DELETE,
 396   SYMBOL_PUSHDEF,
 397   SYMBOL_POPDEF
 398 };
 399
 400 /* Symbol table entry.  */
 401 struct symbol
 402 {
 403   struct symbol *next;
 404   bool_bitfield traced : 1;
 405   bool_bitfield shadowed : 1;
 406   bool_bitfield macro_args : 1;
 407   bool_bitfield blind_no_args : 1;
 408   bool_bitfield deleted : 1;
 409   int pending_expansions;
 410
 411   char *name;
 412   token_data data;  /* Type should be only TOKEN_TEXT or TOKEN_FUNC.  */
 413 };
 414
 415 #define SYMBOL_NEXT(S)          ((S)->next)
 416 #define SYMBOL_TRACED(S)        ((S)->traced)
 417 #define SYMBOL_SHADOWED(S)      ((S)->shadowed)
 418 #define SYMBOL_MACRO_ARGS(S)    ((S)->macro_args)
 419 #define SYMBOL_BLIND_NO_ARGS(S) ((S)->blind_no_args)
 420 #define SYMBOL_DELETED(S)       ((S)->deleted)
 421 #define SYMBOL_PENDING_EXPANSIONS(S) ((S)->pending_expansions)
 422 #define SYMBOL_NAME(S)          ((S)->name)
 423 #define SYMBOL_TYPE(S)          (TOKEN_DATA_TYPE (&(S)->data))
 424 #define SYMBOL_TEXT(S)          (TOKEN_DATA_TEXT (&(S)->data))
 425 #define SYMBOL_FUNC(S)          (TOKEN_DATA_FUNC (&(S)->data))
 426
 427 typedef enum symbol_lookup symbol_lookup;
 428 typedef struct symbol symbol;
 429 typedef void hack_symbol (symbol *, void *);
 430
 431 #define HASHMAX 509             /* default, overridden by -Hsize */
 432
 433 extern symbol **symtab;
 434
 435 void free_symbol (symbol *sym);
 436 void symtab_init (void);
 437 symbol *lookup_symbol (const char *, symbol_lookup);
 438 void hack_all_symbols (hack_symbol *, void *);
 439 \f
 440 /* File: macro.c  --- macro expansion.  */
 441
 442 extern int expansion_level;
 443
 444 void expand_input (void);
 445 void call_macro (symbol *, int, macro_arguments *, struct obstack *);
 446
 447 unsigned int arg_argc (macro_arguments *);
 448 token_data_type arg_type (macro_arguments *, unsigned int);
 449 const char *arg_text (macro_arguments *, unsigned int);
 450 bool arg_equal (macro_arguments *, unsigned int, unsigned int);
 451 bool arg_empty (macro_arguments *, unsigned int);
 452 size_t arg_len (macro_arguments *, unsigned int);
 453 builtin_func *arg_func (macro_arguments *, unsigned int);
 454 struct obstack *arg_scratch (void);
 455 macro_arguments *make_argv_ref (macro_arguments *, const char *, size_t,
 456                                 bool, bool);
 457 void push_arg (struct obstack *, macro_arguments *, unsigned int);
 458 void push_args (struct obstack *, macro_arguments *, bool, bool);
 459 size_t adjust_refcount (int, bool);
 460
 461 \f
 462 /* File: builtin.c  --- builtins.  */
 463
 464 struct builtin
 465 {
 466   const char *name;
 467   bool_bitfield gnu_extension : 1;
 468   bool_bitfield groks_macro_args : 1;
 469   bool_bitfield blind_if_no_args : 1;
 470   builtin_func *func;
 471 };
 472
 473 struct predefined
 474 {
 475   const char *unix_name;
 476   const char *gnu_name;
 477   const char *func;
 478 };
 479
 480 typedef struct builtin builtin;
 481 typedef struct predefined predefined;
 482 struct re_pattern_buffer;
 483 struct re_registers;
 484
 485 /* The default sequence detects multi-digit parameters (obsolete after
 486    1.4.x), and any use of extended arguments with the default ${}
 487    syntax (new in 2.0).  */
 488 #define DEFAULT_MACRO_SEQUENCE "\\$\\({[^}]*}\\|[0-9][0-9]+\\)"
 489
 490 void builtin_init (void);
 491 bool bad_argc (const char *, int, unsigned int, unsigned int);
 492 void define_builtin (const char *, const builtin *, symbol_lookup);
 493 void set_macro_sequence (const char *);
 494 void free_regex (void);
 495 void define_user_macro (const char *, size_t, const char *, symbol_lookup);
 496 void undivert_all (void);
 497 void expand_user_macro (struct obstack *, symbol *, int, macro_arguments *);
 498 void m4_placeholder (struct obstack *, int, macro_arguments *);
 499 void init_pattern_buffer (struct re_pattern_buffer *, struct re_registers *);
 500 const char *ntoa (int32_t, int);
 501
 502 const builtin *find_builtin_by_addr (builtin_func *);
 503 const builtin *find_builtin_by_name (const char *);
 504 \f
 505 /* File: path.c  --- path search for include files.  */
 506
 507 void include_init (void);
 508 void include_env_init (void);
 509 void add_include_directory (const char *);
 510 FILE *m4_path_search (const char *, char **);
 511 \f
 512 /* File: eval.c  --- expression evaluation.  */
 513
 514 bool evaluate (const char *, const char *, int32_t *);
 515 \f
 516 /* File: format.c  --- printf like formatting.  */
 517
 518 void format (struct obstack *, int, macro_arguments *);
 519 \f
 520 /* File: freeze.c --- frozen state files.  */
 521
 522 void produce_frozen_state (const char *);
 523 void reload_frozen_state (const char *);
 524 \f
 525 /* Debugging the memory allocator.  */
 526
 527 #ifdef WITH_DMALLOC
 528 # define DMALLOC_FUNC_CHECK
 529 # include <dmalloc.h>
 530 #endif
 531
 532 /* Other debug stuff.  */
 533
 534 #ifdef DEBUG
 535 # define DEBUG_INCL   1
 536 # define DEBUG_INPUT  1
 537 # define DEBUG_MACRO  1
 538 # define DEBUG_OUTPUT 1
 539 # define DEBUG_REGEX  1
 540 # define DEBUG_STKOVF 1
 541 # define DEBUG_SYM    1
 542 #endif
 543
 544 /* Convert a possibly-signed character to an unsigned character.  This is
 545    a bit safer than casting to unsigned char, since it catches some type
 546    errors that the cast doesn't.  */
 547 #if HAVE_INLINE
 548 static inline unsigned char to_uchar (char ch) { return ch; }
 549 #else
 550 # define to_uchar(C) ((unsigned char) (C))
 551 #endif