usr.bin/unifdef/unifdef.c

   1 /*
   2  * Copyright (c) 1985, 1993
   3  *      The Regents of the University of California.  All rights reserved.
   4  *
   5  * This code is derived from software contributed to Berkeley by
   6  * Dave Yost. Support for #if and #elif was added by Tony Finch.
   7  *
   8  * Redistribution and use in source and binary forms, with or without
   9  * modification, are permitted provided that the following conditions
  10  * are met:
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in the
  15  *    documentation and/or other materials provided with the distribution.
  16  * 3. All advertising materials mentioning features or use of this software
  17  *    must display the following acknowledgement:
  18  *      This product includes software developed by the University of
  19  *      California, Berkeley and its contributors.
  20  * 4. Neither the name of the University nor the names of its contributors
  21  *    may be used to endorse or promote products derived from this software
  22  *    without specific prior written permission.
  23  *
  24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  34  * SUCH DAMAGE.
  35  *
  36  * @(#) Copyright (c) 1985, 1993 The Regents of the University of California.  All rights reserved.
  37  * @(#)unifdef.c        8.1 (Berkeley) 6/6/93
  38  * $NetBSD: unifdef.c,v 1.8 2000/07/03 02:51:36 matt Exp $
  39  * $dotat: things/unifdef.c,v 1.148 2003/01/20 12:05:41 fanf2 Exp $
  40  * $FreeBSD: src/usr.bin/unifdef/unifdef.c,v 1.4.2.5 2003/01/29 21:46:39 fanf Exp $
  41  * $DragonFly: src/usr.bin/unifdef/unifdef.c,v 1.2 2003/06/17 04:29:33 dillon Exp $
  42  */
  43
  44 #include <sys/cdefs.h>
  45
  46 /*
  47  * unifdef - remove ifdef'ed lines
  48  *
  49  *  Wishlist:
  50  *      provide an option which will append the name of the
  51  *        appropriate symbol after #else's and #endif's
  52  *      provide an option which will check symbols after
  53  *        #else's and #endif's to see that they match their
  54  *        corresponding #ifdef or #ifndef
  55  *      generate #line directives in place of deleted code
  56  *
  57  *   The first two items above require better buffer handling, which would
  58  *     also make it possible to handle all "dodgy" directives correctly.
  59  */
  60
  61 #include <ctype.h>
  62 #include <err.h>
  63 #include <stdarg.h>
  64 #include <stdbool.h>
  65 #include <stdio.h>
  66 #include <stdlib.h>
  67 #include <string.h>
  68 #include <unistd.h>
  69
  70 /* types of input lines: */
  71 typedef enum {
  72         LT_TRUEI,               /* a true #if with ignore flag */
  73         LT_FALSEI,              /* a false #if with ignore flag */
  74         LT_IF,                  /* an unknown #if */
  75         LT_TRUE,                /* a true #if */
  76         LT_FALSE,               /* a false #if */
  77         LT_ELIF,                /* an unknown #elif */
  78         LT_ELTRUE,              /* a true #elif */
  79         LT_ELFALSE,             /* a false #elif */
  80         LT_ELSE,                /* #else */
  81         LT_ENDIF,               /* #endif */
  82         LT_DODGY,               /* flag: directive is not on one line */
  83         LT_DODGY_LAST = LT_DODGY + LT_ENDIF,
  84         LT_PLAIN,               /* ordinary line */
  85         LT_EOF,                 /* end of file */
  86         LT_COUNT
  87 } Linetype;
  88
  89 static char const * const linetype_name[] = {
  90         "TRUEI", "FALSEI", "IF", "TRUE", "FALSE",
  91         "ELIF", "ELTRUE", "ELFALSE", "ELSE", "ENDIF",
  92         "DODGY TRUEI", "DODGY FALSEI",
  93         "DODGY IF", "DODGY TRUE", "DODGY FALSE",
  94         "DODGY ELIF", "DODGY ELTRUE", "DODGY ELFALSE",
  95         "DODGY ELSE", "DODGY ENDIF",
  96         "PLAIN", "EOF"
  97 };
  98
  99 /* state of #if processing */
 100 typedef enum {
 101         IS_OUTSIDE,
 102         IS_FALSE_PREFIX,        /* false #if followed by false #elifs */
 103         IS_TRUE_PREFIX,         /* first non-false #(el)if is true */
 104         IS_PASS_MIDDLE,         /* first non-false #(el)if is unknown */
 105         IS_FALSE_MIDDLE,        /* a false #elif after a pass state */
 106         IS_TRUE_MIDDLE,         /* a true #elif after a pass state */
 107         IS_PASS_ELSE,           /* an else after a pass state */
 108         IS_FALSE_ELSE,          /* an else after a true state */
 109         IS_TRUE_ELSE,           /* an else after only false states */
 110         IS_FALSE_TRAILER,       /* #elifs after a true are false */
 111         IS_COUNT
 112 } Ifstate;
 113
 114 static char const * const ifstate_name[] = {
 115         "OUTSIDE", "FALSE_PREFIX", "TRUE_PREFIX",
 116         "PASS_MIDDLE", "FALSE_MIDDLE", "TRUE_MIDDLE",
 117         "PASS_ELSE", "FALSE_ELSE", "TRUE_ELSE",
 118         "FALSE_TRAILER"
 119 };
 120
 121 /* state of comment parser */
 122 typedef enum {
 123         NO_COMMENT = false,     /* outside a comment */
 124         C_COMMENT,              /* in a comment like this one */
 125         CXX_COMMENT,            /* between // and end of line */
 126         STARTING_COMMENT,       /* just after slash-backslash-newline */
 127         FINISHING_COMMENT       /* star-backslash-newline in a C comment */
 128 } Comment_state;
 129
 130 static char const * const comment_name[] = {
 131         "NO", "C", "CXX", "STARTING", "FINISHING"
 132 };
 133
 134 /* state of preprocessor line parser */
 135 typedef enum {
 136         LS_START,               /* only space and comments on this line */
 137         LS_HASH,                /* only space, comments, and a hash */
 138         LS_DIRTY                /* this line can't be a preprocessor line */
 139 } Line_state;
 140
 141 static char const * const linestate_name[] = {
 142         "START", "HASH", "DIRTY"
 143 };
 144
 145 /*
 146  * Minimum translation limits from ISO/IEC 9899:1999 5.2.4.1
 147  */
 148 #define MAXDEPTH        64                      /* maximum #if nesting */
 149 #define MAXLINE         4096                    /* maximum length of line */
 150 #define MAXSYMS         4096                    /* maximum number of symbols */
 151
 152 /*
 153  * Sometimes when editing a keyword the replacement text is longer, so
 154  * we leave some space at the end of the tline buffer to accommodate this.
 155  */
 156 #define EDITSLOP        10
 157
 158 /*
 159  * Globals.
 160  */
 161
 162 static bool             complement;             /* -c: do the complement */
 163 static bool             debugging;              /* -d: debugging reports */
 164 static bool             iocccok;                /* -e: fewer IOCCC errors */
 165 static bool             killconsts;             /* -k: eval constant #ifs */
 166 static bool             lnblank;                /* -l: blank deleted lines */
 167 static bool             symlist;                /* -s: output symbol list */
 168 static bool             text;                   /* -t: this is a text file */
 169
 170 static const char      *symname[MAXSYMS];       /* symbol name */
 171 static const char      *value[MAXSYMS];         /* -Dsym=value */
 172 static bool             ignore[MAXSYMS];        /* -iDsym or -iUsym */
 173 static int              nsyms;                  /* number of symbols */
 174
 175 static FILE            *input;                  /* input file pointer */
 176 static const char      *filename;               /* input file name */
 177 static int              linenum;                /* current line number */
 178
 179 static char             tline[MAXLINE+EDITSLOP];/* input buffer plus space */
 180 static char            *keyword;                /* used for editing #elif's */
 181
 182 static Comment_state    incomment;              /* comment parser state */
 183 static Line_state       linestate;              /* #if line parser state */
 184 static Ifstate          ifstate[MAXDEPTH];      /* #if processor state */
 185 static bool             ignoring[MAXDEPTH];     /* ignore comments state */
 186 static int              stifline[MAXDEPTH];     /* start of current #if */
 187 static int              depth;                  /* current #if nesting */
 188 static bool             keepthis;               /* don't delete constant #if */
 189
 190 static int              exitstat;               /* program exit status */
 191
 192 static void             addsym(bool, bool, char *);
 193 static void             debug(const char *, ...);
 194 static void             error(const char *);
 195 static int              findsym(const char *);
 196 static void             flushline(bool);
 197 static Linetype         getline(void);
 198 static Linetype         ifeval(const char **);
 199 static void             ignoreoff(void);
 200 static void             ignoreon(void);
 201 static void             keywordedit(const char *);
 202 static void             nest(void);
 203 static void             process(void);
 204 static const char      *skipcomment(const char *);
 205 static const char      *skipsym(const char *);
 206 static void             state(Ifstate);
 207 static int              strlcmp(const char *, const char *, size_t);
 208 static void             usage(void);
 209
 210 #define endsym(c) (!isalpha((unsigned char)c) && !isdigit((unsigned char)c) && c != '_')
 211
 212 /*
 213  * The main program.
 214  */
 215 int
 216 main(int argc, char *argv[])
 217 {
 218         int opt;
 219
 220         while ((opt = getopt(argc, argv, "i:D:U:I:cdeklst")) != -1)
 221                 switch (opt) {
 222                 case 'i': /* treat stuff controlled by these symbols as text */
 223                         /*
 224                          * For strict backwards-compatibility the U or D
 225                          * should be immediately after the -i but it doesn't
 226                          * matter much if we relax that requirement.
 227                          */
 228                         opt = *optarg++;
 229                         if (opt == 'D')
 230                                 addsym(true, true, optarg);
 231                         else if (opt == 'U')
 232                                 addsym(true, false, optarg);
 233                         else
 234                                 usage();
 235                         break;
 236                 case 'D': /* define a symbol */
 237                         addsym(false, true, optarg);
 238                         break;
 239                 case 'U': /* undef a symbol */
 240                         addsym(false, false, optarg);
 241                         break;
 242                 case 'I':
 243                         /* no-op for compatibility with cpp */
 244                         break;
 245                 case 'c': /* treat -D as -U and vice versa */
 246                         complement = true;
 247                         break;
 248                 case 'd':
 249                         debugging = true;
 250                         break;
 251                 case 'e': /* fewer errors from dodgy lines */
 252                         iocccok = true;
 253                         break;
 254                 case 'k': /* process constant #ifs */
 255                         killconsts = true;
 256                         break;
 257                 case 'l': /* blank deleted lines instead of omitting them */
 258                         lnblank = true;
 259                         break;
 260                 case 's': /* only output list of symbols that control #ifs */
 261                         symlist = true;
 262                         break;
 263                 case 't': /* don't parse C comments */
 264                         text = true;
 265                         break;
 266                 default:
 267                         usage();
 268                 }
 269         argc -= optind;
 270         argv += optind;
 271         if (nsyms == 0 && !symlist) {
 272                 warnx("must -D or -U at least one symbol");
 273                 usage();
 274         }
 275         if (argc > 1) {
 276                 errx(2, "can only do one file");
 277         } else if (argc == 1 && strcmp(*argv, "-") != 0) {
 278                 filename = *argv;
 279                 if ((input = fopen(filename, "r")) != NULL) {
 280                         process();
 281                         (void) fclose(input);
 282                 } else
 283                         err(2, "can't open %s", *argv);
 284         } else {
 285                 filename = "[stdin]";
 286                 input = stdin;
 287                 process();
 288         }
 289
 290         exit(exitstat);
 291 }
 292
 293 static void
 294 usage(void)
 295 {
 296         fprintf(stderr, "usage: unifdef [-cdeklst]"
 297             " [[-Dsym[=val]] [-Usym] [-iDsym[=val]] [-iUsym]] ... [file]\n");
 298         exit(2);
 299 }
 300
 301 /*
 302  * A state transition function alters the global #if processing state
 303  * in a particular way. The table below is indexed by the current
 304  * processing state and the type of the current line. A NULL entry
 305  * indicate that processing is complete.
 306  *
 307  * Nesting is handled by keeping a stack of states; some transition
 308  * functions increase or decrease the depth. They also maintain the
 309  * ignore state on a stack. In some complicated cases they have to
 310  * alter the preprocessor directive, as follows.
 311  *
 312  * When we have processed a group that starts off with a known-false
 313  * #if/#elif sequence (which has therefore been deleted) followed by a
 314  * #elif that we don't understand and therefore must keep, we edit the
 315  * latter into a #if to keep the nesting correct.
 316  *
 317  * When we find a true #elif in a group, the following block will
 318  * always be kept and the rest of the sequence after the next #elif or
 319  * #else will be discarded. We edit the #elif into a #else and the
 320  * following directive to #endif since this has the desired behaviour.
 321  *
 322  * "Dodgy" directives are split across multiple lines, the most common
 323  * example being a multi-line comment hanging off the right of the
 324  * directive. We can handle them correctly only if there is no change
 325  * from printing to dropping (or vice versa) caused by that directive.
 326  * If the directive is the first of a group we have a choice between
 327  * failing with an error, or passing it through unchanged instead of
 328  * evaluating it. The latter is not the default to avoid questions from
 329  * users about unifdef unexpectedly leaving behind preprocessor directives.
 330  */
 331 typedef void state_fn(void);
 332
 333 /* report an error */
 334 static void Eelif (void) { error("Inappropriate #elif"); }
 335 static void Eelse (void) { error("Inappropriate #else"); }
 336 static void Eendif(void) { error("Inappropriate #endif"); }
 337 static void Eeof  (void) { error("Premature EOF"); }
 338 static void Eioccc(void) { error("Obfuscated preprocessor control line"); }
 339 /* plain line handling */
 340 static void print (void) { flushline(true); }
 341 static void drop  (void) { flushline(false); }
 342 /* output lacks group's start line */
 343 static void Strue (void) { drop();  ignoreoff(); state(IS_TRUE_PREFIX); }
 344 static void Sfalse(void) { drop();  ignoreoff(); state(IS_FALSE_PREFIX); }
 345 static void Selse (void) { drop();               state(IS_TRUE_ELSE); }
 346 /* print/pass this block */
 347 static void Pelif (void) { print(); ignoreoff(); state(IS_PASS_MIDDLE); }
 348 static void Pelse (void) { print();              state(IS_PASS_ELSE); }
 349 static void Pendif(void) { print(); --depth; }
 350 /* discard this block */
 351 static void Dfalse(void) { drop();  ignoreoff(); state(IS_FALSE_TRAILER); }
 352 static void Delif (void) { drop();  ignoreoff(); state(IS_FALSE_MIDDLE); }
 353 static void Delse (void) { drop();               state(IS_FALSE_ELSE); }
 354 static void Dendif(void) { drop();  --depth; }
 355 /* first line of group */
 356 static void Fdrop (void) { nest();  Dfalse(); }
 357 static void Fpass (void) { nest();  Pelif(); }
 358 static void Ftrue (void) { nest();  Strue(); }
 359 static void Ffalse(void) { nest();  Sfalse(); }
 360 /* variable pedantry for obfuscated lines */
 361 static void Oiffy (void) { if (iocccok) Fpass(); else Eioccc(); ignoreon(); }
 362 static void Oif   (void) { if (iocccok) Fpass(); else Eioccc(); }
 363 static void Oelif (void) { if (iocccok) Pelif(); else Eioccc(); }
 364 /* ignore comments in this block */
 365 static void Idrop (void) { Fdrop();  ignoreon(); }
 366 static void Itrue (void) { Ftrue();  ignoreon(); }
 367 static void Ifalse(void) { Ffalse(); ignoreon(); }
 368 /* edit this line */
 369 static void Mpass (void) { strncpy(keyword, "if  ", 4); Pelif(); }
 370 static void Mtrue (void) { keywordedit("else\n");  state(IS_TRUE_MIDDLE); }
 371 static void Melif (void) { keywordedit("endif\n"); state(IS_FALSE_TRAILER); }
 372 static void Melse (void) { keywordedit("endif\n"); state(IS_FALSE_ELSE); }
 373
 374 static state_fn * const trans_table[IS_COUNT][LT_COUNT] = {
 375 /* IS_OUTSIDE */
 376 { Itrue, Ifalse,Fpass, Ftrue, Ffalse,Eelif, Eelif, Eelif, Eelse, Eendif,
 377   Oiffy, Oiffy, Fpass, Oif,   Oif,   Eelif, Eelif, Eelif, Eelse, Eendif,
 378   print, NULL },
 379 /* IS_FALSE_PREFIX */
 380 { Idrop, Idrop, Fdrop, Fdrop, Fdrop, Mpass, Strue, Sfalse,Selse, Dendif,
 381   Idrop, Idrop, Fdrop, Fdrop, Fdrop, Mpass, Eioccc,Eioccc,Eioccc,Eioccc,
 382   drop,  Eeof },
 383 /* IS_TRUE_PREFIX */
 384 { Itrue, Ifalse,Fpass, Ftrue, Ffalse,Dfalse,Dfalse,Dfalse,Delse, Dendif,
 385   Oiffy, Oiffy, Fpass, Oif,   Oif,   Eioccc,Eioccc,Eioccc,Eioccc,Eioccc,
 386   print, Eeof },
 387 /* IS_PASS_MIDDLE */
 388 { Itrue, Ifalse,Fpass, Ftrue, Ffalse,Pelif, Mtrue, Delif, Pelse, Pendif,
 389   Oiffy, Oiffy, Fpass, Oif,   Oif,   Pelif, Oelif, Oelif, Pelse, Pendif,
 390   print, Eeof },
 391 /* IS_FALSE_MIDDLE */
 392 { Idrop, Idrop, Fdrop, Fdrop, Fdrop, Pelif, Mtrue, Delif, Pelse, Pendif,
 393   Idrop, Idrop, Fdrop, Fdrop, Fdrop, Eioccc,Eioccc,Eioccc,Eioccc,Eioccc,
 394   drop,  Eeof },
 395 /* IS_TRUE_MIDDLE */
 396 { Itrue, Ifalse,Fpass, Ftrue, Ffalse,Melif, Melif, Melif, Melse, Pendif,
 397   Oiffy, Oiffy, Fpass, Oif,   Oif,   Eioccc,Eioccc,Eioccc,Eioccc,Pendif,
 398   print, Eeof },
 399 /* IS_PASS_ELSE */
 400 { Itrue, Ifalse,Fpass, Ftrue, Ffalse,Eelif, Eelif, Eelif, Eelse, Pendif,
 401   Oiffy, Oiffy, Fpass, Oif,   Oif,   Eelif, Eelif, Eelif, Eelse, Pendif,
 402   print, Eeof },
 403 /* IS_FALSE_ELSE */
 404 { Idrop, Idrop, Fdrop, Fdrop, Fdrop, Eelif, Eelif, Eelif, Eelse, Dendif,
 405   Idrop, Idrop, Fdrop, Fdrop, Fdrop, Eelif, Eelif, Eelif, Eelse, Eioccc,
 406   drop,  Eeof },
 407 /* IS_TRUE_ELSE */
 408 { Itrue, Ifalse,Fpass, Ftrue, Ffalse,Eelif, Eelif, Eelif, Eelse, Dendif,
 409   Oiffy, Oiffy, Fpass, Oif,   Oif,   Eelif, Eelif, Eelif, Eelse, Eioccc,
 410   print, Eeof },
 411 /* IS_FALSE_TRAILER */
 412 { Idrop, Idrop, Fdrop, Fdrop, Fdrop, Dfalse,Dfalse,Dfalse,Delse, Dendif,
 413   Idrop, Idrop, Fdrop, Fdrop, Fdrop, Dfalse,Dfalse,Dfalse,Delse, Eioccc,
 414   drop,  Eeof }
 415 /*TRUEI  FALSEI IF     TRUE   FALSE  ELIF   ELTRUE ELFALSE ELSE  ENDIF
 416   TRUEI  FALSEI IF     TRUE   FALSE  ELIF   ELTRUE ELFALSE ELSE  ENDIF (DODGY)
 417   PLAIN  EOF */
 418 };
 419
 420 /*
 421  * State machine utility functions
 422  */
 423 static void
 424 ignoreoff(void)
 425 {
 426         ignoring[depth] = ignoring[depth-1];
 427 }
 428 static void
 429 ignoreon(void)
 430 {
 431         ignoring[depth] = true;
 432 }
 433 static void
 434 keywordedit(const char *replacement)
 435 {
 436         strlcpy(keyword, replacement, tline + sizeof(tline) - keyword);
 437         print();
 438 }
 439 static void
 440 nest(void)
 441 {
 442         depth += 1;
 443         if (depth >= MAXDEPTH)
 444                 error("Too many levels of nesting");
 445         stifline[depth] = linenum;
 446 }
 447 static void
 448 state(Ifstate is)
 449 {
 450         ifstate[depth] = is;
 451 }
 452
 453 /*
 454  * Write a line to the output or not, according to command line options.
 455  */
 456 static void
 457 flushline(bool keep)
 458 {
 459         if (symlist)
 460                 return;
 461         if (keep ^ complement)
 462                 fputs(tline, stdout);
 463         else {
 464                 if (lnblank)
 465                         putc('\n', stdout);
 466                 exitstat = 1;
 467         }
 468 }
 469
 470 /*
 471  * The driver for the state machine.
 472  */
 473 static void
 474 process(void)
 475 {
 476         Linetype lineval;
 477         state_fn *trans;
 478
 479         for (;;) {
 480                 linenum++;
 481                 lineval = getline();
 482                 trans = trans_table[ifstate[depth]][lineval];
 483                 if (trans == NULL)
 484                         break;
 485                 trans();
 486                 debug("process %s -> %s depth %d",
 487                     linetype_name[lineval],
 488                     ifstate_name[ifstate[depth]], depth);
 489         }
 490         if (incomment)
 491                 error("EOF in comment");
 492 }
 493
 494 /*
 495  * Parse a line and determine its type. We keep the preprocessor line
 496  * parser state between calls in a global variable.
 497  */
 498 static Linetype
 499 getline(void)
 500 {
 501         const char *cp;
 502         int cursym;
 503         int kwlen;
 504         Linetype retval;
 505         Comment_state wascomment;
 506
 507         if (fgets(tline, MAXLINE, input) == NULL)
 508                 return (LT_EOF);
 509         retval = LT_PLAIN;
 510         wascomment = incomment;
 511         cp = skipcomment(tline);
 512         if (linestate == LS_START) {
 513                 if (*cp == '#') {
 514                         linestate = LS_HASH;
 515                         cp = skipcomment(cp + 1);
 516                 } else if (*cp != '\0')
 517                         linestate = LS_DIRTY;
 518         }
 519         if (!incomment && linestate == LS_HASH) {
 520                 keyword = tline + (cp - tline);
 521                 cp = skipsym(cp);
 522                 kwlen = cp - keyword;
 523                 /* no way can we deal with a continuation inside a keyword */
 524                 if (strncmp(cp, "\\\n", 2) == 0)
 525                         Eioccc();
 526                 if (strlcmp("ifdef", keyword, kwlen) == 0 ||
 527                     strlcmp("ifndef", keyword, kwlen) == 0) {
 528                         cp = skipcomment(cp);
 529                         if ((cursym = findsym(cp)) < 0)
 530                                 retval = LT_IF;
 531                         else {
 532                                 retval = (keyword[2] == 'n')
 533                                     ? LT_FALSE : LT_TRUE;
 534                                 if (value[cursym] == NULL)
 535                                         retval = (retval == LT_TRUE)
 536                                             ? LT_FALSE : LT_TRUE;
 537                                 if (ignore[cursym])
 538                                         retval = (retval == LT_TRUE)
 539                                             ? LT_TRUEI : LT_FALSEI;
 540                         }
 541                         cp = skipsym(cp);
 542                 } else if (strlcmp("if", keyword, kwlen) == 0)
 543                         retval = ifeval(&cp);
 544                 else if (strlcmp("elif", keyword, kwlen) == 0)
 545                         retval = ifeval(&cp) - LT_IF + LT_ELIF;
 546                 else if (strlcmp("else", keyword, kwlen) == 0)
 547                         retval = LT_ELSE;
 548                 else if (strlcmp("endif", keyword, kwlen) == 0)
 549                         retval = LT_ENDIF;
 550                 else {
 551                         linestate = LS_DIRTY;
 552                         retval = LT_PLAIN;
 553                 }
 554                 cp = skipcomment(cp);
 555                 if (*cp != '\0') {
 556                         linestate = LS_DIRTY;
 557                         if (retval == LT_TRUE || retval == LT_FALSE ||
 558                             retval == LT_TRUEI || retval == LT_FALSEI)
 559                                 retval = LT_IF;
 560                         if (retval == LT_ELTRUE || retval == LT_ELFALSE)
 561                                 retval = LT_ELIF;
 562                 }
 563                 if (retval != LT_PLAIN && (wascomment || incomment)) {
 564                         retval += LT_DODGY;
 565                         if (incomment)
 566                                 linestate = LS_DIRTY;
 567                 }
 568                 /* skipcomment should have changed the state */
 569                 if (linestate == LS_HASH)
 570                         abort(); /* bug */
 571         }
 572         if (linestate == LS_DIRTY) {
 573                 while (*cp != '\0')
 574                         cp = skipcomment(cp + 1);
 575         }
 576         debug("parser %s comment %s line",
 577             comment_name[incomment], linestate_name[linestate]);
 578         return (retval);
 579 }
 580
 581 /*
 582  * These are the operators that are supported by the expression evaluator.
 583  */
 584 static int op_lt(int a, int b) { return (a < b); }
 585 static int op_gt(int a, int b) { return (a > b); }
 586 static int op_le(int a, int b) { return (a <= b); }
 587 static int op_ge(int a, int b) { return (a >= b); }
 588 static int op_eq(int a, int b) { return (a == b); }
 589 static int op_ne(int a, int b) { return (a != b); }
 590 static int op_or(int a, int b) { return (a || b); }
 591 static int op_and(int a, int b) { return (a && b); }
 592
 593 /*
 594  * An evaluation function takes three arguments, as follows: (1) a pointer to
 595  * an element of the precedence table which lists the operators at the current
 596  * level of precedence; (2) a pointer to an integer which will receive the
 597  * value of the expression; and (3) a pointer to a char* that points to the
 598  * expression to be evaluated and that is updated to the end of the expression
 599  * when evaluation is complete. The function returns LT_FALSE if the value of
 600  * the expression is zero, LT_TRUE if it is non-zero, or LT_IF if the
 601  * expression could not be evaluated.
 602  */
 603 struct ops;
 604
 605 typedef Linetype eval_fn(const struct ops *, int *, const char **);
 606
 607 static eval_fn eval_table, eval_unary;
 608
 609 /*
 610  * The precedence table. Expressions involving binary operators are evaluated
 611  * in a table-driven way by eval_table. When it evaluates a subexpression it
 612  * calls the inner function with its first argument pointing to the next
 613  * element of the table. Innermost expressions have special non-table-driven
 614  * handling.
 615  */
 616 static const struct ops {
 617         eval_fn *inner;
 618         struct op {
 619                 const char *str;
 620                 int (*fn)(int, int);
 621         } op[5];
 622 } eval_ops[] = {
 623         { eval_table, { { "||", op_or } } },
 624         { eval_table, { { "&&", op_and } } },
 625         { eval_table, { { "==", op_eq },
 626                         { "!=", op_ne } } },
 627         { eval_unary, { { "<=", op_le },
 628                         { ">=", op_ge },
 629                         { "<", op_lt },
 630                         { ">", op_gt } } }
 631 };
 632
 633 /*
 634  * Function for evaluating the innermost parts of expressions,
 635  * viz. !expr (expr) defined(symbol) symbol number
 636  * We reset the keepthis flag when we find a non-constant subexpression.
 637  */
 638 static Linetype
 639 eval_unary(const struct ops *ops, int *valp, const char **cpp)
 640 {
 641         const char *cp;
 642         char *ep;
 643         int sym;
 644
 645         cp = skipcomment(*cpp);
 646         if (*cp == '!') {
 647                 debug("eval%d !", ops - eval_ops);
 648                 cp++;
 649                 if (eval_unary(ops, valp, &cp) == LT_IF)
 650                         return (LT_IF);
 651                 *valp = !*valp;
 652         } else if (*cp == '(') {
 653                 cp++;
 654                 debug("eval%d (", ops - eval_ops);
 655                 if (eval_table(eval_ops, valp, &cp) == LT_IF)
 656                         return (LT_IF);
 657                 cp = skipcomment(cp);
 658                 if (*cp++ != ')')
 659                         return (LT_IF);
 660         } else if (isdigit((unsigned char)*cp)) {
 661                 debug("eval%d number", ops - eval_ops);
 662                 *valp = strtol(cp, &ep, 0);
 663                 cp = skipsym(cp);
 664         } else if (strncmp(cp, "defined", 7) == 0 && endsym(cp[7])) {
 665                 cp = skipcomment(cp+7);
 666                 debug("eval%d defined", ops - eval_ops);
 667                 if (*cp++ != '(')
 668                         return (LT_IF);
 669                 cp = skipcomment(cp);
 670                 sym = findsym(cp);
 671                 if (sym < 0 && !symlist)
 672                         return (LT_IF);
 673                 *valp = (value[sym] != NULL);
 674                 cp = skipsym(cp);
 675                 cp = skipcomment(cp);
 676                 if (*cp++ != ')')
 677                         return (LT_IF);
 678                 keepthis = false;
 679         } else if (!endsym(*cp)) {
 680                 debug("eval%d symbol", ops - eval_ops);
 681                 sym = findsym(cp);
 682                 if (sym < 0 && !symlist)
 683                         return (LT_IF);
 684                 if (value[sym] == NULL)
 685                         *valp = 0;
 686                 else {
 687                         *valp = strtol(value[sym], &ep, 0);
 688                         if (*ep != '\0' || ep == value[sym])
 689                                 return (LT_IF);
 690                 }
 691                 cp = skipsym(cp);
 692                 keepthis = false;
 693         } else
 694                 return (LT_IF);
 695
 696         *cpp = cp;
 697         debug("eval%d = %d", ops - eval_ops, *valp);
 698         return (*valp ? LT_TRUE : LT_FALSE);
 699 }
 700
 701 /*
 702  * Table-driven evaluation of binary operators.
 703  */
 704 static Linetype
 705 eval_table(const struct ops *ops, int *valp, const char **cpp)
 706 {
 707         const struct op *op;
 708         const char *cp;
 709         int val;
 710
 711         debug("eval%d", ops - eval_ops);
 712         cp = *cpp;
 713         if (ops->inner(ops+1, valp, &cp) == LT_IF)
 714                 return (LT_IF);
 715         for (;;) {
 716                 cp = skipcomment(cp);
 717                 for (op = ops->op; op->str != NULL; op++)
 718                         if (strncmp(cp, op->str, strlen(op->str)) == 0)
 719                                 break;
 720                 if (op->str == NULL)
 721                         break;
 722                 cp += strlen(op->str);
 723                 debug("eval%d %s", ops - eval_ops, op->str);
 724                 if (ops->inner(ops+1, &val, &cp) == LT_IF)
 725                         return (LT_IF);
 726                 *valp = op->fn(*valp, val);
 727         }
 728
 729         *cpp = cp;
 730         debug("eval%d = %d", ops - eval_ops, *valp);
 731         return (*valp ? LT_TRUE : LT_FALSE);
 732 }
 733
 734 /*
 735  * Evaluate the expression on a #if or #elif line. If we can work out
 736  * the result we return LT_TRUE or LT_FALSE accordingly, otherwise we
 737  * return just a generic LT_IF.
 738  */
 739 static Linetype
 740 ifeval(const char **cpp)
 741 {
 742         int ret;
 743         int val;
 744
 745         debug("eval %s", *cpp);
 746         keepthis = killconsts ? false : true;
 747         ret = eval_table(eval_ops, &val, cpp);
 748         return (keepthis ? LT_IF : ret);
 749 }
 750
 751 /*
 752  * Skip over comments and stop at the next character position that is
 753  * not whitespace. Between calls we keep the comment state in a global
 754  * variable, and we also make a note when we get a proper end-of-line.
 755  * XXX: doesn't cope with the buffer splitting inside a state transition.
 756  */
 757 static const char *
 758 skipcomment(const char *cp)
 759 {
 760         if (text || ignoring[depth]) {
 761                 while (isspace((unsigned char)*cp))
 762                         cp += 1;
 763                 return (cp);
 764         }
 765         while (*cp != '\0')
 766                 if (strncmp(cp, "\\\n", 2) == 0)
 767                         cp += 2;
 768                 else switch (incomment) {
 769                 case NO_COMMENT:
 770                         if (strncmp(cp, "/\\\n", 3) == 0) {
 771                                 incomment = STARTING_COMMENT;
 772                                 cp += 3;
 773                         } else if (strncmp(cp, "/*", 2) == 0) {
 774                                 incomment = C_COMMENT;
 775                                 cp += 2;
 776                         } else if (strncmp(cp, "//", 2) == 0) {
 777                                 incomment = CXX_COMMENT;
 778                                 cp += 2;
 779                         } else if (strncmp(cp, "\n", 1) == 0) {
 780                                 linestate = LS_START;
 781                                 cp += 1;
 782                         } else if (strchr(" \t", *cp) != NULL) {
 783                                 cp += 1;
 784                         } else
 785                                 return (cp);
 786                         continue;
 787                 case CXX_COMMENT:
 788                         if (strncmp(cp, "\n", 1) == 0) {
 789                                 incomment = NO_COMMENT;
 790                                 linestate = LS_START;
 791                         }
 792                         cp += 1;
 793                         continue;
 794                 case C_COMMENT:
 795                         if (strncmp(cp, "*\\\n", 3) == 0) {
 796                                 incomment = FINISHING_COMMENT;
 797                                 cp += 3;
 798                         } else if (strncmp(cp, "*/", 2) == 0) {
 799                                 incomment = NO_COMMENT;
 800                                 cp += 2;
 801                         } else
 802                                 cp += 1;
 803                         continue;
 804                 case STARTING_COMMENT:
 805                         if (*cp == '*') {
 806                                 incomment = C_COMMENT;
 807                                 cp += 1;
 808                         } else if (*cp == '/') {
 809                                 incomment = CXX_COMMENT;
 810                                 cp += 1;
 811                         } else {
 812                                 incomment = NO_COMMENT;
 813                                 linestate = LS_DIRTY;
 814                         }
 815                         continue;
 816                 case FINISHING_COMMENT:
 817                         if (*cp == '/') {
 818                                 incomment = NO_COMMENT;
 819                                 cp += 1;
 820                         } else
 821                                 incomment = C_COMMENT;
 822                         continue;
 823                 default:
 824                         /* bug */
 825                         abort();
 826                 }
 827         return (cp);
 828 }
 829
 830 /*
 831  * Skip over an identifier.
 832  */
 833 static const char *
 834 skipsym(const char *cp)
 835 {
 836         while (!endsym(*cp))
 837                 ++cp;
 838         return (cp);
 839 }
 840
 841 /*
 842  * Look for the symbol in the symbol table. If is is found, we return
 843  * the symbol table index, else we return -1.
 844  */
 845 static int
 846 findsym(const char *str)
 847 {
 848         const char *cp;
 849         int symind;
 850
 851         cp = skipsym(str);
 852         if (cp == str)
 853                 return (-1);
 854         if (symlist)
 855                 printf("%.*s\n", (int)(cp-str), str);
 856         for (symind = 0; symind < nsyms; ++symind) {
 857                 if (strlcmp(symname[symind], str, cp-str) == 0) {
 858                         debug("findsym %s %s", symname[symind],
 859                             value[symind] ? value[symind] : "");
 860                         return (symind);
 861                 }
 862         }
 863         return (-1);
 864 }
 865
 866 /*
 867  * Add a symbol to the symbol table.
 868  */
 869 static void
 870 addsym(bool ignorethis, bool definethis, char *sym)
 871 {
 872         int symind;
 873         char *val;
 874
 875         symind = findsym(sym);
 876         if (symind < 0) {
 877                 if (nsyms >= MAXSYMS)
 878                         errx(2, "too many symbols");
 879                 symind = nsyms++;
 880         }
 881         symname[symind] = sym;
 882         ignore[symind] = ignorethis;
 883         val = sym + (skipsym(sym) - sym);
 884         if (definethis) {
 885                 if (*val == '=') {
 886                         value[symind] = val+1;
 887                         *val = '\0';
 888                 } else if (*val == '\0')
 889                         value[symind] = "";
 890                 else
 891                         usage();
 892         } else {
 893                 if (*val != '\0')
 894                         usage();
 895                 value[symind] = NULL;
 896         }
 897 }
 898
 899 /*
 900  * Compare s with n characters of t.
 901  * The same as strncmp() except that it checks that s[n] == '\0'.
 902  */
 903 static int
 904 strlcmp(const char *s, const char *t, size_t n)
 905 {
 906         while (n-- && *t != '\0')
 907                 if (*s != *t)
 908                         return ((unsigned char)*s - (unsigned char)*t);
 909                 else
 910                         ++s, ++t;
 911         return ((unsigned char)*s);
 912 }
 913
 914 /*
 915  * Diagnostics.
 916  */
 917 static void
 918 debug(const char *msg, ...)
 919 {
 920         va_list ap;
 921
 922         if (debugging) {
 923                 va_start(ap, msg);
 924                 vwarnx(msg, ap);
 925                 va_end(ap);
 926         }
 927 }
 928
 929 static void
 930 error(const char *msg)
 931 {
 932         if (depth == 0)
 933                 warnx("%s: %d: %s", filename, linenum, msg);
 934         else
 935                 warnx("%s: %d: %s (#if line %d depth %d)",
 936                     filename, linenum, msg, stifline[depth], depth);
 937         errx(2, "output may be truncated");
 938 }