usr.bin/m4/main.c

   1 /*      $OpenBSD: main.c,v 1.53 2002/04/26 16:15:16 espie Exp $ */
   2 /*      $NetBSD: main.c,v 1.12 1997/02/08 23:54:49 cgd Exp $    */
   3
   4 /*-
   5  * Copyright (c) 1989, 1993
   6  *      The Regents of the University of California.  All rights reserved.
   7  *
   8  * This code is derived from software contributed to Berkeley by
   9  * Ozan Yigit at York University.
  10  *
  11  * Redistribution and use in source and binary forms, with or without
  12  * modification, are permitted provided that the following conditions
  13  * are met:
  14  * 1. Redistributions of source code must retain the above copyright
  15  *    notice, this list of conditions and the following disclaimer.
  16  * 2. Redistributions in binary form must reproduce the above copyright
  17  *    notice, this list of conditions and the following disclaimer in the
  18  *    documentation and/or other materials provided with the distribution.
  19  * 3. All advertising materials mentioning features or use of this software
  20  *    must display the following acknowledgement:
  21  *      This product includes software developed by the University of
  22  *      California, Berkeley and its contributors.
  23  * 4. Neither the name of the University nor the names of its contributors
  24  *    may be used to endorse or promote products derived from this software
  25  *    without specific prior written permission.
  26  *
  27  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  28  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  29  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  30  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  31  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  32  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  33  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  34  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  35  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  36  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  37  * SUCH DAMAGE.
  38  *
  39  * @(#) Copyright (c) 1989, 1993 The Regents of the University of California.  All rights reserved.
  40  * @(#)main.c   8.1 (Berkeley) 6/6/93
  41  * $OpenBSD: main.c,v 1.53 2002/04/26 16:15:16 espie Exp $
  42  * $FreeBSD: src/usr.bin/m4/main.c,v 1.26 2004/08/16 14:18:22 tjr Exp $
  43  * $DragonFly: src/usr.bin/m4/main.c,v 1.4 2007/01/01 00:41:58 pavalos Exp $
  44  */
  45
  46 /*
  47  * main.c
  48  * Facility: m4 macro processor
  49  * by: oz
  50  */
  51
  52 #include <sys/types.h>
  53 #include <assert.h>
  54 #include <signal.h>
  55 #include <errno.h>
  56 #include <unistd.h>
  57 #include <stdio.h>
  58 #include <ctype.h>
  59 #include <string.h>
  60 #include <stddef.h>
  61 #include <stdlib.h>
  62 #include <err.h>
  63 #include <locale.h>
  64 #include "mdef.h"
  65 #include "stdd.h"
  66 #include "extern.h"
  67 #include "pathnames.h"
  68
  69 ndptr hashtab[HASHSIZE];        /* hash table for macros etc.  */
  70 stae *mstack;                   /* stack of m4 machine         */
  71 char *sstack;                   /* shadow stack, for string space extension */
  72 static size_t STACKMAX;         /* current maximum size of stack */
  73 int sp;                         /* current m4  stack pointer   */
  74 int fp;                         /* m4 call frame pointer       */
  75 struct input_file infile[MAXINP];/* input file stack (0=stdin)  */
  76 char *inname[MAXINP];           /* names of these input files */
  77 int inlineno[MAXINP];           /* current number in each input file */
  78 FILE **outfile;                 /* diversion array(0=bitbucket)*/
  79 int maxout;
  80 FILE *active;                   /* active output file pointer  */
  81 int ilevel = 0;                 /* input file stack pointer    */
  82 int oindex = 0;                 /* diversion index..           */
  83 char null[] = "";               /* as it says.. just a null..  */
  84 const char *m4wraps = "";       /* m4wrap string default..     */
  85 int m4prefix = 0;               /* prefix keywords with m4_    */
  86 char lquote[MAXCCHARS+1] = {LQUOTE};    /* left quote character  (`)   */
  87 char rquote[MAXCCHARS+1] = {RQUOTE};    /* right quote character (')   */
  88 char scommt[MAXCCHARS+1] = {SCOMMT};    /* start character for comment */
  89 char ecommt[MAXCCHARS+1] = {ECOMMT};    /* end character for comment   */
  90 int synccpp;                    /* Line synchronisation for C preprocessor */
  91
  92 struct keyblk keywrds[] = {     /* m4 keywords to be installed */
  93         { "include",      INCLTYPE },
  94         { "sinclude",     SINCTYPE },
  95         { "define",       DEFITYPE },
  96         { "defn",         DEFNTYPE },
  97         { "divert",       DIVRTYPE | NOARGS },
  98         { "expr",         EXPRTYPE },
  99         { "eval",         EXPRTYPE },
 100         { "substr",       SUBSTYPE },
 101         { "ifelse",       IFELTYPE },
 102         { "ifdef",        IFDFTYPE },
 103         { "len",          LENGTYPE },
 104         { "incr",         INCRTYPE },
 105         { "decr",         DECRTYPE },
 106         { "dnl",          DNLNTYPE | NOARGS },
 107         { "changequote",  CHNQTYPE | NOARGS },
 108         { "changecom",    CHNCTYPE | NOARGS },
 109         { "index",        INDXTYPE },
 110 #ifdef EXTENDED
 111         { "paste",        PASTTYPE },
 112         { "spaste",       SPASTYPE },
 113         /* Newer extensions, needed to handle gnu-m4 scripts */
 114         { "indir",        INDIRTYPE},
 115         { "builtin",      BUILTINTYPE},
 116         { "patsubst",     PATSTYPE},
 117         { "regexp",       REGEXPTYPE},
 118         { "esyscmd",      ESYSCMDTYPE},
 119         { "__file__",     FILENAMETYPE | NOARGS},
 120         { "__line__",     LINETYPE | NOARGS},
 121 #endif
 122         { "popdef",       POPDTYPE },
 123         { "pushdef",      PUSDTYPE },
 124         { "dumpdef",      DUMPTYPE | NOARGS },
 125         { "shift",        SHIFTYPE | NOARGS },
 126         { "translit",     TRNLTYPE },
 127         { "undefine",     UNDFTYPE },
 128         { "undivert",     UNDVTYPE | NOARGS },
 129         { "divnum",       DIVNTYPE | NOARGS },
 130         { "maketemp",     MKTMTYPE },
 131         { "errprint",     ERRPTYPE | NOARGS },
 132         { "m4wrap",       M4WRTYPE | NOARGS },
 133         { "m4exit",       EXITTYPE | NOARGS },
 134         { "syscmd",       SYSCTYPE },
 135         { "sysval",       SYSVTYPE | NOARGS },
 136         { "traceon",      TRACEONTYPE | NOARGS },
 137         { "traceoff",     TRACEOFFTYPE | NOARGS },
 138
 139 #if defined(unix) || defined(__unix__)
 140         { "unix",         SELFTYPE | NOARGS },
 141 #else
 142 #ifdef vms
 143         { "vms",          SELFTYPE | NOARGS },
 144 #endif
 145 #endif
 146 };
 147
 148 #define MAXKEYS (sizeof(keywrds)/sizeof(struct keyblk))
 149
 150 #define MAXRECORD 50
 151 static struct position {
 152         char *name;
 153         unsigned long line;
 154 } quotes[MAXRECORD], paren[MAXRECORD];
 155
 156 static void record(struct position *, int);
 157 static void dump_stack(struct position *, int);
 158
 159 static void macro(void);
 160 static void initkwds(void);
 161 static ndptr inspect(int, char *);
 162 static int do_look_ahead(int, const char *);
 163
 164 static void enlarge_stack(void);
 165
 166 int
 167 main(int argc, char *argv[])
 168 {
 169         int c;
 170         int n;
 171         int rval;
 172         const char *optstr = "D:I:PU:gd:o:st:";
 173         char *p;
 174
 175         setlocale(LC_ALL, "");
 176
 177         traceout = stderr;
 178
 179         if (signal(SIGINT, SIG_IGN) != SIG_IGN)
 180                 signal(SIGINT, onintr);
 181
 182         /*
 183          * We need to know if -P is there before checking -D and -U.
 184          */
 185         while ((c = getopt(argc, argv, optstr)) != -1)
 186                 if (c == 'P')
 187                         m4prefix = 1;
 188         optind = optreset = 1;
 189
 190         initkwds();
 191         initspaces();
 192         STACKMAX = INITSTACKMAX;
 193
 194         mstack = (stae *)xalloc(sizeof(stae) * STACKMAX);
 195         sstack = (char *)xalloc(STACKMAX);
 196
 197         maxout = 0;
 198         outfile = NULL;
 199         resizedivs(MAXOUT);
 200
 201         while ((c = getopt(argc, argv, optstr)) != -1)
 202                 switch(c) {
 203                 case 'D':               /* define something..*/
 204                         for (p = optarg; *p; p++)
 205                                 if (*p == '=')
 206                                         break;
 207                         if (p == optarg)
 208                                 errx(1, "null variable cannot be defined");
 209                         if (*p)
 210                                 *p++ = EOS;
 211                         dodefine(optarg, p);
 212                         break;
 213                 case 'I':
 214                         addtoincludepath(optarg);
 215                         break;
 216                 case 'P':
 217                         break;
 218                 case 'U':               /* undefine...       */
 219                         remhash(optarg, TOP);
 220                         break;
 221                 case 'g':
 222                         mimic_gnu = 1;
 223                         break;
 224                 case 'd':
 225                         set_trace_flags(optarg);
 226                         break;
 227                 case 's':
 228                         synccpp = 1;
 229                         break;
 230                 case 't':
 231                         mark_traced(optarg, 1);
 232                         break;
 233                 case 'o':
 234                         trace_file(optarg);
 235                         break;
 236                 case '?':
 237                 default:
 238                         usage();
 239                 }
 240
 241         argc -= optind;
 242         argv += optind;
 243
 244         rval = 0;
 245         active = stdout;                /* default active output     */
 246         bbase[0] = bufbase;
 247         if (!argc) {
 248                 sp = -1;                /* stack pointer initialized */
 249                 fp = 0;                 /* frame pointer initialized */
 250                 set_input(infile+0, stdin, "stdin");
 251                                         /* default input (naturally) */
 252                 if ((inname[0] = strdup("-")) == NULL)
 253                         err(1, NULL);
 254                 inlineno[0] = 1;
 255                 emitline();
 256                 macro();
 257         } else
 258                 for (; argc--; ++argv) {
 259                         p = *argv;
 260                         if (p[0] == '-' && p[1] == EOS)
 261                                 set_input(infile, stdin, "stdin");
 262                         else if (fopen_trypath(infile, p) == NULL) {
 263                                 warn("%s", p);
 264                                 rval = 1;
 265                                 continue;
 266                         }
 267                         sp = -1;
 268                         fp = 0;
 269                         if ((inname[0] = strdup(p)) == NULL)
 270                                 err(1, NULL);
 271                         inlineno[0] = 1;
 272                         emitline();
 273                         macro();
 274                         release_input(infile);
 275                 }
 276
 277         if (*m4wraps) {                 /* anything for rundown ??   */
 278                 ilevel = 0;             /* in case m4wrap includes.. */
 279                 bufbase = bp = buf;     /* use the entire buffer   */
 280                 pbstr(m4wraps);         /* user-defined wrapup act   */
 281                 macro();                /* last will and testament   */
 282         }
 283
 284         if (active != stdout)
 285                 active = stdout;        /* reset output just in case */
 286         for (n = 1; n < maxout; n++)    /* default wrap-up: undivert */
 287                 if (outfile[n] != NULL)
 288                         getdiv(n);
 289                                         /* remove bitbucket if used  */
 290         if (outfile[0] != NULL) {
 291                 (void) fclose(outfile[0]);
 292         }
 293
 294         exit(rval);
 295 }
 296
 297 /*
 298  * Look ahead for `token'.
 299  * (on input `t == token[0]')
 300  * Used for comment and quoting delimiters.
 301  * Returns 1 if `token' present; copied to output.
 302  *         0 if `token' not found; all characters pushed back
 303  */
 304 static int
 305 do_look_ahead(int t, const char *token)
 306 {
 307         int i;
 308
 309         assert((unsigned char)t == (unsigned char)token[0]);
 310
 311         for (i = 1; *++token; i++) {
 312                 t = gpbc();
 313                 if (t == EOF || (unsigned char)t != (unsigned char)*token) {
 314                         putback(t);
 315                         while (--i)
 316                                 putback(*--token);
 317                         return 0;
 318                 }
 319         }
 320         return 1;
 321 }
 322
 323 #define LOOK_AHEAD(t, token) (t != EOF &&               \
 324     (unsigned char)(t)==(unsigned char)(token)[0] &&    \
 325     do_look_ahead(t,token))
 326
 327 /*
 328  * macro - the work horse..
 329  */
 330 static void
 331 macro(void)
 332 {
 333         char token[MAXTOK+1];
 334         int t, l;
 335         ndptr p;
 336         int  nlpar;
 337
 338         cycle {
 339                 t = gpbc();
 340                 if (t == '_' || isalpha(t)) {
 341                         p = inspect(t, token);
 342                         if (p != nil)
 343                                 putback(l = gpbc());
 344                         if (p == nil || (l != LPAREN &&
 345                             (p->type & NEEDARGS) != 0))
 346                                 outputstr(token);
 347                         else {
 348                 /*
 349                  * real thing.. First build a call frame:
 350                  */
 351                                 pushf(fp);      /* previous call frm */
 352                                 pushf(p->type); /* type of the call  */
 353                                 pushf(0);       /* parenthesis level */
 354                                 fp = sp;        /* new frame pointer */
 355                 /*
 356                  * now push the string arguments:
 357                  */
 358                                 pushs1(p->defn);        /* defn string */
 359                                 pushs1(p->name);        /* macro name  */
 360                                 pushs(ep);              /* start next..*/
 361
 362                                 if (l != LPAREN && PARLEV == 0)  {
 363                                     /* no bracks  */
 364                                         chrsave(EOS);
 365
 366                                         if ((uintptr_t)sp == STACKMAX)
 367                                                 errx(1, "internal stack overflow");
 368                                         eval((const char **) mstack+fp+1, 2,
 369                                             CALTYP);
 370
 371                                         ep = PREVEP;    /* flush strspace */
 372                                         sp = PREVSP;    /* previous sp..  */
 373                                         fp = PREVFP;    /* rewind stack...*/
 374                                 }
 375                         }
 376                 } else if (t == EOF) {
 377                         if (sp > -1) {
 378                                 warnx( "unexpected end of input, unclosed parenthesis:");
 379                                 dump_stack(paren, PARLEV);
 380                                 exit(1);
 381                         }
 382                         if (ilevel <= 0)
 383                                 break;                  /* all done thanks.. */
 384                         release_input(infile+ilevel--);
 385                         free(inname[ilevel+1]);
 386                         bufbase = bbase[ilevel];
 387                         emitline();
 388                         continue;
 389                 }
 390         /*
 391          * non-alpha token possibly seen..
 392          * [the order of else if .. stmts is important.]
 393          */
 394                 else if (LOOK_AHEAD(t,lquote)) {        /* strip quotes */
 395                         nlpar = 0;
 396                         record(quotes, nlpar++);
 397                         /*
 398                          * Opening quote: scan forward until matching
 399                          * closing quote has been found.
 400                          */
 401                         do {
 402
 403                                 l = gpbc();
 404                                 if (LOOK_AHEAD(l,rquote)) {
 405                                         if (--nlpar > 0)
 406                                                 outputstr(rquote);
 407                                 } else if (LOOK_AHEAD(l,lquote)) {
 408                                         record(quotes, nlpar++);
 409                                         outputstr(lquote);
 410                                 } else if (l == EOF) {
 411                                         if (nlpar == 1)
 412                                                 warnx("unclosed quote:");
 413                                         else
 414                                                 warnx("%d unclosed quotes:", nlpar);
 415                                         dump_stack(quotes, nlpar);
 416                                         exit(1);
 417                                 } else {
 418                                         if (nlpar > 0) {
 419                                                 if (sp < 0)
 420                                                         putc(l, active);
 421                                                 else
 422                                                         CHRSAVE(l);
 423                                         }
 424                                 }
 425                         }
 426                         while (nlpar != 0);
 427                 }
 428
 429                 else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
 430                         fputs(scommt, active);
 431
 432                         for(;;) {
 433                                 t = gpbc();
 434                                 if (LOOK_AHEAD(t, ecommt)) {
 435                                         fputs(ecommt, active);
 436                                         break;
 437                                 }
 438                                 if (t == EOF)
 439                                         break;
 440                                 putc(t, active);
 441                         }
 442                 }
 443
 444                 else if (sp < 0) {              /* not in a macro at all */
 445                         putc(t, active);        /* output directly..     */
 446                 }
 447
 448                 else switch(t) {
 449
 450                 case LPAREN:
 451                         if (PARLEV > 0)
 452                                 chrsave(t);
 453                         while (isspace(l = gpbc()))
 454                                 ;               /* skip blank, tab, nl.. */
 455                         putback(l);
 456                         record(paren, PARLEV++);
 457                         break;
 458
 459                 case RPAREN:
 460                         if (--PARLEV > 0)
 461                                 chrsave(t);
 462                         else {                  /* end of argument list */
 463                                 chrsave(EOS);
 464
 465                                 if ((uintptr_t)sp == STACKMAX)
 466                                         errx(1, "internal stack overflow");
 467
 468                                 eval((const char **) mstack+fp+1, sp-fp,
 469                                     CALTYP);
 470
 471                                 ep = PREVEP;    /* flush strspace */
 472                                 sp = PREVSP;    /* previous sp..  */
 473                                 fp = PREVFP;    /* rewind stack...*/
 474                         }
 475                         break;
 476
 477                 case COMMA:
 478                         if (PARLEV == 1) {
 479                                 chrsave(EOS);           /* new argument   */
 480                                 while (isspace(l = gpbc()))
 481                                         ;
 482                                 putback(l);
 483                                 pushs(ep);
 484                         } else
 485                                 chrsave(t);
 486                         break;
 487
 488                 default:
 489                         if (LOOK_AHEAD(t, scommt)) {
 490                                 char *pc;
 491                                 for (pc = scommt; *pc; pc++)
 492                                         chrsave(*pc);
 493                                 for(;;) {
 494                                         t = gpbc();
 495                                         if (LOOK_AHEAD(t, ecommt)) {
 496                                                 for (pc = ecommt; *pc; pc++)
 497                                                         chrsave(*pc);
 498                                                 break;
 499                                         }
 500                                         if (t == EOF)
 501                                             break;
 502                                         CHRSAVE(t);
 503                                 }
 504                         } else
 505                                 CHRSAVE(t);             /* stack the char */
 506                         break;
 507                 }
 508         }
 509 }
 510
 511 /*
 512  * output string directly, without pushing it for reparses.
 513  */
 514 void
 515 outputstr(const char *s)
 516 {
 517         if (sp < 0)
 518                 while (*s)
 519                         putc(*s++, active);
 520         else
 521                 while (*s)
 522                         CHRSAVE(*s++);
 523 }
 524
 525 /*
 526  * build an input token..
 527  * consider only those starting with _ or A-Za-z. This is a
 528  * combo with lookup to speed things up.
 529  */
 530 static ndptr
 531 inspect(int c, char *tp)
 532 {
 533         char *name = tp;
 534         char *etp = tp+MAXTOK;
 535         ndptr p;
 536         unsigned int h;
 537
 538         h = *tp++ = c;
 539
 540         while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
 541                 h = (h << 5) + h + (*tp++ = c);
 542         if (c != EOF)
 543                 PUTBACK(c);
 544         *tp = EOS;
 545         /* token is too long, it won't match anything, but it can still
 546          * be output. */
 547         if (tp == ep) {
 548                 outputstr(name);
 549                 while (isalnum(c = gpbc()) || c == '_') {
 550                         if (sp < 0)
 551                                 putc(c, active);
 552                         else
 553                                 CHRSAVE(c);
 554                 }
 555                 *name = EOS;
 556                 return nil;
 557         }
 558
 559         for (p = hashtab[h % HASHSIZE]; p != nil; p = p->nxtptr)
 560                 if (h == p->hv && STREQ(name, p->name))
 561                         break;
 562         return p;
 563 }
 564
 565 /*
 566  * initkwds - initialise m4 keywords as fast as possible.
 567  * This very similar to install, but without certain overheads,
 568  * such as calling lookup. Malloc is not used for storing the
 569  * keyword strings, since we simply use the static pointers
 570  * within keywrds block.
 571  */
 572 static void
 573 initkwds(void)
 574 {
 575         size_t i;
 576         unsigned int h;
 577         ndptr p;
 578         char *k;
 579
 580         for (i = 0; i < MAXKEYS; i++) {
 581                 k = (char *)keywrds[i].knam;
 582                 if (m4prefix) {
 583                         if (asprintf(&k, "m4_%s", k) == -1)
 584                                 err(1, "asprintf");
 585                         keywrds[i].knam = k;
 586                 }
 587                 h = hash(k);
 588                 p = (ndptr) xalloc(sizeof(struct ndblock));
 589                 p->nxtptr = hashtab[h % HASHSIZE];
 590                 hashtab[h % HASHSIZE] = p;
 591                 p->name = xstrdup(keywrds[i].knam);
 592                 p->defn = null;
 593                 p->hv = h;
 594                 p->type = keywrds[i].ktyp & TYPEMASK;
 595                 if ((keywrds[i].ktyp & NOARGS) == 0)
 596                         p->type |= NEEDARGS;
 597         }
 598 }
 599
 600 /* Look up a builtin type, even if overridden by the user */
 601 int
 602 builtin_type(const char *key)
 603 {
 604         int i;
 605
 606         for (i = 0; i != MAXKEYS; i++)
 607                 if (STREQ(keywrds[i].knam, key))
 608                         return keywrds[i].ktyp;
 609         return -1;
 610 }
 611
 612 const char *
 613 builtin_realname(int n)
 614 {
 615         int i;
 616
 617         for (i = 0; i != MAXKEYS; i++)
 618                 if (((keywrds[i].ktyp ^ n) & TYPEMASK) == 0)
 619                         return keywrds[i].knam;
 620         return NULL;
 621 }
 622
 623 static void
 624 record(struct position *t, int lev)
 625 {
 626         if (lev < MAXRECORD) {
 627                 t[lev].name = CURRENT_NAME;
 628                 t[lev].line = CURRENT_LINE;
 629         }
 630 }
 631
 632 static void
 633 dump_stack(struct position *t, int lev)
 634 {
 635         int i;
 636
 637         for (i = 0; i < lev; i++) {
 638                 if (i == MAXRECORD) {
 639                         fprintf(stderr, "   ...\n");
 640                         break;
 641                 }
 642                 fprintf(stderr, "   %s at line %lu\n",
 643                         t[i].name, t[i].line);
 644         }
 645 }
 646
 647
 648 static void
 649 enlarge_stack(void)
 650 {
 651         STACKMAX *= 2;
 652         mstack = realloc(mstack, sizeof(stae) * STACKMAX);
 653         sstack = realloc(sstack, STACKMAX);
 654         if (mstack == NULL || sstack == NULL)
 655                 errx(1, "Evaluation stack overflow (%lu)",
 656                     (unsigned long)STACKMAX);
 657 }
 658
 659 /* Emit preprocessor #line directive if -s option used. */
 660 void
 661 emitline(void)
 662 {
 663
 664         if (synccpp)
 665                 fprintf(active, "#line %d \"%s\"\n", inlineno[ilevel],
 666                         inname[ilevel]);
 667 }