usr.bin/m4/main.c

   1 /*      $OpenBSD: main.c,v 1.77 2009/10/14 17:19:47 sthen Exp $ */
   2 /*      $NetBSD: main.c,v 1.39 2009/11/06 15:13:27 joerg 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. Neither the name of the University nor the names of its contributors
  20  *    may be used to endorse or promote products derived from this software
  21  *    without specific prior written permission.
  22  *
  23  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  24  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  26  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  27  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  28  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  29  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  30  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  31  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  32  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  33  * SUCH DAMAGE.
  34  */
  35
  36 /*
  37  * main.c
  38  * Facility: m4 macro processor
  39  * by: oz
  40  */
  41 #if HAVE_NBTOOL_CONFIG_H
  42 #include "nbtool_config.h"
  43 #endif
  44 #include <sys/cdefs.h>
  45 __RCSID("$NetBSD: main.c,v 1.39 2009/11/06 15:13:27 joerg Exp $");
  46 #include <assert.h>
  47 #include <signal.h>
  48 #include <err.h>
  49 #include <errno.h>
  50 #include <unistd.h>
  51 #include <stdio.h>
  52 #include <ctype.h>
  53 #include <string.h>
  54 #include <stddef.h>
  55 #include <stdint.h>
  56 #include <stdlib.h>
  57 #include <ohash.h>
  58 #include "mdef.h"
  59 #include "stdd.h"
  60 #include "extern.h"
  61 #include "pathnames.h"
  62
  63 ndptr hashtab[HASHSIZE];        /* hash table for macros etc.  */
  64 stae *mstack;                   /* stack of m4 machine         */
  65 char *sstack;                   /* shadow stack, for string space extension */
  66 static size_t STACKMAX;         /* current maximum size of stack */
  67 int sp;                         /* current m4  stack pointer   */
  68 int fp;                         /* m4 call frame pointer       */
  69 struct input_file infile[MAXINP];/* input file stack (0=stdin)  */
  70 FILE **outfile;                 /* diversion array(0=bitbucket)*/
  71 int maxout;
  72 FILE *active;                   /* active output file pointer  */
  73 int ilevel = 0;                 /* input file stack pointer    */
  74 int oindex = 0;                 /* diversion index..           */
  75 const char *null = "";          /* as it says.. just a null..  */
  76 char **m4wraps = NULL;          /* m4wraps array.              */
  77 int maxwraps = 0;               /* size of m4wraps array       */
  78 int wrapindex = 0;              /* current offset in m4wraps   */
  79 char lquote[MAXCCHARS+1] = {LQUOTE};    /* left quote character  (`)   */
  80 char rquote[MAXCCHARS+1] = {RQUOTE};    /* right quote character (')   */
  81 char scommt[MAXCCHARS+1] = {SCOMMT};    /* start character for comment */
  82 char ecommt[MAXCCHARS+1] = {ECOMMT};    /* end character for comment   */
  83 int  synch_lines = 0;           /* line synchronisation for C preprocessor */
  84 int  prefix_builtins = 0;       /* -P option to prefix builtin keywords */
  85
  86 struct keyblk {
  87         const char *knam;       /* keyword name */
  88         int     ktyp;           /* keyword type */
  89 };
  90
  91 struct keyblk keywrds[] = {     /* m4 keywords to be installed */
  92         { "include",      INCLTYPE },
  93         { "sinclude",     SINCTYPE },
  94         { "define",       DEFITYPE },
  95         { "defn",         DEFNTYPE },
  96         { "divert",       DIVRTYPE | NOARGS },
  97         { "expr",         EXPRTYPE },
  98         { "eval",         EXPRTYPE },
  99         { "substr",       SUBSTYPE },
 100         { "ifelse",       IFELTYPE },
 101         { "ifdef",        IFDFTYPE },
 102         { "len",          LENGTYPE },
 103         { "incr",         INCRTYPE },
 104         { "decr",         DECRTYPE },
 105         { "dnl",          DNLNTYPE | NOARGS },
 106         { "changequote",  CHNQTYPE | NOARGS },
 107         { "changecom",    CHNCTYPE | NOARGS },
 108         { "index",        INDXTYPE },
 109 #ifdef EXTENDED
 110         { "paste",        PASTTYPE },
 111         { "spaste",       SPASTYPE },
 112         /* Newer extensions, needed to handle gnu-m4 scripts */
 113         { "indir",        INDIRTYPE},
 114         { "builtin",      BUILTINTYPE},
 115         { "patsubst",     PATSTYPE},
 116         { "regexp",       REGEXPTYPE},
 117         { "esyscmd",      ESYSCMDTYPE},
 118         { "__file__",     FILENAMETYPE | NOARGS},
 119         { "__line__",     LINETYPE | NOARGS},
 120 #endif
 121         { "popdef",       POPDTYPE },
 122         { "pushdef",      PUSDTYPE },
 123         { "dumpdef",      DUMPTYPE | NOARGS },
 124         { "shift",        SHIFTYPE | NOARGS },
 125         { "translit",     TRNLTYPE },
 126         { "undefine",     UNDFTYPE },
 127         { "undivert",     UNDVTYPE | NOARGS },
 128         { "divnum",       DIVNTYPE | NOARGS },
 129         { "maketemp",     MKTMTYPE },
 130         { "errprint",     ERRPTYPE | NOARGS },
 131         { "m4wrap",       M4WRTYPE | NOARGS },
 132         { "m4exit",       EXITTYPE | NOARGS },
 133         { "syscmd",       SYSCTYPE },
 134         { "sysval",       SYSVTYPE | NOARGS },
 135         { "traceon",      TRACEONTYPE | NOARGS },
 136         { "traceoff",     TRACEOFFTYPE | NOARGS },
 137
 138 #if defined(unix) || defined(__unix__)
 139         { "unix",         SELFTYPE | NOARGS },
 140 #else
 141 #ifdef vms
 142         { "vms",          SELFTYPE | NOARGS },
 143 #endif
 144 #endif
 145 };
 146
 147 #define MAXKEYS (sizeof(keywrds)/sizeof(struct keyblk))
 148
 149 extern int optind;
 150 extern char *optarg;
 151
 152 #define MAXRECORD 50
 153 static struct position {
 154         char *name;
 155         unsigned long line;
 156 } quotes[MAXRECORD], paren[MAXRECORD];
 157
 158 static void record(struct position *, int);
 159 static void dump_stack(struct position *, int);
 160
 161 static void macro(void);
 162 static void initkwds(void);
 163 static ndptr inspect(int, char *);
 164 static int do_look_ahead(int, const char *);
 165 static void reallyoutputstr(const char *);
 166 static void reallyputchar(int);
 167
 168 static void enlarge_stack(void);
 169
 170 int main(int, char *[]);
 171
 172 int
 173 main(int argc, char *argv[])
 174 {
 175         int c;
 176         int n;
 177         char *p;
 178
 179         setprogname(argv[0]);
 180
 181         if (signal(SIGINT, SIG_IGN) != SIG_IGN)
 182                 signal(SIGINT, onintr);
 183
 184         init_macros();
 185         initspaces();
 186         STACKMAX = INITSTACKMAX;
 187
 188         mstack = (stae *)xalloc(sizeof(stae) * STACKMAX, NULL);
 189         sstack = (char *)xalloc(STACKMAX, NULL);
 190
 191         maxout = 0;
 192         outfile = NULL;
 193         resizedivs(MAXOUT);
 194
 195         while ((c = getopt(argc, argv, "gst:d:D:U:o:I:P")) != -1)
 196                 switch(c) {
 197
 198                 case 'D':               /* define something..*/
 199                         for (p = optarg; *p; p++)
 200                                 if (*p == '=')
 201                                         break;
 202                         if (*p)
 203                                 *p++ = EOS;
 204                         dodefine(optarg, p);
 205                         break;
 206                 case 'I':
 207                         addtoincludepath(optarg);
 208                         break;
 209                 case 'P':
 210                         prefix_builtins = 1;
 211                         break;
 212                 case 'U':               /* undefine...       */
 213                         macro_popdef(optarg);
 214                         break;
 215                 case 'g':
 216                         mimic_gnu = 1;
 217                         break;
 218                 case 'd':
 219                         set_trace_flags(optarg);
 220                         break;
 221                 case 's':
 222                         synch_lines = 1;
 223                         break;
 224                 case 't':
 225                         mark_traced(optarg, 1);
 226                         break;
 227                 case 'o':
 228                         trace_file(optarg);
 229                         break;
 230                 case '?':
 231                         usage();
 232                 }
 233
 234         argc -= optind;
 235         argv += optind;
 236
 237         initkwds();
 238         if (mimic_gnu)
 239                 setup_builtin("format", FORMATTYPE);
 240
 241         active = stdout;                /* default active output     */
 242         bbase[0] = bufbase;
 243         if (!argc) {
 244                 sp = -1;                /* stack pointer initialized */
 245                 fp = 0;                 /* frame pointer initialized */
 246                 set_input(infile+0, stdin, "stdin");
 247                                         /* default input (naturally) */
 248                 macro();
 249         } else
 250                 for (; argc--; ++argv) {
 251                         p = *argv;
 252                         if (p[0] == '-' && p[1] == EOS)
 253                                 set_input(infile, stdin, "stdin");
 254                         else if (fopen_trypath(infile, p) == NULL)
 255                                 err(1, "%s", p);
 256                         sp = -1;
 257                         fp = 0;
 258                         macro();
 259                         release_input(infile);
 260                 }
 261
 262         if (wrapindex) {
 263                 int i;
 264
 265                 ilevel = 0;             /* in case m4wrap includes.. */
 266                 bufbase = bp = buf;     /* use the entire buffer   */
 267                 if (mimic_gnu) {
 268                         while (wrapindex != 0) {
 269                                 for (i = 0; i < wrapindex; i++)
 270                                         pbstr(m4wraps[i]);
 271                                 wrapindex =0;
 272                                 macro();
 273                         }
 274                 } else {
 275                         for (i = 0; i < wrapindex; i++) {
 276                                 pbstr(m4wraps[i]);
 277                                 macro();
 278                         }
 279                 }
 280         }
 281
 282         if (active != stdout)
 283                 active = stdout;        /* reset output just in case */
 284         for (n = 1; n < maxout; n++)    /* default wrap-up: undivert */
 285                 if (outfile[n] != NULL)
 286                         getdiv(n);
 287                                         /* remove bitbucket if used  */
 288         if (outfile[0] != NULL) {
 289                 (void) fclose(outfile[0]);
 290         }
 291
 292         return 0;
 293 }
 294
 295 /*
 296  * Look ahead for `token'.
 297  * (on input `t == token[0]')
 298  * Used for comment and quoting delimiters.
 299  * Returns 1 if `token' present; copied to output.
 300  *         0 if `token' not found; all characters pushed back
 301  */
 302 static int
 303 do_look_ahead(int t, const char *token)
 304 {
 305         int i;
 306
 307         assert((unsigned char)t == (unsigned char)token[0]);
 308
 309         for (i = 1; *++token; i++) {
 310                 t = gpbc();
 311                 if (t == EOF || (unsigned char)t != (unsigned char)*token) {
 312                         pushback(t);
 313                         while (--i)
 314                                 pushback(*--token);
 315                         return 0;
 316                 }
 317         }
 318         return 1;
 319 }
 320
 321 #define LOOK_AHEAD(t, token) (t != EOF &&               \
 322     (unsigned char)(t)==(unsigned char)(token)[0] &&    \
 323     do_look_ahead(t,token))
 324
 325 /*
 326  * macro - the work horse..
 327  */
 328 static void
 329 macro(void)
 330 {
 331         char token[MAXTOK+1];
 332         int t, l;
 333         ndptr p;
 334         int  nlpar;
 335
 336         cycle {
 337                 t = gpbc();
 338
 339                 if (LOOK_AHEAD(t,lquote)) {     /* strip quotes */
 340                         nlpar = 0;
 341                         record(quotes, nlpar++);
 342                         /*
 343                          * Opening quote: scan forward until matching
 344                          * closing quote has been found.
 345                          */
 346                         do {
 347
 348                                 l = gpbc();
 349                                 if (LOOK_AHEAD(l,rquote)) {
 350                                         if (--nlpar > 0)
 351                                                 outputstr(rquote);
 352                                 } else if (LOOK_AHEAD(l,lquote)) {
 353                                         record(quotes, nlpar++);
 354                                         outputstr(lquote);
 355                                 } else if (l == EOF) {
 356                                         if (nlpar == 1)
 357                                                 warnx("unclosed quote:");
 358                                         else
 359                                                 warnx("%d unclosed quotes:", nlpar);
 360                                         dump_stack(quotes, nlpar);
 361                                         exit(1);
 362                                 } else {
 363                                         if (nlpar > 0) {
 364                                                 if (sp < 0)
 365                                                         reallyputchar(l);
 366                                                 else
 367                                                         CHRSAVE(l);
 368                                         }
 369                                 }
 370                         }
 371                         while (nlpar != 0);
 372                 } else if (sp < 0 && LOOK_AHEAD(t, scommt)) {
 373                         reallyoutputstr(scommt);
 374
 375                         for(;;) {
 376                                 t = gpbc();
 377                                 if (LOOK_AHEAD(t, ecommt)) {
 378                                         reallyoutputstr(ecommt);
 379                                         break;
 380                                 }
 381                                 if (t == EOF)
 382                                         break;
 383                                 reallyputchar(t);
 384                         }
 385                 } else if (t == '_' || isalpha(t)) {
 386                         p = inspect(t, token);
 387                         if (p != NULL)
 388                                 pushback(l = gpbc());
 389                         if (p == NULL || (l != LPAREN &&
 390                             (macro_getdef(p)->type & NEEDARGS) != 0))
 391                                 outputstr(token);
 392                         else {
 393                 /*
 394                  * real thing.. First build a call frame:
 395                  */
 396                                 pushf(fp);      /* previous call frm */
 397                                 pushf(macro_getdef(p)->type); /* type of the call  */
 398                                 pushf(is_traced(p));
 399                                 pushf(0);       /* parenthesis level */
 400                                 fp = sp;        /* new frame pointer */
 401                 /*
 402                  * now push the string arguments:
 403                  */
 404                                 pushs1(macro_getdef(p)->defn);  /* defn string */
 405                                 pushs1((char *)macro_name(p));  /* macro name  */
 406                                 pushs(ep);              /* start next..*/
 407
 408                                 if (l != LPAREN && PARLEV == 0)  {
 409                                     /* no bracks  */
 410                                         chrsave(EOS);
 411
 412                                         if ((size_t)sp == STACKMAX)
 413                                                 errx(1, "internal stack overflow");
 414                                         eval((const char **) mstack+fp+1, 2,
 415                                             CALTYP, TRACESTATUS);
 416
 417                                         ep = PREVEP;    /* flush strspace */
 418                                         sp = PREVSP;    /* previous sp..  */
 419                                         fp = PREVFP;    /* rewind stack...*/
 420                                 }
 421                         }
 422                 } else if (t == EOF) {
 423                         if (sp > -1 && ilevel <= 0) {
 424                                 warnx( "unexpected end of input, unclosed parenthesis:");
 425                                 dump_stack(paren, PARLEV);
 426                                 exit(1);
 427                         }
 428                         if (ilevel <= 0)
 429                                 break;                  /* all done thanks.. */
 430                         release_input(infile+ilevel--);
 431                         emit_synchline();
 432                         bufbase = bbase[ilevel];
 433                         continue;
 434                 } else if (sp < 0) {            /* not in a macro at all */
 435                         reallyputchar(t);       /* output directly..     */
 436                 }
 437
 438                 else switch(t) {
 439
 440                 case LPAREN:
 441                         if (PARLEV > 0)
 442                                 chrsave(t);
 443                         while (isspace(l = gpbc())) /* skip blank, tab, nl.. */
 444                                 if (PARLEV > 0)
 445                                         chrsave(l);
 446                         pushback(l);
 447                         record(paren, PARLEV++);
 448                         break;
 449
 450                 case RPAREN:
 451                         if (--PARLEV > 0)
 452                                 chrsave(t);
 453                         else {                  /* end of argument list */
 454                                 chrsave(EOS);
 455
 456                                 if ((size_t)sp == STACKMAX)
 457                                         errx(1, "internal stack overflow");
 458
 459                                 eval((const char **) mstack+fp+1, sp-fp,
 460                                     CALTYP, TRACESTATUS);
 461
 462                                 ep = PREVEP;    /* flush strspace */
 463                                 sp = PREVSP;    /* previous sp..  */
 464                                 fp = PREVFP;    /* rewind stack...*/
 465                         }
 466                         break;
 467
 468                 case COMMA:
 469                         if (PARLEV == 1) {
 470                                 chrsave(EOS);           /* new argument   */
 471                                 while (isspace(l = gpbc()))
 472                                         ;
 473                                 pushback(l);
 474                                 pushs(ep);
 475                         } else
 476                                 chrsave(t);
 477                         break;
 478
 479                 default:
 480                         if (LOOK_AHEAD(t, scommt)) {
 481                                 char *q;
 482                                 for (q = scommt; *q; p++)
 483                                         chrsave(*q);
 484                                 for(;;) {
 485                                         t = gpbc();
 486                                         if (LOOK_AHEAD(t, ecommt)) {
 487                                                 for (q = ecommt; *q; q++)
 488                                                         chrsave(*q);
 489                                                 break;
 490                                         }
 491                                         if (t == EOF)
 492                                             break;
 493                                         CHRSAVE(t);
 494                                 }
 495                         } else
 496                                 CHRSAVE(t);             /* stack the char */
 497                         break;
 498                 }
 499         }
 500 }
 501
 502 /*
 503  * output string directly, without pushing it for reparses.
 504  */
 505 void
 506 outputstr(const char *s)
 507 {
 508         if (sp < 0)
 509                 reallyoutputstr(s);
 510         else
 511                 while (*s)
 512                         CHRSAVE(*s++);
 513 }
 514
 515 void
 516 reallyoutputstr(const char *s)
 517 {
 518         if (synch_lines) {
 519                 while (*s) {
 520                         fputc(*s, active);
 521                         if (*s++ == '\n') {
 522                                 infile[ilevel].synch_lineno++;
 523                                 if (infile[ilevel].synch_lineno !=
 524                                     infile[ilevel].lineno)
 525                                         do_emit_synchline();
 526                         }
 527                 }
 528         } else
 529                 fputs(s, active);
 530 }
 531
 532 void
 533 reallyputchar(int c)
 534 {
 535         putc(c, active);
 536         if (synch_lines && c == '\n') {
 537                 infile[ilevel].synch_lineno++;
 538                 if (infile[ilevel].synch_lineno != infile[ilevel].lineno)
 539                         do_emit_synchline();
 540         }
 541 }
 542
 543 /*
 544  * build an input token..
 545  * consider only those starting with _ or A-Za-z.
 546  */
 547 static ndptr
 548 inspect(int c, char *tp)
 549 {
 550         char *name = tp;
 551         char *etp = tp+MAXTOK;
 552         ndptr p;
 553
 554         *tp++ = c;
 555
 556         while ((isalnum(c = gpbc()) || c == '_') && tp < etp)
 557                 *tp++ = c;
 558         if (c != EOF)
 559                 PUSHBACK(c);
 560         *tp = EOS;
 561         /* token is too long, it won't match anything, but it can still
 562          * be output. */
 563         if (tp == ep) {
 564                 outputstr(name);
 565                 while (isalnum(c = gpbc()) || c == '_') {
 566                         if (sp < 0)
 567                                 reallyputchar(c);
 568                         else
 569                                 CHRSAVE(c);
 570                 }
 571                 *name = EOS;
 572                 return NULL;
 573         }
 574
 575         p = ohash_find(&macros, ohash_qlookupi(&macros, name, (const char **)&tp));
 576         if (p == NULL)
 577                 return NULL;
 578         if (macro_getdef(p) == NULL)
 579                 return NULL;
 580         return p;
 581 }
 582
 583 /*
 584  * initkwds - initialise m4 keywords as fast as possible.
 585  * This very similar to install, but without certain overheads,
 586  * such as calling lookup. Malloc is not used for storing the
 587  * keyword strings, since we simply use the static pointers
 588  * within keywrds block.
 589  */
 590 static void
 591 initkwds(void)
 592 {
 593         unsigned int type;
 594         size_t i;
 595
 596         for (i = 0; i < MAXKEYS; i++) {
 597                 type = keywrds[i].ktyp & TYPEMASK;
 598                 if ((keywrds[i].ktyp & NOARGS) == 0)
 599                         type |= NEEDARGS;
 600                 setup_builtin(keywrds[i].knam, type);
 601         }
 602 }
 603
 604 static void
 605 record(struct position *t, int lev)
 606 {
 607         if (lev < MAXRECORD) {
 608                 t[lev].name = CURRENT_NAME;
 609                 t[lev].line = CURRENT_LINE;
 610         }
 611 }
 612
 613 static void
 614 dump_stack(struct position *t, int lev)
 615 {
 616         int i;
 617
 618         for (i = 0; i < lev; i++) {
 619                 if (i == MAXRECORD) {
 620                         fprintf(stderr, "   ...\n");
 621                         break;
 622                 }
 623                 fprintf(stderr, "   %s at line %lu\n",
 624                         t[i].name, t[i].line);
 625         }
 626 }
 627
 628
 629 static void
 630 enlarge_stack(void)
 631 {
 632         STACKMAX += STACKMAX/2;
 633         mstack = xrealloc(mstack, sizeof(stae) * STACKMAX,
 634             "Evaluation stack overflow (%lu)",
 635             (unsigned long)STACKMAX);
 636         sstack = xrealloc(sstack, STACKMAX,
 637             "Evaluation stack overflow (%lu)",
 638             (unsigned long)STACKMAX);
 639 }