main.c

   1 /* $Id: main.c,v 1.24 2009/04/12 19:19:57 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@openbsd.org>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the
   7  * above copyright notice and this permission notice appear in all
   8  * copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL
  11  * WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED
  12  * WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE
  13  * AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
  14  * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
  15  * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
  16  * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
  17  * PERFORMANCE OF THIS SOFTWARE.
  18  */
  19 #include <sys/stat.h>
  20
  21 #include <assert.h>
  22 #include <err.h>
  23 #include <fcntl.h>
  24 #include <stdio.h>
  25 #include <stdlib.h>
  26 #include <string.h>
  27 #include <unistd.h>
  28
  29 #include "mdoc.h"
  30 #include "man.h"
  31
  32 /* Account for FreeBSD and Linux in our declarations. */
  33
  34 #ifdef __linux__
  35 extern  int               getsubopt(char **, char * const *, char **);
  36 # ifndef __dead
  37 #  define __dead __attribute__((__noreturn__))
  38 # endif
  39 #elif defined(__dead2)
  40 # ifndef __dead
  41 #  define __dead __dead2
  42 # endif
  43 #endif
  44
  45 typedef int             (*out_mdoc)(void *, const struct mdoc *);
  46 typedef int             (*out_man)(void *, const struct man *);
  47 typedef void            (*out_free)(void *);
  48
  49 struct  buf {
  50         char             *buf;
  51         size_t            sz;
  52 };
  53
  54 enum    intt {
  55         INTT_AUTO,
  56         INTT_MDOC,
  57         INTT_MAN
  58 };
  59
  60 enum    outt {
  61         OUTT_ASCII = 0,
  62         OUTT_TREE,
  63         OUTT_LINT
  64 };
  65
  66 struct  curparse {
  67         const char       *file;         /* Current parse. */
  68         int               fd;           /* Current parse. */
  69         int               wflags;
  70 #define WARN_WALL         0x03          /* All-warnings mask. */
  71 #define WARN_WCOMPAT     (1 << 0)       /* Compatibility warnings. */
  72 #define WARN_WSYNTAX     (1 << 1)       /* Syntax warnings. */
  73 #define WARN_WERR        (1 << 2)       /* Warnings->errors. */
  74         int               fflags;
  75 #define IGN_SCOPE        (1 << 0)       /* Ignore scope errors. */
  76 #define NO_IGN_ESCAPE    (1 << 1)       /* Don't ignore bad escapes. */
  77 #define NO_IGN_MACRO     (1 << 2)       /* Don't ignore bad macros. */
  78 #define NO_IGN_CHARS     (1 << 3)       /* Don't ignore bad chars. */
  79         enum intt         inttype;      /* Input parsers. */
  80         struct man       *man;
  81         struct man       *lastman;
  82         struct mdoc      *mdoc;
  83         struct mdoc      *lastmdoc;
  84         enum outt         outtype;      /* Output devices. */
  85         out_mdoc          outmdoc;
  86         out_man           outman;
  87         out_free          outfree;
  88         void             *outdata;
  89 };
  90
  91 extern  void             *ascii_alloc(void);
  92 extern  int               tree_mdoc(void *, const struct mdoc *);
  93 extern  int               tree_man(void *, const struct man *);
  94 extern  int               terminal_mdoc(void *, const struct mdoc *);
  95 extern  int               terminal_man(void *, const struct man *);
  96 extern  void              terminal_free(void *);
  97
  98 static  int               foptions(int *, char *);
  99 static  int               toptions(enum outt *, char *);
 100 static  int               moptions(enum intt *, char *);
 101 static  int               woptions(int *, char *);
 102 static  int               merr(void *, int, int, const char *);
 103 static  int               manwarn(void *, int, int, const char *);
 104 static  int               mdocwarn(void *, int, int,
 105                                 enum mdoc_warn, const char *);
 106 static  int               fstdin(struct buf *, struct buf *,
 107                                 struct curparse *);
 108 static  int               ffile(struct buf *, struct buf *,
 109                                 const char *, struct curparse *);
 110 static  int               fdesc(struct buf *, struct buf *,
 111                                 struct curparse *);
 112 static  int               pset(const char *, int, struct curparse *,
 113                                 struct man **, struct mdoc **);
 114 static  struct man       *man_init(struct curparse *);
 115 static  struct mdoc      *mdoc_init(struct curparse *);
 116 __dead  static void       version(void);
 117 __dead  static void       usage(void);
 118
 119 extern  char             *__progname;
 120
 121
 122 int
 123 main(int argc, char *argv[])
 124 {
 125         int              c, rc;
 126         struct buf       ln, blk;
 127         struct curparse  curp;
 128
 129         bzero(&curp, sizeof(struct curparse));
 130
 131         curp.inttype = INTT_AUTO;
 132         curp.outtype = OUTT_ASCII;
 133
 134         /* LINTED */
 135         while (-1 != (c = getopt(argc, argv, "f:m:VW:T:")))
 136                 switch (c) {
 137                 case ('f'):
 138                         if ( ! foptions(&curp.fflags, optarg))
 139                                 return(0);
 140                         break;
 141                 case ('m'):
 142                         if ( ! moptions(&curp.inttype, optarg))
 143                                 return(0);
 144                         break;
 145                 case ('T'):
 146                         if ( ! toptions(&curp.outtype, optarg))
 147                                 return(0);
 148                         break;
 149                 case ('W'):
 150                         if ( ! woptions(&curp.wflags, optarg))
 151                                 return(0);
 152                         break;
 153                 case ('V'):
 154                         version();
 155                         /* NOTREACHED */
 156                 default:
 157                         usage();
 158                         /* NOTREACHED */
 159                 }
 160
 161         argc -= optind;
 162         argv += optind;
 163
 164         /* Configure buffers. */
 165
 166         bzero(&ln, sizeof(struct buf));
 167         bzero(&blk, sizeof(struct buf));
 168
 169         rc = 1;
 170
 171         if (NULL == *argv)
 172                 if ( ! fstdin(&blk, &ln, &curp))
 173                         rc = 0;
 174
 175         while (rc && *argv) {
 176                 if ( ! ffile(&blk, &ln, *argv, &curp))
 177                         rc = 0;
 178                 argv++;
 179                 if (*argv && rc) {
 180                         if (curp.lastman)
 181                                 if ( ! man_reset(curp.lastman))
 182                                         rc = 0;
 183                         if (curp.lastmdoc)
 184                                 if ( ! mdoc_reset(curp.lastmdoc))
 185                                         rc = 0;
 186                         curp.lastman = NULL;
 187                         curp.lastmdoc = NULL;
 188                 }
 189         }
 190
 191         if (blk.buf)
 192                 free(blk.buf);
 193         if (ln.buf)
 194                 free(ln.buf);
 195         if (curp.outfree)
 196                 (*curp.outfree)(curp.outdata);
 197         if (curp.mdoc)
 198                 mdoc_free(curp.mdoc);
 199         if (curp.man)
 200                 man_free(curp.man);
 201
 202         return(rc ? EXIT_SUCCESS : EXIT_FAILURE);
 203 }
 204
 205
 206 __dead static void
 207 version(void)
 208 {
 209
 210         (void)printf("%s %s\n", __progname, VERSION);
 211         exit(EXIT_SUCCESS);
 212 }
 213
 214
 215 __dead static void
 216 usage(void)
 217 {
 218
 219         (void)fprintf(stderr, "usage: %s [-V] [-foption...] "
 220                         "[-mformat] [-Toutput] [-Werr...]\n",
 221                         __progname);
 222         exit(EXIT_FAILURE);
 223 }
 224
 225
 226 static struct man *
 227 man_init(struct curparse *curp)
 228 {
 229         int              pflags;
 230         struct man      *man;
 231         struct man_cb    mancb;
 232
 233         mancb.man_err = merr;
 234         mancb.man_warn = manwarn;
 235
 236         pflags = MAN_IGN_MACRO; /* XXX */
 237
 238         if (curp->fflags & NO_IGN_MACRO)
 239                 pflags &= ~MAN_IGN_MACRO;
 240
 241         if (NULL == (man = man_alloc(curp, pflags, &mancb)))
 242                 warnx("memory exhausted");
 243
 244         return(man);
 245 }
 246
 247
 248 static struct mdoc *
 249 mdoc_init(struct curparse *curp)
 250 {
 251         int              pflags;
 252         struct mdoc     *mdoc;
 253         struct mdoc_cb   mdoccb;
 254
 255         mdoccb.mdoc_msg = NULL;
 256         mdoccb.mdoc_err = merr;
 257         mdoccb.mdoc_warn = mdocwarn;
 258
 259         pflags = MDOC_IGN_MACRO | MDOC_IGN_ESCAPE | MDOC_IGN_CHARS;
 260
 261         if (curp->fflags & IGN_SCOPE)
 262                 pflags |= MDOC_IGN_SCOPE;
 263         if (curp->fflags & NO_IGN_ESCAPE)
 264                 pflags &= ~MDOC_IGN_ESCAPE;
 265         if (curp->fflags & NO_IGN_MACRO)
 266                 pflags &= ~MDOC_IGN_MACRO;
 267         if (curp->fflags & NO_IGN_CHARS)
 268                 pflags &= ~MDOC_IGN_CHARS;
 269
 270         if (NULL == (mdoc = mdoc_alloc(curp, pflags, &mdoccb)))
 271                 warnx("memory exhausted");
 272
 273         return(mdoc);
 274 }
 275
 276
 277 static int
 278 fstdin(struct buf *blk, struct buf *ln, struct curparse *curp)
 279 {
 280
 281         curp->file = "<stdin>";
 282         curp->fd = STDIN_FILENO;
 283         return(fdesc(blk, ln, curp));
 284 }
 285
 286
 287 static int
 288 ffile(struct buf *blk, struct buf *ln,
 289                 const char *file, struct curparse *curp)
 290 {
 291         int              c;
 292
 293         curp->file = file;
 294         if (-1 == (curp->fd = open(curp->file, O_RDONLY, 0))) {
 295                 warn("%s", curp->file);
 296                 return(0);
 297         }
 298
 299         c = fdesc(blk, ln, curp);
 300
 301         if (-1 == close(curp->fd))
 302                 warn("%s", curp->file);
 303
 304         return(c);
 305 }
 306
 307
 308 static int
 309 fdesc(struct buf *blk, struct buf *ln, struct curparse *curp)
 310 {
 311         size_t           sz;
 312         ssize_t          ssz;
 313         struct stat      st;
 314         int              j, i, pos, lnn;
 315         struct man      *man;
 316         struct mdoc     *mdoc;
 317
 318         sz = BUFSIZ;
 319         man = NULL;
 320         mdoc = NULL;
 321
 322         /*
 323          * Two buffers: ln and buf.  buf is the input buffer optimised
 324          * here for each file's block size.  ln is a line buffer.  Both
 325          * growable, hence passed in by ptr-ptr.
 326          */
 327
 328         if (-1 == fstat(curp->fd, &st))
 329                 warnx("%s", curp->file);
 330         else if ((size_t)st.st_blksize > sz)
 331                 sz = st.st_blksize;
 332
 333         if (sz > blk->sz) {
 334                 blk->buf = realloc(blk->buf, sz);
 335                 if (NULL == blk->buf) {
 336                         warn("realloc");
 337                         return(0);
 338                 }
 339                 blk->sz = sz;
 340         }
 341
 342         /* Fill buf with file blocksize. */
 343
 344         for (lnn = 0, pos = 0; ; ) {
 345                 if (-1 == (ssz = read(curp->fd, blk->buf, sz))) {
 346                         warn("%s", curp->file);
 347                         return(0);
 348                 } else if (0 == ssz)
 349                         break;
 350
 351                 /* Parse the read block into partial or full lines. */
 352
 353                 for (i = 0; i < (int)ssz; i++) {
 354                         if (pos >= (int)ln->sz) {
 355                                 ln->sz += 256; /* Step-size. */
 356                                 ln->buf = realloc(ln->buf, ln->sz);
 357                                 if (NULL == ln->buf) {
 358                                         warn("realloc");
 359                                         return(0);
 360                                 }
 361                         }
 362
 363                         if ('\n' != blk->buf[i]) {
 364                                 ln->buf[pos++] = blk->buf[i];
 365                                 continue;
 366                         }
 367
 368                         /* Check for CPP-escaped newline.  */
 369
 370                         if (pos > 0 && '\\' == ln->buf[pos - 1]) {
 371                                 for (j = pos - 1; j >= 0; j--)
 372                                         if ('\\' != ln->buf[j])
 373                                                 break;
 374
 375                                 if ( ! ((pos - j) % 2)) {
 376                                         pos--;
 377                                         lnn++;
 378                                         continue;
 379                                 }
 380                         }
 381
 382                         ln->buf[pos] = 0;
 383                         lnn++;
 384
 385                         /*
 386                          * If no manual parser has been assigned, then
 387                          * try to assign one in pset(), which may do
 388                          * nothing at all.  After this, parse the manual
 389                          * line accordingly.
 390                          */
 391
 392                         if ( ! (man || mdoc) && ! pset(ln->buf,
 393                                                 pos, curp, &man, &mdoc))
 394                                 return(0);
 395
 396                         pos = 0;
 397
 398                         if (man && ! man_parseln(man, lnn, ln->buf))
 399                                 return(0);
 400                         if (mdoc && ! mdoc_parseln(mdoc, lnn, ln->buf))
 401                                 return(0);
 402                 }
 403         }
 404
 405         /* Note that a parser may not have been assigned, yet. */
 406
 407         if ( ! (man || mdoc)) {
 408                 warnx("%s: not a manual", curp->file);
 409                 return(0);
 410         }
 411
 412         if (mdoc && ! mdoc_endparse(mdoc))
 413                 return(0);
 414         if (man && ! man_endparse(man))
 415                 return(0);
 416
 417         /*
 418          * If an output device hasn't been allocated, see if we should
 419          * do so now.  Note that not all outtypes have functions, so
 420          * this switch statement may be superfluous, but it's
 421          * low-overhead enough not to matter very much.
 422          */
 423
 424         if ( ! (curp->outman && curp->outmdoc)) {
 425                 switch (curp->outtype) {
 426                 case (OUTT_TREE):
 427                         curp->outman = tree_man;
 428                         curp->outmdoc = tree_mdoc;
 429                         break;
 430                 case (OUTT_LINT):
 431                         break;
 432                 default:
 433                         curp->outdata = ascii_alloc();
 434                         curp->outman = terminal_man;
 435                         curp->outmdoc = terminal_mdoc;
 436                         curp->outfree = terminal_free;
 437                         break;
 438                 }
 439         }
 440
 441         /* Execute the out device, if it exists. */
 442
 443         if (man && curp->outman)
 444                 if ( ! (*curp->outman)(curp->outdata, man))
 445                         return(0);
 446         if (mdoc && curp->outmdoc)
 447                 if ( ! (*curp->outmdoc)(curp->outdata, mdoc))
 448                         return(0);
 449
 450         return(1);
 451 }
 452
 453
 454 static int
 455 pset(const char *buf, int pos, struct curparse *curp,
 456                 struct man **man, struct mdoc **mdoc)
 457 {
 458
 459         /*
 460          * Try to intuit which kind of manual parser should be used.  If
 461          * passed in by command-line (-man, -mdoc), then use that
 462          * explicitly.  If passed as -mandoc, then try to guess from the
 463          * line: either skip comments, use -mdoc when finding `.Dt', or
 464          * default to -man, which is more lenient.
 465          */
 466
 467         if (pos >= 3 && 0 == memcmp(buf, ".\\\"", 3))
 468                 return(1);
 469
 470         switch (curp->inttype) {
 471         case (INTT_MDOC):
 472                 if (NULL == curp->mdoc)
 473                         curp->mdoc = mdoc_init(curp);
 474                 if (NULL == (*mdoc = curp->mdoc))
 475                         return(0);
 476                 curp->lastmdoc = *mdoc;
 477                 return(1);
 478         case (INTT_MAN):
 479                 if (NULL == curp->man)
 480                         curp->man = man_init(curp);
 481                 if (NULL == (*man = curp->man))
 482                         return(0);
 483                 curp->lastman = *man;
 484                 return(1);
 485         default:
 486                 break;
 487         }
 488
 489         if (pos >= 3 && 0 == memcmp(buf, ".Dd", 3))  {
 490                 if (NULL == curp->mdoc)
 491                         curp->mdoc = mdoc_init(curp);
 492                 if (NULL == (*mdoc = curp->mdoc))
 493                         return(0);
 494                 curp->lastmdoc = *mdoc;
 495                 return(1);
 496         }
 497
 498         if (NULL == curp->man)
 499                 curp->man = man_init(curp);
 500         if (NULL == (*man = curp->man))
 501                 return(0);
 502         curp->lastman = *man;
 503         return(1);
 504 }
 505
 506
 507 static int
 508 moptions(enum intt *tflags, char *arg)
 509 {
 510
 511         if (0 == strcmp(arg, "doc"))
 512                 *tflags = INTT_MDOC;
 513         else if (0 == strcmp(arg, "andoc"))
 514                 *tflags = INTT_AUTO;
 515         else if (0 == strcmp(arg, "an"))
 516                 *tflags = INTT_MAN;
 517         else {
 518                 warnx("bad argument: -m%s", arg);
 519                 return(0);
 520         }
 521
 522         return(1);
 523 }
 524
 525
 526 static int
 527 toptions(enum outt *tflags, char *arg)
 528 {
 529
 530         if (0 == strcmp(arg, "ascii"))
 531                 *tflags = OUTT_ASCII;
 532         else if (0 == strcmp(arg, "lint"))
 533                 *tflags = OUTT_LINT;
 534         else if (0 == strcmp(arg, "tree"))
 535                 *tflags = OUTT_TREE;
 536         else {
 537                 warnx("bad argument: -T%s", arg);
 538                 return(0);
 539         }
 540
 541         return(1);
 542 }
 543
 544
 545 /*
 546  * Parse out the options for [-fopt...] setting compiler options.  These
 547  * can be comma-delimited or called again.
 548  */
 549 static int
 550 foptions(int *fflags, char *arg)
 551 {
 552         char            *v;
 553         char            *toks[6];
 554
 555         toks[0] = "ign-scope";
 556         toks[1] = "no-ign-escape";
 557         toks[2] = "no-ign-macro";
 558         toks[3] = "no-ign-chars";
 559         toks[4] = "strict";
 560         toks[5] = NULL;
 561
 562         while (*arg)
 563                 switch (getsubopt(&arg, toks, &v)) {
 564                 case (0):
 565                         *fflags |= IGN_SCOPE;
 566                         break;
 567                 case (1):
 568                         *fflags |= NO_IGN_ESCAPE;
 569                         break;
 570                 case (2):
 571                         *fflags |= NO_IGN_MACRO;
 572                         break;
 573                 case (3):
 574                         *fflags |= NO_IGN_CHARS;
 575                         break;
 576                 case (4):
 577                         *fflags |= NO_IGN_ESCAPE |
 578                                    NO_IGN_MACRO | NO_IGN_CHARS;
 579                         break;
 580                 default:
 581                         warnx("bad argument: -f%s", arg);
 582                         return(0);
 583                 }
 584
 585         return(1);
 586 }
 587
 588
 589 /*
 590  * Parse out the options for [-Werr...], which sets warning modes.
 591  * These can be comma-delimited or called again.
 592  */
 593 static int
 594 woptions(int *wflags, char *arg)
 595 {
 596         char            *v;
 597         char            *toks[5];
 598
 599         toks[0] = "all";
 600         toks[1] = "compat";
 601         toks[2] = "syntax";
 602         toks[3] = "error";
 603         toks[4] = NULL;
 604
 605         while (*arg)
 606                 switch (getsubopt(&arg, toks, &v)) {
 607                 case (0):
 608                         *wflags |= WARN_WALL;
 609                         break;
 610                 case (1):
 611                         *wflags |= WARN_WCOMPAT;
 612                         break;
 613                 case (2):
 614                         *wflags |= WARN_WSYNTAX;
 615                         break;
 616                 case (3):
 617                         *wflags |= WARN_WERR;
 618                         break;
 619                 default:
 620                         warnx("bad argument: -W%s", arg);
 621                         return(0);
 622                 }
 623
 624         return(1);
 625 }
 626
 627
 628 /* ARGSUSED */
 629 static int
 630 merr(void *arg, int line, int col, const char *msg)
 631 {
 632         struct curparse *curp;
 633
 634         curp = (struct curparse *)arg;
 635
 636         warnx("%s:%d: error: %s (column %d)",
 637                         curp->file, line, msg, col);
 638         return(0);
 639 }
 640
 641
 642 static int
 643 mdocwarn(void *arg, int line, int col,
 644                 enum mdoc_warn type, const char *msg)
 645 {
 646         struct curparse *curp;
 647         char            *wtype;
 648
 649         curp = (struct curparse *)arg;
 650         wtype = NULL;
 651
 652         switch (type) {
 653         case (WARN_COMPAT):
 654                 wtype = "compat";
 655                 if (curp->wflags & WARN_WCOMPAT)
 656                         break;
 657                 return(1);
 658         case (WARN_SYNTAX):
 659                 wtype = "syntax";
 660                 if (curp->wflags & WARN_WSYNTAX)
 661                         break;
 662                 return(1);
 663         }
 664
 665         assert(wtype);
 666         warnx("%s:%d: %s warning: %s (column %d)",
 667                         curp->file, line, wtype, msg, col);
 668
 669         if ( ! (curp->wflags & WARN_WERR))
 670                 return(1);
 671
 672         warnx("%s: considering warnings as errors",
 673                         __progname);
 674         return(0);
 675 }
 676
 677
 678 static int
 679 manwarn(void *arg, int line, int col, const char *msg)
 680 {
 681         struct curparse *curp;
 682
 683         curp = (struct curparse *)arg;
 684
 685         if ( ! (curp->wflags & WARN_WSYNTAX))
 686                 return(1);
 687
 688         warnx("%s:%d: syntax warning: %s (column %d)",
 689                         curp->file, line, msg, col);
 690
 691         if ( ! (curp->wflags & WARN_WERR))
 692                 return(1);
 693
 694         warnx("%s: considering warnings as errors",
 695                         __progname);
 696         return(0);
 697 }