main.c

   1 /* $Id: main.c,v 1.8 2009/03/22 19:01:11 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
  31 #ifdef __linux__
  32 extern  int               getsubopt(char **, char * const *, char **);
  33 # ifndef __dead
  34 #  define __dead __attribute__((__noreturn__))
  35 # endif
  36 #endif
  37
  38 struct  buf {
  39         char             *buf;
  40         size_t            sz;
  41 };
  42
  43 struct  curparse {
  44         const char       *file;
  45         int               wflags;
  46 #define WARN_WALL         0x03          /* All-warnings mask. */
  47 #define WARN_WCOMPAT     (1 << 0)       /* Compatibility warnings. */
  48 #define WARN_WSYNTAX     (1 << 1)       /* Syntax warnings. */
  49 #define WARN_WERR        (1 << 2)       /* Warnings->errors. */
  50 };
  51
  52 enum outt {
  53         OUTT_ASCII,
  54         OUTT_LATIN1,
  55         OUTT_UTF8,
  56         OUTT_TREE,
  57         OUTT_LINT
  58 };
  59
  60 typedef int             (*out_run)(void *, const struct mdoc *);
  61 typedef void            (*out_free)(void *);
  62
  63 extern  char             *__progname;
  64
  65 extern  void             *ascii_alloc(void);
  66 extern  void             *latin1_alloc(void);
  67 extern  void             *utf8_alloc(void);
  68 extern  int               terminal_run(void *, const struct mdoc *);
  69 extern  int               tree_run(void *, const struct mdoc *);
  70 extern  void              terminal_free(void *);
  71
  72 __dead  static void       version(void);
  73 __dead  static void       usage(void);
  74 static  int               foptions(int *, char *);
  75 static  int               toptions(enum outt *, char *);
  76 static  int               woptions(int *, char *);
  77 static  int               merr(void *, int, int, const char *);
  78 static  int               mwarn(void *, int, int,
  79                                 enum mdoc_warn, const char *);
  80 static  int               file(struct buf *, struct buf *,
  81                                 const char *, struct mdoc *);
  82 static  int               fdesc(struct buf *, struct buf *,
  83                                 const char *, int, struct mdoc *);
  84
  85
  86 int
  87 main(int argc, char *argv[])
  88 {
  89         int              c, rc, fflags;
  90         struct mdoc_cb   cb;
  91         struct mdoc     *mdoc;
  92         void            *outdata;
  93         enum outt        outtype;
  94         struct buf       ln, blk;
  95         out_run          outrun;
  96         out_free         outfree;
  97         struct curparse  curp;
  98
  99         fflags = 0;
 100         outtype = OUTT_ASCII;
 101
 102         bzero(&curp, sizeof(struct curparse));
 103
 104         /* LINTED */
 105         while (-1 != (c = getopt(argc, argv, "f:VW:T:")))
 106                 switch (c) {
 107                 case ('f'):
 108                         if ( ! foptions(&fflags, optarg))
 109                                 return(0);
 110                         break;
 111                 case ('T'):
 112                         if ( ! toptions(&outtype, optarg))
 113                                 return(0);
 114                         break;
 115                 case ('W'):
 116                         if ( ! woptions(&curp.wflags, optarg))
 117                                 return(0);
 118                         break;
 119                 case ('V'):
 120                         version();
 121                         /* NOTREACHED */
 122                 default:
 123                         usage();
 124                         /* NOTREACHED */
 125                 }
 126
 127         argc -= optind;
 128         argv += optind;
 129
 130         /*
 131          * Allocate the appropriate front-end.  Note that utf8, ascii
 132          * and latin1 all resolve to the terminal front-end with
 133          * different encodings (see terminal.c).  Not all frontends have
 134          * cleanup or alloc routines.
 135          */
 136
 137         switch (outtype) {
 138         case (OUTT_LATIN1):
 139                 outdata = latin1_alloc();
 140                 outrun = terminal_run;
 141                 outfree = terminal_free;
 142                 break;
 143         case (OUTT_UTF8):
 144                 outdata = utf8_alloc();
 145                 outrun = terminal_run;
 146                 outfree = terminal_free;
 147                 break;
 148         case (OUTT_TREE):
 149                 outdata = NULL;
 150                 outrun = tree_run;
 151                 outfree = NULL;
 152                 break;
 153         case (OUTT_LINT):
 154                 outdata = NULL;
 155                 outrun = NULL;
 156                 outfree = NULL;
 157                 break;
 158         default:
 159                 outdata = ascii_alloc();
 160                 outrun = terminal_run;
 161                 outfree = terminal_free;
 162                 break;
 163         }
 164
 165         /*
 166          * All callbacks route into here, where we print them onto the
 167          * screen.  XXX - for now, no path for debugging messages.
 168          */
 169
 170         cb.mdoc_msg = NULL;
 171         cb.mdoc_err = merr;
 172         cb.mdoc_warn = mwarn;
 173
 174         bzero(&ln, sizeof(struct buf));
 175         bzero(&blk, sizeof(struct buf));
 176
 177         mdoc = mdoc_alloc(&curp, fflags, &cb);
 178
 179         /*
 180          * Loop around available files.
 181          */
 182
 183         if (NULL == *argv) {
 184                 curp.file = "<stdin>";
 185                 c = fdesc(&blk, &ln, "stdin", STDIN_FILENO, mdoc);
 186                 rc = 0;
 187                 if (c && NULL == outrun)
 188                         rc = 1;
 189                 else if (c && outrun && (*outrun)(outdata, mdoc))
 190                         rc = 1;
 191         } else {
 192                 while (*argv) {
 193                         curp.file = *argv;
 194                         c = file(&blk, &ln, *argv, mdoc);
 195                         if ( ! c)
 196                                 break;
 197                         if (outrun && ! (*outrun)(outdata, mdoc))
 198                                 break;
 199                         /* Reset the parser for another file. */
 200                         mdoc_reset(mdoc);
 201                         argv++;
 202                 }
 203                 rc = NULL == *argv;
 204         }
 205
 206         if (blk.buf)
 207                 free(blk.buf);
 208         if (ln.buf)
 209                 free(ln.buf);
 210         if (outfree)
 211                 (*outfree)(outdata);
 212
 213         mdoc_free(mdoc);
 214
 215         return(rc ? EXIT_SUCCESS : EXIT_FAILURE);
 216 }
 217
 218
 219 __dead static void
 220 version(void)
 221 {
 222
 223         (void)printf("%s %s\n", __progname, VERSION);
 224         exit(0);
 225         /* NOTREACHED */
 226 }
 227
 228
 229 __dead static void
 230 usage(void)
 231 {
 232
 233         (void)fprintf(stderr, "usage: %s\n", __progname);
 234         exit(1);
 235         /* NOTREACHED */
 236 }
 237
 238
 239 static int
 240 file(struct buf *blk, struct buf *ln,
 241                 const char *file, struct mdoc *mdoc)
 242 {
 243         int              fd, c;
 244
 245         if (-1 == (fd = open(file, O_RDONLY, 0))) {
 246                 warn("%s", file);
 247                 return(0);
 248         }
 249
 250         c = fdesc(blk, ln, file, fd, mdoc);
 251
 252         if (-1 == close(fd))
 253                 warn("%s", file);
 254
 255         return(c);
 256 }
 257
 258
 259 static int
 260 fdesc(struct buf *blk, struct buf *ln,
 261                 const char *f, int fd, struct mdoc *mdoc)
 262 {
 263         size_t           sz;
 264         ssize_t          ssz;
 265         struct stat      st;
 266         int              j, i, pos, lnn;
 267 #ifdef  STRIP_XO
 268         int              macro, xo, xeoln;
 269 #endif
 270
 271         /*
 272          * Two buffers: ln and buf.  buf is the input buffer, optimised
 273          * for each file's block size.  ln is a line buffer.  Both
 274          * growable, hence passed in by ptr-ptr.
 275          */
 276
 277         sz = BUFSIZ;
 278
 279         if (-1 == fstat(fd, &st))
 280                 warnx("%s", f);
 281         else if ((size_t)st.st_blksize > sz)
 282                 sz = st.st_blksize;
 283
 284         if (sz > blk->sz) {
 285                 blk->buf = realloc(blk->buf, sz);
 286                 if (NULL == blk->buf)
 287                         err(1, "realloc");
 288                 blk->sz = sz;
 289         }
 290
 291         /*
 292          * Fill buf with file blocksize and parse newlines into ln.
 293          */
 294 #ifdef  STRIP_XO
 295         macro = xo = xeoln = 0;
 296 #endif
 297
 298         for (lnn = 1, pos = 0; ; ) {
 299                 if (-1 == (ssz = read(fd, blk->buf, sz))) {
 300                         warn("%s", f);
 301                         return(0);
 302                 } else if (0 == ssz)
 303                         break;
 304
 305                 for (i = 0; i < (int)ssz; i++) {
 306                         if (pos >= (int)ln->sz) {
 307                                 ln->sz += 256; /* Step-size. */
 308                                 ln->buf = realloc(ln->buf, ln->sz);
 309                                 if (NULL == ln->buf)
 310                                         err(1, "realloc");
 311                         }
 312
 313                         if ('\n' != blk->buf[i]) {
 314                                 /*
 315                                  * Ugly of uglies.  Here we handle the
 316                                  * dreaded `Xo/Xc' scoping.  Cover the
 317                                  * eyes of any nearby children.  This
 318                                  * makes `Xo/Xc' enclosures look like
 319                                  * one huge line.
 320                                  */
 321 #ifdef  STRIP_XO
 322                                 /*
 323                                  * First, note whether we're in a macro
 324                                  * line.
 325                                  */
 326                                 if (0 == pos && '.' == blk->buf[i])
 327                                         macro = 1;
 328
 329                                 /*
 330                                  * If we're in an `Xo' context and just
 331                                  * nixed a newline, remove the control
 332                                  * character for new macro lines:
 333                                  * they're going to show up as all part
 334                                  * of the same line.
 335                                  */
 336                                 if (xo && xeoln && '.' == blk->buf[i]) {
 337                                         xeoln = 0;
 338                                         continue;
 339                                 }
 340                                 xeoln = 0;
 341
 342                                 /*
 343                                  * If we've parsed `Xo', enter an xo
 344                                  * context.  `Xo' must be in a parsable
 345                                  * state.  This is the ugly part.  IT IS
 346                                  * NOT SMART ENOUGH TO HANDLE ESCAPED
 347                                  * WHITESPACE.
 348                                  */
 349                                 if (macro && pos && 'o' == blk->buf[i]) {
 350                                         if (xo && 'X' == ln->buf[pos - 1])  {
 351                                                 if (' ' == ln->buf[pos - 2])
 352                                                         xo++;
 353                                         } else if ('X' == ln->buf[pos - 1]) {
 354                                                 if (2 == pos && '.' == ln->buf[pos - 2])
 355                                                         xo++;
 356                                                 else if (' ' == ln->buf[pos - 2])
 357                                                         xo++;
 358                                         }
 359                                 }
 360
 361                                 /*
 362                                  * If we're parsed `Xc', leave an xo
 363                                  * context if one's already pending.
 364                                  * `Xc' must be in a parsable state.
 365                                  * THIS IS NOT SMART ENOUGH TO HANDLE
 366                                  * ESCAPED WHITESPACE.
 367                                  */
 368                                 if (macro && pos && 'c' == blk->buf[i])
 369                                         if (xo && 'X' == ln->buf[pos - 1])
 370                                                 if (' ' == ln->buf[pos - 2])
 371                                                         xo--;
 372 #endif  /* STRIP_XO */
 373
 374                                 ln->buf[pos++] = blk->buf[i];
 375                                 continue;
 376                         }
 377
 378                         /* Check for CPP-escaped newline.  */
 379
 380                         if (pos > 0 && '\\' == ln->buf[pos - 1]) {
 381                                 for (j = pos - 1; j >= 0; j--)
 382                                         if ('\\' != ln->buf[j])
 383                                                 break;
 384
 385                                 if ( ! ((pos - j) % 2)) {
 386                                         pos--;
 387                                         lnn++;
 388                                         continue;
 389                                 }
 390                         }
 391
 392 #ifdef  STRIP_XO
 393                         /*
 394                          * If we're in an xo context, put a space in
 395                          * place of the newline and continue parsing.
 396                          * Mark that we just did a newline.
 397                          */
 398                         if (xo) {
 399                                 xeoln = 1;
 400                                 ln->buf[pos++] = ' ';
 401                                 lnn++;
 402                                 continue;
 403                         }
 404 #endif  /* STRIP_XO */
 405
 406                         ln->buf[pos] = 0;
 407                         if ( ! mdoc_parseln(mdoc, lnn, ln->buf))
 408                                 return(0);
 409                         lnn++;
 410                         macro = pos = 0;
 411                 }
 412         }
 413
 414         return(mdoc_endparse(mdoc));
 415 }
 416
 417
 418 static int
 419 toptions(enum outt *tflags, char *arg)
 420 {
 421
 422         if (0 == strcmp(arg, "ascii"))
 423                 *tflags = OUTT_ASCII;
 424         else if (0 == strcmp(arg, "latin1"))
 425                 *tflags = OUTT_LATIN1;
 426         else if (0 == strcmp(arg, "utf8"))
 427                 *tflags = OUTT_UTF8;
 428         else if (0 == strcmp(arg, "lint"))
 429                 *tflags = OUTT_LINT;
 430         else if (0 == strcmp(arg, "tree"))
 431                 *tflags = OUTT_TREE;
 432         else {
 433                 warnx("bad argument: -T%s", arg);
 434                 return(0);
 435         }
 436
 437         return(1);
 438 }
 439
 440
 441 /*
 442  * Parse out the options for [-fopt...] setting compiler options.  These
 443  * can be comma-delimited or called again.
 444  */
 445 static int
 446 foptions(int *fflags, char *arg)
 447 {
 448         char            *v;
 449         char            *toks[4];
 450
 451         toks[0] = "ign-scope";
 452         toks[1] = "ign-escape";
 453         toks[2] = "ign-macro";
 454         toks[3] = NULL;
 455
 456         while (*arg)
 457                 switch (getsubopt(&arg, toks, &v)) {
 458                 case (0):
 459                         *fflags |= MDOC_IGN_SCOPE;
 460                         break;
 461                 case (1):
 462                         *fflags |= MDOC_IGN_ESCAPE;
 463                         break;
 464                 case (2):
 465                         *fflags |= MDOC_IGN_MACRO;
 466                         break;
 467                 default:
 468                         warnx("bad argument: -f%s", arg);
 469                         return(0);
 470                 }
 471
 472         return(1);
 473 }
 474
 475
 476 /*
 477  * Parse out the options for [-Werr...], which sets warning modes.
 478  * These can be comma-delimited or called again.
 479  */
 480 static int
 481 woptions(int *wflags, char *arg)
 482 {
 483         char            *v;
 484         char            *toks[5];
 485
 486         toks[0] = "all";
 487         toks[1] = "compat";
 488         toks[2] = "syntax";
 489         toks[3] = "error";
 490         toks[4] = NULL;
 491
 492         while (*arg)
 493                 switch (getsubopt(&arg, toks, &v)) {
 494                 case (0):
 495                         *wflags |= WARN_WALL;
 496                         break;
 497                 case (1):
 498                         *wflags |= WARN_WCOMPAT;
 499                         break;
 500                 case (2):
 501                         *wflags |= WARN_WSYNTAX;
 502                         break;
 503                 case (3):
 504                         *wflags |= WARN_WERR;
 505                         break;
 506                 default:
 507                         warnx("bad argument: -W%s", arg);
 508                         return(0);
 509                 }
 510
 511         return(1);
 512 }
 513
 514
 515 /* ARGSUSED */
 516 static int
 517 merr(void *arg, int line, int col, const char *msg)
 518 {
 519         struct curparse *curp;
 520
 521         curp = (struct curparse *)arg;
 522
 523         warnx("%s:%d: error: %s (column %d)",
 524                         curp->file, line, msg, col);
 525         return(0);
 526 }
 527
 528
 529 static int
 530 mwarn(void *arg, int line, int col,
 531                 enum mdoc_warn type, const char *msg)
 532 {
 533         struct curparse *curp;
 534         char            *wtype;
 535
 536         curp = (struct curparse *)arg;
 537         wtype = NULL;
 538
 539         switch (type) {
 540         case (WARN_COMPAT):
 541                 wtype = "compat";
 542                 if (curp->wflags & WARN_WCOMPAT)
 543                         break;
 544                 return(1);
 545         case (WARN_SYNTAX):
 546                 wtype = "syntax";
 547                 if (curp->wflags & WARN_WSYNTAX)
 548                         break;
 549                 return(1);
 550         }
 551
 552         assert(wtype);
 553         warnx("%s:%d: %s warning: %s (column %d)",
 554                         curp->file, line, wtype, msg, col);
 555
 556         if ( ! (curp->wflags & WARN_WERR))
 557                 return(1);
 558
 559         warnx("%s: considering warnings as errors",
 560                         __progname);
 561         return(0);
 562 }
 563
 564