mdocml.c

   1 /* $Id: mdocml.c,v 1.39 2009/01/12 10:31:53 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008 Kristaps Dzonsons <kristaps@kth.se>
   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 #include <sys/param.h>
  21
  22 #include <assert.h>
  23 #include <fcntl.h>
  24 #include <err.h>
  25 #include <getopt.h>
  26 #include <stdio.h>
  27 #include <stdlib.h>
  28 #include <string.h>
  29 #include <unistd.h>
  30
  31 #include "mdoc.h"
  32
  33 #define MD_LINE_SZ      (256)
  34
  35 struct  md_parse {
  36         int              warn;
  37 #define MD_WARN_ALL     (1 << 0)
  38 #define MD_WARN_ERR     (1 << 1)
  39         int              dbg;
  40         struct mdoc     *mdoc;
  41         char            *buf;
  42         u_long           bufsz;
  43         char            *name;
  44         int              fd;
  45         int              lnn;
  46         char            *line;
  47 };
  48
  49 static  void             usage(void);
  50
  51 static  int              parse_begin(struct md_parse *);
  52 static  int              parse_leave(struct md_parse *, int);
  53 static  int              io_begin(struct md_parse *);
  54 static  int              io_leave(struct md_parse *, int);
  55 static  int              buf_begin(struct md_parse *);
  56 static  int              buf_leave(struct md_parse *, int);
  57
  58 static  int              msg_err(void *, int, int, enum mdoc_err);
  59 static  int              msg_warn(void *, int, int, enum mdoc_warn);
  60 static  void             msg_msg(void *, int, int, const char *);
  61
  62 #ifdef __linux__
  63 extern  int              getsubopt(char **, char *const *, char **);
  64 #endif
  65
  66 int
  67 main(int argc, char *argv[])
  68 {
  69         int              c;
  70         struct md_parse  parser;
  71         char            *opts, *v;
  72 #define ALL              0
  73 #define ERROR            1
  74         char            *toks[] = { "all", "error", NULL };
  75
  76         extern char     *optarg;
  77         extern int       optind;
  78
  79         (void)memset(&parser, 0, sizeof(struct md_parse));
  80
  81         while (-1 != (c = getopt(argc, argv, "vW:")))
  82                 switch (c) {
  83                 case ('v'):
  84                         parser.dbg++;
  85                         break;
  86                 case ('W'):
  87                         opts = optarg;
  88                         while (*opts)
  89                                 switch (getsubopt(&opts, toks, &v)) {
  90                                 case (ALL):
  91                                         parser.warn |= MD_WARN_ALL;
  92                                         break;
  93                                 case (ERROR):
  94                                         parser.warn |= MD_WARN_ERR;
  95                                         break;
  96                                 default:
  97                                         usage();
  98                                         return(1);
  99                                 }
 100                         break;
 101                 default:
 102                         usage();
 103                         return(1);
 104                 }
 105
 106         argv += optind;
 107         argc -= optind;
 108
 109         parser.name = "-";
 110         if (1 == argc)
 111                 parser.name = *argv++;
 112
 113         if ( ! io_begin(&parser))
 114                 return(EXIT_FAILURE);
 115
 116         return(EXIT_SUCCESS);
 117 }
 118
 119
 120 static int
 121 io_leave(struct md_parse *p, int code)
 122 {
 123
 124         if (-1 == p->fd || STDIN_FILENO == p->fd)
 125                 return(code);
 126
 127         if (-1 == close(p->fd)) {
 128                 warn("%s", p->name);
 129                 code = 0;
 130         }
 131         return(code);
 132 }
 133
 134
 135 static int
 136 io_begin(struct md_parse *p)
 137 {
 138
 139         p->fd = STDIN_FILENO;
 140         if (0 != strncmp(p->name, "-", 1))
 141                 if (-1 == (p->fd = open(p->name, O_RDONLY, 0))) {
 142                         warn("%s", p->name);
 143                         return(io_leave(p, 0));
 144                 }
 145
 146         return(io_leave(p, buf_begin(p)));
 147 }
 148
 149
 150 static int
 151 buf_leave(struct md_parse *p, int code)
 152 {
 153
 154         if (p->buf)
 155                 free(p->buf);
 156         return(code);
 157 }
 158
 159
 160 static int
 161 buf_begin(struct md_parse *p)
 162 {
 163         struct stat      st;
 164
 165         if (-1 == fstat(p->fd, &st)) {
 166                 warn("%s", p->name);
 167                 return(1);
 168         }
 169
 170         p->bufsz = MAX(st.st_blksize, BUFSIZ);
 171
 172         if (NULL == (p->buf = malloc(p->bufsz))) {
 173                 warn("malloc");
 174                 return(buf_leave(p, 0));
 175         }
 176
 177         return(buf_leave(p, parse_begin(p)));
 178 }
 179
 180
 181 static void
 182 print_node(const struct mdoc_node *n, int indent)
 183 {
 184         const char       *p, *t;
 185         int               i, j;
 186         size_t            argc, sz;
 187         char            **params;
 188         struct mdoc_arg  *argv;
 189
 190         argv = NULL;
 191         argc = 0;
 192         params = NULL;
 193         sz = 0;
 194
 195         switch (n->type) {
 196         case (MDOC_TEXT):
 197                 assert(NULL == n->child);
 198                 p = n->data.text.string;
 199                 t = "text";
 200                 break;
 201         case (MDOC_BODY):
 202                 p = mdoc_macronames[n->data.body.tok];
 203                 t = "block-body";
 204                 break;
 205         case (MDOC_HEAD):
 206                 p = mdoc_macronames[n->data.head.tok];
 207                 t = "block-head";
 208                 break;
 209         case (MDOC_TAIL):
 210                 p = mdoc_macronames[n->data.tail.tok];
 211                 t = "block-tail";
 212                 break;
 213         case (MDOC_ELEM):
 214                 p = mdoc_macronames[n->data.elem.tok];
 215                 t = "element";
 216                 argv = n->data.elem.argv;
 217                 argc = n->data.elem.argc;
 218                 break;
 219         case (MDOC_BLOCK):
 220                 p = mdoc_macronames[n->data.block.tok];
 221                 t = "block";
 222                 argv = n->data.block.argv;
 223                 argc = n->data.block.argc;
 224                 break;
 225         case (MDOC_ROOT):
 226                 p = "root";
 227                 t = "root";
 228                 break;
 229         default:
 230                 abort();
 231                 /* NOTREACHED */
 232         }
 233
 234         for (i = 0; i < indent; i++)
 235                 (void)printf("    ");
 236         (void)printf("%s (%s)", p, t);
 237
 238         for (i = 0; i < (int)argc; i++) {
 239                 (void)printf(" -%s", mdoc_argnames[argv[i].arg]);
 240                 for (j = 0; j < (int)argv[i].sz; j++)
 241                         (void)printf(" \"%s\"", argv[i].value[j]);
 242         }
 243
 244         for (i = 0; i < (int)sz; i++)
 245                 (void)printf(" \"%s\"", params[i]);
 246
 247         (void)printf(" %d:%d\n", n->line, n->pos);
 248
 249         if (n->child)
 250                 print_node(n->child, indent + 1);
 251         if (n->next)
 252                 print_node(n->next, indent);
 253 }
 254
 255
 256 static int
 257 parse_leave(struct md_parse *p, int code)
 258 {
 259         const struct mdoc_node *n;
 260
 261         if (NULL == p->mdoc)
 262                 return(code);
 263
 264         if ( ! mdoc_endparse(p->mdoc))
 265                 code = 0;
 266         if ((n = mdoc_result(p->mdoc)))
 267                 print_node(n, 0);
 268
 269         mdoc_free(p->mdoc);
 270
 271         return(code);
 272 }
 273
 274
 275 static int
 276 parse_begin(struct md_parse *p)
 277 {
 278         ssize_t          sz, i;
 279         size_t           pos;
 280         char             line[256], sv[256];
 281         struct mdoc_cb   cb;
 282
 283         cb.mdoc_err = msg_err;
 284         cb.mdoc_warn = msg_warn;
 285         cb.mdoc_msg = msg_msg;
 286
 287         if (NULL == (p->mdoc = mdoc_alloc(p, &cb)))
 288                 return(parse_leave(p, 0));
 289
 290         p->lnn = 1;
 291         p->line = sv;
 292
 293         for (pos = 0; ; ) {
 294                 if (-1 == (sz = read(p->fd, p->buf, p->bufsz))) {
 295                         warn("%s", p->name);
 296                         return(parse_leave(p, 0));
 297                 } else if (0 == sz)
 298                         break;
 299
 300                 for (i = 0; i < sz; i++) {
 301                         if ('\n' != p->buf[i]) {
 302                                 if (pos < sizeof(line)) {
 303                                         sv[(int)pos] = p->buf[(int)i];
 304                                         line[(int)pos++] =
 305                                                 p->buf[(int)i];
 306                                         continue;
 307                                 }
 308                                 warnx("%s: line %d too long",
 309                                                 p->name, p->lnn);
 310                                 return(parse_leave(p, 0));
 311                         }
 312
 313                         line[(int)pos] = sv[(int)pos] = 0;
 314                         if ( ! mdoc_parseln(p->mdoc, p->lnn, line))
 315                                 return(parse_leave(p, 0));
 316
 317                         p->lnn++;
 318                         pos = 0;
 319                 }
 320         }
 321
 322         return(parse_leave(p, 1));
 323 }
 324
 325
 326 static int
 327 msg_err(void *arg, int line, int col, enum mdoc_err type)
 328 {
 329         char             *lit;
 330         struct md_parse  *p;
 331
 332         p = (struct md_parse *)arg;
 333
 334         lit = NULL;
 335
 336         switch (type) {
 337         case (ERR_SYNTAX_NOTEXT):
 338                 lit = "syntax: context-free text disallowed";
 339                 break;
 340         case (ERR_SYNTAX_QUOTE):
 341                 lit = "syntax: disallowed argument quotation";
 342                 break;
 343         case (ERR_SYNTAX_UNQUOTE):
 344                 lit = "syntax: unterminated quotation";
 345                 break;
 346         case (ERR_SYNTAX_WS):
 347                 lit = "syntax: whitespace in argument";
 348                 break;
 349         case (ERR_SYNTAX_ARGFORM):
 350                 lit = "syntax: macro arguments malformed";
 351                 break;
 352         case (ERR_SYNTAX_NOPUNCT):
 353                 lit = "syntax: macro doesn't understand punctuation";
 354                 break;
 355         case (ERR_SYNTAX_ARG):
 356                 lit = "syntax: unknown argument for macro";
 357                 break;
 358         case (ERR_SCOPE_BREAK):
 359                 /* Which scope is broken? */
 360                 lit = "scope: macro breaks prior explicit scope";
 361                 break;
 362         case (ERR_SCOPE_NOCTX):
 363                 lit = "scope: closure macro has no context";
 364                 break;
 365         case (ERR_SCOPE_NONEST):
 366                 lit = "scope: macro may not be nested in the current context";
 367                 break;
 368         case (ERR_MACRO_NOTSUP):
 369                 lit = "macro not supported";
 370                 break;
 371         case (ERR_MACRO_NOTCALL):
 372                 lit = "macro not callable";
 373                 break;
 374         case (ERR_SEC_PROLOGUE):
 375                 lit = "macro cannot be called in the prologue";
 376                 break;
 377         case (ERR_SEC_NPROLOGUE):
 378                 lit = "macro called outside of prologue";
 379                 break;
 380         case (ERR_ARGS_EQ0):
 381                 lit = "macro expects zero arguments";
 382                 break;
 383         case (ERR_ARGS_EQ1):
 384                 lit = "macro expects one argument";
 385                 break;
 386         case (ERR_ARGS_GE1):
 387                 lit = "macro expects one or more arguments";
 388                 break;
 389         case (ERR_ARGS_LE2):
 390                 lit = "macro expects two or fewer arguments";
 391                 break;
 392         case (ERR_ARGS_LE8):
 393                 lit = "macro expects eight or fewer arguments";
 394                 break;
 395         case (ERR_ARGS_MANY):
 396                 lit = "macro has too many arguments";
 397                 break;
 398         case (ERR_SEC_PROLOGUE_OO):
 399                 lit = "prologue macro is out-of-order";
 400                 break;
 401         case (ERR_SEC_PROLOGUE_REP):
 402                 lit = "prologue macro repeated";
 403                 break;
 404         case (ERR_SEC_NAME):
 405                 lit = "`NAME' section must be first";
 406                 break;
 407         case (ERR_SYNTAX_ARGVAL):
 408                 lit = "syntax: expected value for macro argument";
 409                 break;
 410         case (ERR_SYNTAX_ARGBAD):
 411                 lit = "syntax: invalid value(s) for macro argument";
 412                 break;
 413         case (ERR_SYNTAX_ARGMISS):
 414                 lit = "syntax: missing required argument(s) for macro";
 415                 break;
 416         case (ERR_SYNTAX_ARGMANY):
 417                 lit = "syntax: too many values for macro argument";
 418                 break;
 419         case (ERR_SYNTAX_CHILDBAD):
 420                 lit = "syntax: invalid child for parent macro";
 421                 break;
 422         case (ERR_SYNTAX_CHILDHEAD):
 423                 lit = "syntax: expected only block-header section";
 424                 break;
 425         case (ERR_SYNTAX_CHILDBODY):
 426                 lit = "syntax: expected only a block-body section";
 427                 break;
 428         case (ERR_SYNTAX_EMPTYHEAD):
 429                 lit = "syntax: block-header section may not be empty";
 430                 break;
 431         case (ERR_SYNTAX_EMPTYBODY):
 432                 lit = "syntax: block-body section may not be empty";
 433                 break;
 434         default:
 435                 abort();
 436                 /* NOTREACHED */
 437         }
 438
 439         (void)fprintf(stderr, "%s:%d: error: %s (column %d)\n",
 440                         p->name, line, lit, col);
 441         return(0);
 442 }
 443
 444
 445 static void
 446 msg_msg(void *arg, int line, int col, const char *msg)
 447 {
 448         struct md_parse  *p;
 449
 450         p = (struct md_parse *)arg;
 451
 452         if (p->dbg < 2)
 453                 return;
 454
 455         (void)printf("%s:%d: %s (column %d)\n",
 456                         p->name, line, msg, col);
 457 }
 458
 459
 460 static int
 461 msg_warn(void *arg, int line, int col, enum mdoc_warn type)
 462 {
 463         char             *lit;
 464         struct md_parse  *p;
 465         extern char      *__progname;
 466
 467         p = (struct md_parse *)arg;
 468
 469         if ( ! (p->warn & MD_WARN_ALL))
 470                 return(1);
 471
 472         lit = NULL;
 473
 474         switch (type) {
 475         case (WARN_SYNTAX_WS_EOLN):
 476                 lit = "syntax: whitespace at end-of-line";
 477                 break;
 478         case (WARN_SYNTAX_QUOTED):
 479                 lit = "syntax: quotation mark starting string";
 480                 break;
 481         case (WARN_SYNTAX_MACLIKE):
 482                 lit = "syntax: macro-like argument";
 483                 break;
 484         case (WARN_SYNTAX_ARGLIKE):
 485                 lit = "syntax: argument-like value";
 486                 break;
 487         case (WARN_SYNTAX_EMPTYBODY):
 488                 lit = "syntax: empty block-body section";
 489                 break;
 490         case (WARN_SEC_OO):
 491                 lit = "section is out of conventional order";
 492                 break;
 493         case (WARN_SEC_REP):
 494                 lit = "section repeated";
 495                 break;
 496         case (WARN_ARGS_GE1):
 497                 lit = "macro suggests one or more arguments";
 498                 break;
 499         case (WARN_ARGS_EQ0):
 500                 lit = "macro suggests zero arguments";
 501                 break;
 502         case (WARN_IGN_AFTER_BLK):
 503                 lit = "ignore: macro ignored after block macro";
 504                 break;
 505         case (WARN_IGN_OBSOLETE):
 506                 lit = "ignore: macro is obsolete";
 507                 break;
 508         case (WARN_IGN_BEFORE_BLK):
 509                 lit = "ignore: macro before block macro ignored";
 510                 break;
 511         case (WARN_COMPAT_TROFF):
 512                 lit = "compat: macro behaves differently in troff and nroff";
 513                 break;
 514         default:
 515                 abort();
 516                 /* NOTREACHED */
 517         }
 518
 519
 520         (void)fprintf(stderr, "%s:%d: warning: %s (column %d)\n",
 521                         p->name, line, lit, col);
 522
 523         if (p->warn & MD_WARN_ERR) {
 524                 (void)fprintf(stderr, "%s: considering warnings as "
 525                                 "errors\n", __progname);
 526                 return(0);
 527         }
 528
 529         return(1);
 530 }
 531
 532
 533 static void
 534 usage(void)
 535 {
 536         extern char     *__progname;
 537
 538         (void)fprintf(stderr, "usage: %s [-v] [-Wwarn...] [infile]\n",
 539                         __progname);
 540 }
 541