mdocml.c

   1 /* $Id: mdocml.c,v 1.34 2009/01/05 16:11:14 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, 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                 params = n->data.elem.args;
 219                 sz = n->data.elem.sz;
 220                 break;
 221         case (MDOC_BLOCK):
 222                 p = mdoc_macronames[n->data.block.tok];
 223                 t = "block";
 224                 argv = n->data.block.argv;
 225                 argc = n->data.block.argc;
 226                 break;
 227         default:
 228                 abort();
 229                 /* NOTREACHED */
 230         }
 231
 232         for (i = 0; i < indent; i++)
 233                 (void)printf("    ");
 234         (void)printf("%s (%s)", p, t);
 235
 236         for (i = 0; i < (int)argc; i++) {
 237                 (void)printf(" -%s", mdoc_argnames[argv[i].arg]);
 238                 for (j = 0; j < (int)argv[i].sz; j++)
 239                         (void)printf(" \"%s\"", argv[i].value[j]);
 240         }
 241
 242         for (i = 0; i < (int)sz; i++)
 243                 (void)printf(" \"%s\"", params[i]);
 244
 245         (void)printf("\n");
 246
 247         if (n->child)
 248                 print_node(n->child, indent + 1);
 249         if (n->next)
 250                 print_node(n->next, indent);
 251 }
 252
 253
 254 static int
 255 parse_leave(struct md_parse *p, int code)
 256 {
 257         const struct mdoc_node *n;
 258
 259         if (p->mdoc) {
 260                 if ((n = mdoc_result(p->mdoc)))
 261                         print_node(n, 0);
 262                 mdoc_free(p->mdoc);
 263         }
 264         return(code);
 265 }
 266
 267
 268 static int
 269 parse_begin(struct md_parse *p)
 270 {
 271         ssize_t          sz, i;
 272         size_t           pos;
 273         char             line[256], sv[256];
 274         struct mdoc_cb   cb;
 275
 276         cb.mdoc_err = msg_err;
 277         cb.mdoc_warn = msg_warn;
 278         cb.mdoc_msg = msg_msg;
 279
 280         if (NULL == (p->mdoc = mdoc_alloc(p, &cb)))
 281                 return(parse_leave(p, 0));
 282
 283         p->lnn = 1;
 284         p->line = sv;
 285
 286         for (pos = 0; ; ) {
 287                 if (-1 == (sz = read(p->fd, p->buf, p->bufsz))) {
 288                         warn("%s", p->name);
 289                         return(parse_leave(p, 0));
 290                 } else if (0 == sz)
 291                         break;
 292
 293                 for (i = 0; i < sz; i++) {
 294                         if ('\n' != p->buf[i]) {
 295                                 if (pos < sizeof(line)) {
 296                                         sv[(int)pos] = p->buf[(int)i];
 297                                         line[(int)pos++] =
 298                                                 p->buf[(int)i];
 299                                         continue;
 300                                 }
 301                                 warnx("%s: line %d too long",
 302                                                 p->name, p->lnn);
 303                                 return(parse_leave(p, 0));
 304                         }
 305
 306                         line[(int)pos] = sv[(int)pos] = 0;
 307                         if ( ! mdoc_parseln(p->mdoc, line))
 308                                 return(parse_leave(p, 0));
 309
 310                         p->lnn++;
 311                         pos = 0;
 312                 }
 313         }
 314
 315         return(parse_leave(p, 1));
 316 }
 317
 318
 319 static int
 320 msg_err(void *arg, int tok, int col, enum mdoc_err type)
 321 {
 322         char             *fmt, *lit;
 323         struct md_parse  *p;
 324         int               i;
 325
 326         p = (struct md_parse *)arg;
 327
 328         fmt = lit = NULL;
 329
 330         switch (type) {
 331         case (ERR_SYNTAX_QUOTE):
 332                 lit = "syntax: disallowed argument quotation";
 333                 break;
 334         case (ERR_SYNTAX_UNQUOTE):
 335                 lit = "syntax: unterminated quotation";
 336                 break;
 337         case (ERR_SYNTAX_WS):
 338                 lit = "syntax: whitespace in argument";
 339                 break;
 340         case (ERR_SYNTAX_ARGFORM):
 341                 fmt = "syntax: macro `%s' arguments malformed";
 342                 break;
 343         case (ERR_SYNTAX_NOPUNCT):
 344                 fmt = "syntax: macro `%s' doesn't understand punctuation";
 345                 break;
 346         case (ERR_SYNTAX_ARG):
 347                 fmt = "syntax: unknown argument for macro `%s'";
 348                 break;
 349         case (ERR_SCOPE_BREAK):
 350                 /* Which scope is broken? */
 351                 fmt = "scope: macro `%s' breaks prior explicit scope";
 352                 break;
 353         case (ERR_SCOPE_NOCTX):
 354                 fmt = "scope: closure macro `%s' has no context";
 355                 break;
 356         case (ERR_SCOPE_NONEST):
 357                 fmt = "scope: macro `%s' may not be nested in the current context";
 358                 break;
 359         case (ERR_MACRO_NOTSUP):
 360                 fmt = "macro `%s' not supported";
 361                 break;
 362         case (ERR_MACRO_NOTCALL):
 363                 fmt = "macro `%s' not callable";
 364                 break;
 365         case (ERR_SEC_PROLOGUE):
 366                 fmt = "macro `%s' cannot be called in the prologue";
 367                 break;
 368         case (ERR_SEC_NPROLOGUE):
 369                 fmt = "macro `%s' called outside of prologue";
 370                 break;
 371         case (ERR_ARGS_EQ0):
 372                 fmt = "macro `%s' expects zero arguments";
 373                 break;
 374         case (ERR_ARGS_EQ1):
 375                 fmt = "macro `%s' expects one argument";
 376                 break;
 377         case (ERR_ARGS_GE1):
 378                 fmt = "macro `%s' expects one or more arguments";
 379                 break;
 380         case (ERR_ARGS_LE2):
 381                 fmt = "macro `%s' expects two or fewer arguments";
 382                 break;
 383         case (ERR_ARGS_MANY):
 384                 fmt = "macro `%s' has too many arguments";
 385                 break;
 386         case (ERR_SEC_PROLOGUE_OO):
 387                 fmt = "prologue macro `%s' is out-of-order";
 388                 break;
 389         case (ERR_SEC_PROLOGUE_REP):
 390                 fmt = "prologue macro `%s' repeated";
 391                 break;
 392         case (ERR_SEC_NAME):
 393                 lit = "`NAME' section must be first";
 394                 break;
 395         case (ERR_SYNTAX_ARGVAL):
 396                 lit = "syntax: expected value for macro argument";
 397                 break;
 398         case (ERR_SYNTAX_ARGBAD):
 399                 lit = "syntax: invalid value for macro argument";
 400                 break;
 401         case (ERR_SYNTAX_ARGMANY):
 402                 lit = "syntax: too many values for macro argument";
 403                 break;
 404         case (ERR_SYNTAX_CHILDHEAD):
 405                 lit = "syntax: expected only block-header section";
 406                 break;
 407         case (ERR_SYNTAX_CHILDBODY):
 408                 lit = "syntax: expected only a block-body section";
 409                 break;
 410         case (ERR_SYNTAX_EMPTYHEAD):
 411                 lit = "syntax: block-header section may not be empty";
 412                 break;
 413         case (ERR_SYNTAX_EMPTYBODY):
 414                 lit = "syntax: block-body section may not be empty";
 415                 break;
 416         default:
 417                 abort();
 418                 /* NOTREACHED */
 419         }
 420
 421         if (fmt) {
 422                 (void)fprintf(stderr, "%s:%d: error: ",
 423                                 p->name, p->lnn);
 424                 (void)fprintf(stderr, fmt, mdoc_macronames[tok]);
 425         } else
 426                 (void)fprintf(stderr, "%s:%d: error: %s",
 427                                 p->name, p->lnn, lit);
 428
 429         if (p->dbg < 1) {
 430                 if (-1 != col)
 431                         (void)fprintf(stderr, " (column %d)\n", col);
 432                 return(0);
 433         } else if (-1 == col) {
 434                 (void)fprintf(stderr, "\nFrom: %s", p->line);
 435                 return(0);
 436         }
 437
 438         (void)fprintf(stderr, "\nFrom: %s\n      ", p->line);
 439         for (i = 0; i < col; i++)
 440                 (void)fprintf(stderr, " ");
 441         (void)fprintf(stderr, "^\n");
 442
 443         return(0);
 444 }
 445
 446
 447 static void
 448 msg_msg(void *arg, int col, const char *msg)
 449 {
 450         struct md_parse  *p;
 451         int               i;
 452
 453         p = (struct md_parse *)arg;
 454
 455         if (p->dbg < 2)
 456                 return;
 457
 458         (void)printf("%s:%d: %s", p->name, p->lnn, msg);
 459
 460         if (p->dbg < 3) {
 461                 if (-1 != col)
 462                         (void)printf(" (column %d)\n", col);
 463                 return;
 464         } else if (-1 == col) {
 465                 (void)printf("\nFrom %s\n", p->line);
 466                 return;
 467         }
 468
 469         (void)printf("\nFrom: %s\n      ", p->line);
 470         for (i = 0; i < col; i++)
 471                 (void)printf(" ");
 472         (void)printf("^\n");
 473 }
 474
 475
 476 static int
 477 msg_warn(void *arg, int tok, int col, enum mdoc_warn type)
 478 {
 479         char             *fmt, *lit;
 480         struct md_parse  *p;
 481         int               i;
 482         extern char      *__progname;
 483
 484         p = (struct md_parse *)arg;
 485
 486         if ( ! (p->warn & MD_WARN_ALL))
 487                 return(1);
 488
 489         fmt = lit = NULL;
 490
 491         switch (type) {
 492         case (WARN_SYNTAX_WS_EOLN):
 493                 lit = "syntax: whitespace at end-of-line";
 494                 break;
 495         case (WARN_SYNTAX_QUOTED):
 496                 lit = "syntax: quotation mark starting string";
 497                 break;
 498         case (WARN_SYNTAX_MACLIKE):
 499                 lit = "syntax: macro-like argument";
 500                 break;
 501         case (WARN_SYNTAX_ARGLIKE):
 502                 lit = "syntax: argument-like value";
 503                 break;
 504         case (WARN_SYNTAX_EMPTYBODY):
 505                 lit = "syntax: empty block-body section";
 506                 break;
 507         case (WARN_SEC_OO):
 508                 lit = "section is out of conventional order";
 509                 break;
 510         case (WARN_ARGS_GE1):
 511                 fmt = "macro `%s' suggests one or more arguments";
 512                 break;
 513         case (WARN_ARGS_EQ0):
 514                 fmt = "macro `%s' suggests zero arguments";
 515                 break;
 516         case (WARN_IGN_AFTER_BLK):
 517                 fmt = "ignore: macro `%s' ignored after block macro";
 518                 break;
 519         case (WARN_IGN_OBSOLETE):
 520                 fmt = "ignore: macro `%s' is obsolete";
 521                 break;
 522         case (WARN_IGN_BEFORE_BLK):
 523                 fmt = "ignore: macro before block macro `%s' ignored";
 524                 break;
 525         case (WARN_COMPAT_TROFF):
 526                 fmt = "compat: macro `%s' behaves differently in troff and nroff";
 527                 break;
 528         default:
 529                 abort();
 530                 /* NOTREACHED */
 531         }
 532
 533         if (fmt) {
 534                 (void)fprintf(stderr, "%s:%d: warning: ",
 535                                 p->name, p->lnn);
 536                 (void)fprintf(stderr, fmt, mdoc_macronames[tok]);
 537         } else
 538                 (void)fprintf(stderr, "%s:%d: warning: %s",
 539                                 p->name, p->lnn, lit);
 540
 541         if (p->dbg >= 1) {
 542                 (void)fprintf(stderr, "\nFrom: %s\n      ", p->line);
 543                 for (i = 0; i < col; i++)
 544                         (void)fprintf(stderr, " ");
 545                 (void)fprintf(stderr, "^\n");
 546         } else
 547                 (void)fprintf(stderr, " (column %d)\n", col);
 548
 549         if (p->warn & MD_WARN_ERR) {
 550                 (void)fprintf(stderr, "%s: considering warnings as "
 551                                 "errors\n", __progname);
 552                 return(0);
 553         }
 554
 555         return(1);
 556 }
 557
 558
 559 static void
 560 usage(void)
 561 {
 562         extern char     *__progname;
 563
 564         (void)fprintf(stderr, "usage: %s [-v] [-Wwarn...] [infile]\n",
 565                         __progname);
 566 }
 567