mdocml.c

   1 /* $Id: mdocml.c,v 1.33 2009/01/03 22:10:22 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_ELEM):
 210                 p = mdoc_macronames[n->data.elem.tok];
 211                 t = "element";
 212                 argv = n->data.elem.argv;
 213                 argc = n->data.elem.argc;
 214                 params = n->data.elem.args;
 215                 sz = n->data.elem.sz;
 216                 break;
 217         case (MDOC_BLOCK):
 218                 p = mdoc_macronames[n->data.block.tok];
 219                 t = "block";
 220                 argv = n->data.block.argv;
 221                 argc = n->data.block.argc;
 222                 break;
 223         default:
 224                 abort();
 225                 /* NOTREACHED */
 226         }
 227
 228         for (i = 0; i < indent; i++)
 229                 (void)printf("    ");
 230         (void)printf("%s (%s)", p, t);
 231
 232         for (i = 0; i < (int)argc; i++) {
 233                 (void)printf(" -%s", mdoc_argnames[argv[i].arg]);
 234                 for (j = 0; j < (int)argv[i].sz; j++)
 235                         (void)printf(" \"%s\"", argv[i].value[j]);
 236         }
 237
 238         for (i = 0; i < (int)sz; i++)
 239                 (void)printf(" \"%s\"", params[i]);
 240
 241         (void)printf("\n");
 242
 243         if (n->child)
 244                 print_node(n->child, indent + 1);
 245         if (n->next)
 246                 print_node(n->next, indent);
 247 }
 248
 249
 250 static int
 251 parse_leave(struct md_parse *p, int code)
 252 {
 253         const struct mdoc_node *n;
 254
 255         if (p->mdoc) {
 256                 if ((n = mdoc_result(p->mdoc)))
 257                         print_node(n, 0);
 258                 mdoc_free(p->mdoc);
 259         }
 260         return(code);
 261 }
 262
 263
 264 static int
 265 parse_begin(struct md_parse *p)
 266 {
 267         ssize_t          sz, i;
 268         size_t           pos;
 269         char             line[256], sv[256];
 270         struct mdoc_cb   cb;
 271
 272         cb.mdoc_err = msg_err;
 273         cb.mdoc_warn = msg_warn;
 274         cb.mdoc_msg = msg_msg;
 275
 276         if (NULL == (p->mdoc = mdoc_alloc(p, &cb)))
 277                 return(parse_leave(p, 0));
 278
 279         p->lnn = 1;
 280         p->line = sv;
 281
 282         for (pos = 0; ; ) {
 283                 if (-1 == (sz = read(p->fd, p->buf, p->bufsz))) {
 284                         warn("%s", p->name);
 285                         return(parse_leave(p, 0));
 286                 } else if (0 == sz)
 287                         break;
 288
 289                 for (i = 0; i < sz; i++) {
 290                         if ('\n' != p->buf[i]) {
 291                                 if (pos < sizeof(line)) {
 292                                         sv[(int)pos] = p->buf[(int)i];
 293                                         line[(int)pos++] =
 294                                                 p->buf[(int)i];
 295                                         continue;
 296                                 }
 297                                 warnx("%s: line %d too long",
 298                                                 p->name, p->lnn);
 299                                 return(parse_leave(p, 0));
 300                         }
 301
 302                         line[(int)pos] = sv[(int)pos] = 0;
 303                         if ( ! mdoc_parseln(p->mdoc, line))
 304                                 return(parse_leave(p, 0));
 305
 306                         p->lnn++;
 307                         pos = 0;
 308                 }
 309         }
 310
 311         return(parse_leave(p, 1));
 312 }
 313
 314
 315 static int
 316 msg_err(void *arg, int tok, int col, enum mdoc_err type)
 317 {
 318         char             *fmt, *lit;
 319         struct md_parse  *p;
 320         int               i;
 321
 322         p = (struct md_parse *)arg;
 323
 324         fmt = lit = NULL;
 325
 326         switch (type) {
 327         case (ERR_SYNTAX_QUOTE):
 328                 lit = "syntax: disallowed argument quotation";
 329                 break;
 330         case (ERR_SYNTAX_UNQUOTE):
 331                 lit = "syntax: unterminated quotation";
 332                 break;
 333         case (ERR_SYNTAX_WS):
 334                 lit = "syntax: whitespace in argument";
 335                 break;
 336         case (ERR_SYNTAX_ARGFORM):
 337                 fmt = "syntax: macro `%s' arguments malformed";
 338                 break;
 339         case (ERR_SYNTAX_NOPUNCT):
 340                 fmt = "syntax: macro `%s' doesn't understand punctuation";
 341                 break;
 342         case (ERR_SYNTAX_ARG):
 343                 fmt = "syntax: unknown argument for macro `%s'";
 344                 break;
 345         case (ERR_SCOPE_BREAK):
 346                 /* Which scope is broken? */
 347                 fmt = "scope: macro `%s' breaks prior explicit scope";
 348                 break;
 349         case (ERR_SCOPE_NOCTX):
 350                 fmt = "scope: closure macro `%s' has no context";
 351                 break;
 352         case (ERR_SCOPE_NONEST):
 353                 fmt = "scope: macro `%s' may not be nested in the current context";
 354                 break;
 355         case (ERR_MACRO_NOTSUP):
 356                 fmt = "macro `%s' not supported";
 357                 break;
 358         case (ERR_MACRO_NOTCALL):
 359                 fmt = "macro `%s' not callable";
 360                 break;
 361         case (ERR_SEC_PROLOGUE):
 362                 fmt = "macro `%s' cannot be called in the prologue";
 363                 break;
 364         case (ERR_SEC_NPROLOGUE):
 365                 fmt = "macro `%s' called outside of prologue";
 366                 break;
 367         case (ERR_ARGS_EQ0):
 368                 fmt = "macro `%s' expects zero arguments";
 369                 break;
 370         case (ERR_ARGS_EQ1):
 371                 fmt = "macro `%s' expects one argument";
 372                 break;
 373         case (ERR_ARGS_GE1):
 374                 fmt = "macro `%s' expects one or more arguments";
 375                 break;
 376         case (ERR_ARGS_LE2):
 377                 fmt = "macro `%s' expects two or fewer arguments";
 378                 break;
 379         case (ERR_ARGS_MANY):
 380                 fmt = "macro `%s' has too many arguments";
 381                 break;
 382         case (ERR_SEC_PROLOGUE_OO):
 383                 fmt = "prologue macro `%s' is out-of-order";
 384                 break;
 385         case (ERR_SEC_PROLOGUE_REP):
 386                 fmt = "prologue macro `%s' repeated";
 387                 break;
 388         case (ERR_SEC_NAME):
 389                 lit = "`NAME' section must be first";
 390                 break;
 391         case (ERR_SYNTAX_ARGVAL):
 392                 lit = "syntax: expected value for macro argument";
 393                 break;
 394         case (ERR_SYNTAX_ARGBAD):
 395                 lit = "syntax: invalid value for macro argument";
 396                 break;
 397         case (ERR_SYNTAX_ARGMANY):
 398                 lit = "syntax: too many values for macro argument";
 399                 break;
 400         case (ERR_SYNTAX_CHILDHEAD):
 401                 lit = "syntax: expected only block-header section";
 402                 break;
 403         case (ERR_SYNTAX_CHILDBODY):
 404                 lit = "syntax: expected only a block-body section";
 405                 break;
 406         case (ERR_SYNTAX_EMPTYHEAD):
 407                 lit = "syntax: block-header section may not be empty";
 408                 break;
 409         case (ERR_SYNTAX_EMPTYBODY):
 410                 lit = "syntax: block-body section may not be empty";
 411                 break;
 412         default:
 413                 abort();
 414                 /* NOTREACHED */
 415         }
 416
 417         if (fmt) {
 418                 (void)fprintf(stderr, "%s:%d: error: ",
 419                                 p->name, p->lnn);
 420                 (void)fprintf(stderr, fmt, mdoc_macronames[tok]);
 421         } else
 422                 (void)fprintf(stderr, "%s:%d: error: %s",
 423                                 p->name, p->lnn, lit);
 424
 425         if (p->dbg < 1) {
 426                 if (-1 != col)
 427                         (void)fprintf(stderr, " (column %d)\n", col);
 428                 return(0);
 429         } else if (-1 == col) {
 430                 (void)fprintf(stderr, "\nFrom: %s", p->line);
 431                 return(0);
 432         }
 433
 434         (void)fprintf(stderr, "\nFrom: %s\n      ", p->line);
 435         for (i = 0; i < col; i++)
 436                 (void)fprintf(stderr, " ");
 437         (void)fprintf(stderr, "^\n");
 438
 439         return(0);
 440 }
 441
 442
 443 static void
 444 msg_msg(void *arg, int col, const char *msg)
 445 {
 446         struct md_parse  *p;
 447         int               i;
 448
 449         p = (struct md_parse *)arg;
 450
 451         if (p->dbg < 2)
 452                 return;
 453
 454         (void)printf("%s:%d: %s", p->name, p->lnn, msg);
 455
 456         if (p->dbg < 3) {
 457                 if (-1 != col)
 458                         (void)printf(" (column %d)\n", col);
 459                 return;
 460         } else if (-1 == col) {
 461                 (void)printf("\nFrom %s\n", p->line);
 462                 return;
 463         }
 464
 465         (void)printf("\nFrom: %s\n      ", p->line);
 466         for (i = 0; i < col; i++)
 467                 (void)printf(" ");
 468         (void)printf("^\n");
 469 }
 470
 471
 472 static int
 473 msg_warn(void *arg, int tok, int col, enum mdoc_warn type)
 474 {
 475         char             *fmt, *lit;
 476         struct md_parse  *p;
 477         int               i;
 478         extern char      *__progname;
 479
 480         p = (struct md_parse *)arg;
 481
 482         if ( ! (p->warn & MD_WARN_ALL))
 483                 return(1);
 484
 485         fmt = lit = NULL;
 486
 487         switch (type) {
 488         case (WARN_SYNTAX_WS_EOLN):
 489                 lit = "syntax: whitespace at end-of-line";
 490                 break;
 491         case (WARN_SYNTAX_QUOTED):
 492                 lit = "syntax: quotation mark starting string";
 493                 break;
 494         case (WARN_SYNTAX_MACLIKE):
 495                 lit = "syntax: macro-like argument";
 496                 break;
 497         case (WARN_SYNTAX_ARGLIKE):
 498                 lit = "syntax: argument-like value";
 499                 break;
 500         case (WARN_SYNTAX_EMPTYBODY):
 501                 lit = "syntax: empty block-body section";
 502                 break;
 503         case (WARN_SEC_OO):
 504                 lit = "section is out of conventional order";
 505                 break;
 506         case (WARN_ARGS_GE1):
 507                 fmt = "macro `%s' suggests one or more arguments";
 508                 break;
 509         case (WARN_ARGS_EQ0):
 510                 fmt = "macro `%s' suggests zero arguments";
 511                 break;
 512         case (WARN_IGN_AFTER_BLK):
 513                 fmt = "ignore: macro `%s' ignored after block macro";
 514                 break;
 515         case (WARN_IGN_OBSOLETE):
 516                 fmt = "ignore: macro `%s' is obsolete";
 517                 break;
 518         case (WARN_IGN_BEFORE_BLK):
 519                 fmt = "ignore: macro before block macro `%s' ignored";
 520                 break;
 521         case (WARN_COMPAT_TROFF):
 522                 fmt = "compat: macro `%s' behaves differently in troff and nroff";
 523                 break;
 524         default:
 525                 abort();
 526                 /* NOTREACHED */
 527         }
 528
 529         if (fmt) {
 530                 (void)fprintf(stderr, "%s:%d: warning: ",
 531                                 p->name, p->lnn);
 532                 (void)fprintf(stderr, fmt, mdoc_macronames[tok]);
 533         } else
 534                 (void)fprintf(stderr, "%s:%d: warning: %s",
 535                                 p->name, p->lnn, lit);
 536
 537         if (p->dbg >= 1) {
 538                 (void)fprintf(stderr, "\nFrom: %s\n      ", p->line);
 539                 for (i = 0; i < col; i++)
 540                         (void)fprintf(stderr, " ");
 541                 (void)fprintf(stderr, "^\n");
 542         } else
 543                 (void)fprintf(stderr, " (column %d)\n", col);
 544
 545         if (p->warn & MD_WARN_ERR) {
 546                 (void)fprintf(stderr, "%s: considering warnings as "
 547                                 "errors\n", __progname);
 548                 return(0);
 549         }
 550
 551         return(1);
 552 }
 553
 554
 555 static void
 556 usage(void)
 557 {
 558         extern char     *__progname;
 559
 560         (void)fprintf(stderr, "usage: %s [-v] [-Wwarn...] [infile]\n",
 561                         __progname);
 562 }
 563