argv.c

   1 /* $Id: argv.c,v 1.17 2009/01/19 17:02:58 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 <assert.h>
  20 #include <ctype.h>
  21 #include <err.h>
  22 #include <stdlib.h>
  23 #include <stdio.h>
  24 #include <string.h>
  25
  26 #include "private.h"
  27
  28
  29 /* FIXME: .It called with -column and quoted arguments. */
  30
  31 static  int              lookup(int, const char *);
  32 static  int              parse(struct mdoc *, int,
  33                                 struct mdoc_arg *, int *, char *);
  34 static  int              parse_single(struct mdoc *, int,
  35                                 struct mdoc_arg *, int *, char *);
  36 static  int              parse_multi(struct mdoc *, int,
  37                                 struct mdoc_arg *, int *, char *);
  38 static  int              postparse(struct mdoc *, int,
  39                                 const struct mdoc_arg *, int);
  40
  41
  42 int
  43 mdoc_args(struct mdoc *mdoc, int line, int *pos, char *buf, int fl, char **v)
  44 {
  45         int              i;
  46
  47         if (0 == buf[*pos])
  48                 return(ARGS_EOLN);
  49
  50         if ('\"' == buf[*pos] && ! (fl & ARGS_QUOTED))
  51                 if ( ! mdoc_pwarn(mdoc, line, *pos, WARN_SYNTAX, "unexpected quoted parameter"))
  52                         return(ARGS_ERROR);
  53
  54         if ('-' == buf[*pos])
  55                 if ( ! mdoc_pwarn(mdoc, line, *pos, WARN_SYNTAX, "argument-like parameter"))
  56                         return(ARGS_ERROR);
  57
  58         if ((fl & ARGS_DELIM) && mdoc_iscdelim(buf[*pos])) {
  59                 /*
  60                  * If ARGS_DELIM, return ARGS_PUNCT if only space-separated
  61                  * punctuation remains.
  62                  */
  63                 for (i = *pos; buf[i]; ) {
  64                         if ( ! mdoc_iscdelim(buf[i]))
  65                                 break;
  66                         i++;
  67                         if (0 == buf[i] || ! isspace(buf[i]))
  68                                 break;
  69                         i++;
  70                         while (buf[i] && isspace(buf[i]))
  71                                 i++;
  72                 }
  73                 if (0 == buf[i]) {
  74                         *v = &buf[*pos];
  75                         return(ARGS_PUNCT);
  76                 }
  77         }
  78
  79         /* Parse routine for non-quoted string. */
  80
  81         assert(*pos > 0);
  82         if ('\"' != buf[*pos] || ! (ARGS_QUOTED & fl)) {
  83                 *v = &buf[*pos];
  84
  85                 /* FIXME: UGLY tab-sep processing. */
  86
  87                 if (ARGS_TABSEP & fl)
  88                         while (buf[*pos]) {
  89                                 if ('\t' == buf[*pos])
  90                                         break;
  91                                 if ('T' == buf[*pos]) {
  92                                         (*pos)++;
  93                                         if (0 == buf[*pos])
  94                                                 break;
  95                                         if ('a' == buf[*pos]) {
  96                                                 buf[*pos - 1] = 0;
  97                                                 break;
  98                                         }
  99                                 }
 100                                 (*pos)++;
 101                         }
 102                 else {
 103                         while (buf[*pos]) {
 104                                 if (isspace(buf[*pos]))
 105                                         if ('\\' != buf[*pos - 1])
 106                                                 break;
 107                                 (*pos)++;
 108                         }
 109                 }
 110
 111                 if (0 == buf[*pos])
 112                         return(ARGS_WORD);
 113
 114                 buf[(*pos)++] = 0;
 115
 116                 if (0 == buf[*pos])
 117                         return(ARGS_WORD);
 118
 119                 if ( ! (ARGS_TABSEP & fl))
 120                         while (buf[*pos] && isspace(buf[*pos]))
 121                                 (*pos)++;
 122
 123                 if (buf[*pos])
 124                         return(ARGS_WORD);
 125
 126                 if ( ! mdoc_pwarn(mdoc, line, *pos, WARN_COMPAT, "whitespace at end-of-line"))
 127                         return(ARGS_ERROR);
 128
 129                 return(ARGS_WORD);
 130         }
 131
 132         /*
 133          * If we're a quoted string (and quoted strings are allowed),
 134          * then parse ahead to the next quote.  If none's found, it's an
 135          * error.  After, parse to the next word.
 136          */
 137
 138         *v = &buf[++(*pos)];
 139
 140         while (buf[*pos] && '\"' != buf[*pos])
 141                 (*pos)++;
 142
 143         if (0 == buf[*pos]) {
 144                 (void)mdoc_perr(mdoc, line, *pos, "unterminated quoted parameter");
 145                 return(ARGS_ERROR);
 146         }
 147
 148         buf[(*pos)++] = 0;
 149         if (0 == buf[*pos])
 150                 return(ARGS_WORD);
 151
 152         while (buf[*pos] && isspace(buf[*pos]))
 153                 (*pos)++;
 154
 155         if (buf[*pos])
 156                 return(ARGS_WORD);
 157
 158         if ( ! mdoc_pwarn(mdoc, line, *pos, WARN_COMPAT, "whitespace at end-of-line"))
 159                 return(ARGS_ERROR);
 160
 161         return(ARGS_WORD);
 162 }
 163
 164
 165 static int
 166 lookup(int tok, const char *argv)
 167 {
 168
 169         switch (tok) {
 170         case (MDOC_An):
 171                 if (xstrcmp(argv, "split"))
 172                         return(MDOC_Split);
 173                 else if (xstrcmp(argv, "nosplit"))
 174                         return(MDOC_Nosplit);
 175                 break;
 176
 177         case (MDOC_Bd):
 178                 if (xstrcmp(argv, "ragged"))
 179                         return(MDOC_Ragged);
 180                 else if (xstrcmp(argv, "unfilled"))
 181                         return(MDOC_Unfilled);
 182                 else if (xstrcmp(argv, "filled"))
 183                         return(MDOC_Filled);
 184                 else if (xstrcmp(argv, "literal"))
 185                         return(MDOC_Literal);
 186                 else if (xstrcmp(argv, "file"))
 187                         return(MDOC_File);
 188                 else if (xstrcmp(argv, "offset"))
 189                         return(MDOC_Offset);
 190                 break;
 191
 192         case (MDOC_Bf):
 193                 if (xstrcmp(argv, "emphasis"))
 194                         return(MDOC_Emphasis);
 195                 else if (xstrcmp(argv, "literal"))
 196                         return(MDOC_Literal);
 197                 else if (xstrcmp(argv, "symbolic"))
 198                         return(MDOC_Symbolic);
 199                 break;
 200
 201         case (MDOC_Bk):
 202                 if (xstrcmp(argv, "words"))
 203                         return(MDOC_Words);
 204                 break;
 205
 206         case (MDOC_Bl):
 207                 if (xstrcmp(argv, "bullet"))
 208                         return(MDOC_Bullet);
 209                 else if (xstrcmp(argv, "dash"))
 210                         return(MDOC_Dash);
 211                 else if (xstrcmp(argv, "hyphen"))
 212                         return(MDOC_Hyphen);
 213                 else if (xstrcmp(argv, "item"))
 214                         return(MDOC_Item);
 215                 else if (xstrcmp(argv, "enum"))
 216                         return(MDOC_Enum);
 217                 else if (xstrcmp(argv, "tag"))
 218                         return(MDOC_Tag);
 219                 else if (xstrcmp(argv, "diag"))
 220                         return(MDOC_Diag);
 221                 else if (xstrcmp(argv, "hang"))
 222                         return(MDOC_Hang);
 223                 else if (xstrcmp(argv, "ohang"))
 224                         return(MDOC_Ohang);
 225                 else if (xstrcmp(argv, "inset"))
 226                         return(MDOC_Inset);
 227                 else if (xstrcmp(argv, "column"))
 228                         return(MDOC_Column);
 229                 else if (xstrcmp(argv, "width"))
 230                         return(MDOC_Width);
 231                 else if (xstrcmp(argv, "offset"))
 232                         return(MDOC_Offset);
 233                 else if (xstrcmp(argv, "compact"))
 234                         return(MDOC_Compact);
 235                 break;
 236
 237         case (MDOC_Rv):
 238                 /* FALLTHROUGH */
 239         case (MDOC_Ex):
 240                 if (xstrcmp(argv, "std"))
 241                         return(MDOC_Std);
 242                 break;
 243
 244         case (MDOC_St):
 245                 if (xstrcmp(argv, "p1003.1-88"))
 246                         return(MDOC_p1003_1_88);
 247                 else if (xstrcmp(argv, "p1003.1-90"))
 248                         return(MDOC_p1003_1_90);
 249                 else if (xstrcmp(argv, "p1003.1-96"))
 250                         return(MDOC_p1003_1_96);
 251                 else if (xstrcmp(argv, "p1003.1-2001"))
 252                         return(MDOC_p1003_1_2001);
 253                 else if (xstrcmp(argv, "p1003.1-2004"))
 254                         return(MDOC_p1003_1_2004);
 255                 else if (xstrcmp(argv, "p1003.1"))
 256                         return(MDOC_p1003_1);
 257                 else if (xstrcmp(argv, "p1003.1b"))
 258                         return(MDOC_p1003_1b);
 259                 else if (xstrcmp(argv, "p1003.1b-93"))
 260                         return(MDOC_p1003_1b_93);
 261                 else if (xstrcmp(argv, "p1003.1c-95"))
 262                         return(MDOC_p1003_1c_95);
 263                 else if (xstrcmp(argv, "p1003.1g-2000"))
 264                         return(MDOC_p1003_1g_2000);
 265                 else if (xstrcmp(argv, "p1003.2-92"))
 266                         return(MDOC_p1003_2_92);
 267                 else if (xstrcmp(argv, "p1003.2-95"))
 268                         return(MDOC_p1387_2_95);
 269                 else if (xstrcmp(argv, "p1003.2"))
 270                         return(MDOC_p1003_2);
 271                 else if (xstrcmp(argv, "p1387.2-95"))
 272                         return(MDOC_p1387_2);
 273                 else if (xstrcmp(argv, "isoC-90"))
 274                         return(MDOC_isoC_90);
 275                 else if (xstrcmp(argv, "isoC-amd1"))
 276                         return(MDOC_isoC_amd1);
 277                 else if (xstrcmp(argv, "isoC-tcor1"))
 278                         return(MDOC_isoC_tcor1);
 279                 else if (xstrcmp(argv, "isoC-tcor2"))
 280                         return(MDOC_isoC_tcor2);
 281                 else if (xstrcmp(argv, "isoC-99"))
 282                         return(MDOC_isoC_99);
 283                 else if (xstrcmp(argv, "ansiC"))
 284                         return(MDOC_ansiC);
 285                 else if (xstrcmp(argv, "ansiC-89"))
 286                         return(MDOC_ansiC_89);
 287                 else if (xstrcmp(argv, "ansiC-99"))
 288                         return(MDOC_ansiC_99);
 289                 else if (xstrcmp(argv, "ieee754"))
 290                         return(MDOC_ieee754);
 291                 else if (xstrcmp(argv, "iso8802-3"))
 292                         return(MDOC_iso8802_3);
 293                 else if (xstrcmp(argv, "xpg3"))
 294                         return(MDOC_xpg3);
 295                 else if (xstrcmp(argv, "xpg4"))
 296                         return(MDOC_xpg4);
 297                 else if (xstrcmp(argv, "xpg4.2"))
 298                         return(MDOC_xpg4_2);
 299                 else if (xstrcmp(argv, "xpg4.3"))
 300                         return(MDOC_xpg4_3);
 301                 else if (xstrcmp(argv, "xbd5"))
 302                         return(MDOC_xbd5);
 303                 else if (xstrcmp(argv, "xcu5"))
 304                         return(MDOC_xcu5);
 305                 else if (xstrcmp(argv, "xsh5"))
 306                         return(MDOC_xsh5);
 307                 else if (xstrcmp(argv, "xns5"))
 308                         return(MDOC_xns5);
 309                 else if (xstrcmp(argv, "xns5.2d2.0"))
 310                         return(MDOC_xns5_2d2_0);
 311                 else if (xstrcmp(argv, "xcurses4.2"))
 312                         return(MDOC_xcurses4_2);
 313                 else if (xstrcmp(argv, "susv2"))
 314                         return(MDOC_susv2);
 315                 else if (xstrcmp(argv, "susv3"))
 316                         return(MDOC_susv3);
 317                 else if (xstrcmp(argv, "svid4"))
 318                         return(MDOC_svid4);
 319                 break;
 320
 321         default:
 322                 break;
 323         }
 324
 325         return(MDOC_ARG_MAX);
 326 }
 327
 328
 329 static int
 330 postparse(struct mdoc *mdoc, int line, const struct mdoc_arg *v, int pos)
 331 {
 332
 333         switch (v->arg) {
 334         case (MDOC_Offset):
 335                 assert(v->value);
 336                 assert(v->value[0]);
 337                 if (xstrcmp(v->value[0], "left"))
 338                         break;
 339                 if (xstrcmp(v->value[0], "right"))
 340                         break;
 341                 if (xstrcmp(v->value[0], "center"))
 342                         break;
 343                 if (xstrcmp(v->value[0], "indent"))
 344                         break;
 345                 if (xstrcmp(v->value[0], "indent-two"))
 346                         break;
 347                 return(mdoc_perr(mdoc, line, pos, "invalid offset value"));
 348         default:
 349                 break;
 350         }
 351
 352         return(1);
 353 }
 354
 355
 356 static int
 357 parse_multi(struct mdoc *mdoc, int line,
 358                 struct mdoc_arg *v, int *pos, char *buf)
 359 {
 360         int              c, ppos;
 361         char            *p;
 362
 363         v->sz = 0;
 364         v->value = xcalloc(MDOC_LINEARG_MAX, sizeof(char *));
 365
 366         ppos = *pos;
 367
 368         for (v->sz = 0; v->sz < MDOC_LINEARG_MAX; v->sz++) {
 369                 if ('-' == buf[*pos])
 370                         break;
 371                 c = mdoc_args(mdoc, line, pos, buf, ARGS_QUOTED, &p);
 372                 if (ARGS_ERROR == c) {
 373                         free(v->value);
 374                         return(0);
 375                 } else if (ARGS_EOLN == c)
 376                         break;
 377                 v->value[v->sz] = p;
 378         }
 379
 380         if (0 < v->sz && v->sz < MDOC_LINEARG_MAX)
 381                 return(1);
 382
 383         free(v->value);
 384         return(mdoc_perr(mdoc, line, ppos, 0 == v->sz ?
 385                                 "argument requires a value" :
 386                                 "too many values to argument"));
 387 }
 388
 389
 390 static int
 391 parse_single(struct mdoc *mdoc, int line,
 392                 struct mdoc_arg *v, int *pos, char *buf)
 393 {
 394         int              c, ppos;
 395         char            *p;
 396
 397         ppos = *pos;
 398
 399         c = mdoc_args(mdoc, line, pos, buf, ARGS_QUOTED, &p);
 400         if (ARGS_ERROR == c)
 401                 return(0);
 402         if (ARGS_EOLN == c)
 403                 return(mdoc_perr(mdoc, line, ppos,  "argument requires a value"));
 404
 405         v->sz = 1;
 406         v->value = xcalloc(1, sizeof(char *));
 407         v->value[0] = p;
 408         return(1);
 409 }
 410
 411
 412 static int
 413 parse(struct mdoc *mdoc, int line,
 414                 struct mdoc_arg *v, int *pos, char *buf)
 415 {
 416
 417         v->sz = 0;
 418         v->value = NULL;
 419
 420         switch (v->arg) {
 421         case(MDOC_Std):
 422                 /* FALLTHROUGH */
 423         case(MDOC_Width):
 424                 /* FALLTHROUGH */
 425         case(MDOC_Offset):
 426                 return(parse_single(mdoc, line, v, pos, buf));
 427         case(MDOC_Column):
 428                 return(parse_multi(mdoc, line, v, pos, buf));
 429         default:
 430                 break;
 431         }
 432
 433         return(1);
 434 }
 435
 436
 437 int
 438 mdoc_argv(struct mdoc *mdoc, int line, int tok,
 439                 struct mdoc_arg *v, int *pos, char *buf)
 440 {
 441         int              i, ppos;
 442         char            *argv;
 443
 444         (void)memset(v, 0, sizeof(struct mdoc_arg));
 445
 446         if (0 == buf[*pos])
 447                 return(ARGV_EOLN);
 448
 449         assert( ! isspace(buf[*pos]));
 450
 451         if ('-' != buf[*pos])
 452                 return(ARGV_WORD);
 453
 454         i = *pos;
 455         argv = &buf[++(*pos)];
 456
 457         v->line = line;
 458         v->pos = *pos;
 459
 460         assert(*pos > 0);
 461         while (buf[*pos]) {
 462                 if (isspace(buf[*pos]))
 463                         if ('\\' != buf[*pos - 1])
 464                                 break;
 465                 (*pos)++;
 466         }
 467
 468         if (buf[*pos])
 469                 buf[(*pos)++] = 0;
 470
 471         if (MDOC_ARG_MAX == (v->arg = lookup(tok, argv))) {
 472                 if ( ! mdoc_pwarn(mdoc, line, i, WARN_SYNTAX, "argument-like parameter"))
 473                         return(ARGV_ERROR);
 474                 return(ARGV_WORD);
 475         }
 476
 477         while (buf[*pos] && isspace(buf[*pos]))
 478                 (*pos)++;
 479
 480         /* FIXME: whitespace if no value. */
 481
 482         ppos = *pos;
 483         if ( ! parse(mdoc, line, v, pos, buf))
 484                 return(ARGV_ERROR);
 485         if ( ! postparse(mdoc, line, v, ppos))
 486                 return(ARGV_ERROR);
 487
 488         return(ARGV_ARG);
 489 }
 490
 491
 492 void
 493 mdoc_argv_free(int sz, struct mdoc_arg *arg)
 494 {
 495         int              i;
 496
 497         for (i = 0; i < sz; i++) {
 498                 if (0 == arg[i].sz) {
 499                         assert(NULL == arg[i].value);
 500                         continue;
 501                 }
 502                 assert(arg[i].value);
 503                 free(arg[i].value);
 504         }
 505 }
 506