mdoc.c

   1 /*      $Id: mdoc.c,v 1.266 2017/06/07 20:58:49 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012-2017 Ingo Schwarze <schwarze@openbsd.org>
   5  *
   6  * Permission to use, copy, modify, and distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18 #include "config.h"
  19
  20 #include <sys/types.h>
  21
  22 #include <assert.h>
  23 #include <ctype.h>
  24 #include <stdarg.h>
  25 #include <stdio.h>
  26 #include <stdlib.h>
  27 #include <string.h>
  28 #include <time.h>
  29
  30 #include "mandoc_aux.h"
  31 #include "mandoc.h"
  32 #include "roff.h"
  33 #include "mdoc.h"
  34 #include "libmandoc.h"
  35 #include "roff_int.h"
  36 #include "libmdoc.h"
  37
  38 const   char *const __mdoc_argnames[MDOC_ARG_MAX] = {
  39         "split",                "nosplit",              "ragged",
  40         "unfilled",             "literal",              "file",
  41         "offset",               "bullet",               "dash",
  42         "hyphen",               "item",                 "enum",
  43         "tag",                  "diag",                 "hang",
  44         "ohang",                "inset",                "column",
  45         "width",                "compact",              "std",
  46         "filled",               "words",                "emphasis",
  47         "symbolic",             "nested",               "centered"
  48 };
  49 const   char * const *mdoc_argnames = __mdoc_argnames;
  50
  51 static  int               mdoc_ptext(struct roff_man *, int, char *, int);
  52 static  int               mdoc_pmacro(struct roff_man *, int, char *, int);
  53
  54
  55 /*
  56  * Main parse routine.  Parses a single line -- really just hands off to
  57  * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
  58  */
  59 int
  60 mdoc_parseln(struct roff_man *mdoc, int ln, char *buf, int offs)
  61 {
  62
  63         if (mdoc->last->type != ROFFT_EQN || ln > mdoc->last->line)
  64                 mdoc->flags |= MDOC_NEWLINE;
  65
  66         /*
  67          * Let the roff nS register switch SYNOPSIS mode early,
  68          * such that the parser knows at all times
  69          * whether this mode is on or off.
  70          * Note that this mode is also switched by the Sh macro.
  71          */
  72         if (roff_getreg(mdoc->roff, "nS"))
  73                 mdoc->flags |= MDOC_SYNOPSIS;
  74         else
  75                 mdoc->flags &= ~MDOC_SYNOPSIS;
  76
  77         return roff_getcontrol(mdoc->roff, buf, &offs) ?
  78             mdoc_pmacro(mdoc, ln, buf, offs) :
  79             mdoc_ptext(mdoc, ln, buf, offs);
  80 }
  81
  82 void
  83 mdoc_macro(MACRO_PROT_ARGS)
  84 {
  85         assert(tok >= MDOC_Dd && tok < MDOC_MAX);
  86         (*mdoc_macros[tok].fp)(mdoc, tok, line, ppos, pos, buf);
  87 }
  88
  89 void
  90 mdoc_tail_alloc(struct roff_man *mdoc, int line, int pos, enum roff_tok tok)
  91 {
  92         struct roff_node *p;
  93
  94         p = roff_node_alloc(mdoc, line, pos, ROFFT_TAIL, tok);
  95         roff_node_append(mdoc, p);
  96         mdoc->next = ROFF_NEXT_CHILD;
  97 }
  98
  99 struct roff_node *
 100 mdoc_endbody_alloc(struct roff_man *mdoc, int line, int pos,
 101     enum roff_tok tok, struct roff_node *body)
 102 {
 103         struct roff_node *p;
 104
 105         body->flags |= NODE_ENDED;
 106         body->parent->flags |= NODE_ENDED;
 107         p = roff_node_alloc(mdoc, line, pos, ROFFT_BODY, tok);
 108         p->body = body;
 109         p->norm = body->norm;
 110         p->end = ENDBODY_SPACE;
 111         roff_node_append(mdoc, p);
 112         mdoc->next = ROFF_NEXT_SIBLING;
 113         return p;
 114 }
 115
 116 struct roff_node *
 117 mdoc_block_alloc(struct roff_man *mdoc, int line, int pos,
 118     enum roff_tok tok, struct mdoc_arg *args)
 119 {
 120         struct roff_node *p;
 121
 122         p = roff_node_alloc(mdoc, line, pos, ROFFT_BLOCK, tok);
 123         p->args = args;
 124         if (p->args)
 125                 (args->refcnt)++;
 126
 127         switch (tok) {
 128         case MDOC_Bd:
 129         case MDOC_Bf:
 130         case MDOC_Bl:
 131         case MDOC_En:
 132         case MDOC_Rs:
 133                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 134                 break;
 135         default:
 136                 break;
 137         }
 138         roff_node_append(mdoc, p);
 139         mdoc->next = ROFF_NEXT_CHILD;
 140         return p;
 141 }
 142
 143 void
 144 mdoc_elem_alloc(struct roff_man *mdoc, int line, int pos,
 145      enum roff_tok tok, struct mdoc_arg *args)
 146 {
 147         struct roff_node *p;
 148
 149         p = roff_node_alloc(mdoc, line, pos, ROFFT_ELEM, tok);
 150         p->args = args;
 151         if (p->args)
 152                 (args->refcnt)++;
 153
 154         switch (tok) {
 155         case MDOC_An:
 156                 p->norm = mandoc_calloc(1, sizeof(union mdoc_data));
 157                 break;
 158         default:
 159                 break;
 160         }
 161         roff_node_append(mdoc, p);
 162         mdoc->next = ROFF_NEXT_CHILD;
 163 }
 164
 165 void
 166 mdoc_node_relink(struct roff_man *mdoc, struct roff_node *p)
 167 {
 168
 169         roff_node_unlink(mdoc, p);
 170         p->prev = p->next = NULL;
 171         roff_node_append(mdoc, p);
 172 }
 173
 174 /*
 175  * Parse free-form text, that is, a line that does not begin with the
 176  * control character.
 177  */
 178 static int
 179 mdoc_ptext(struct roff_man *mdoc, int line, char *buf, int offs)
 180 {
 181         struct roff_node *n;
 182         char             *c, *ws, *end;
 183
 184         n = mdoc->last;
 185
 186         /*
 187          * If a column list contains plain text, assume an implicit item
 188          * macro.  This can happen one or more times at the beginning
 189          * of such a list, intermixed with non-It mdoc macros and with
 190          * nodes generated on the roff level, for example by tbl.
 191          */
 192
 193         if ((n->tok == MDOC_Bl && n->type == ROFFT_BODY &&
 194              n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) ||
 195             (n->parent != NULL && n->parent->tok == MDOC_Bl &&
 196              n->parent->norm->Bl.type == LIST_column)) {
 197                 mdoc->flags |= MDOC_FREECOL;
 198                 mdoc_macro(mdoc, MDOC_It, line, offs, &offs, buf);
 199                 return 1;
 200         }
 201
 202         /*
 203          * Search for the beginning of unescaped trailing whitespace (ws)
 204          * and for the first character not to be output (end).
 205          */
 206
 207         /* FIXME: replace with strcspn(). */
 208         ws = NULL;
 209         for (c = end = buf + offs; *c; c++) {
 210                 switch (*c) {
 211                 case ' ':
 212                         if (NULL == ws)
 213                                 ws = c;
 214                         continue;
 215                 case '\t':
 216                         /*
 217                          * Always warn about trailing tabs,
 218                          * even outside literal context,
 219                          * where they should be put on the next line.
 220                          */
 221                         if (NULL == ws)
 222                                 ws = c;
 223                         /*
 224                          * Strip trailing tabs in literal context only;
 225                          * outside, they affect the next line.
 226                          */
 227                         if (MDOC_LITERAL & mdoc->flags)
 228                                 continue;
 229                         break;
 230                 case '\\':
 231                         /* Skip the escaped character, too, if any. */
 232                         if (c[1])
 233                                 c++;
 234                         /* FALLTHROUGH */
 235                 default:
 236                         ws = NULL;
 237                         break;
 238                 }
 239                 end = c + 1;
 240         }
 241         *end = '\0';
 242
 243         if (ws)
 244                 mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
 245                     line, (int)(ws-buf), NULL);
 246
 247         if (buf[offs] == '\0' && ! (mdoc->flags & MDOC_LITERAL)) {
 248                 mandoc_msg(MANDOCERR_FI_BLANK, mdoc->parse,
 249                     line, (int)(c - buf), NULL);
 250
 251                 /*
 252                  * Insert a `sp' in the case of a blank line.  Technically,
 253                  * blank lines aren't allowed, but enough manuals assume this
 254                  * behaviour that we want to work around it.
 255                  */
 256                 roff_elem_alloc(mdoc, line, offs, ROFF_sp);
 257                 mdoc->last->flags |= NODE_VALID | NODE_ENDED;
 258                 mdoc->next = ROFF_NEXT_SIBLING;
 259                 return 1;
 260         }
 261
 262         roff_word_alloc(mdoc, line, offs, buf+offs);
 263
 264         if (mdoc->flags & MDOC_LITERAL)
 265                 return 1;
 266
 267         /*
 268          * End-of-sentence check.  If the last character is an unescaped
 269          * EOS character, then flag the node as being the end of a
 270          * sentence.  The front-end will know how to interpret this.
 271          */
 272
 273         assert(buf < end);
 274
 275         if (mandoc_eos(buf+offs, (size_t)(end-buf-offs)))
 276                 mdoc->last->flags |= NODE_EOS;
 277
 278         for (c = buf + offs; c != NULL; c = strchr(c + 1, '.')) {
 279                 if (c - buf < offs + 2)
 280                         continue;
 281                 if (end - c < 3)
 282                         break;
 283                 if (c[1] != ' ' ||
 284                     isalpha((unsigned char)c[-2]) == 0 ||
 285                     isalpha((unsigned char)c[-1]) == 0 ||
 286                     (c[-2] == 'n' && c[-1] == 'c') ||
 287                     (c[-2] == 'v' && c[-1] == 's'))
 288                         continue;
 289                 c += 2;
 290                 if (*c == ' ')
 291                         c++;
 292                 if (*c == ' ')
 293                         c++;
 294                 if (isupper((unsigned char)(*c)))
 295                         mandoc_msg(MANDOCERR_EOS, mdoc->parse,
 296                             line, (int)(c - buf), NULL);
 297         }
 298
 299         return 1;
 300 }
 301
 302 /*
 303  * Parse a macro line, that is, a line beginning with the control
 304  * character.
 305  */
 306 static int
 307 mdoc_pmacro(struct roff_man *mdoc, int ln, char *buf, int offs)
 308 {
 309         struct roff_node *n;
 310         const char       *cp;
 311         size_t            sz;
 312         enum roff_tok     tok;
 313         int               sv;
 314
 315         /* Determine the line macro. */
 316
 317         sv = offs;
 318         tok = TOKEN_NONE;
 319         for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
 320                 offs++;
 321         if (sz == 2 || sz == 3)
 322                 tok = roffhash_find(mdoc->mdocmac, buf + sv, sz);
 323         if (tok == TOKEN_NONE) {
 324                 mandoc_msg(MANDOCERR_MACRO, mdoc->parse,
 325                     ln, sv, buf + sv - 1);
 326                 return 1;
 327         }
 328
 329         /* Skip a leading escape sequence or tab. */
 330
 331         switch (buf[offs]) {
 332         case '\\':
 333                 cp = buf + offs + 1;
 334                 mandoc_escape(&cp, NULL, NULL);
 335                 offs = cp - buf;
 336                 break;
 337         case '\t':
 338                 offs++;
 339                 break;
 340         default:
 341                 break;
 342         }
 343
 344         /* Jump to the next non-whitespace word. */
 345
 346         while (buf[offs] == ' ')
 347                 offs++;
 348
 349         /*
 350          * Trailing whitespace.  Note that tabs are allowed to be passed
 351          * into the parser as "text", so we only warn about spaces here.
 352          */
 353
 354         if ('\0' == buf[offs] && ' ' == buf[offs - 1])
 355                 mandoc_msg(MANDOCERR_SPACE_EOL, mdoc->parse,
 356                     ln, offs - 1, NULL);
 357
 358         /*
 359          * If an initial macro or a list invocation, divert directly
 360          * into macro processing.
 361          */
 362
 363         n = mdoc->last;
 364         if (n == NULL || tok == MDOC_It || tok == MDOC_El) {
 365                 mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
 366                 return 1;
 367         }
 368
 369         /*
 370          * If a column list contains a non-It macro, assume an implicit
 371          * item macro.  This can happen one or more times at the
 372          * beginning of such a list, intermixed with text lines and
 373          * with nodes generated on the roff level, for example by tbl.
 374          */
 375
 376         if ((n->tok == MDOC_Bl && n->type == ROFFT_BODY &&
 377              n->end == ENDBODY_NOT && n->norm->Bl.type == LIST_column) ||
 378             (n->parent != NULL && n->parent->tok == MDOC_Bl &&
 379              n->parent->norm->Bl.type == LIST_column)) {
 380                 mdoc->flags |= MDOC_FREECOL;
 381                 mdoc_macro(mdoc, MDOC_It, ln, sv, &sv, buf);
 382                 return 1;
 383         }
 384
 385         /* Normal processing of a macro. */
 386
 387         mdoc_macro(mdoc, tok, ln, sv, &offs, buf);
 388
 389         /* In quick mode (for mandocdb), abort after the NAME section. */
 390
 391         if (mdoc->quick && MDOC_Sh == tok &&
 392             SEC_NAME != mdoc->last->sec)
 393                 return 2;
 394
 395         return 1;
 396 }
 397
 398 enum mdelim
 399 mdoc_isdelim(const char *p)
 400 {
 401
 402         if ('\0' == p[0])
 403                 return DELIM_NONE;
 404
 405         if ('\0' == p[1])
 406                 switch (p[0]) {
 407                 case '(':
 408                 case '[':
 409                         return DELIM_OPEN;
 410                 case '|':
 411                         return DELIM_MIDDLE;
 412                 case '.':
 413                 case ',':
 414                 case ';':
 415                 case ':':
 416                 case '?':
 417                 case '!':
 418                 case ')':
 419                 case ']':
 420                         return DELIM_CLOSE;
 421                 default:
 422                         return DELIM_NONE;
 423                 }
 424
 425         if ('\\' != p[0])
 426                 return DELIM_NONE;
 427
 428         if (0 == strcmp(p + 1, "."))
 429                 return DELIM_CLOSE;
 430         if (0 == strcmp(p + 1, "fR|\\fP"))
 431                 return DELIM_MIDDLE;
 432
 433         return DELIM_NONE;
 434 }
 435
 436 void
 437 mdoc_validate(struct roff_man *mdoc)
 438 {
 439
 440         mdoc->last = mdoc->first;
 441         mdoc_node_validate(mdoc);
 442         mdoc_state_reset(mdoc);
 443 }