man_macro.c

   1 /*      $Id: man_macro.c,v 1.120 2017/05/05 15:17:32 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2012-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
   5  * Copyright (c) 2013 Franco Fichtner <franco@lastsummer.de>
   6  *
   7  * Permission to use, copy, modify, and distribute this software for any
   8  * purpose with or without fee is hereby granted, provided that the above
   9  * copyright notice and this permission notice appear in all copies.
  10  *
  11  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
  12  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  13  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
  14  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  15  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  16  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  17  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  18  */
  19 #include "config.h"
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <ctype.h>
  25 #include <stdlib.h>
  26 #include <string.h>
  27
  28 #include "mandoc.h"
  29 #include "roff.h"
  30 #include "man.h"
  31 #include "libmandoc.h"
  32 #include "roff_int.h"
  33 #include "libman.h"
  34
  35 static  void             blk_close(MACRO_PROT_ARGS);
  36 static  void             blk_exp(MACRO_PROT_ARGS);
  37 static  void             blk_imp(MACRO_PROT_ARGS);
  38 static  void             in_line_eoln(MACRO_PROT_ARGS);
  39 static  int              man_args(struct roff_man *, int,
  40                                 int *, char *, char **);
  41 static  void             rew_scope(struct roff_man *, enum roff_tok);
  42
  43 const   struct man_macro __man_macros[MAN_MAX - MAN_TH] = {
  44         { in_line_eoln, MAN_BSCOPE }, /* TH */
  45         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SH */
  46         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* SS */
  47         { blk_imp, MAN_BSCOPE | MAN_SCOPED }, /* TP */
  48         { blk_imp, MAN_BSCOPE }, /* LP */
  49         { blk_imp, MAN_BSCOPE }, /* PP */
  50         { blk_imp, MAN_BSCOPE }, /* P */
  51         { blk_imp, MAN_BSCOPE }, /* IP */
  52         { blk_imp, MAN_BSCOPE }, /* HP */
  53         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SM */
  54         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* SB */
  55         { in_line_eoln, 0 }, /* BI */
  56         { in_line_eoln, 0 }, /* IB */
  57         { in_line_eoln, 0 }, /* BR */
  58         { in_line_eoln, 0 }, /* RB */
  59         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* R */
  60         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* B */
  61         { in_line_eoln, MAN_SCOPED | MAN_JOIN }, /* I */
  62         { in_line_eoln, 0 }, /* IR */
  63         { in_line_eoln, 0 }, /* RI */
  64         { in_line_eoln, MAN_NSCOPED }, /* nf */
  65         { in_line_eoln, MAN_NSCOPED }, /* fi */
  66         { blk_close, MAN_BSCOPE }, /* RE */
  67         { blk_exp, MAN_BSCOPE }, /* RS */
  68         { in_line_eoln, 0 }, /* DT */
  69         { in_line_eoln, 0 }, /* UC */
  70         { in_line_eoln, MAN_NSCOPED }, /* PD */
  71         { in_line_eoln, 0 }, /* AT */
  72         { in_line_eoln, 0 }, /* in */
  73         { in_line_eoln, 0 }, /* OP */
  74         { in_line_eoln, MAN_BSCOPE }, /* EX */
  75         { in_line_eoln, MAN_BSCOPE }, /* EE */
  76         { blk_exp, MAN_BSCOPE }, /* UR */
  77         { blk_close, MAN_BSCOPE }, /* UE */
  78 };
  79 const   struct man_macro *const man_macros = __man_macros - MAN_TH;
  80
  81
  82 void
  83 man_unscope(struct roff_man *man, const struct roff_node *to)
  84 {
  85         struct roff_node *n;
  86
  87         to = to->parent;
  88         n = man->last;
  89         while (n != to) {
  90
  91                 /* Reached the end of the document? */
  92
  93                 if (to == NULL && ! (n->flags & NODE_VALID)) {
  94                         if (man->flags & (MAN_BLINE | MAN_ELINE) &&
  95                             man_macros[n->tok].flags & MAN_SCOPED) {
  96                                 mandoc_vmsg(MANDOCERR_BLK_LINE,
  97                                     man->parse, n->line, n->pos,
  98                                     "EOF breaks %s", roff_name[n->tok]);
  99                                 if (man->flags & MAN_ELINE)
 100                                         man->flags &= ~MAN_ELINE;
 101                                 else {
 102                                         assert(n->type == ROFFT_HEAD);
 103                                         n = n->parent;
 104                                         man->flags &= ~MAN_BLINE;
 105                                 }
 106                                 man->last = n;
 107                                 n = n->parent;
 108                                 roff_node_delete(man, man->last);
 109                                 continue;
 110                         }
 111                         if (n->type == ROFFT_BLOCK &&
 112                             man_macros[n->tok].fp == blk_exp)
 113                                 mandoc_msg(MANDOCERR_BLK_NOEND,
 114                                     man->parse, n->line, n->pos,
 115                                     roff_name[n->tok]);
 116                 }
 117
 118                 /*
 119                  * We might delete the man->last node
 120                  * in the post-validation phase.
 121                  * Save a pointer to the parent such that
 122                  * we know where to continue the iteration.
 123                  */
 124
 125                 man->last = n;
 126                 n = n->parent;
 127                 man->last->flags |= NODE_VALID;
 128         }
 129
 130         /*
 131          * If we ended up at the parent of the node we were
 132          * supposed to rewind to, that means the target node
 133          * got deleted, so add the next node we parse as a child
 134          * of the parent instead of as a sibling of the target.
 135          */
 136
 137         man->next = (man->last == to) ?
 138             ROFF_NEXT_CHILD : ROFF_NEXT_SIBLING;
 139 }
 140
 141 /*
 142  * Rewinding entails ascending the parse tree until a coherent point,
 143  * for example, the `SH' macro will close out any intervening `SS'
 144  * scopes.  When a scope is closed, it must be validated and actioned.
 145  */
 146 static void
 147 rew_scope(struct roff_man *man, enum roff_tok tok)
 148 {
 149         struct roff_node *n;
 150
 151         /* Preserve empty paragraphs before RS. */
 152
 153         n = man->last;
 154         if (tok == MAN_RS && n->child == NULL &&
 155             (n->tok == MAN_P || n->tok == MAN_PP || n->tok == MAN_LP))
 156                 return;
 157
 158         for (;;) {
 159                 if (n->type == ROFFT_ROOT)
 160                         return;
 161                 if (n->flags & NODE_VALID) {
 162                         n = n->parent;
 163                         continue;
 164                 }
 165                 if (n->type != ROFFT_BLOCK) {
 166                         if (n->parent->type == ROFFT_ROOT) {
 167                                 man_unscope(man, n);
 168                                 return;
 169                         } else {
 170                                 n = n->parent;
 171                                 continue;
 172                         }
 173                 }
 174                 if (tok != MAN_SH && (n->tok == MAN_SH ||
 175                     (tok != MAN_SS && (n->tok == MAN_SS ||
 176                      man_macros[n->tok].fp == blk_exp))))
 177                         return;
 178                 man_unscope(man, n);
 179                 n = man->last;
 180         }
 181 }
 182
 183
 184 /*
 185  * Close out a generic explicit macro.
 186  */
 187 void
 188 blk_close(MACRO_PROT_ARGS)
 189 {
 190         enum roff_tok            ntok;
 191         const struct roff_node  *nn;
 192         char                    *p;
 193         int                      nrew, target;
 194
 195         nrew = 1;
 196         switch (tok) {
 197         case MAN_RE:
 198                 ntok = MAN_RS;
 199                 if ( ! man_args(man, line, pos, buf, &p))
 200                         break;
 201                 for (nn = man->last->parent; nn; nn = nn->parent)
 202                         if (nn->tok == ntok && nn->type == ROFFT_BLOCK)
 203                                 nrew++;
 204                 target = strtol(p, &p, 10);
 205                 if (*p != '\0')
 206                         mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 207                             line, p - buf, "RE ... %s", p);
 208                 if (target == 0)
 209                         target = 1;
 210                 nrew -= target;
 211                 if (nrew < 1) {
 212                         mandoc_vmsg(MANDOCERR_RE_NOTOPEN, man->parse,
 213                             line, ppos, "RE %d", target);
 214                         return;
 215                 }
 216                 break;
 217         case MAN_UE:
 218                 ntok = MAN_UR;
 219                 break;
 220         default:
 221                 abort();
 222         }
 223
 224         for (nn = man->last->parent; nn; nn = nn->parent)
 225                 if (nn->tok == ntok && nn->type == ROFFT_BLOCK && ! --nrew)
 226                         break;
 227
 228         if (nn == NULL) {
 229                 mandoc_msg(MANDOCERR_BLK_NOTOPEN, man->parse,
 230                     line, ppos, roff_name[tok]);
 231                 rew_scope(man, MAN_PP);
 232         } else {
 233                 line = man->last->line;
 234                 ppos = man->last->pos;
 235                 ntok = man->last->tok;
 236                 man_unscope(man, nn);
 237
 238                 /* Move a trailing paragraph behind the block. */
 239
 240                 if (ntok == MAN_LP || ntok == MAN_PP || ntok == MAN_P) {
 241                         *pos = strlen(buf);
 242                         blk_imp(man, ntok, line, ppos, pos, buf);
 243                 }
 244         }
 245 }
 246
 247 void
 248 blk_exp(MACRO_PROT_ARGS)
 249 {
 250         struct roff_node *head;
 251         char            *p;
 252         int              la;
 253
 254         rew_scope(man, tok);
 255         roff_block_alloc(man, line, ppos, tok);
 256         head = roff_head_alloc(man, line, ppos, tok);
 257
 258         la = *pos;
 259         if (man_args(man, line, pos, buf, &p))
 260                 roff_word_alloc(man, line, la, p);
 261
 262         if (buf[*pos] != '\0')
 263                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse, line,
 264                     *pos, "%s ... %s", roff_name[tok], buf + *pos);
 265
 266         man_unscope(man, head);
 267         roff_body_alloc(man, line, ppos, tok);
 268 }
 269
 270 /*
 271  * Parse an implicit-block macro.  These contain a ROFFT_HEAD and a
 272  * ROFFT_BODY contained within a ROFFT_BLOCK.  Rules for closing out other
 273  * scopes, such as `SH' closing out an `SS', are defined in the rew
 274  * routines.
 275  */
 276 void
 277 blk_imp(MACRO_PROT_ARGS)
 278 {
 279         int              la;
 280         char            *p;
 281         struct roff_node *n;
 282
 283         rew_scope(man, tok);
 284         n = roff_block_alloc(man, line, ppos, tok);
 285         if (n->tok == MAN_SH || n->tok == MAN_SS)
 286                 man->flags &= ~MAN_LITERAL;
 287         n = roff_head_alloc(man, line, ppos, tok);
 288
 289         /* Add line arguments. */
 290
 291         for (;;) {
 292                 la = *pos;
 293                 if ( ! man_args(man, line, pos, buf, &p))
 294                         break;
 295                 roff_word_alloc(man, line, la, p);
 296         }
 297
 298         /*
 299          * For macros having optional next-line scope,
 300          * keep the head open if there were no arguments.
 301          * For `TP', always keep the head open.
 302          */
 303
 304         if (man_macros[tok].flags & MAN_SCOPED &&
 305             (tok == MAN_TP || n == man->last)) {
 306                 man->flags |= MAN_BLINE;
 307                 return;
 308         }
 309
 310         /* Close out the head and open the body. */
 311
 312         man_unscope(man, n);
 313         roff_body_alloc(man, line, ppos, tok);
 314 }
 315
 316 void
 317 in_line_eoln(MACRO_PROT_ARGS)
 318 {
 319         int              la;
 320         char            *p;
 321         struct roff_node *n;
 322
 323         roff_elem_alloc(man, line, ppos, tok);
 324         n = man->last;
 325
 326         for (;;) {
 327                 if (buf[*pos] != '\0' && (tok == MAN_fi || tok == MAN_nf)) {
 328                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 329                             man->parse, line, *pos, "%s %s",
 330                             roff_name[tok], buf + *pos);
 331                         break;
 332                 }
 333                 if (buf[*pos] != '\0' && man->last != n && tok == MAN_PD) {
 334                         mandoc_vmsg(MANDOCERR_ARG_EXCESS,
 335                             man->parse, line, *pos, "%s ... %s",
 336                             roff_name[tok], buf + *pos);
 337                         break;
 338                 }
 339                 la = *pos;
 340                 if ( ! man_args(man, line, pos, buf, &p))
 341                         break;
 342                 if (man_macros[tok].flags & MAN_JOIN &&
 343                     man->last->type == ROFFT_TEXT)
 344                         roff_word_append(man, p);
 345                 else
 346                         roff_word_alloc(man, line, la, p);
 347         }
 348
 349         /*
 350          * Append NODE_EOS in case the last snipped argument
 351          * ends with a dot, e.g. `.IR syslog (3).'
 352          */
 353
 354         if (n != man->last &&
 355             mandoc_eos(man->last->string, strlen(man->last->string)))
 356                 man->last->flags |= NODE_EOS;
 357
 358         /*
 359          * If no arguments are specified and this is MAN_SCOPED (i.e.,
 360          * next-line scoped), then set our mode to indicate that we're
 361          * waiting for terms to load into our context.
 362          */
 363
 364         if (n == man->last && man_macros[tok].flags & MAN_SCOPED) {
 365                 assert( ! (man_macros[tok].flags & MAN_NSCOPED));
 366                 man->flags |= MAN_ELINE;
 367                 return;
 368         }
 369
 370         assert(man->last->type != ROFFT_ROOT);
 371         man->next = ROFF_NEXT_SIBLING;
 372
 373         /* Rewind our element scope. */
 374
 375         for ( ; man->last; man->last = man->last->parent) {
 376                 man_state(man, man->last);
 377                 if (man->last == n)
 378                         break;
 379         }
 380 }
 381
 382 void
 383 man_endparse(struct roff_man *man)
 384 {
 385
 386         man_unscope(man, man->first);
 387         man->flags &= ~MAN_LITERAL;
 388 }
 389
 390 static int
 391 man_args(struct roff_man *man, int line, int *pos, char *buf, char **v)
 392 {
 393         char     *start;
 394
 395         assert(*pos);
 396         *v = start = buf + *pos;
 397         assert(' ' != *start);
 398
 399         if ('\0' == *start)
 400                 return 0;
 401
 402         *v = mandoc_getarg(man->parse, v, line, pos);
 403         return 1;
 404 }