man_macro.c

   1 /*      $Id: man_macro.c,v 1.133 2018/08/26 16:21:24 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2012-2015, 2017, 2018 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 static const struct man_macro man_macros[MAN_MAX - MAN_TH] = {
  44         { in_line_eoln, MAN_XSCOPE }, /* TH */
  45         { blk_imp, MAN_XSCOPE | MAN_BSCOPED }, /* SH */
  46         { blk_imp, MAN_XSCOPE | MAN_BSCOPED }, /* SS */
  47         { blk_imp, MAN_XSCOPE | MAN_BSCOPED }, /* TP */
  48         { blk_imp, MAN_XSCOPE | MAN_BSCOPED }, /* TQ */
  49         { blk_imp, MAN_XSCOPE }, /* LP */
  50         { blk_imp, MAN_XSCOPE }, /* PP */
  51         { blk_imp, MAN_XSCOPE }, /* P */
  52         { blk_imp, MAN_XSCOPE }, /* IP */
  53         { blk_imp, MAN_XSCOPE }, /* HP */
  54         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* SM */
  55         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* SB */
  56         { in_line_eoln, 0 }, /* BI */
  57         { in_line_eoln, 0 }, /* IB */
  58         { in_line_eoln, 0 }, /* BR */
  59         { in_line_eoln, 0 }, /* RB */
  60         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* R */
  61         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* B */
  62         { in_line_eoln, MAN_NSCOPED | MAN_ESCOPED | MAN_JOIN }, /* I */
  63         { in_line_eoln, 0 }, /* IR */
  64         { in_line_eoln, 0 }, /* RI */
  65         { in_line_eoln, MAN_NSCOPED }, /* nf */
  66         { in_line_eoln, MAN_NSCOPED }, /* fi */
  67         { blk_close, MAN_XSCOPE }, /* RE */
  68         { blk_exp, MAN_XSCOPE }, /* RS */
  69         { in_line_eoln, 0 }, /* DT */
  70         { in_line_eoln, 0 }, /* UC */
  71         { in_line_eoln, MAN_NSCOPED }, /* PD */
  72         { in_line_eoln, 0 }, /* AT */
  73         { in_line_eoln, MAN_NSCOPED }, /* in */
  74         { blk_imp, MAN_XSCOPE }, /* SY */
  75         { blk_close, MAN_XSCOPE }, /* YS */
  76         { in_line_eoln, 0 }, /* OP */
  77         { in_line_eoln, MAN_XSCOPE }, /* EX */
  78         { in_line_eoln, MAN_XSCOPE }, /* EE */
  79         { blk_exp, MAN_XSCOPE }, /* UR */
  80         { blk_close, MAN_XSCOPE }, /* UE */
  81         { blk_exp, MAN_XSCOPE }, /* MT */
  82         { blk_close, MAN_XSCOPE }, /* ME */
  83 };
  84
  85
  86 const struct man_macro *
  87 man_macro(enum roff_tok tok)
  88 {
  89         assert(tok >= MAN_TH && tok <= MAN_MAX);
  90         return man_macros + (tok - MAN_TH);
  91 }
  92
  93 void
  94 man_unscope(struct roff_man *man, const struct roff_node *to)
  95 {
  96         struct roff_node *n;
  97
  98         to = to->parent;
  99         n = man->last;
 100         while (n != to) {
 101
 102                 /* Reached the end of the document? */
 103
 104                 if (to == NULL && ! (n->flags & NODE_VALID)) {
 105                         if (man->flags & (MAN_BLINE | MAN_ELINE) &&
 106                             man_macro(n->tok)->flags &
 107                              (MAN_BSCOPED | MAN_NSCOPED)) {
 108                                 mandoc_vmsg(MANDOCERR_BLK_LINE,
 109                                     man->parse, n->line, n->pos,
 110                                     "EOF breaks %s", roff_name[n->tok]);
 111                                 if (man->flags & MAN_ELINE)
 112                                         man->flags &= ~MAN_ELINE;
 113                                 else {
 114                                         assert(n->type == ROFFT_HEAD);
 115                                         n = n->parent;
 116                                         man->flags &= ~MAN_BLINE;
 117                                 }
 118                                 man->last = n;
 119                                 n = n->parent;
 120                                 roff_node_delete(man, man->last);
 121                                 continue;
 122                         }
 123                         if (n->type == ROFFT_BLOCK &&
 124                             man_macro(n->tok)->fp == blk_exp)
 125                                 mandoc_msg(MANDOCERR_BLK_NOEND,
 126                                     man->parse, n->line, n->pos,
 127                                     roff_name[n->tok]);
 128                 }
 129
 130                 /*
 131                  * We might delete the man->last node
 132                  * in the post-validation phase.
 133                  * Save a pointer to the parent such that
 134                  * we know where to continue the iteration.
 135                  */
 136
 137                 man->last = n;
 138                 n = n->parent;
 139                 man->last->flags |= NODE_VALID;
 140         }
 141
 142         /*
 143          * If we ended up at the parent of the node we were
 144          * supposed to rewind to, that means the target node
 145          * got deleted, so add the next node we parse as a child
 146          * of the parent instead of as a sibling of the target.
 147          */
 148
 149         man->next = (man->last == to) ?
 150             ROFF_NEXT_CHILD : ROFF_NEXT_SIBLING;
 151 }
 152
 153 /*
 154  * Rewinding entails ascending the parse tree until a coherent point,
 155  * for example, the `SH' macro will close out any intervening `SS'
 156  * scopes.  When a scope is closed, it must be validated and actioned.
 157  */
 158 static void
 159 rew_scope(struct roff_man *man, enum roff_tok tok)
 160 {
 161         struct roff_node *n;
 162
 163         /* Preserve empty paragraphs before RS. */
 164
 165         n = man->last;
 166         if (tok == MAN_RS && n->child == NULL &&
 167             (n->tok == MAN_P || n->tok == MAN_PP || n->tok == MAN_LP))
 168                 return;
 169
 170         for (;;) {
 171                 if (n->type == ROFFT_ROOT)
 172                         return;
 173                 if (n->flags & NODE_VALID) {
 174                         n = n->parent;
 175                         continue;
 176                 }
 177                 if (n->type != ROFFT_BLOCK) {
 178                         if (n->parent->type == ROFFT_ROOT) {
 179                                 man_unscope(man, n);
 180                                 return;
 181                         } else {
 182                                 n = n->parent;
 183                                 continue;
 184                         }
 185                 }
 186                 if (tok != MAN_SH && (n->tok == MAN_SH ||
 187                     (tok != MAN_SS && (n->tok == MAN_SS ||
 188                      man_macro(n->tok)->fp == blk_exp))))
 189                         return;
 190                 man_unscope(man, n);
 191                 n = man->last;
 192         }
 193 }
 194
 195
 196 /*
 197  * Close out a generic explicit macro.
 198  */
 199 void
 200 blk_close(MACRO_PROT_ARGS)
 201 {
 202         enum roff_tok            ctok, ntok;
 203         const struct roff_node  *nn;
 204         char                    *p;
 205         int                      cline, cpos, nrew, target;
 206
 207         nrew = 1;
 208         switch (tok) {
 209         case MAN_RE:
 210                 ntok = MAN_RS;
 211                 if ( ! man_args(man, line, pos, buf, &p))
 212                         break;
 213                 for (nn = man->last->parent; nn; nn = nn->parent)
 214                         if (nn->tok == ntok && nn->type == ROFFT_BLOCK)
 215                                 nrew++;
 216                 target = strtol(p, &p, 10);
 217                 if (*p != '\0')
 218                         mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse,
 219                             line, p - buf, "RE ... %s", p);
 220                 if (target == 0)
 221                         target = 1;
 222                 nrew -= target;
 223                 if (nrew < 1) {
 224                         mandoc_vmsg(MANDOCERR_RE_NOTOPEN, man->parse,
 225                             line, ppos, "RE %d", target);
 226                         return;
 227                 }
 228                 break;
 229         case MAN_YS:
 230                 ntok = MAN_SY;
 231                 break;
 232         case MAN_UE:
 233                 ntok = MAN_UR;
 234                 break;
 235         case MAN_ME:
 236                 ntok = MAN_MT;
 237                 break;
 238         default:
 239                 abort();
 240         }
 241
 242         for (nn = man->last->parent; nn; nn = nn->parent)
 243                 if (nn->tok == ntok && nn->type == ROFFT_BLOCK && ! --nrew)
 244                         break;
 245
 246         if (nn == NULL) {
 247                 mandoc_msg(MANDOCERR_BLK_NOTOPEN, man->parse,
 248                     line, ppos, roff_name[tok]);
 249                 rew_scope(man, MAN_PP);
 250                 if (tok == MAN_RE) {
 251                         roff_elem_alloc(man, line, ppos, ROFF_br);
 252                         man->last->flags |= NODE_LINE |
 253                             NODE_VALID | NODE_ENDED;
 254                         man->next = ROFF_NEXT_SIBLING;
 255                 }
 256                 return;
 257         }
 258
 259         cline = man->last->line;
 260         cpos = man->last->pos;
 261         ctok = man->last->tok;
 262         man_unscope(man, nn);
 263
 264         if (tok == MAN_RE && nn->head->aux > 0)
 265                 roff_setreg(man->roff, "an-margin", nn->head->aux, '-');
 266
 267         /* Trailing text. */
 268
 269         if (buf[*pos] != '\0') {
 270                 roff_word_alloc(man, line, ppos, buf + *pos);
 271                 man->last->flags |= NODE_DELIMC;
 272                 if (mandoc_eos(man->last->string, strlen(man->last->string)))
 273                         man->last->flags |= NODE_EOS;
 274         }
 275
 276         /* Move a trailing paragraph behind the block. */
 277
 278         if (ctok == MAN_LP || ctok == MAN_PP || ctok == MAN_P) {
 279                 *pos = strlen(buf);
 280                 blk_imp(man, ctok, cline, cpos, pos, buf);
 281         }
 282
 283         /* Synopsis blocks need an explicit end marker for spacing. */
 284
 285         if (tok == MAN_YS && man->last == nn) {
 286                 roff_elem_alloc(man, line, ppos, tok);
 287                 man_unscope(man, man->last);
 288         }
 289 }
 290
 291 void
 292 blk_exp(MACRO_PROT_ARGS)
 293 {
 294         struct roff_node *head;
 295         char            *p;
 296         int              la;
 297
 298         if (tok == MAN_RS)
 299                 rew_scope(man, tok);
 300         roff_block_alloc(man, line, ppos, tok);
 301         head = roff_head_alloc(man, line, ppos, tok);
 302
 303         la = *pos;
 304         if (man_args(man, line, pos, buf, &p)) {
 305                 roff_word_alloc(man, line, la, p);
 306                 if (tok == MAN_RS) {
 307                         if (roff_getreg(man->roff, "an-margin") == 0)
 308                                 roff_setreg(man->roff, "an-margin",
 309                                     7 * 24, '=');
 310                         if ((head->aux = strtod(p, NULL) * 24.0) > 0)
 311                                 roff_setreg(man->roff, "an-margin",
 312                                     head->aux, '+');
 313                 }
 314         }
 315
 316         if (buf[*pos] != '\0')
 317                 mandoc_vmsg(MANDOCERR_ARG_EXCESS, man->parse, line,
 318                     *pos, "%s ... %s", roff_name[tok], buf + *pos);
 319
 320         man_unscope(man, head);
 321         roff_body_alloc(man, line, ppos, tok);
 322 }
 323
 324 /*
 325  * Parse an implicit-block macro.  These contain a ROFFT_HEAD and a
 326  * ROFFT_BODY contained within a ROFFT_BLOCK.  Rules for closing out other
 327  * scopes, such as `SH' closing out an `SS', are defined in the rew
 328  * routines.
 329  */
 330 void
 331 blk_imp(MACRO_PROT_ARGS)
 332 {
 333         int              la;
 334         char            *p;
 335         struct roff_node *n;
 336
 337         rew_scope(man, tok);
 338         n = roff_block_alloc(man, line, ppos, tok);
 339         if (n->tok == MAN_SH || n->tok == MAN_SS)
 340                 man->flags &= ~MAN_LITERAL;
 341         n = roff_head_alloc(man, line, ppos, tok);
 342
 343         /* Add line arguments. */
 344
 345         for (;;) {
 346                 la = *pos;
 347                 if ( ! man_args(man, line, pos, buf, &p))
 348                         break;
 349                 roff_word_alloc(man, line, la, p);
 350         }
 351
 352         /*
 353          * For macros having optional next-line scope,
 354          * keep the head open if there were no arguments.
 355          * For `TP' and `TQ', always keep the head open.
 356          */
 357
 358         if (man_macro(tok)->flags & MAN_BSCOPED &&
 359             (tok == MAN_TP || tok == MAN_TQ || n == man->last)) {
 360                 man->flags |= MAN_BLINE;
 361                 return;
 362         }
 363
 364         /* Close out the head and open the body. */
 365
 366         man_unscope(man, n);
 367         roff_body_alloc(man, line, ppos, tok);
 368 }
 369
 370 void
 371 in_line_eoln(MACRO_PROT_ARGS)
 372 {
 373         int              la;
 374         char            *p;
 375         struct roff_node *n;
 376
 377         roff_elem_alloc(man, line, ppos, tok);
 378         n = man->last;
 379
 380         for (;;) {
 381                 if (buf[*pos] != '\0' && (tok == MAN_fi || tok == MAN_nf)) {
 382                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 383                             man->parse, line, *pos, "%s %s",
 384                             roff_name[tok], buf + *pos);
 385                         break;
 386                 }
 387                 if (buf[*pos] != '\0' && man->last != n && tok == MAN_PD) {
 388                         mandoc_vmsg(MANDOCERR_ARG_EXCESS,
 389                             man->parse, line, *pos, "%s ... %s",
 390                             roff_name[tok], buf + *pos);
 391                         break;
 392                 }
 393                 la = *pos;
 394                 if ( ! man_args(man, line, pos, buf, &p))
 395                         break;
 396                 if (man_macro(tok)->flags & MAN_JOIN &&
 397                     man->last->type == ROFFT_TEXT)
 398                         roff_word_append(man, p);
 399                 else
 400                         roff_word_alloc(man, line, la, p);
 401         }
 402
 403         /*
 404          * Append NODE_EOS in case the last snipped argument
 405          * ends with a dot, e.g. `.IR syslog (3).'
 406          */
 407
 408         if (n != man->last &&
 409             mandoc_eos(man->last->string, strlen(man->last->string)))
 410                 man->last->flags |= NODE_EOS;
 411
 412         /*
 413          * If no arguments are specified and this is MAN_ESCOPED (i.e.,
 414          * next-line scoped), then set our mode to indicate that we're
 415          * waiting for terms to load into our context.
 416          */
 417
 418         if (n == man->last && man_macro(tok)->flags & MAN_ESCOPED) {
 419                 man->flags |= MAN_ELINE;
 420                 return;
 421         }
 422
 423         assert(man->last->type != ROFFT_ROOT);
 424         man->next = ROFF_NEXT_SIBLING;
 425
 426         /* Rewind our element scope. */
 427
 428         for ( ; man->last; man->last = man->last->parent) {
 429                 man_state(man, man->last);
 430                 if (man->last == n)
 431                         break;
 432         }
 433
 434         /* Rewind next-line scoped ancestors, if any. */
 435
 436         if (man_macro(tok)->flags & MAN_ESCOPED)
 437                 man_descope(man, line, ppos, NULL);
 438 }
 439
 440 void
 441 man_endparse(struct roff_man *man)
 442 {
 443
 444         man_unscope(man, man->first);
 445         man->flags &= ~MAN_LITERAL;
 446 }
 447
 448 static int
 449 man_args(struct roff_man *man, int line, int *pos, char *buf, char **v)
 450 {
 451         char     *start;
 452
 453         assert(*pos);
 454         *v = start = buf + *pos;
 455         assert(' ' != *start);
 456
 457         if ('\0' == *start)
 458                 return 0;
 459
 460         *v = mandoc_getarg(man->parse, v, line, pos);
 461         return 1;
 462 }