man.c

   1 /*      $Id: man.c,v 1.180 2018/08/26 16:21:23 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2013,2014,2015,2017,2018 Ingo Schwarze <schwarze@openbsd.org>
   5  * Copyright (c) 2011 Joerg Sonnenberger <joerg@netbsd.org>
   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 <stdarg.h>
  26 #include <stdlib.h>
  27 #include <stdio.h>
  28 #include <string.h>
  29
  30 #include "mandoc_aux.h"
  31 #include "mandoc.h"
  32 #include "roff.h"
  33 #include "man.h"
  34 #include "libmandoc.h"
  35 #include "roff_int.h"
  36 #include "libman.h"
  37
  38 static  char            *man_hasc(char *);
  39 static  int              man_ptext(struct roff_man *, int, char *, int);
  40 static  int              man_pmacro(struct roff_man *, int, char *, int);
  41
  42
  43 int
  44 man_parseln(struct roff_man *man, int ln, char *buf, int offs)
  45 {
  46
  47         if (man->last->type != ROFFT_EQN || ln > man->last->line)
  48                 man->flags |= MAN_NEWLINE;
  49
  50         return roff_getcontrol(man->roff, buf, &offs) ?
  51             man_pmacro(man, ln, buf, offs) :
  52             man_ptext(man, ln, buf, offs);
  53 }
  54
  55 /*
  56  * If the string ends with \c, return a pointer to the backslash.
  57  * Otherwise, return NULL.
  58  */
  59 static char *
  60 man_hasc(char *start)
  61 {
  62         char    *cp, *ep;
  63
  64         ep = strchr(start, '\0') - 2;
  65         if (ep < start || ep[0] != '\\' || ep[1] != 'c')
  66                 return NULL;
  67         for (cp = ep; cp > start; cp--)
  68                 if (cp[-1] != '\\')
  69                         break;
  70         return (ep - cp) % 2 ? NULL : ep;
  71 }
  72
  73 void
  74 man_descope(struct roff_man *man, int line, int offs, char *start)
  75 {
  76         /* Trailing \c keeps next-line scope open. */
  77
  78         if (start != NULL && man_hasc(start) != NULL)
  79                 return;
  80
  81         /*
  82          * Co-ordinate what happens with having a next-line scope open:
  83          * first close out the element scopes (if applicable),
  84          * then close out the block scope (also if applicable).
  85          */
  86
  87         if (man->flags & MAN_ELINE) {
  88                 while (man->last->parent->type != ROFFT_ROOT &&
  89                     man_macro(man->last->parent->tok)->flags & MAN_ESCOPED)
  90                         man_unscope(man, man->last->parent);
  91                 man->flags &= ~MAN_ELINE;
  92         }
  93         if ( ! (man->flags & MAN_BLINE))
  94                 return;
  95         man->flags &= ~MAN_BLINE;
  96         man_unscope(man, man->last->parent);
  97         roff_body_alloc(man, line, offs, man->last->tok);
  98 }
  99
 100 static int
 101 man_ptext(struct roff_man *man, int line, char *buf, int offs)
 102 {
 103         int              i;
 104         char            *ep;
 105
 106         /* Literal free-form text whitespace is preserved. */
 107
 108         if (man->flags & MAN_LITERAL) {
 109                 roff_word_alloc(man, line, offs, buf + offs);
 110                 man_descope(man, line, offs, buf + offs);
 111                 return 1;
 112         }
 113
 114         for (i = offs; buf[i] == ' '; i++)
 115                 /* Skip leading whitespace. */ ;
 116
 117         /*
 118          * Blank lines are ignored in next line scope
 119          * and right after headings and cancel preceding \c,
 120          * but add a single vertical space elsewhere.
 121          */
 122
 123         if (buf[i] == '\0') {
 124                 if (man->flags & (MAN_ELINE | MAN_BLINE)) {
 125                         mandoc_msg(MANDOCERR_BLK_BLANK, man->parse,
 126                             line, 0, NULL);
 127                         return 1;
 128                 }
 129                 if (man->last->tok == MAN_SH || man->last->tok == MAN_SS)
 130                         return 1;
 131                 if (man->last->type == ROFFT_TEXT &&
 132                     ((ep = man_hasc(man->last->string)) != NULL)) {
 133                         *ep = '\0';
 134                         return 1;
 135                 }
 136                 roff_elem_alloc(man, line, offs, ROFF_sp);
 137                 man->next = ROFF_NEXT_SIBLING;
 138                 return 1;
 139         }
 140
 141         /*
 142          * Warn if the last un-escaped character is whitespace. Then
 143          * strip away the remaining spaces (tabs stay!).
 144          */
 145
 146         i = (int)strlen(buf);
 147         assert(i);
 148
 149         if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
 150                 if (i > 1 && '\\' != buf[i - 2])
 151                         mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 152                             line, i - 1, NULL);
 153
 154                 for (--i; i && ' ' == buf[i]; i--)
 155                         /* Spin back to non-space. */ ;
 156
 157                 /* Jump ahead of escaped whitespace. */
 158                 i += '\\' == buf[i] ? 2 : 1;
 159
 160                 buf[i] = '\0';
 161         }
 162         roff_word_alloc(man, line, offs, buf + offs);
 163
 164         /*
 165          * End-of-sentence check.  If the last character is an unescaped
 166          * EOS character, then flag the node as being the end of a
 167          * sentence.  The front-end will know how to interpret this.
 168          */
 169
 170         assert(i);
 171         if (mandoc_eos(buf, (size_t)i))
 172                 man->last->flags |= NODE_EOS;
 173
 174         man_descope(man, line, offs, buf + offs);
 175         return 1;
 176 }
 177
 178 static int
 179 man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
 180 {
 181         struct roff_node *n;
 182         const char      *cp;
 183         size_t           sz;
 184         enum roff_tok    tok;
 185         int              ppos;
 186         int              bline;
 187
 188         /* Determine the line macro. */
 189
 190         ppos = offs;
 191         tok = TOKEN_NONE;
 192         for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
 193                 offs++;
 194         if (sz > 0 && sz < 4)
 195                 tok = roffhash_find(man->manmac, buf + ppos, sz);
 196         if (tok == TOKEN_NONE) {
 197                 mandoc_msg(MANDOCERR_MACRO, man->parse,
 198                     ln, ppos, buf + ppos - 1);
 199                 return 1;
 200         }
 201
 202         /* Skip a leading escape sequence or tab. */
 203
 204         switch (buf[offs]) {
 205         case '\\':
 206                 cp = buf + offs + 1;
 207                 mandoc_escape(&cp, NULL, NULL);
 208                 offs = cp - buf;
 209                 break;
 210         case '\t':
 211                 offs++;
 212                 break;
 213         default:
 214                 break;
 215         }
 216
 217         /* Jump to the next non-whitespace word. */
 218
 219         while (buf[offs] == ' ')
 220                 offs++;
 221
 222         /*
 223          * Trailing whitespace.  Note that tabs are allowed to be passed
 224          * into the parser as "text", so we only warn about spaces here.
 225          */
 226
 227         if (buf[offs] == '\0' && buf[offs - 1] == ' ')
 228                 mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 229                     ln, offs - 1, NULL);
 230
 231         /*
 232          * Some macros break next-line scopes; otherwise, remember
 233          * whether we are in next-line scope for a block head.
 234          */
 235
 236         man_breakscope(man, tok);
 237         bline = man->flags & MAN_BLINE;
 238
 239         /*
 240          * If the line in next-line scope ends with \c, keep the
 241          * next-line scope open for the subsequent input line.
 242          * That is not at all portable, only groff >= 1.22.4
 243          * does it, but *if* this weird idiom occurs in a manual
 244          * page, that's very likely what the author intended.
 245          */
 246
 247         if (bline && man_hasc(buf + offs))
 248                 bline = 0;
 249
 250         /* Call to handler... */
 251
 252         (*man_macro(tok)->fp)(man, tok, ln, ppos, &offs, buf);
 253
 254         /* In quick mode (for mandocdb), abort after the NAME section. */
 255
 256         if (man->quick && tok == MAN_SH) {
 257                 n = man->last;
 258                 if (n->type == ROFFT_BODY &&
 259                     strcmp(n->prev->child->string, "NAME"))
 260                         return 2;
 261         }
 262
 263         /*
 264          * If we are in a next-line scope for a block head,
 265          * close it out now and switch to the body,
 266          * unless the next-line scope is allowed to continue.
 267          */
 268
 269         if (bline == 0 ||
 270             (man->flags & MAN_BLINE) == 0 ||
 271             man->flags & MAN_ELINE ||
 272             man_macro(tok)->flags & MAN_NSCOPED)
 273                 return 1;
 274
 275         man->flags &= ~MAN_BLINE;
 276         man_unscope(man, man->last->parent);
 277         roff_body_alloc(man, ln, ppos, man->last->tok);
 278         return 1;
 279 }
 280
 281 void
 282 man_breakscope(struct roff_man *man, int tok)
 283 {
 284         struct roff_node *n;
 285
 286         /*
 287          * An element next line scope is open,
 288          * and the new macro is not allowed inside elements.
 289          * Delete the element that is being broken.
 290          */
 291
 292         if (man->flags & MAN_ELINE && (tok < MAN_TH ||
 293             (man_macro(tok)->flags & MAN_NSCOPED) == 0)) {
 294                 n = man->last;
 295                 if (n->type == ROFFT_TEXT)
 296                         n = n->parent;
 297                 if (n->tok < MAN_TH ||
 298                     (man_macro(n->tok)->flags & (MAN_NSCOPED | MAN_ESCOPED))
 299                      == MAN_NSCOPED)
 300                         n = n->parent;
 301
 302                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 303                     n->line, n->pos, "%s breaks %s",
 304                     roff_name[tok], roff_name[n->tok]);
 305
 306                 roff_node_delete(man, n);
 307                 man->flags &= ~MAN_ELINE;
 308         }
 309
 310         /*
 311          * Weird special case:
 312          * Switching fill mode closes section headers.
 313          */
 314
 315         if (man->flags & MAN_BLINE &&
 316             (tok == MAN_nf || tok == MAN_fi) &&
 317             (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
 318                 n = man->last;
 319                 man_unscope(man, n);
 320                 roff_body_alloc(man, n->line, n->pos, n->tok);
 321                 man->flags &= ~MAN_BLINE;
 322         }
 323
 324         /*
 325          * A block header next line scope is open,
 326          * and the new macro is not allowed inside block headers.
 327          * Delete the block that is being broken.
 328          */
 329
 330         if (man->flags & MAN_BLINE && (tok < MAN_TH ||
 331             man_macro(tok)->flags & MAN_XSCOPE)) {
 332                 n = man->last;
 333                 if (n->type == ROFFT_TEXT)
 334                         n = n->parent;
 335                 if (n->tok < MAN_TH ||
 336                     (man_macro(n->tok)->flags & MAN_XSCOPE) == 0)
 337                         n = n->parent;
 338
 339                 assert(n->type == ROFFT_HEAD);
 340                 n = n->parent;
 341                 assert(n->type == ROFFT_BLOCK);
 342                 assert(man_macro(n->tok)->flags & MAN_BSCOPED);
 343
 344                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 345                     n->line, n->pos, "%s breaks %s",
 346                     roff_name[tok], roff_name[n->tok]);
 347
 348                 roff_node_delete(man, n);
 349                 man->flags &= ~MAN_BLINE;
 350         }
 351 }
 352
 353 void
 354 man_state(struct roff_man *man, struct roff_node *n)
 355 {
 356
 357         switch(n->tok) {
 358         case MAN_nf:
 359         case MAN_EX:
 360                 if (man->flags & MAN_LITERAL && ! (n->flags & NODE_VALID))
 361                         mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 362                             n->line, n->pos, "nf");
 363                 man->flags |= MAN_LITERAL;
 364                 break;
 365         case MAN_fi:
 366         case MAN_EE:
 367                 if ( ! (man->flags & MAN_LITERAL) &&
 368                      ! (n->flags & NODE_VALID))
 369                         mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 370                             n->line, n->pos, "fi");
 371                 man->flags &= ~MAN_LITERAL;
 372                 break;
 373         default:
 374                 break;
 375         }
 376         man->last->flags |= NODE_VALID;
 377 }
 378
 379 void
 380 man_validate(struct roff_man *man)
 381 {
 382
 383         man->last = man->first;
 384         man_node_validate(man);
 385         man->flags &= ~MAN_LITERAL;
 386 }