man.c

   1 /*      $Id: man.c,v 1.181 2018/12/14 05:18:02 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, line, 0, NULL);
 126                         return 1;
 127                 }
 128                 if (man->last->tok == MAN_SH || man->last->tok == MAN_SS)
 129                         return 1;
 130                 if (man->last->type == ROFFT_TEXT &&
 131                     ((ep = man_hasc(man->last->string)) != NULL)) {
 132                         *ep = '\0';
 133                         return 1;
 134                 }
 135                 roff_elem_alloc(man, line, offs, ROFF_sp);
 136                 man->next = ROFF_NEXT_SIBLING;
 137                 return 1;
 138         }
 139
 140         /*
 141          * Warn if the last un-escaped character is whitespace. Then
 142          * strip away the remaining spaces (tabs stay!).
 143          */
 144
 145         i = (int)strlen(buf);
 146         assert(i);
 147
 148         if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
 149                 if (i > 1 && '\\' != buf[i - 2])
 150                         mandoc_msg(MANDOCERR_SPACE_EOL, line, i - 1, NULL);
 151
 152                 for (--i; i && ' ' == buf[i]; i--)
 153                         /* Spin back to non-space. */ ;
 154
 155                 /* Jump ahead of escaped whitespace. */
 156                 i += '\\' == buf[i] ? 2 : 1;
 157
 158                 buf[i] = '\0';
 159         }
 160         roff_word_alloc(man, line, offs, buf + offs);
 161
 162         /*
 163          * End-of-sentence check.  If the last character is an unescaped
 164          * EOS character, then flag the node as being the end of a
 165          * sentence.  The front-end will know how to interpret this.
 166          */
 167
 168         assert(i);
 169         if (mandoc_eos(buf, (size_t)i))
 170                 man->last->flags |= NODE_EOS;
 171
 172         man_descope(man, line, offs, buf + offs);
 173         return 1;
 174 }
 175
 176 static int
 177 man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
 178 {
 179         struct roff_node *n;
 180         const char      *cp;
 181         size_t           sz;
 182         enum roff_tok    tok;
 183         int              ppos;
 184         int              bline;
 185
 186         /* Determine the line macro. */
 187
 188         ppos = offs;
 189         tok = TOKEN_NONE;
 190         for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
 191                 offs++;
 192         if (sz > 0 && sz < 4)
 193                 tok = roffhash_find(man->manmac, buf + ppos, sz);
 194         if (tok == TOKEN_NONE) {
 195                 mandoc_msg(MANDOCERR_MACRO, ln, ppos, "%s", buf + ppos - 1);
 196                 return 1;
 197         }
 198
 199         /* Skip a leading escape sequence or tab. */
 200
 201         switch (buf[offs]) {
 202         case '\\':
 203                 cp = buf + offs + 1;
 204                 mandoc_escape(&cp, NULL, NULL);
 205                 offs = cp - buf;
 206                 break;
 207         case '\t':
 208                 offs++;
 209                 break;
 210         default:
 211                 break;
 212         }
 213
 214         /* Jump to the next non-whitespace word. */
 215
 216         while (buf[offs] == ' ')
 217                 offs++;
 218
 219         /*
 220          * Trailing whitespace.  Note that tabs are allowed to be passed
 221          * into the parser as "text", so we only warn about spaces here.
 222          */
 223
 224         if (buf[offs] == '\0' && buf[offs - 1] == ' ')
 225                 mandoc_msg(MANDOCERR_SPACE_EOL, ln, offs - 1, NULL);
 226
 227         /*
 228          * Some macros break next-line scopes; otherwise, remember
 229          * whether we are in next-line scope for a block head.
 230          */
 231
 232         man_breakscope(man, tok);
 233         bline = man->flags & MAN_BLINE;
 234
 235         /*
 236          * If the line in next-line scope ends with \c, keep the
 237          * next-line scope open for the subsequent input line.
 238          * That is not at all portable, only groff >= 1.22.4
 239          * does it, but *if* this weird idiom occurs in a manual
 240          * page, that's very likely what the author intended.
 241          */
 242
 243         if (bline && man_hasc(buf + offs))
 244                 bline = 0;
 245
 246         /* Call to handler... */
 247
 248         (*man_macro(tok)->fp)(man, tok, ln, ppos, &offs, buf);
 249
 250         /* In quick mode (for mandocdb), abort after the NAME section. */
 251
 252         if (man->quick && tok == MAN_SH) {
 253                 n = man->last;
 254                 if (n->type == ROFFT_BODY &&
 255                     strcmp(n->prev->child->string, "NAME"))
 256                         return 2;
 257         }
 258
 259         /*
 260          * If we are in a next-line scope for a block head,
 261          * close it out now and switch to the body,
 262          * unless the next-line scope is allowed to continue.
 263          */
 264
 265         if (bline == 0 ||
 266             (man->flags & MAN_BLINE) == 0 ||
 267             man->flags & MAN_ELINE ||
 268             man_macro(tok)->flags & MAN_NSCOPED)
 269                 return 1;
 270
 271         man->flags &= ~MAN_BLINE;
 272         man_unscope(man, man->last->parent);
 273         roff_body_alloc(man, ln, ppos, man->last->tok);
 274         return 1;
 275 }
 276
 277 void
 278 man_breakscope(struct roff_man *man, int tok)
 279 {
 280         struct roff_node *n;
 281
 282         /*
 283          * An element next line scope is open,
 284          * and the new macro is not allowed inside elements.
 285          * Delete the element that is being broken.
 286          */
 287
 288         if (man->flags & MAN_ELINE && (tok < MAN_TH ||
 289             (man_macro(tok)->flags & MAN_NSCOPED) == 0)) {
 290                 n = man->last;
 291                 if (n->type == ROFFT_TEXT)
 292                         n = n->parent;
 293                 if (n->tok < MAN_TH ||
 294                     (man_macro(n->tok)->flags & (MAN_NSCOPED | MAN_ESCOPED))
 295                      == MAN_NSCOPED)
 296                         n = n->parent;
 297
 298                 mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
 299                     "%s breaks %s", roff_name[tok], roff_name[n->tok]);
 300
 301                 roff_node_delete(man, n);
 302                 man->flags &= ~MAN_ELINE;
 303         }
 304
 305         /*
 306          * Weird special case:
 307          * Switching fill mode closes section headers.
 308          */
 309
 310         if (man->flags & MAN_BLINE &&
 311             (tok == MAN_nf || tok == MAN_fi) &&
 312             (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
 313                 n = man->last;
 314                 man_unscope(man, n);
 315                 roff_body_alloc(man, n->line, n->pos, n->tok);
 316                 man->flags &= ~MAN_BLINE;
 317         }
 318
 319         /*
 320          * A block header next line scope is open,
 321          * and the new macro is not allowed inside block headers.
 322          * Delete the block that is being broken.
 323          */
 324
 325         if (man->flags & MAN_BLINE && (tok < MAN_TH ||
 326             man_macro(tok)->flags & MAN_XSCOPE)) {
 327                 n = man->last;
 328                 if (n->type == ROFFT_TEXT)
 329                         n = n->parent;
 330                 if (n->tok < MAN_TH ||
 331                     (man_macro(n->tok)->flags & MAN_XSCOPE) == 0)
 332                         n = n->parent;
 333
 334                 assert(n->type == ROFFT_HEAD);
 335                 n = n->parent;
 336                 assert(n->type == ROFFT_BLOCK);
 337                 assert(man_macro(n->tok)->flags & MAN_BSCOPED);
 338
 339                 mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
 340                     "%s breaks %s", roff_name[tok], roff_name[n->tok]);
 341
 342                 roff_node_delete(man, n);
 343                 man->flags &= ~MAN_BLINE;
 344         }
 345 }
 346
 347 void
 348 man_state(struct roff_man *man, struct roff_node *n)
 349 {
 350
 351         switch(n->tok) {
 352         case MAN_nf:
 353         case MAN_EX:
 354                 if (man->flags & MAN_LITERAL && ! (n->flags & NODE_VALID))
 355                         mandoc_msg(MANDOCERR_NF_SKIP, n->line, n->pos, "nf");
 356                 man->flags |= MAN_LITERAL;
 357                 break;
 358         case MAN_fi:
 359         case MAN_EE:
 360                 if ( ! (man->flags & MAN_LITERAL) &&
 361                      ! (n->flags & NODE_VALID))
 362                         mandoc_msg(MANDOCERR_FI_SKIP, n->line, n->pos, "fi");
 363                 man->flags &= ~MAN_LITERAL;
 364                 break;
 365         default:
 366                 break;
 367         }
 368         man->last->flags |= NODE_VALID;
 369 }
 370
 371 void
 372 man_validate(struct roff_man *man)
 373 {
 374
 375         man->last = man->first;
 376         man_node_validate(man);
 377         man->flags &= ~MAN_LITERAL;
 378 }