usr/src/cmd/mandoc/man.c

   1 /*      $Id: man.c,v 1.166 2015/10/22 21:54:23 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2013, 2014, 2015 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 const   char *const __man_macronames[MAN_MAX] = {
  39         "br",           "TH",           "SH",           "SS",
  40         "TP",           "LP",           "PP",           "P",
  41         "IP",           "HP",           "SM",           "SB",
  42         "BI",           "IB",           "BR",           "RB",
  43         "R",            "B",            "I",            "IR",
  44         "RI",           "sp",           "nf",
  45         "fi",           "RE",           "RS",           "DT",
  46         "UC",           "PD",           "AT",           "in",
  47         "ft",           "OP",           "EX",           "EE",
  48         "UR",           "UE",           "ll"
  49         };
  50
  51 const   char * const *man_macronames = __man_macronames;
  52
  53 static  void             man_descope(struct roff_man *, int, int);
  54 static  int              man_ptext(struct roff_man *, int, char *, int);
  55 static  int              man_pmacro(struct roff_man *, int, char *, int);
  56
  57
  58 int
  59 man_parseln(struct roff_man *man, int ln, char *buf, int offs)
  60 {
  61
  62         if (man->last->type != ROFFT_EQN || ln > man->last->line)
  63                 man->flags |= MAN_NEWLINE;
  64
  65         return roff_getcontrol(man->roff, buf, &offs) ?
  66             man_pmacro(man, ln, buf, offs) :
  67             man_ptext(man, ln, buf, offs);
  68 }
  69
  70 static void
  71 man_descope(struct roff_man *man, int line, int offs)
  72 {
  73         /*
  74          * Co-ordinate what happens with having a next-line scope open:
  75          * first close out the element scope (if applicable), then close
  76          * out the block scope (also if applicable).
  77          */
  78
  79         if (man->flags & MAN_ELINE) {
  80                 man->flags &= ~MAN_ELINE;
  81                 man_unscope(man, man->last->parent);
  82         }
  83         if ( ! (man->flags & MAN_BLINE))
  84                 return;
  85         man->flags &= ~MAN_BLINE;
  86         man_unscope(man, man->last->parent);
  87         roff_body_alloc(man, line, offs, man->last->tok);
  88 }
  89
  90 static int
  91 man_ptext(struct roff_man *man, int line, char *buf, int offs)
  92 {
  93         int              i;
  94
  95         /* Literal free-form text whitespace is preserved. */
  96
  97         if (man->flags & MAN_LITERAL) {
  98                 roff_word_alloc(man, line, offs, buf + offs);
  99                 man_descope(man, line, offs);
 100                 return 1;
 101         }
 102
 103         for (i = offs; buf[i] == ' '; i++)
 104                 /* Skip leading whitespace. */ ;
 105
 106         /*
 107          * Blank lines are ignored right after headings
 108          * but add a single vertical space elsewhere.
 109          */
 110
 111         if (buf[i] == '\0') {
 112                 /* Allocate a blank entry. */
 113                 if (man->last->tok != MAN_SH &&
 114                     man->last->tok != MAN_SS) {
 115                         roff_elem_alloc(man, line, offs, MAN_sp);
 116                         man->next = ROFF_NEXT_SIBLING;
 117                 }
 118                 return 1;
 119         }
 120
 121         /*
 122          * Warn if the last un-escaped character is whitespace. Then
 123          * strip away the remaining spaces (tabs stay!).
 124          */
 125
 126         i = (int)strlen(buf);
 127         assert(i);
 128
 129         if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
 130                 if (i > 1 && '\\' != buf[i - 2])
 131                         mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 132                             line, i - 1, NULL);
 133
 134                 for (--i; i && ' ' == buf[i]; i--)
 135                         /* Spin back to non-space. */ ;
 136
 137                 /* Jump ahead of escaped whitespace. */
 138                 i += '\\' == buf[i] ? 2 : 1;
 139
 140                 buf[i] = '\0';
 141         }
 142         roff_word_alloc(man, line, offs, buf + offs);
 143
 144         /*
 145          * End-of-sentence check.  If the last character is an unescaped
 146          * EOS character, then flag the node as being the end of a
 147          * sentence.  The front-end will know how to interpret this.
 148          */
 149
 150         assert(i);
 151         if (mandoc_eos(buf, (size_t)i))
 152                 man->last->flags |= MAN_EOS;
 153
 154         man_descope(man, line, offs);
 155         return 1;
 156 }
 157
 158 static int
 159 man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
 160 {
 161         struct roff_node *n;
 162         const char      *cp;
 163         int              tok;
 164         int              i, ppos;
 165         int              bline;
 166         char             mac[5];
 167
 168         ppos = offs;
 169
 170         /*
 171          * Copy the first word into a nil-terminated buffer.
 172          * Stop when a space, tab, escape, or eoln is encountered.
 173          */
 174
 175         i = 0;
 176         while (i < 4 && strchr(" \t\\", buf[offs]) == NULL)
 177                 mac[i++] = buf[offs++];
 178
 179         mac[i] = '\0';
 180
 181         tok = (i > 0 && i < 4) ? man_hash_find(mac) : TOKEN_NONE;
 182
 183         if (tok == TOKEN_NONE) {
 184                 mandoc_msg(MANDOCERR_MACRO, man->parse,
 185                     ln, ppos, buf + ppos - 1);
 186                 return 1;
 187         }
 188
 189         /* Skip a leading escape sequence or tab. */
 190
 191         switch (buf[offs]) {
 192         case '\\':
 193                 cp = buf + offs + 1;
 194                 mandoc_escape(&cp, NULL, NULL);
 195                 offs = cp - buf;
 196                 break;
 197         case '\t':
 198                 offs++;
 199                 break;
 200         default:
 201                 break;
 202         }
 203
 204         /* Jump to the next non-whitespace word. */
 205
 206         while (buf[offs] && buf[offs] == ' ')
 207                 offs++;
 208
 209         /*
 210          * Trailing whitespace.  Note that tabs are allowed to be passed
 211          * into the parser as "text", so we only warn about spaces here.
 212          */
 213
 214         if (buf[offs] == '\0' && buf[offs - 1] == ' ')
 215                 mandoc_msg(MANDOCERR_SPACE_EOL, man->parse,
 216                     ln, offs - 1, NULL);
 217
 218         /*
 219          * Some macros break next-line scopes; otherwise, remember
 220          * whether we are in next-line scope for a block head.
 221          */
 222
 223         man_breakscope(man, tok);
 224         bline = man->flags & MAN_BLINE;
 225
 226         /* Call to handler... */
 227
 228         assert(man_macros[tok].fp);
 229         (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf);
 230
 231         /* In quick mode (for mandocdb), abort after the NAME section. */
 232
 233         if (man->quick && tok == MAN_SH) {
 234                 n = man->last;
 235                 if (n->type == ROFFT_BODY &&
 236                     strcmp(n->prev->child->string, "NAME"))
 237                         return 2;
 238         }
 239
 240         /*
 241          * If we are in a next-line scope for a block head,
 242          * close it out now and switch to the body,
 243          * unless the next-line scope is allowed to continue.
 244          */
 245
 246         if ( ! bline || man->flags & MAN_ELINE ||
 247             man_macros[tok].flags & MAN_NSCOPED)
 248                 return 1;
 249
 250         assert(man->flags & MAN_BLINE);
 251         man->flags &= ~MAN_BLINE;
 252
 253         man_unscope(man, man->last->parent);
 254         roff_body_alloc(man, ln, ppos, man->last->tok);
 255         return 1;
 256 }
 257
 258 void
 259 man_breakscope(struct roff_man *man, int tok)
 260 {
 261         struct roff_node *n;
 262
 263         /*
 264          * An element next line scope is open,
 265          * and the new macro is not allowed inside elements.
 266          * Delete the element that is being broken.
 267          */
 268
 269         if (man->flags & MAN_ELINE && (tok == TOKEN_NONE ||
 270             ! (man_macros[tok].flags & MAN_NSCOPED))) {
 271                 n = man->last;
 272                 assert(n->type != ROFFT_TEXT);
 273                 if (man_macros[n->tok].flags & MAN_NSCOPED)
 274                         n = n->parent;
 275
 276                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 277                     n->line, n->pos, "%s breaks %s",
 278                     tok == TOKEN_NONE ? "TS" : man_macronames[tok],
 279                     man_macronames[n->tok]);
 280
 281                 roff_node_delete(man, n);
 282                 man->flags &= ~MAN_ELINE;
 283         }
 284
 285         /*
 286          * Weird special case:
 287          * Switching fill mode closes section headers.
 288          */
 289
 290         if (man->flags & MAN_BLINE &&
 291             (tok == MAN_nf || tok == MAN_fi) &&
 292             (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
 293                 n = man->last;
 294                 man_unscope(man, n);
 295                 roff_body_alloc(man, n->line, n->pos, n->tok);
 296                 man->flags &= ~MAN_BLINE;
 297         }
 298
 299         /*
 300          * A block header next line scope is open,
 301          * and the new macro is not allowed inside block headers.
 302          * Delete the block that is being broken.
 303          */
 304
 305         if (man->flags & MAN_BLINE && (tok == TOKEN_NONE ||
 306             man_macros[tok].flags & MAN_BSCOPE)) {
 307                 n = man->last;
 308                 if (n->type == ROFFT_TEXT)
 309                         n = n->parent;
 310                 if ( ! (man_macros[n->tok].flags & MAN_BSCOPE))
 311                         n = n->parent;
 312
 313                 assert(n->type == ROFFT_HEAD);
 314                 n = n->parent;
 315                 assert(n->type == ROFFT_BLOCK);
 316                 assert(man_macros[n->tok].flags & MAN_SCOPED);
 317
 318                 mandoc_vmsg(MANDOCERR_BLK_LINE, man->parse,
 319                     n->line, n->pos, "%s breaks %s",
 320                     tok == TOKEN_NONE ? "TS" : man_macronames[tok],
 321                     man_macronames[n->tok]);
 322
 323                 roff_node_delete(man, n);
 324                 man->flags &= ~MAN_BLINE;
 325         }
 326 }
 327
 328 const struct mparse *
 329 man_mparse(const struct roff_man *man)
 330 {
 331
 332         assert(man && man->parse);
 333         return man->parse;
 334 }
 335
 336 void
 337 man_state(struct roff_man *man, struct roff_node *n)
 338 {
 339
 340         switch(n->tok) {
 341         case MAN_nf:
 342         case MAN_EX:
 343                 if (man->flags & MAN_LITERAL && ! (n->flags & MAN_VALID))
 344                         mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 345                             n->line, n->pos, "nf");
 346                 man->flags |= MAN_LITERAL;
 347                 break;
 348         case MAN_fi:
 349         case MAN_EE:
 350                 if ( ! (man->flags & MAN_LITERAL) &&
 351                      ! (n->flags & MAN_VALID))
 352                         mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 353                             n->line, n->pos, "fi");
 354                 man->flags &= ~MAN_LITERAL;
 355                 break;
 356         default:
 357                 break;
 358         }
 359         man->last->flags |= MAN_VALID;
 360 }
 361
 362 void
 363 man_validate(struct roff_man *man)
 364 {
 365
 366         man->last = man->first;
 367         man_node_validate(man);
 368         man->flags &= ~MAN_LITERAL;
 369 }