man.c

   1 /* $Id: man.c,v 1.188 2022/04/28 10:26:37 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2013-2015,2017-2019,2022 Ingo Schwarze <schwarze@openbsd.org>
   4  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   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 /*
  74  * Rewind all open next-line scopes.
  75  */
  76 void
  77 man_descope(struct roff_man *man, int line, int offs, char *start)
  78 {
  79         /* Trailing \c keeps next-line scope open. */
  80
  81         if (start != NULL && man_hasc(start) != NULL)
  82                 return;
  83
  84         /*
  85          * Co-ordinate what happens with having a next-line scope open:
  86          * first close out the element scopes (if applicable),
  87          * then close out the block scope (also if applicable).
  88          */
  89
  90         if (man->flags & MAN_ELINE) {
  91                 while (man->last->parent->type != ROFFT_ROOT &&
  92                     man_macro(man->last->parent->tok)->flags & MAN_ESCOPED)
  93                         man_unscope(man, man->last->parent);
  94                 man->flags &= ~MAN_ELINE;
  95         }
  96         if ( ! (man->flags & MAN_BLINE))
  97                 return;
  98         man_unscope(man, man->last->parent);
  99         roff_body_alloc(man, line, offs, man->last->tok);
 100         man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
 101 }
 102
 103 static int
 104 man_ptext(struct roff_man *man, int line, char *buf, int offs)
 105 {
 106         int              i;
 107         char            *ep;
 108
 109         /* In no-fill mode, whitespace is preserved on text lines. */
 110
 111         if (man->flags & ROFF_NOFILL) {
 112                 roff_word_alloc(man, line, offs, buf + offs);
 113                 man_descope(man, line, offs, buf + offs);
 114                 return 1;
 115         }
 116
 117         for (i = offs; buf[i] == ' '; i++)
 118                 /* Skip leading whitespace. */ ;
 119
 120         /*
 121          * Blank lines are ignored in next line scope
 122          * and right after headings and cancel preceding \c,
 123          * but add a single vertical space elsewhere.
 124          */
 125
 126         if (buf[i] == '\0') {
 127                 if (man->flags & (MAN_ELINE | MAN_BLINE)) {
 128                         mandoc_msg(MANDOCERR_BLK_BLANK, line, 0, NULL);
 129                         return 1;
 130                 }
 131                 if (man->last->tok == MAN_SH || man->last->tok == MAN_SS)
 132                         return 1;
 133                 if (man->last->type == ROFFT_TEXT &&
 134                     ((ep = man_hasc(man->last->string)) != NULL)) {
 135                         *ep = '\0';
 136                         return 1;
 137                 }
 138                 roff_elem_alloc(man, line, offs, ROFF_sp);
 139                 man->next = ROFF_NEXT_SIBLING;
 140                 return 1;
 141         }
 142
 143         /*
 144          * Warn if the last un-escaped character is whitespace. Then
 145          * strip away the remaining spaces (tabs stay!).
 146          */
 147
 148         i = (int)strlen(buf);
 149         assert(i);
 150
 151         if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
 152                 if (i > 1 && '\\' != buf[i - 2])
 153                         mandoc_msg(MANDOCERR_SPACE_EOL, line, i - 1, NULL);
 154
 155                 for (--i; i && ' ' == buf[i]; i--)
 156                         /* Spin back to non-space. */ ;
 157
 158                 /* Jump ahead of escaped whitespace. */
 159                 i += '\\' == buf[i] ? 2 : 1;
 160
 161                 buf[i] = '\0';
 162         }
 163         roff_word_alloc(man, line, offs, buf + offs);
 164
 165         /*
 166          * End-of-sentence check.  If the last character is an unescaped
 167          * EOS character, then flag the node as being the end of a
 168          * sentence.  The front-end will know how to interpret this.
 169          */
 170
 171         assert(i);
 172         if (mandoc_eos(buf, (size_t)i))
 173                 man->last->flags |= NODE_EOS;
 174
 175         man_descope(man, line, offs, buf + offs);
 176         return 1;
 177 }
 178
 179 static int
 180 man_pmacro(struct roff_man *man, int ln, char *buf, int offs)
 181 {
 182         struct roff_node *n;
 183         const char      *cp;
 184         size_t           sz;
 185         enum roff_tok    tok;
 186         int              ppos;
 187         int              bline;
 188
 189         /* Determine the line macro. */
 190
 191         ppos = offs;
 192         tok = TOKEN_NONE;
 193         for (sz = 0; sz < 4 && strchr(" \t\\", buf[offs]) == NULL; sz++)
 194                 offs++;
 195         if (sz > 0 && sz < 4)
 196                 tok = roffhash_find(man->manmac, buf + ppos, sz);
 197         if (tok == TOKEN_NONE) {
 198                 mandoc_msg(MANDOCERR_MACRO, ln, ppos, "%s", 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, ln, offs - 1, NULL);
 229
 230         /*
 231          * Some macros break next-line scopes; otherwise, remember
 232          * whether we are in next-line scope for a block head.
 233          */
 234
 235         man_breakscope(man, tok);
 236         bline = man->flags & MAN_BLINE;
 237
 238         /*
 239          * If the line in next-line scope ends with \c, keep the
 240          * next-line scope open for the subsequent input line.
 241          * That is not at all portable, only groff >= 1.22.4
 242          * does it, but *if* this weird idiom occurs in a manual
 243          * page, that's very likely what the author intended.
 244          */
 245
 246         if (bline && man_hasc(buf + offs))
 247                 bline = 0;
 248
 249         /* Call to handler... */
 250
 251         (*man_macro(tok)->fp)(man, tok, ln, ppos, &offs, buf);
 252
 253         /* In quick mode (for mandocdb), abort after the NAME section. */
 254
 255         if (man->quick && tok == MAN_SH) {
 256                 n = man->last;
 257                 if (n->type == ROFFT_BODY &&
 258                     strcmp(n->prev->child->string, "NAME"))
 259                         return 2;
 260         }
 261
 262         /*
 263          * If we are in a next-line scope for a block head,
 264          * close it out now and switch to the body,
 265          * unless the next-line scope is allowed to continue.
 266          */
 267
 268         if (bline == 0 ||
 269             (man->flags & MAN_BLINE) == 0 ||
 270             man->flags & MAN_ELINE ||
 271             man_macro(tok)->flags & MAN_NSCOPED)
 272                 return 1;
 273
 274         man_unscope(man, man->last->parent);
 275         roff_body_alloc(man, ln, ppos, man->last->tok);
 276         man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
 277         return 1;
 278 }
 279
 280 /*
 281  * Rewind open next-line scopes
 282  * unless the tok request or macro is allowed inside them.
 283  */
 284 void
 285 man_breakscope(struct roff_man *man, int tok)
 286 {
 287         struct roff_node *n;
 288
 289         /*
 290          * An element next line scope is open,
 291          * and the new macro is not allowed inside elements.
 292          * Delete the element that is being broken.
 293          */
 294
 295         if (man->flags & MAN_ELINE && (tok < MAN_TH ||
 296             (man_macro(tok)->flags & MAN_NSCOPED) == 0)) {
 297                 n = man->last;
 298                 if (n->type == ROFFT_TEXT)
 299                         n = n->parent;
 300                 if (n->tok < MAN_TH ||
 301                     (man_macro(n->tok)->flags & (MAN_NSCOPED | MAN_ESCOPED))
 302                      == MAN_NSCOPED)
 303                         n = n->parent;
 304                 for (;;) {
 305                         mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
 306                             "%s breaks %s", roff_name[tok], roff_name[n->tok]);
 307                         if (n->parent->type != ROFFT_ELEM ||
 308                             (man_macro(n->parent->tok)->flags &
 309                              MAN_ESCOPED) == 0)
 310                                 break;
 311                         n = n->parent;
 312                 }
 313                 roff_node_delete(man, n);
 314                 man->flags &= ~MAN_ELINE;
 315         }
 316
 317         /*
 318          * Weird special case:
 319          * Switching fill mode closes section headers.
 320          */
 321
 322         if (man->flags & MAN_BLINE &&
 323             (tok == ROFF_nf || tok == ROFF_fi) &&
 324             (man->last->tok == MAN_SH || man->last->tok == MAN_SS)) {
 325                 n = man->last;
 326                 man_unscope(man, n);
 327                 roff_body_alloc(man, n->line, n->pos, n->tok);
 328                 man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
 329         }
 330
 331         /*
 332          * A block header next line scope is open,
 333          * and the new macro is not allowed inside block headers.
 334          * Delete the block that is being broken.
 335          */
 336
 337         if (man->flags & MAN_BLINE && tok != ROFF_nf && tok != ROFF_fi &&
 338             (tok < MAN_TH || man_macro(tok)->flags & MAN_XSCOPE)) {
 339                 n = man->last;
 340                 if (n->type == ROFFT_TEXT)
 341                         n = n->parent;
 342                 if (n->tok < MAN_TH ||
 343                     (man_macro(n->tok)->flags & MAN_XSCOPE) == 0)
 344                         n = n->parent;
 345
 346                 assert(n->type == ROFFT_HEAD);
 347                 n = n->parent;
 348                 assert(n->type == ROFFT_BLOCK);
 349                 assert(man_macro(n->tok)->flags & MAN_BSCOPED);
 350
 351                 mandoc_msg(MANDOCERR_BLK_LINE, n->line, n->pos,
 352                     "%s breaks %s", roff_name[tok], roff_name[n->tok]);
 353
 354                 roff_node_delete(man, n);
 355                 man->flags &= ~(MAN_BLINE | ROFF_NONOFILL);
 356         }
 357 }