tbl_layout.c

   1 /*      $Id: tbl_layout.c,v 1.25 2014/03/28 23:26:25 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2012, 2014 Ingo Schwarze <schwarze@openbsd.org>
   5  *
   6  * Permission to use, copy, modify, and distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18 #ifdef HAVE_CONFIG_H
  19 #include "config.h"
  20 #endif
  21
  22 #include <ctype.h>
  23 #include <stdlib.h>
  24 #include <string.h>
  25 #include <time.h>
  26
  27 #include "mandoc.h"
  28 #include "mandoc_aux.h"
  29 #include "libmandoc.h"
  30 #include "libroff.h"
  31
  32 struct  tbl_phrase {
  33         char             name;
  34         enum tbl_cellt   key;
  35 };
  36
  37 /*
  38  * FIXME: we can make this parse a lot nicer by, when an error is
  39  * encountered in a layout key, bailing to the next key (i.e. to the
  40  * next whitespace then continuing).
  41  */
  42
  43 #define KEYS_MAX         11
  44
  45 static  const struct tbl_phrase keys[KEYS_MAX] = {
  46         { 'c',           TBL_CELL_CENTRE },
  47         { 'r',           TBL_CELL_RIGHT },
  48         { 'l',           TBL_CELL_LEFT },
  49         { 'n',           TBL_CELL_NUMBER },
  50         { 's',           TBL_CELL_SPAN },
  51         { 'a',           TBL_CELL_LONG },
  52         { '^',           TBL_CELL_DOWN },
  53         { '-',           TBL_CELL_HORIZ },
  54         { '_',           TBL_CELL_HORIZ },
  55         { '=',           TBL_CELL_DHORIZ }
  56 };
  57
  58 static  int              mods(struct tbl_node *, struct tbl_cell *,
  59                                 int, const char *, int *);
  60 static  int              cell(struct tbl_node *, struct tbl_row *,
  61                                 int, const char *, int *);
  62 static  void             row(struct tbl_node *, int, const char *, int *);
  63 static  struct tbl_cell *cell_alloc(struct tbl_node *, struct tbl_row *,
  64                                 enum tbl_cellt, int vert);
  65
  66 static int
  67 mods(struct tbl_node *tbl, struct tbl_cell *cp,
  68                 int ln, const char *p, int *pos)
  69 {
  70         char             buf[5];
  71         int              i;
  72
  73         /* Not all types accept modifiers. */
  74
  75         switch (cp->pos) {
  76         case (TBL_CELL_DOWN):
  77                 /* FALLTHROUGH */
  78         case (TBL_CELL_HORIZ):
  79                 /* FALLTHROUGH */
  80         case (TBL_CELL_DHORIZ):
  81                 return(1);
  82         default:
  83                 break;
  84         }
  85
  86 mod:
  87         /*
  88          * XXX: since, at least for now, modifiers are non-conflicting
  89          * (are separable by value, regardless of position), we let
  90          * modifiers come in any order.  The existing tbl doesn't let
  91          * this happen.
  92          */
  93         switch (p[*pos]) {
  94         case ('\0'):
  95                 /* FALLTHROUGH */
  96         case (' '):
  97                 /* FALLTHROUGH */
  98         case ('\t'):
  99                 /* FALLTHROUGH */
 100         case (','):
 101                 /* FALLTHROUGH */
 102         case ('.'):
 103                 /* FALLTHROUGH */
 104         case ('|'):
 105                 return(1);
 106         default:
 107                 break;
 108         }
 109
 110         /* Throw away parenthesised expression. */
 111
 112         if ('(' == p[*pos]) {
 113                 (*pos)++;
 114                 while (p[*pos] && ')' != p[*pos])
 115                         (*pos)++;
 116                 if (')' == p[*pos]) {
 117                         (*pos)++;
 118                         goto mod;
 119                 }
 120                 mandoc_msg(MANDOCERR_TBLLAYOUT,
 121                                 tbl->parse, ln, *pos, NULL);
 122                 return(0);
 123         }
 124
 125         /* Parse numerical spacing from modifier string. */
 126
 127         if (isdigit((unsigned char)p[*pos])) {
 128                 for (i = 0; i < 4; i++) {
 129                         if ( ! isdigit((unsigned char)p[*pos + i]))
 130                                 break;
 131                         buf[i] = p[*pos + i];
 132                 }
 133                 buf[i] = '\0';
 134
 135                 /* No greater than 4 digits. */
 136
 137                 if (4 == i) {
 138                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 139                                         ln, *pos, NULL);
 140                         return(0);
 141                 }
 142
 143                 *pos += i;
 144                 cp->spacing = (size_t)atoi(buf);
 145
 146                 goto mod;
 147                 /* NOTREACHED */
 148         }
 149
 150         /* TODO: GNU has many more extensions. */
 151
 152         switch (tolower((unsigned char)p[(*pos)++])) {
 153         case ('z'):
 154                 cp->flags |= TBL_CELL_WIGN;
 155                 goto mod;
 156         case ('u'):
 157                 cp->flags |= TBL_CELL_UP;
 158                 goto mod;
 159         case ('e'):
 160                 cp->flags |= TBL_CELL_EQUAL;
 161                 goto mod;
 162         case ('t'):
 163                 cp->flags |= TBL_CELL_TALIGN;
 164                 goto mod;
 165         case ('d'):
 166                 cp->flags |= TBL_CELL_BALIGN;
 167                 goto mod;
 168         case ('w'):  /* XXX for now, ignore minimal column width */
 169                 goto mod;
 170         case ('f'):
 171                 break;
 172         case ('r'):
 173                 /* FALLTHROUGH */
 174         case ('b'):
 175                 /* FALLTHROUGH */
 176         case ('i'):
 177                 (*pos)--;
 178                 break;
 179         default:
 180                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 181                                 ln, *pos - 1, NULL);
 182                 return(0);
 183         }
 184
 185         switch (tolower((unsigned char)p[(*pos)++])) {
 186         case ('3'):
 187                 /* FALLTHROUGH */
 188         case ('b'):
 189                 cp->flags |= TBL_CELL_BOLD;
 190                 goto mod;
 191         case ('2'):
 192                 /* FALLTHROUGH */
 193         case ('i'):
 194                 cp->flags |= TBL_CELL_ITALIC;
 195                 goto mod;
 196         case ('1'):
 197                 /* FALLTHROUGH */
 198         case ('r'):
 199                 goto mod;
 200         default:
 201                 break;
 202         }
 203
 204         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 205                         ln, *pos - 1, NULL);
 206         return(0);
 207 }
 208
 209 static int
 210 cell(struct tbl_node *tbl, struct tbl_row *rp,
 211                 int ln, const char *p, int *pos)
 212 {
 213         int              vert, i;
 214         enum tbl_cellt   c;
 215
 216         /* Handle vertical lines. */
 217
 218         for (vert = 0; '|' == p[*pos]; ++*pos)
 219                 vert++;
 220         while (' ' == p[*pos])
 221                 (*pos)++;
 222
 223         /* Handle trailing vertical lines */
 224
 225         if ('.' == p[*pos] || '\0' == p[*pos]) {
 226                 rp->vert = vert;
 227                 return(1);
 228         }
 229
 230         /* Parse the column position (`c', `l', `r', ...). */
 231
 232         for (i = 0; i < KEYS_MAX; i++)
 233                 if (tolower((unsigned char)p[*pos]) == keys[i].name)
 234                         break;
 235
 236         if (KEYS_MAX == i) {
 237                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 238                                 ln, *pos, NULL);
 239                 return(0);
 240         }
 241
 242         c = keys[i].key;
 243
 244         /*
 245          * If a span cell is found first, raise a warning and abort the
 246          * parse.  If a span cell is found and the last layout element
 247          * isn't a "normal" layout, bail.
 248          *
 249          * FIXME: recover from this somehow?
 250          */
 251
 252         if (TBL_CELL_SPAN == c) {
 253                 if (NULL == rp->first) {
 254                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 255                                         ln, *pos, NULL);
 256                         return(0);
 257                 } else if (rp->last)
 258                         switch (rp->last->pos) {
 259                         case (TBL_CELL_HORIZ):
 260                         case (TBL_CELL_DHORIZ):
 261                                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 262                                                 ln, *pos, NULL);
 263                                 return(0);
 264                         default:
 265                                 break;
 266                         }
 267         }
 268
 269         /*
 270          * If a vertical spanner is found, we may not be in the first
 271          * row.
 272          */
 273
 274         if (TBL_CELL_DOWN == c && rp == tbl->first_row) {
 275                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos, NULL);
 276                 return(0);
 277         }
 278
 279         (*pos)++;
 280
 281         /* Disallow adjacent spacers. */
 282
 283         if (vert > 2) {
 284                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos - 1, NULL);
 285                 return(0);
 286         }
 287
 288         /* Allocate cell then parse its modifiers. */
 289
 290         return(mods(tbl, cell_alloc(tbl, rp, c, vert), ln, p, pos));
 291 }
 292
 293
 294 static void
 295 row(struct tbl_node *tbl, int ln, const char *p, int *pos)
 296 {
 297         struct tbl_row  *rp;
 298
 299 row:    /*
 300          * EBNF describing this section:
 301          *
 302          * row          ::= row_list [:space:]* [.]?[\n]
 303          * row_list     ::= [:space:]* row_elem row_tail
 304          * row_tail     ::= [:space:]*[,] row_list |
 305          *                  epsilon
 306          * row_elem     ::= [\t\ ]*[:alpha:]+
 307          */
 308
 309         rp = mandoc_calloc(1, sizeof(struct tbl_row));
 310         if (tbl->last_row)
 311                 tbl->last_row->next = rp;
 312         else
 313                 tbl->first_row = rp;
 314         tbl->last_row = rp;
 315
 316 cell:
 317         while (isspace((unsigned char)p[*pos]))
 318                 (*pos)++;
 319
 320         /* Safely exit layout context. */
 321
 322         if ('.' == p[*pos]) {
 323                 tbl->part = TBL_PART_DATA;
 324                 if (NULL == tbl->first_row)
 325                         mandoc_msg(MANDOCERR_TBLNOLAYOUT, tbl->parse,
 326                                         ln, *pos, NULL);
 327                 (*pos)++;
 328                 return;
 329         }
 330
 331         /* End (and possibly restart) a row. */
 332
 333         if (',' == p[*pos]) {
 334                 (*pos)++;
 335                 goto row;
 336         } else if ('\0' == p[*pos])
 337                 return;
 338
 339         if ( ! cell(tbl, rp, ln, p, pos))
 340                 return;
 341
 342         goto cell;
 343         /* NOTREACHED */
 344 }
 345
 346 int
 347 tbl_layout(struct tbl_node *tbl, int ln, const char *p)
 348 {
 349         int              pos;
 350
 351         pos = 0;
 352         row(tbl, ln, p, &pos);
 353
 354         /* Always succeed. */
 355         return(1);
 356 }
 357
 358 static struct tbl_cell *
 359 cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos,
 360                 int vert)
 361 {
 362         struct tbl_cell *p, *pp;
 363         struct tbl_head *h, *hp;
 364
 365         p = mandoc_calloc(1, sizeof(struct tbl_cell));
 366
 367         if (NULL != (pp = rp->last)) {
 368                 pp->next = p;
 369                 h = pp->head->next;
 370         } else {
 371                 rp->first = p;
 372                 h = tbl->first_head;
 373         }
 374         rp->last = p;
 375
 376         p->pos = pos;
 377         p->vert = vert;
 378
 379         /* Re-use header. */
 380
 381         if (h) {
 382                 p->head = h;
 383                 return(p);
 384         }
 385
 386         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 387         hp->ident = tbl->opts.cols++;
 388         hp->vert = vert;
 389
 390         if (tbl->last_head) {
 391                 hp->prev = tbl->last_head;
 392                 tbl->last_head->next = hp;
 393         } else
 394                 tbl->first_head = hp;
 395         tbl->last_head = hp;
 396
 397         p->head = hp;
 398         return(p);
 399 }