tbl_layout.c

   1 /*      $Id: tbl_layout.c,v 1.29 2014/10/14 02:16:06 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 #include "config.h"
  19
  20 #include <sys/types.h>
  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
  67 static int
  68 mods(struct tbl_node *tbl, struct tbl_cell *cp,
  69                 int ln, const char *p, int *pos)
  70 {
  71         char             buf[5];
  72         int              i;
  73
  74         /* Not all types accept modifiers. */
  75
  76         switch (cp->pos) {
  77         case TBL_CELL_DOWN:
  78                 /* FALLTHROUGH */
  79         case TBL_CELL_HORIZ:
  80                 /* FALLTHROUGH */
  81         case TBL_CELL_DHORIZ:
  82                 return(1);
  83         default:
  84                 break;
  85         }
  86
  87 mod:
  88         /*
  89          * XXX: since, at least for now, modifiers are non-conflicting
  90          * (are separable by value, regardless of position), we let
  91          * modifiers come in any order.  The existing tbl doesn't let
  92          * this happen.
  93          */
  94         switch (p[*pos]) {
  95         case '\0':
  96                 /* FALLTHROUGH */
  97         case ' ':
  98                 /* FALLTHROUGH */
  99         case '\t':
 100                 /* FALLTHROUGH */
 101         case ',':
 102                 /* FALLTHROUGH */
 103         case '.':
 104                 /* FALLTHROUGH */
 105         case '|':
 106                 return(1);
 107         default:
 108                 break;
 109         }
 110
 111         /* Throw away parenthesised expression. */
 112
 113         if ('(' == p[*pos]) {
 114                 (*pos)++;
 115                 while (p[*pos] && ')' != p[*pos])
 116                         (*pos)++;
 117                 if (')' == p[*pos]) {
 118                         (*pos)++;
 119                         goto mod;
 120                 }
 121                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 122                     ln, *pos, NULL);
 123                 return(0);
 124         }
 125
 126         /* Parse numerical spacing from modifier string. */
 127
 128         if (isdigit((unsigned char)p[*pos])) {
 129                 for (i = 0; i < 4; i++) {
 130                         if ( ! isdigit((unsigned char)p[*pos + i]))
 131                                 break;
 132                         buf[i] = p[*pos + i];
 133                 }
 134                 buf[i] = '\0';
 135
 136                 /* No greater than 4 digits. */
 137
 138                 if (4 == i) {
 139                         mandoc_msg(MANDOCERR_TBLLAYOUT,
 140                             tbl->parse, ln, *pos, NULL);
 141                         return(0);
 142                 }
 143
 144                 *pos += i;
 145                 cp->spacing = (size_t)atoi(buf);
 146
 147                 goto mod;
 148                 /* NOTREACHED */
 149         }
 150
 151         /* TODO: GNU has many more extensions. */
 152
 153         switch (tolower((unsigned char)p[(*pos)++])) {
 154         case 'z':
 155                 cp->flags |= TBL_CELL_WIGN;
 156                 goto mod;
 157         case 'u':
 158                 cp->flags |= TBL_CELL_UP;
 159                 goto mod;
 160         case 'e':
 161                 cp->flags |= TBL_CELL_EQUAL;
 162                 goto mod;
 163         case 't':
 164                 cp->flags |= TBL_CELL_TALIGN;
 165                 goto mod;
 166         case 'd':
 167                 cp->flags |= TBL_CELL_BALIGN;
 168                 goto mod;
 169         case 'w':  /* XXX for now, ignore minimal column width */
 170                 goto mod;
 171         case 'x':
 172                 cp->flags |= TBL_CELL_WMAX;
 173                 goto mod;
 174         case 'f':
 175                 break;
 176         case 'r':
 177                 /* FALLTHROUGH */
 178         case 'b':
 179                 /* FALLTHROUGH */
 180         case 'i':
 181                 (*pos)--;
 182                 break;
 183         default:
 184                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 185                     ln, *pos - 1, NULL);
 186                 return(0);
 187         }
 188
 189         switch (tolower((unsigned char)p[(*pos)++])) {
 190         case '3':
 191                 /* FALLTHROUGH */
 192         case 'b':
 193                 cp->flags |= TBL_CELL_BOLD;
 194                 goto mod;
 195         case '2':
 196                 /* FALLTHROUGH */
 197         case 'i':
 198                 cp->flags |= TBL_CELL_ITALIC;
 199                 goto mod;
 200         case '1':
 201                 /* FALLTHROUGH */
 202         case 'r':
 203                 goto mod;
 204         default:
 205                 break;
 206         }
 207         if (isalnum((unsigned char)p[*pos - 1])) {
 208                 mandoc_vmsg(MANDOCERR_FT_BAD, tbl->parse,
 209                     ln, *pos - 1, "TS f%c", p[*pos - 1]);
 210                 goto mod;
 211         }
 212
 213         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 214             ln, *pos - 1, NULL);
 215         return(0);
 216 }
 217
 218 static int
 219 cell(struct tbl_node *tbl, struct tbl_row *rp,
 220                 int ln, const char *p, int *pos)
 221 {
 222         int              vert, i;
 223         enum tbl_cellt   c;
 224
 225         /* Handle vertical lines. */
 226
 227         for (vert = 0; '|' == p[*pos]; ++*pos)
 228                 vert++;
 229         while (' ' == p[*pos])
 230                 (*pos)++;
 231
 232         /* Handle trailing vertical lines */
 233
 234         if ('.' == p[*pos] || '\0' == p[*pos]) {
 235                 rp->vert = vert;
 236                 return(1);
 237         }
 238
 239         /* Parse the column position (`c', `l', `r', ...). */
 240
 241         for (i = 0; i < KEYS_MAX; i++)
 242                 if (tolower((unsigned char)p[*pos]) == keys[i].name)
 243                         break;
 244
 245         if (KEYS_MAX == i) {
 246                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 247                     ln, *pos, NULL);
 248                 return(0);
 249         }
 250
 251         c = keys[i].key;
 252
 253         /*
 254          * If a span cell is found first, raise a warning and abort the
 255          * parse.  If a span cell is found and the last layout element
 256          * isn't a "normal" layout, bail.
 257          *
 258          * FIXME: recover from this somehow?
 259          */
 260
 261         if (TBL_CELL_SPAN == c) {
 262                 if (NULL == rp->first) {
 263                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 264                             ln, *pos, NULL);
 265                         return(0);
 266                 } else if (rp->last)
 267                         switch (rp->last->pos) {
 268                         case TBL_CELL_HORIZ:
 269                                 /* FALLTHROUGH */
 270                         case TBL_CELL_DHORIZ:
 271                                 mandoc_msg(MANDOCERR_TBLLAYOUT,
 272                                     tbl->parse, ln, *pos, NULL);
 273                                 return(0);
 274                         default:
 275                                 break;
 276                         }
 277         }
 278
 279         /*
 280          * If a vertical spanner is found, we may not be in the first
 281          * row.
 282          */
 283
 284         if (TBL_CELL_DOWN == c && rp == tbl->first_row) {
 285                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos, NULL);
 286                 return(0);
 287         }
 288
 289         (*pos)++;
 290
 291         /* Disallow adjacent spacers. */
 292
 293         if (vert > 2) {
 294                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos - 1, NULL);
 295                 return(0);
 296         }
 297
 298         /* Allocate cell then parse its modifiers. */
 299
 300         return(mods(tbl, cell_alloc(tbl, rp, c, vert), ln, p, pos));
 301 }
 302
 303 static void
 304 row(struct tbl_node *tbl, int ln, const char *p, int *pos)
 305 {
 306         struct tbl_row  *rp;
 307
 308 row:    /*
 309          * EBNF describing this section:
 310          *
 311          * row          ::= row_list [:space:]* [.]?[\n]
 312          * row_list     ::= [:space:]* row_elem row_tail
 313          * row_tail     ::= [:space:]*[,] row_list |
 314          *                  epsilon
 315          * row_elem     ::= [\t\ ]*[:alpha:]+
 316          */
 317
 318         rp = mandoc_calloc(1, sizeof(struct tbl_row));
 319         if (tbl->last_row)
 320                 tbl->last_row->next = rp;
 321         else
 322                 tbl->first_row = rp;
 323         tbl->last_row = rp;
 324
 325 cell:
 326         while (isspace((unsigned char)p[*pos]))
 327                 (*pos)++;
 328
 329         /* Safely exit layout context. */
 330
 331         if ('.' == p[*pos]) {
 332                 tbl->part = TBL_PART_DATA;
 333                 if (NULL == tbl->first_row)
 334                         mandoc_msg(MANDOCERR_TBLNOLAYOUT,
 335                             tbl->parse, ln, *pos, NULL);
 336                 (*pos)++;
 337                 return;
 338         }
 339
 340         /* End (and possibly restart) a row. */
 341
 342         if (',' == p[*pos]) {
 343                 (*pos)++;
 344                 goto row;
 345         } else if ('\0' == p[*pos])
 346                 return;
 347
 348         if ( ! cell(tbl, rp, ln, p, pos))
 349                 return;
 350
 351         goto cell;
 352         /* NOTREACHED */
 353 }
 354
 355 int
 356 tbl_layout(struct tbl_node *tbl, int ln, const char *p)
 357 {
 358         int              pos;
 359
 360         pos = 0;
 361         row(tbl, ln, p, &pos);
 362
 363         /* Always succeed. */
 364         return(1);
 365 }
 366
 367 static struct tbl_cell *
 368 cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos,
 369                 int vert)
 370 {
 371         struct tbl_cell *p, *pp;
 372         struct tbl_head *h, *hp;
 373
 374         p = mandoc_calloc(1, sizeof(struct tbl_cell));
 375
 376         if (NULL != (pp = rp->last)) {
 377                 pp->next = p;
 378                 h = pp->head->next;
 379         } else {
 380                 rp->first = p;
 381                 h = tbl->first_head;
 382         }
 383         rp->last = p;
 384
 385         p->pos = pos;
 386         p->vert = vert;
 387
 388         /* Re-use header. */
 389
 390         if (h) {
 391                 p->head = h;
 392                 return(p);
 393         }
 394
 395         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 396         hp->ident = tbl->opts.cols++;
 397         hp->vert = vert;
 398
 399         if (tbl->last_head) {
 400                 hp->prev = tbl->last_head;
 401                 tbl->last_head->next = hp;
 402         } else
 403                 tbl->first_head = hp;
 404         tbl->last_head = hp;
 405
 406         p->head = hp;
 407         return(p);
 408 }