tbl_layout.c

   1 /*      $Id: tbl_layout.c,v 1.17 2011/03/20 16:02:05 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #include <assert.h>
  18 #include <ctype.h>
  19 #include <stdlib.h>
  20 #include <string.h>
  21 #include <time.h>
  22
  23 #include "mandoc.h"
  24 #include "libmandoc.h"
  25 #include "libroff.h"
  26
  27 struct  tbl_phrase {
  28         char             name;
  29         enum tbl_cellt   key;
  30 };
  31
  32 /*
  33  * FIXME: we can make this parse a lot nicer by, when an error is
  34  * encountered in a layout key, bailing to the next key (i.e. to the
  35  * next whitespace then continuing).
  36  */
  37
  38 #define KEYS_MAX         11
  39
  40 static  const struct tbl_phrase keys[KEYS_MAX] = {
  41         { 'c',           TBL_CELL_CENTRE },
  42         { 'r',           TBL_CELL_RIGHT },
  43         { 'l',           TBL_CELL_LEFT },
  44         { 'n',           TBL_CELL_NUMBER },
  45         { 's',           TBL_CELL_SPAN },
  46         { 'a',           TBL_CELL_LONG },
  47         { '^',           TBL_CELL_DOWN },
  48         { '-',           TBL_CELL_HORIZ },
  49         { '_',           TBL_CELL_HORIZ },
  50         { '=',           TBL_CELL_DHORIZ },
  51         { '|',           TBL_CELL_VERT }
  52 };
  53
  54 static  int              mods(struct tbl_node *, struct tbl_cell *,
  55                                 int, const char *, int *);
  56 static  int              cell(struct tbl_node *, struct tbl_row *,
  57                                 int, const char *, int *);
  58 static  void             row(struct tbl_node *, int, const char *, int *);
  59 static  struct tbl_cell *cell_alloc(struct tbl_node *,
  60                                 struct tbl_row *, enum tbl_cellt);
  61 static  void             head_adjust(const struct tbl_cell *,
  62                                 struct tbl_head *);
  63
  64 static int
  65 mods(struct tbl_node *tbl, struct tbl_cell *cp,
  66                 int ln, const char *p, int *pos)
  67 {
  68         char             buf[5];
  69         int              i;
  70
  71 mod:
  72         /*
  73          * XXX: since, at least for now, modifiers are non-conflicting
  74          * (are separable by value, regardless of position), we let
  75          * modifiers come in any order.  The existing tbl doesn't let
  76          * this happen.
  77          */
  78         switch (p[*pos]) {
  79         case ('\0'):
  80                 /* FALLTHROUGH */
  81         case (' '):
  82                 /* FALLTHROUGH */
  83         case ('\t'):
  84                 /* FALLTHROUGH */
  85         case (','):
  86                 /* FALLTHROUGH */
  87         case ('.'):
  88                 return(1);
  89         default:
  90                 break;
  91         }
  92
  93         /* Throw away parenthesised expression. */
  94
  95         if ('(' == p[*pos]) {
  96                 (*pos)++;
  97                 while (p[*pos] && ')' != p[*pos])
  98                         (*pos)++;
  99                 if (')' == p[*pos]) {
 100                         (*pos)++;
 101                         goto mod;
 102                 }
 103                 mandoc_msg(MANDOCERR_TBLLAYOUT,
 104                                 tbl->parse, ln, *pos, NULL);
 105                 return(0);
 106         }
 107
 108         /* Parse numerical spacing from modifier string. */
 109
 110         if (isdigit((unsigned char)p[*pos])) {
 111                 for (i = 0; i < 4; i++) {
 112                         if ( ! isdigit((unsigned char)p[*pos + i]))
 113                                 break;
 114                         buf[i] = p[*pos + i];
 115                 }
 116                 buf[i] = '\0';
 117
 118                 /* No greater than 4 digits. */
 119
 120                 if (4 == i) {
 121                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 122                                         ln, *pos, NULL);
 123                         return(0);
 124                 }
 125
 126                 *pos += i;
 127                 cp->spacing = (size_t)atoi(buf);
 128
 129                 goto mod;
 130                 /* NOTREACHED */
 131         }
 132
 133         /* TODO: GNU has many more extensions. */
 134
 135         switch (tolower((unsigned char)p[(*pos)++])) {
 136         case ('z'):
 137                 cp->flags |= TBL_CELL_WIGN;
 138                 goto mod;
 139         case ('u'):
 140                 cp->flags |= TBL_CELL_UP;
 141                 goto mod;
 142         case ('e'):
 143                 cp->flags |= TBL_CELL_EQUAL;
 144                 goto mod;
 145         case ('t'):
 146                 cp->flags |= TBL_CELL_TALIGN;
 147                 goto mod;
 148         case ('d'):
 149                 cp->flags |= TBL_CELL_BALIGN;
 150                 goto mod;
 151         case ('w'):  /* XXX for now, ignore minimal column width */
 152                 goto mod;
 153         case ('f'):
 154                 break;
 155         case ('b'):
 156                 /* FALLTHROUGH */
 157         case ('i'):
 158                 (*pos)--;
 159                 break;
 160         default:
 161                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 162                                 ln, *pos - 1, NULL);
 163                 return(0);
 164         }
 165
 166         switch (tolower((unsigned char)p[(*pos)++])) {
 167         case ('b'):
 168                 cp->flags |= TBL_CELL_BOLD;
 169                 goto mod;
 170         case ('i'):
 171                 cp->flags |= TBL_CELL_ITALIC;
 172                 goto mod;
 173         default:
 174                 break;
 175         }
 176
 177         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 178                         ln, *pos - 1, NULL);
 179         return(0);
 180 }
 181
 182 static int
 183 cell(struct tbl_node *tbl, struct tbl_row *rp,
 184                 int ln, const char *p, int *pos)
 185 {
 186         int              i;
 187         enum tbl_cellt   c;
 188
 189         /* Parse the column position (`r', `R', `|', ...). */
 190
 191         for (i = 0; i < KEYS_MAX; i++)
 192                 if (tolower((unsigned char)p[*pos]) == keys[i].name)
 193                         break;
 194
 195         if (KEYS_MAX == i) {
 196                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 197                                 ln, *pos, NULL);
 198                 return(0);
 199         }
 200
 201         c = keys[i].key;
 202
 203         /*
 204          * If a span cell is found first, raise a warning and abort the
 205          * parse.  If a span cell is found and the last layout element
 206          * isn't a "normal" layout, bail.
 207          *
 208          * FIXME: recover from this somehow?
 209          */
 210
 211         if (TBL_CELL_SPAN == c) {
 212                 if (NULL == rp->first) {
 213                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 214                                         ln, *pos, NULL);
 215                         return(0);
 216                 } else if (rp->last)
 217                         switch (rp->last->pos) {
 218                         case (TBL_CELL_VERT):
 219                         case (TBL_CELL_DVERT):
 220                         case (TBL_CELL_HORIZ):
 221                         case (TBL_CELL_DHORIZ):
 222                                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 223                                                 ln, *pos, NULL);
 224                                 return(0);
 225                         default:
 226                                 break;
 227                         }
 228         }
 229
 230         /*
 231          * If a vertical spanner is found, we may not be in the first
 232          * row.
 233          */
 234
 235         if (TBL_CELL_DOWN == c && rp == tbl->first_row) {
 236                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos, NULL);
 237                 return(0);
 238         }
 239
 240         (*pos)++;
 241
 242         /* Extra check for the double-vertical. */
 243
 244         if (TBL_CELL_VERT == c && '|' == p[*pos]) {
 245                 (*pos)++;
 246                 c = TBL_CELL_DVERT;
 247         }
 248
 249         /* Disallow adjacent spacers. */
 250
 251         if (rp->last && (TBL_CELL_VERT == c || TBL_CELL_DVERT == c) &&
 252                         (TBL_CELL_VERT == rp->last->pos ||
 253                          TBL_CELL_DVERT == rp->last->pos)) {
 254                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos - 1, NULL);
 255                 return(0);
 256         }
 257
 258         /* Allocate cell then parse its modifiers. */
 259
 260         return(mods(tbl, cell_alloc(tbl, rp, c), ln, p, pos));
 261 }
 262
 263
 264 static void
 265 row(struct tbl_node *tbl, int ln, const char *p, int *pos)
 266 {
 267         struct tbl_row  *rp;
 268
 269 row:    /*
 270          * EBNF describing this section:
 271          *
 272          * row          ::= row_list [:space:]* [.]?[\n]
 273          * row_list     ::= [:space:]* row_elem row_tail
 274          * row_tail     ::= [:space:]*[,] row_list |
 275          *                  epsilon
 276          * row_elem     ::= [\t\ ]*[:alpha:]+
 277          */
 278
 279         rp = mandoc_calloc(1, sizeof(struct tbl_row));
 280         if (tbl->last_row) {
 281                 tbl->last_row->next = rp;
 282                 tbl->last_row = rp;
 283         } else
 284                 tbl->last_row = tbl->first_row = rp;
 285
 286 cell:
 287         while (isspace((unsigned char)p[*pos]))
 288                 (*pos)++;
 289
 290         /* Safely exit layout context. */
 291
 292         if ('.' == p[*pos]) {
 293                 tbl->part = TBL_PART_DATA;
 294                 if (NULL == tbl->first_row)
 295                         mandoc_msg(MANDOCERR_TBLNOLAYOUT, tbl->parse,
 296                                         ln, *pos, NULL);
 297                 (*pos)++;
 298                 return;
 299         }
 300
 301         /* End (and possibly restart) a row. */
 302
 303         if (',' == p[*pos]) {
 304                 (*pos)++;
 305                 goto row;
 306         } else if ('\0' == p[*pos])
 307                 return;
 308
 309         if ( ! cell(tbl, rp, ln, p, pos))
 310                 return;
 311
 312         goto cell;
 313         /* NOTREACHED */
 314 }
 315
 316 int
 317 tbl_layout(struct tbl_node *tbl, int ln, const char *p)
 318 {
 319         int              pos;
 320
 321         pos = 0;
 322         row(tbl, ln, p, &pos);
 323
 324         /* Always succeed. */
 325         return(1);
 326 }
 327
 328 static struct tbl_cell *
 329 cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos)
 330 {
 331         struct tbl_cell *p, *pp;
 332         struct tbl_head *h, *hp;
 333
 334         p = mandoc_calloc(1, sizeof(struct tbl_cell));
 335
 336         if (NULL != (pp = rp->last)) {
 337                 rp->last->next = p;
 338                 rp->last = p;
 339         } else
 340                 rp->last = rp->first = p;
 341
 342         p->pos = pos;
 343
 344         /*
 345          * This is a little bit complicated.  Here we determine the
 346          * header the corresponds to a cell.  We add headers dynamically
 347          * when need be or re-use them, otherwise.  As an example, given
 348          * the following:
 349          *
 350          *      1  c || l
 351          *      2  | c | l
 352          *      3  l l
 353          *      3  || c | l |.
 354          *
 355          * We first add the new headers (as there are none) in (1); then
 356          * in (2) we insert the first spanner (as it doesn't match up
 357          * with the header); then we re-use the prior data headers,
 358          * skipping over the spanners; then we re-use everything and add
 359          * a last spanner.  Note that VERT headers are made into DVERT
 360          * ones.
 361          */
 362
 363         h = pp ? pp->head->next : tbl->first_head;
 364
 365         if (h) {
 366                 /* Re-use data header. */
 367                 if (TBL_HEAD_DATA == h->pos &&
 368                                 (TBL_CELL_VERT != p->pos &&
 369                                  TBL_CELL_DVERT != p->pos)) {
 370                         p->head = h;
 371                         return(p);
 372                 }
 373
 374                 /* Re-use spanner header. */
 375                 if (TBL_HEAD_DATA != h->pos &&
 376                                 (TBL_CELL_VERT == p->pos ||
 377                                  TBL_CELL_DVERT == p->pos)) {
 378                         head_adjust(p, h);
 379                         p->head = h;
 380                         return(p);
 381                 }
 382
 383                 /* Right-shift headers with a new spanner. */
 384                 if (TBL_HEAD_DATA == h->pos &&
 385                                 (TBL_CELL_VERT == p->pos ||
 386                                  TBL_CELL_DVERT == p->pos)) {
 387                         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 388                         hp->ident = tbl->opts.cols++;
 389                         hp->prev = h->prev;
 390                         if (h->prev)
 391                                 h->prev->next = hp;
 392                         if (h == tbl->first_head)
 393                                 tbl->first_head = hp;
 394                         h->prev = hp;
 395                         hp->next = h;
 396                         head_adjust(p, hp);
 397                         p->head = hp;
 398                         return(p);
 399                 }
 400
 401                 if (NULL != (h = h->next)) {
 402                         head_adjust(p, h);
 403                         p->head = h;
 404                         return(p);
 405                 }
 406
 407                 /* Fall through to default case... */
 408         }
 409
 410         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 411         hp->ident = tbl->opts.cols++;
 412
 413         if (tbl->last_head) {
 414                 hp->prev = tbl->last_head;
 415                 tbl->last_head->next = hp;
 416                 tbl->last_head = hp;
 417         } else
 418                 tbl->last_head = tbl->first_head = hp;
 419
 420         head_adjust(p, hp);
 421         p->head = hp;
 422         return(p);
 423 }
 424
 425 static void
 426 head_adjust(const struct tbl_cell *cell, struct tbl_head *head)
 427 {
 428         if (TBL_CELL_VERT != cell->pos &&
 429                         TBL_CELL_DVERT != cell->pos) {
 430                 head->pos = TBL_HEAD_DATA;
 431                 return;
 432         }
 433
 434         if (TBL_CELL_VERT == cell->pos)
 435                 if (TBL_HEAD_DVERT != head->pos)
 436                         head->pos = TBL_HEAD_VERT;
 437
 438         if (TBL_CELL_DVERT == cell->pos)
 439                 head->pos = TBL_HEAD_DVERT;
 440 }
 441