tbl_layout.c

   1 /*      $Id: tbl_layout.c,v 1.16 2011/01/11 14:12:01 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                 TBL_MSG(tbl, MANDOCERR_TBLLAYOUT, ln, *pos);
 104                 return(0);
 105         }
 106
 107         /* Parse numerical spacing from modifier string. */
 108
 109         if (isdigit((unsigned char)p[*pos])) {
 110                 for (i = 0; i < 4; i++) {
 111                         if ( ! isdigit((unsigned char)p[*pos + i]))
 112                                 break;
 113                         buf[i] = p[*pos + i];
 114                 }
 115                 buf[i] = '\0';
 116
 117                 /* No greater than 4 digits. */
 118
 119                 if (4 == i) {
 120                         TBL_MSG(tbl, MANDOCERR_TBLLAYOUT, ln, *pos);
 121                         return(0);
 122                 }
 123
 124                 *pos += i;
 125                 cp->spacing = (size_t)atoi(buf);
 126
 127                 goto mod;
 128                 /* NOTREACHED */
 129         }
 130
 131         /* TODO: GNU has many more extensions. */
 132
 133         switch (tolower((unsigned char)p[(*pos)++])) {
 134         case ('z'):
 135                 cp->flags |= TBL_CELL_WIGN;
 136                 goto mod;
 137         case ('u'):
 138                 cp->flags |= TBL_CELL_UP;
 139                 goto mod;
 140         case ('e'):
 141                 cp->flags |= TBL_CELL_EQUAL;
 142                 goto mod;
 143         case ('t'):
 144                 cp->flags |= TBL_CELL_TALIGN;
 145                 goto mod;
 146         case ('d'):
 147                 cp->flags |= TBL_CELL_BALIGN;
 148                 goto mod;
 149         case ('w'):  /* XXX for now, ignore minimal column width */
 150                 goto mod;
 151         case ('f'):
 152                 break;
 153         case ('b'):
 154                 /* FALLTHROUGH */
 155         case ('i'):
 156                 (*pos)--;
 157                 break;
 158         default:
 159                 TBL_MSG(tbl, MANDOCERR_TBLLAYOUT, ln, *pos - 1);
 160                 return(0);
 161         }
 162
 163         switch (tolower((unsigned char)p[(*pos)++])) {
 164         case ('b'):
 165                 cp->flags |= TBL_CELL_BOLD;
 166                 goto mod;
 167         case ('i'):
 168                 cp->flags |= TBL_CELL_ITALIC;
 169                 goto mod;
 170         default:
 171                 break;
 172         }
 173
 174         TBL_MSG(tbl, MANDOCERR_TBLLAYOUT, ln, *pos - 1);
 175         return(0);
 176 }
 177
 178 static int
 179 cell(struct tbl_node *tbl, struct tbl_row *rp,
 180                 int ln, const char *p, int *pos)
 181 {
 182         int              i;
 183         enum tbl_cellt   c;
 184
 185         /* Parse the column position (`r', `R', `|', ...). */
 186
 187         for (i = 0; i < KEYS_MAX; i++)
 188                 if (tolower((unsigned char)p[*pos]) == keys[i].name)
 189                         break;
 190
 191         if (KEYS_MAX == i) {
 192                 TBL_MSG(tbl, MANDOCERR_TBLLAYOUT, ln, *pos);
 193                 return(0);
 194         }
 195
 196         c = keys[i].key;
 197
 198         /*
 199          * If a span cell is found first, raise a warning and abort the
 200          * parse.  If a span cell is found and the last layout element
 201          * isn't a "normal" layout, bail.
 202          *
 203          * FIXME: recover from this somehow?
 204          */
 205
 206         if (TBL_CELL_SPAN == c) {
 207                 if (NULL == rp->first) {
 208                         TBL_MSG(tbl, MANDOCERR_TBLLAYOUT, ln, *pos);
 209                         return(0);
 210                 } else if (rp->last)
 211                         switch (rp->last->pos) {
 212                         case (TBL_CELL_VERT):
 213                         case (TBL_CELL_DVERT):
 214                         case (TBL_CELL_HORIZ):
 215                         case (TBL_CELL_DHORIZ):
 216                                 TBL_MSG(tbl, MANDOCERR_TBLLAYOUT, ln, *pos);
 217                                 return(0);
 218                         default:
 219                                 break;
 220                         }
 221         }
 222
 223         /*
 224          * If a vertical spanner is found, we may not be in the first
 225          * row.
 226          */
 227
 228         if (TBL_CELL_DOWN == c && rp == tbl->first_row) {
 229                 TBL_MSG(tbl, MANDOCERR_TBLLAYOUT, ln, *pos);
 230                 return(0);
 231         }
 232
 233         (*pos)++;
 234
 235         /* Extra check for the double-vertical. */
 236
 237         if (TBL_CELL_VERT == c && '|' == p[*pos]) {
 238                 (*pos)++;
 239                 c = TBL_CELL_DVERT;
 240         }
 241
 242         /* Disallow adjacent spacers. */
 243
 244         if (rp->last && (TBL_CELL_VERT == c || TBL_CELL_DVERT == c) &&
 245                         (TBL_CELL_VERT == rp->last->pos ||
 246                          TBL_CELL_DVERT == rp->last->pos)) {
 247                 TBL_MSG(tbl, MANDOCERR_TBLLAYOUT, ln, *pos - 1);
 248                 return(0);
 249         }
 250
 251         /* Allocate cell then parse its modifiers. */
 252
 253         return(mods(tbl, cell_alloc(tbl, rp, c), ln, p, pos));
 254 }
 255
 256
 257 static void
 258 row(struct tbl_node *tbl, int ln, const char *p, int *pos)
 259 {
 260         struct tbl_row  *rp;
 261
 262 row:    /*
 263          * EBNF describing this section:
 264          *
 265          * row          ::= row_list [:space:]* [.]?[\n]
 266          * row_list     ::= [:space:]* row_elem row_tail
 267          * row_tail     ::= [:space:]*[,] row_list |
 268          *                  epsilon
 269          * row_elem     ::= [\t\ ]*[:alpha:]+
 270          */
 271
 272         rp = mandoc_calloc(1, sizeof(struct tbl_row));
 273         if (tbl->last_row) {
 274                 tbl->last_row->next = rp;
 275                 tbl->last_row = rp;
 276         } else
 277                 tbl->last_row = tbl->first_row = rp;
 278
 279 cell:
 280         while (isspace((unsigned char)p[*pos]))
 281                 (*pos)++;
 282
 283         /* Safely exit layout context. */
 284
 285         if ('.' == p[*pos]) {
 286                 tbl->part = TBL_PART_DATA;
 287                 if (NULL == tbl->first_row)
 288                         TBL_MSG(tbl, MANDOCERR_TBLNOLAYOUT, ln, *pos);
 289                 (*pos)++;
 290                 return;
 291         }
 292
 293         /* End (and possibly restart) a row. */
 294
 295         if (',' == p[*pos]) {
 296                 (*pos)++;
 297                 goto row;
 298         } else if ('\0' == p[*pos])
 299                 return;
 300
 301         if ( ! cell(tbl, rp, ln, p, pos))
 302                 return;
 303
 304         goto cell;
 305         /* NOTREACHED */
 306 }
 307
 308 int
 309 tbl_layout(struct tbl_node *tbl, int ln, const char *p)
 310 {
 311         int              pos;
 312
 313         pos = 0;
 314         row(tbl, ln, p, &pos);
 315
 316         /* Always succeed. */
 317         return(1);
 318 }
 319
 320 static struct tbl_cell *
 321 cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos)
 322 {
 323         struct tbl_cell *p, *pp;
 324         struct tbl_head *h, *hp;
 325
 326         p = mandoc_calloc(1, sizeof(struct tbl_cell));
 327
 328         if (NULL != (pp = rp->last)) {
 329                 rp->last->next = p;
 330                 rp->last = p;
 331         } else
 332                 rp->last = rp->first = p;
 333
 334         p->pos = pos;
 335
 336         /*
 337          * This is a little bit complicated.  Here we determine the
 338          * header the corresponds to a cell.  We add headers dynamically
 339          * when need be or re-use them, otherwise.  As an example, given
 340          * the following:
 341          *
 342          *      1  c || l
 343          *      2  | c | l
 344          *      3  l l
 345          *      3  || c | l |.
 346          *
 347          * We first add the new headers (as there are none) in (1); then
 348          * in (2) we insert the first spanner (as it doesn't match up
 349          * with the header); then we re-use the prior data headers,
 350          * skipping over the spanners; then we re-use everything and add
 351          * a last spanner.  Note that VERT headers are made into DVERT
 352          * ones.
 353          */
 354
 355         h = pp ? pp->head->next : tbl->first_head;
 356
 357         if (h) {
 358                 /* Re-use data header. */
 359                 if (TBL_HEAD_DATA == h->pos &&
 360                                 (TBL_CELL_VERT != p->pos &&
 361                                  TBL_CELL_DVERT != p->pos)) {
 362                         p->head = h;
 363                         return(p);
 364                 }
 365
 366                 /* Re-use spanner header. */
 367                 if (TBL_HEAD_DATA != h->pos &&
 368                                 (TBL_CELL_VERT == p->pos ||
 369                                  TBL_CELL_DVERT == p->pos)) {
 370                         head_adjust(p, h);
 371                         p->head = h;
 372                         return(p);
 373                 }
 374
 375                 /* Right-shift headers with a new spanner. */
 376                 if (TBL_HEAD_DATA == h->pos &&
 377                                 (TBL_CELL_VERT == p->pos ||
 378                                  TBL_CELL_DVERT == p->pos)) {
 379                         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 380                         hp->ident = tbl->opts.cols++;
 381                         hp->prev = h->prev;
 382                         if (h->prev)
 383                                 h->prev->next = hp;
 384                         if (h == tbl->first_head)
 385                                 tbl->first_head = hp;
 386                         h->prev = hp;
 387                         hp->next = h;
 388                         head_adjust(p, hp);
 389                         p->head = hp;
 390                         return(p);
 391                 }
 392
 393                 if (NULL != (h = h->next)) {
 394                         head_adjust(p, h);
 395                         p->head = h;
 396                         return(p);
 397                 }
 398
 399                 /* Fall through to default case... */
 400         }
 401
 402         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 403         hp->ident = tbl->opts.cols++;
 404
 405         if (tbl->last_head) {
 406                 hp->prev = tbl->last_head;
 407                 tbl->last_head->next = hp;
 408                 tbl->last_head = hp;
 409         } else
 410                 tbl->last_head = tbl->first_head = hp;
 411
 412         head_adjust(p, hp);
 413         p->head = hp;
 414         return(p);
 415 }
 416
 417 static void
 418 head_adjust(const struct tbl_cell *cell, struct tbl_head *head)
 419 {
 420         if (TBL_CELL_VERT != cell->pos &&
 421                         TBL_CELL_DVERT != cell->pos) {
 422                 head->pos = TBL_HEAD_DATA;
 423                 return;
 424         }
 425
 426         if (TBL_CELL_VERT == cell->pos)
 427                 if (TBL_HEAD_DVERT != head->pos)
 428                         head->pos = TBL_HEAD_VERT;
 429
 430         if (TBL_CELL_DVERT == cell->pos)
 431                 head->pos = TBL_HEAD_DVERT;
 432 }
 433