tbl_layout.c

   1 /*      $Id: tbl_layout.c,v 1.20 2011/05/17 13:11:40 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 #ifdef HAVE_CONFIG_H
  18 #include "config.h"
  19 #endif
  20
  21 #include <assert.h>
  22 #include <ctype.h>
  23 #include <stdlib.h>
  24 #include <string.h>
  25 #include <time.h>
  26
  27 #include "mandoc.h"
  28 #include "libmandoc.h"
  29 #include "libroff.h"
  30
  31 struct  tbl_phrase {
  32         char             name;
  33         enum tbl_cellt   key;
  34 };
  35
  36 /*
  37  * FIXME: we can make this parse a lot nicer by, when an error is
  38  * encountered in a layout key, bailing to the next key (i.e. to the
  39  * next whitespace then continuing).
  40  */
  41
  42 #define KEYS_MAX         11
  43
  44 static  const struct tbl_phrase keys[KEYS_MAX] = {
  45         { 'c',           TBL_CELL_CENTRE },
  46         { 'r',           TBL_CELL_RIGHT },
  47         { 'l',           TBL_CELL_LEFT },
  48         { 'n',           TBL_CELL_NUMBER },
  49         { 's',           TBL_CELL_SPAN },
  50         { 'a',           TBL_CELL_LONG },
  51         { '^',           TBL_CELL_DOWN },
  52         { '-',           TBL_CELL_HORIZ },
  53         { '_',           TBL_CELL_HORIZ },
  54         { '=',           TBL_CELL_DHORIZ },
  55         { '|',           TBL_CELL_VERT }
  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 *,
  64                                 struct tbl_row *, enum tbl_cellt);
  65 static  void             head_adjust(const struct tbl_cell *,
  66                                 struct tbl_head *);
  67
  68 static int
  69 mods(struct tbl_node *tbl, struct tbl_cell *cp,
  70                 int ln, const char *p, int *pos)
  71 {
  72         char             buf[5];
  73         int              i;
  74
  75         /* Not all types accept modifiers. */
  76
  77         switch (cp->pos) {
  78         case (TBL_CELL_DOWN):
  79                 /* FALLTHROUGH */
  80         case (TBL_CELL_HORIZ):
  81                 /* FALLTHROUGH */
  82         case (TBL_CELL_DHORIZ):
  83                 /* FALLTHROUGH */
  84         case (TBL_CELL_VERT):
  85                 /* FALLTHROUGH */
  86         case (TBL_CELL_DVERT):
  87                 return(1);
  88         default:
  89                 break;
  90         }
  91
  92 mod:
  93         /*
  94          * XXX: since, at least for now, modifiers are non-conflicting
  95          * (are separable by value, regardless of position), we let
  96          * modifiers come in any order.  The existing tbl doesn't let
  97          * this happen.
  98          */
  99         switch (p[*pos]) {
 100         case ('\0'):
 101                 /* FALLTHROUGH */
 102         case (' '):
 103                 /* FALLTHROUGH */
 104         case ('\t'):
 105                 /* FALLTHROUGH */
 106         case (','):
 107                 /* FALLTHROUGH */
 108         case ('.'):
 109                 return(1);
 110         default:
 111                 break;
 112         }
 113
 114         /* Throw away parenthesised expression. */
 115
 116         if ('(' == p[*pos]) {
 117                 (*pos)++;
 118                 while (p[*pos] && ')' != p[*pos])
 119                         (*pos)++;
 120                 if (')' == p[*pos]) {
 121                         (*pos)++;
 122                         goto mod;
 123                 }
 124                 mandoc_msg(MANDOCERR_TBLLAYOUT,
 125                                 tbl->parse, ln, *pos, NULL);
 126                 return(0);
 127         }
 128
 129         /* Parse numerical spacing from modifier string. */
 130
 131         if (isdigit((unsigned char)p[*pos])) {
 132                 for (i = 0; i < 4; i++) {
 133                         if ( ! isdigit((unsigned char)p[*pos + i]))
 134                                 break;
 135                         buf[i] = p[*pos + i];
 136                 }
 137                 buf[i] = '\0';
 138
 139                 /* No greater than 4 digits. */
 140
 141                 if (4 == i) {
 142                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 143                                         ln, *pos, NULL);
 144                         return(0);
 145                 }
 146
 147                 *pos += i;
 148                 cp->spacing = (size_t)atoi(buf);
 149
 150                 goto mod;
 151                 /* NOTREACHED */
 152         }
 153
 154         /* TODO: GNU has many more extensions. */
 155
 156         switch (tolower((unsigned char)p[(*pos)++])) {
 157         case ('z'):
 158                 cp->flags |= TBL_CELL_WIGN;
 159                 goto mod;
 160         case ('u'):
 161                 cp->flags |= TBL_CELL_UP;
 162                 goto mod;
 163         case ('e'):
 164                 cp->flags |= TBL_CELL_EQUAL;
 165                 goto mod;
 166         case ('t'):
 167                 cp->flags |= TBL_CELL_TALIGN;
 168                 goto mod;
 169         case ('d'):
 170                 cp->flags |= TBL_CELL_BALIGN;
 171                 goto mod;
 172         case ('w'):  /* XXX for now, ignore minimal column width */
 173                 goto mod;
 174         case ('f'):
 175                 break;
 176         case ('b'):
 177                 /* FALLTHROUGH */
 178         case ('i'):
 179                 (*pos)--;
 180                 break;
 181         default:
 182                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 183                                 ln, *pos - 1, NULL);
 184                 return(0);
 185         }
 186
 187         switch (tolower((unsigned char)p[(*pos)++])) {
 188         case ('b'):
 189                 cp->flags |= TBL_CELL_BOLD;
 190                 goto mod;
 191         case ('i'):
 192                 cp->flags |= TBL_CELL_ITALIC;
 193                 goto mod;
 194         default:
 195                 break;
 196         }
 197
 198         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 199                         ln, *pos - 1, NULL);
 200         return(0);
 201 }
 202
 203 static int
 204 cell(struct tbl_node *tbl, struct tbl_row *rp,
 205                 int ln, const char *p, int *pos)
 206 {
 207         int              i;
 208         enum tbl_cellt   c;
 209
 210         /* Parse the column position (`r', `R', `|', ...). */
 211
 212         for (i = 0; i < KEYS_MAX; i++)
 213                 if (tolower((unsigned char)p[*pos]) == keys[i].name)
 214                         break;
 215
 216         if (KEYS_MAX == i) {
 217                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 218                                 ln, *pos, NULL);
 219                 return(0);
 220         }
 221
 222         c = keys[i].key;
 223
 224         /*
 225          * If a span cell is found first, raise a warning and abort the
 226          * parse.  If a span cell is found and the last layout element
 227          * isn't a "normal" layout, bail.
 228          *
 229          * FIXME: recover from this somehow?
 230          */
 231
 232         if (TBL_CELL_SPAN == c) {
 233                 if (NULL == rp->first) {
 234                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 235                                         ln, *pos, NULL);
 236                         return(0);
 237                 } else if (rp->last)
 238                         switch (rp->last->pos) {
 239                         case (TBL_CELL_VERT):
 240                         case (TBL_CELL_DVERT):
 241                         case (TBL_CELL_HORIZ):
 242                         case (TBL_CELL_DHORIZ):
 243                                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 244                                                 ln, *pos, NULL);
 245                                 return(0);
 246                         default:
 247                                 break;
 248                         }
 249         }
 250
 251         /*
 252          * If a vertical spanner is found, we may not be in the first
 253          * row.
 254          */
 255
 256         if (TBL_CELL_DOWN == c && rp == tbl->first_row) {
 257                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos, NULL);
 258                 return(0);
 259         }
 260
 261         (*pos)++;
 262
 263         /* Extra check for the double-vertical. */
 264
 265         if (TBL_CELL_VERT == c && '|' == p[*pos]) {
 266                 (*pos)++;
 267                 c = TBL_CELL_DVERT;
 268         }
 269
 270         /* Disallow adjacent spacers. */
 271
 272         if (rp->last && (TBL_CELL_VERT == c || TBL_CELL_DVERT == c) &&
 273                         (TBL_CELL_VERT == rp->last->pos ||
 274                          TBL_CELL_DVERT == rp->last->pos)) {
 275                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos - 1, NULL);
 276                 return(0);
 277         }
 278
 279         /* Allocate cell then parse its modifiers. */
 280
 281         return(mods(tbl, cell_alloc(tbl, rp, c), ln, p, pos));
 282 }
 283
 284
 285 static void
 286 row(struct tbl_node *tbl, int ln, const char *p, int *pos)
 287 {
 288         struct tbl_row  *rp;
 289
 290 row:    /*
 291          * EBNF describing this section:
 292          *
 293          * row          ::= row_list [:space:]* [.]?[\n]
 294          * row_list     ::= [:space:]* row_elem row_tail
 295          * row_tail     ::= [:space:]*[,] row_list |
 296          *                  epsilon
 297          * row_elem     ::= [\t\ ]*[:alpha:]+
 298          */
 299
 300         rp = mandoc_calloc(1, sizeof(struct tbl_row));
 301         if (tbl->last_row) {
 302                 tbl->last_row->next = rp;
 303                 tbl->last_row = rp;
 304         } else
 305                 tbl->last_row = tbl->first_row = rp;
 306
 307 cell:
 308         while (isspace((unsigned char)p[*pos]))
 309                 (*pos)++;
 310
 311         /* Safely exit layout context. */
 312
 313         if ('.' == p[*pos]) {
 314                 tbl->part = TBL_PART_DATA;
 315                 if (NULL == tbl->first_row)
 316                         mandoc_msg(MANDOCERR_TBLNOLAYOUT, tbl->parse,
 317                                         ln, *pos, NULL);
 318                 (*pos)++;
 319                 return;
 320         }
 321
 322         /* End (and possibly restart) a row. */
 323
 324         if (',' == p[*pos]) {
 325                 (*pos)++;
 326                 goto row;
 327         } else if ('\0' == p[*pos])
 328                 return;
 329
 330         if ( ! cell(tbl, rp, ln, p, pos))
 331                 return;
 332
 333         goto cell;
 334         /* NOTREACHED */
 335 }
 336
 337 int
 338 tbl_layout(struct tbl_node *tbl, int ln, const char *p)
 339 {
 340         int              pos;
 341
 342         pos = 0;
 343         row(tbl, ln, p, &pos);
 344
 345         /* Always succeed. */
 346         return(1);
 347 }
 348
 349 static struct tbl_cell *
 350 cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos)
 351 {
 352         struct tbl_cell *p, *pp;
 353         struct tbl_head *h, *hp;
 354
 355         p = mandoc_calloc(1, sizeof(struct tbl_cell));
 356
 357         if (NULL != (pp = rp->last)) {
 358                 rp->last->next = p;
 359                 rp->last = p;
 360         } else
 361                 rp->last = rp->first = p;
 362
 363         p->pos = pos;
 364
 365         /*
 366          * This is a little bit complicated.  Here we determine the
 367          * header the corresponds to a cell.  We add headers dynamically
 368          * when need be or re-use them, otherwise.  As an example, given
 369          * the following:
 370          *
 371          *      1  c || l
 372          *      2  | c | l
 373          *      3  l l
 374          *      3  || c | l |.
 375          *
 376          * We first add the new headers (as there are none) in (1); then
 377          * in (2) we insert the first spanner (as it doesn't match up
 378          * with the header); then we re-use the prior data headers,
 379          * skipping over the spanners; then we re-use everything and add
 380          * a last spanner.  Note that VERT headers are made into DVERT
 381          * ones.
 382          */
 383
 384         h = pp ? pp->head->next : tbl->first_head;
 385
 386         if (h) {
 387                 /* Re-use data header. */
 388                 if (TBL_HEAD_DATA == h->pos &&
 389                                 (TBL_CELL_VERT != p->pos &&
 390                                  TBL_CELL_DVERT != p->pos)) {
 391                         p->head = h;
 392                         return(p);
 393                 }
 394
 395                 /* Re-use spanner header. */
 396                 if (TBL_HEAD_DATA != h->pos &&
 397                                 (TBL_CELL_VERT == p->pos ||
 398                                  TBL_CELL_DVERT == p->pos)) {
 399                         head_adjust(p, h);
 400                         p->head = h;
 401                         return(p);
 402                 }
 403
 404                 /* Right-shift headers with a new spanner. */
 405                 if (TBL_HEAD_DATA == h->pos &&
 406                                 (TBL_CELL_VERT == p->pos ||
 407                                  TBL_CELL_DVERT == p->pos)) {
 408                         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 409                         hp->ident = tbl->opts.cols++;
 410                         hp->prev = h->prev;
 411                         if (h->prev)
 412                                 h->prev->next = hp;
 413                         if (h == tbl->first_head)
 414                                 tbl->first_head = hp;
 415                         h->prev = hp;
 416                         hp->next = h;
 417                         head_adjust(p, hp);
 418                         p->head = hp;
 419                         return(p);
 420                 }
 421
 422                 if (NULL != (h = h->next)) {
 423                         head_adjust(p, h);
 424                         p->head = h;
 425                         return(p);
 426                 }
 427
 428                 /* Fall through to default case... */
 429         }
 430
 431         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 432         hp->ident = tbl->opts.cols++;
 433
 434         if (tbl->last_head) {
 435                 hp->prev = tbl->last_head;
 436                 tbl->last_head->next = hp;
 437                 tbl->last_head = hp;
 438         } else
 439                 tbl->last_head = tbl->first_head = hp;
 440
 441         head_adjust(p, hp);
 442         p->head = hp;
 443         return(p);
 444 }
 445
 446 static void
 447 head_adjust(const struct tbl_cell *cellp, struct tbl_head *head)
 448 {
 449         if (TBL_CELL_VERT != cellp->pos &&
 450                         TBL_CELL_DVERT != cellp->pos) {
 451                 head->pos = TBL_HEAD_DATA;
 452                 return;
 453         }
 454
 455         if (TBL_CELL_VERT == cellp->pos)
 456                 if (TBL_HEAD_DVERT != head->pos)
 457                         head->pos = TBL_HEAD_VERT;
 458
 459         if (TBL_CELL_DVERT == cellp->pos)
 460                 head->pos = TBL_HEAD_DVERT;
 461 }
 462