tbl_layout.c

   1 /*      $Id: tbl_layout.c,v 1.19 2011/04/07 01:08:42 joerg 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 mod:
  76         /*
  77          * XXX: since, at least for now, modifiers are non-conflicting
  78          * (are separable by value, regardless of position), we let
  79          * modifiers come in any order.  The existing tbl doesn't let
  80          * this happen.
  81          */
  82         switch (p[*pos]) {
  83         case ('\0'):
  84                 /* FALLTHROUGH */
  85         case (' '):
  86                 /* FALLTHROUGH */
  87         case ('\t'):
  88                 /* FALLTHROUGH */
  89         case (','):
  90                 /* FALLTHROUGH */
  91         case ('.'):
  92                 return(1);
  93         default:
  94                 break;
  95         }
  96
  97         /* Throw away parenthesised expression. */
  98
  99         if ('(' == p[*pos]) {
 100                 (*pos)++;
 101                 while (p[*pos] && ')' != p[*pos])
 102                         (*pos)++;
 103                 if (')' == p[*pos]) {
 104                         (*pos)++;
 105                         goto mod;
 106                 }
 107                 mandoc_msg(MANDOCERR_TBLLAYOUT,
 108                                 tbl->parse, ln, *pos, NULL);
 109                 return(0);
 110         }
 111
 112         /* Parse numerical spacing from modifier string. */
 113
 114         if (isdigit((unsigned char)p[*pos])) {
 115                 for (i = 0; i < 4; i++) {
 116                         if ( ! isdigit((unsigned char)p[*pos + i]))
 117                                 break;
 118                         buf[i] = p[*pos + i];
 119                 }
 120                 buf[i] = '\0';
 121
 122                 /* No greater than 4 digits. */
 123
 124                 if (4 == i) {
 125                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 126                                         ln, *pos, NULL);
 127                         return(0);
 128                 }
 129
 130                 *pos += i;
 131                 cp->spacing = (size_t)atoi(buf);
 132
 133                 goto mod;
 134                 /* NOTREACHED */
 135         }
 136
 137         /* TODO: GNU has many more extensions. */
 138
 139         switch (tolower((unsigned char)p[(*pos)++])) {
 140         case ('z'):
 141                 cp->flags |= TBL_CELL_WIGN;
 142                 goto mod;
 143         case ('u'):
 144                 cp->flags |= TBL_CELL_UP;
 145                 goto mod;
 146         case ('e'):
 147                 cp->flags |= TBL_CELL_EQUAL;
 148                 goto mod;
 149         case ('t'):
 150                 cp->flags |= TBL_CELL_TALIGN;
 151                 goto mod;
 152         case ('d'):
 153                 cp->flags |= TBL_CELL_BALIGN;
 154                 goto mod;
 155         case ('w'):  /* XXX for now, ignore minimal column width */
 156                 goto mod;
 157         case ('f'):
 158                 break;
 159         case ('b'):
 160                 /* FALLTHROUGH */
 161         case ('i'):
 162                 (*pos)--;
 163                 break;
 164         default:
 165                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 166                                 ln, *pos - 1, NULL);
 167                 return(0);
 168         }
 169
 170         switch (tolower((unsigned char)p[(*pos)++])) {
 171         case ('b'):
 172                 cp->flags |= TBL_CELL_BOLD;
 173                 goto mod;
 174         case ('i'):
 175                 cp->flags |= TBL_CELL_ITALIC;
 176                 goto mod;
 177         default:
 178                 break;
 179         }
 180
 181         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 182                         ln, *pos - 1, NULL);
 183         return(0);
 184 }
 185
 186 static int
 187 cell(struct tbl_node *tbl, struct tbl_row *rp,
 188                 int ln, const char *p, int *pos)
 189 {
 190         int              i;
 191         enum tbl_cellt   c;
 192
 193         /* Parse the column position (`r', `R', `|', ...). */
 194
 195         for (i = 0; i < KEYS_MAX; i++)
 196                 if (tolower((unsigned char)p[*pos]) == keys[i].name)
 197                         break;
 198
 199         if (KEYS_MAX == i) {
 200                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 201                                 ln, *pos, NULL);
 202                 return(0);
 203         }
 204
 205         c = keys[i].key;
 206
 207         /*
 208          * If a span cell is found first, raise a warning and abort the
 209          * parse.  If a span cell is found and the last layout element
 210          * isn't a "normal" layout, bail.
 211          *
 212          * FIXME: recover from this somehow?
 213          */
 214
 215         if (TBL_CELL_SPAN == c) {
 216                 if (NULL == rp->first) {
 217                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 218                                         ln, *pos, NULL);
 219                         return(0);
 220                 } else if (rp->last)
 221                         switch (rp->last->pos) {
 222                         case (TBL_CELL_VERT):
 223                         case (TBL_CELL_DVERT):
 224                         case (TBL_CELL_HORIZ):
 225                         case (TBL_CELL_DHORIZ):
 226                                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 227                                                 ln, *pos, NULL);
 228                                 return(0);
 229                         default:
 230                                 break;
 231                         }
 232         }
 233
 234         /*
 235          * If a vertical spanner is found, we may not be in the first
 236          * row.
 237          */
 238
 239         if (TBL_CELL_DOWN == c && rp == tbl->first_row) {
 240                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos, NULL);
 241                 return(0);
 242         }
 243
 244         (*pos)++;
 245
 246         /* Extra check for the double-vertical. */
 247
 248         if (TBL_CELL_VERT == c && '|' == p[*pos]) {
 249                 (*pos)++;
 250                 c = TBL_CELL_DVERT;
 251         }
 252
 253         /* Disallow adjacent spacers. */
 254
 255         if (rp->last && (TBL_CELL_VERT == c || TBL_CELL_DVERT == c) &&
 256                         (TBL_CELL_VERT == rp->last->pos ||
 257                          TBL_CELL_DVERT == rp->last->pos)) {
 258                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos - 1, NULL);
 259                 return(0);
 260         }
 261
 262         /* Allocate cell then parse its modifiers. */
 263
 264         return(mods(tbl, cell_alloc(tbl, rp, c), ln, p, pos));
 265 }
 266
 267
 268 static void
 269 row(struct tbl_node *tbl, int ln, const char *p, int *pos)
 270 {
 271         struct tbl_row  *rp;
 272
 273 row:    /*
 274          * EBNF describing this section:
 275          *
 276          * row          ::= row_list [:space:]* [.]?[\n]
 277          * row_list     ::= [:space:]* row_elem row_tail
 278          * row_tail     ::= [:space:]*[,] row_list |
 279          *                  epsilon
 280          * row_elem     ::= [\t\ ]*[:alpha:]+
 281          */
 282
 283         rp = mandoc_calloc(1, sizeof(struct tbl_row));
 284         if (tbl->last_row) {
 285                 tbl->last_row->next = rp;
 286                 tbl->last_row = rp;
 287         } else
 288                 tbl->last_row = tbl->first_row = rp;
 289
 290 cell:
 291         while (isspace((unsigned char)p[*pos]))
 292                 (*pos)++;
 293
 294         /* Safely exit layout context. */
 295
 296         if ('.' == p[*pos]) {
 297                 tbl->part = TBL_PART_DATA;
 298                 if (NULL == tbl->first_row)
 299                         mandoc_msg(MANDOCERR_TBLNOLAYOUT, tbl->parse,
 300                                         ln, *pos, NULL);
 301                 (*pos)++;
 302                 return;
 303         }
 304
 305         /* End (and possibly restart) a row. */
 306
 307         if (',' == p[*pos]) {
 308                 (*pos)++;
 309                 goto row;
 310         } else if ('\0' == p[*pos])
 311                 return;
 312
 313         if ( ! cell(tbl, rp, ln, p, pos))
 314                 return;
 315
 316         goto cell;
 317         /* NOTREACHED */
 318 }
 319
 320 int
 321 tbl_layout(struct tbl_node *tbl, int ln, const char *p)
 322 {
 323         int              pos;
 324
 325         pos = 0;
 326         row(tbl, ln, p, &pos);
 327
 328         /* Always succeed. */
 329         return(1);
 330 }
 331
 332 static struct tbl_cell *
 333 cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos)
 334 {
 335         struct tbl_cell *p, *pp;
 336         struct tbl_head *h, *hp;
 337
 338         p = mandoc_calloc(1, sizeof(struct tbl_cell));
 339
 340         if (NULL != (pp = rp->last)) {
 341                 rp->last->next = p;
 342                 rp->last = p;
 343         } else
 344                 rp->last = rp->first = p;
 345
 346         p->pos = pos;
 347
 348         /*
 349          * This is a little bit complicated.  Here we determine the
 350          * header the corresponds to a cell.  We add headers dynamically
 351          * when need be or re-use them, otherwise.  As an example, given
 352          * the following:
 353          *
 354          *      1  c || l
 355          *      2  | c | l
 356          *      3  l l
 357          *      3  || c | l |.
 358          *
 359          * We first add the new headers (as there are none) in (1); then
 360          * in (2) we insert the first spanner (as it doesn't match up
 361          * with the header); then we re-use the prior data headers,
 362          * skipping over the spanners; then we re-use everything and add
 363          * a last spanner.  Note that VERT headers are made into DVERT
 364          * ones.
 365          */
 366
 367         h = pp ? pp->head->next : tbl->first_head;
 368
 369         if (h) {
 370                 /* Re-use data header. */
 371                 if (TBL_HEAD_DATA == h->pos &&
 372                                 (TBL_CELL_VERT != p->pos &&
 373                                  TBL_CELL_DVERT != p->pos)) {
 374                         p->head = h;
 375                         return(p);
 376                 }
 377
 378                 /* Re-use spanner header. */
 379                 if (TBL_HEAD_DATA != h->pos &&
 380                                 (TBL_CELL_VERT == p->pos ||
 381                                  TBL_CELL_DVERT == p->pos)) {
 382                         head_adjust(p, h);
 383                         p->head = h;
 384                         return(p);
 385                 }
 386
 387                 /* Right-shift headers with a new spanner. */
 388                 if (TBL_HEAD_DATA == h->pos &&
 389                                 (TBL_CELL_VERT == p->pos ||
 390                                  TBL_CELL_DVERT == p->pos)) {
 391                         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 392                         hp->ident = tbl->opts.cols++;
 393                         hp->prev = h->prev;
 394                         if (h->prev)
 395                                 h->prev->next = hp;
 396                         if (h == tbl->first_head)
 397                                 tbl->first_head = hp;
 398                         h->prev = hp;
 399                         hp->next = h;
 400                         head_adjust(p, hp);
 401                         p->head = hp;
 402                         return(p);
 403                 }
 404
 405                 if (NULL != (h = h->next)) {
 406                         head_adjust(p, h);
 407                         p->head = h;
 408                         return(p);
 409                 }
 410
 411                 /* Fall through to default case... */
 412         }
 413
 414         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 415         hp->ident = tbl->opts.cols++;
 416
 417         if (tbl->last_head) {
 418                 hp->prev = tbl->last_head;
 419                 tbl->last_head->next = hp;
 420                 tbl->last_head = hp;
 421         } else
 422                 tbl->last_head = tbl->first_head = hp;
 423
 424         head_adjust(p, hp);
 425         p->head = hp;
 426         return(p);
 427 }
 428
 429 static void
 430 head_adjust(const struct tbl_cell *cellp, struct tbl_head *head)
 431 {
 432         if (TBL_CELL_VERT != cellp->pos &&
 433                         TBL_CELL_DVERT != cellp->pos) {
 434                 head->pos = TBL_HEAD_DATA;
 435                 return;
 436         }
 437
 438         if (TBL_CELL_VERT == cellp->pos)
 439                 if (TBL_HEAD_DVERT != head->pos)
 440                         head->pos = TBL_HEAD_VERT;
 441
 442         if (TBL_CELL_DVERT == cellp->pos)
 443                 head->pos = TBL_HEAD_DVERT;
 444 }
 445