tbl_layout.c

   1 /*      $Id: tbl_layout.c,v 1.28 2014/10/07 14:07:03 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 'f':
 172                 break;
 173         case 'r':
 174                 /* FALLTHROUGH */
 175         case 'b':
 176                 /* FALLTHROUGH */
 177         case 'i':
 178                 (*pos)--;
 179                 break;
 180         default:
 181                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 182                     ln, *pos - 1, NULL);
 183                 return(0);
 184         }
 185
 186         switch (tolower((unsigned char)p[(*pos)++])) {
 187         case '3':
 188                 /* FALLTHROUGH */
 189         case 'b':
 190                 cp->flags |= TBL_CELL_BOLD;
 191                 goto mod;
 192         case '2':
 193                 /* FALLTHROUGH */
 194         case 'i':
 195                 cp->flags |= TBL_CELL_ITALIC;
 196                 goto mod;
 197         case '1':
 198                 /* FALLTHROUGH */
 199         case 'r':
 200                 goto mod;
 201         default:
 202                 break;
 203         }
 204         if (isalnum((unsigned char)p[*pos - 1])) {
 205                 mandoc_vmsg(MANDOCERR_FT_BAD, tbl->parse,
 206                     ln, *pos - 1, "TS f%c", p[*pos - 1]);
 207                 goto mod;
 208         }
 209
 210         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 211             ln, *pos - 1, NULL);
 212         return(0);
 213 }
 214
 215 static int
 216 cell(struct tbl_node *tbl, struct tbl_row *rp,
 217                 int ln, const char *p, int *pos)
 218 {
 219         int              vert, i;
 220         enum tbl_cellt   c;
 221
 222         /* Handle vertical lines. */
 223
 224         for (vert = 0; '|' == p[*pos]; ++*pos)
 225                 vert++;
 226         while (' ' == p[*pos])
 227                 (*pos)++;
 228
 229         /* Handle trailing vertical lines */
 230
 231         if ('.' == p[*pos] || '\0' == p[*pos]) {
 232                 rp->vert = vert;
 233                 return(1);
 234         }
 235
 236         /* Parse the column position (`c', `l', `r', ...). */
 237
 238         for (i = 0; i < KEYS_MAX; i++)
 239                 if (tolower((unsigned char)p[*pos]) == keys[i].name)
 240                         break;
 241
 242         if (KEYS_MAX == i) {
 243                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 244                     ln, *pos, NULL);
 245                 return(0);
 246         }
 247
 248         c = keys[i].key;
 249
 250         /*
 251          * If a span cell is found first, raise a warning and abort the
 252          * parse.  If a span cell is found and the last layout element
 253          * isn't a "normal" layout, bail.
 254          *
 255          * FIXME: recover from this somehow?
 256          */
 257
 258         if (TBL_CELL_SPAN == c) {
 259                 if (NULL == rp->first) {
 260                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 261                             ln, *pos, NULL);
 262                         return(0);
 263                 } else if (rp->last)
 264                         switch (rp->last->pos) {
 265                         case TBL_CELL_HORIZ:
 266                                 /* FALLTHROUGH */
 267                         case TBL_CELL_DHORIZ:
 268                                 mandoc_msg(MANDOCERR_TBLLAYOUT,
 269                                     tbl->parse, ln, *pos, NULL);
 270                                 return(0);
 271                         default:
 272                                 break;
 273                         }
 274         }
 275
 276         /*
 277          * If a vertical spanner is found, we may not be in the first
 278          * row.
 279          */
 280
 281         if (TBL_CELL_DOWN == c && rp == tbl->first_row) {
 282                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos, NULL);
 283                 return(0);
 284         }
 285
 286         (*pos)++;
 287
 288         /* Disallow adjacent spacers. */
 289
 290         if (vert > 2) {
 291                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos - 1, NULL);
 292                 return(0);
 293         }
 294
 295         /* Allocate cell then parse its modifiers. */
 296
 297         return(mods(tbl, cell_alloc(tbl, rp, c, vert), ln, p, pos));
 298 }
 299
 300 static void
 301 row(struct tbl_node *tbl, int ln, const char *p, int *pos)
 302 {
 303         struct tbl_row  *rp;
 304
 305 row:    /*
 306          * EBNF describing this section:
 307          *
 308          * row          ::= row_list [:space:]* [.]?[\n]
 309          * row_list     ::= [:space:]* row_elem row_tail
 310          * row_tail     ::= [:space:]*[,] row_list |
 311          *                  epsilon
 312          * row_elem     ::= [\t\ ]*[:alpha:]+
 313          */
 314
 315         rp = mandoc_calloc(1, sizeof(struct tbl_row));
 316         if (tbl->last_row)
 317                 tbl->last_row->next = rp;
 318         else
 319                 tbl->first_row = rp;
 320         tbl->last_row = rp;
 321
 322 cell:
 323         while (isspace((unsigned char)p[*pos]))
 324                 (*pos)++;
 325
 326         /* Safely exit layout context. */
 327
 328         if ('.' == p[*pos]) {
 329                 tbl->part = TBL_PART_DATA;
 330                 if (NULL == tbl->first_row)
 331                         mandoc_msg(MANDOCERR_TBLNOLAYOUT,
 332                             tbl->parse, ln, *pos, NULL);
 333                 (*pos)++;
 334                 return;
 335         }
 336
 337         /* End (and possibly restart) a row. */
 338
 339         if (',' == p[*pos]) {
 340                 (*pos)++;
 341                 goto row;
 342         } else if ('\0' == p[*pos])
 343                 return;
 344
 345         if ( ! cell(tbl, rp, ln, p, pos))
 346                 return;
 347
 348         goto cell;
 349         /* NOTREACHED */
 350 }
 351
 352 int
 353 tbl_layout(struct tbl_node *tbl, int ln, const char *p)
 354 {
 355         int              pos;
 356
 357         pos = 0;
 358         row(tbl, ln, p, &pos);
 359
 360         /* Always succeed. */
 361         return(1);
 362 }
 363
 364 static struct tbl_cell *
 365 cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos,
 366                 int vert)
 367 {
 368         struct tbl_cell *p, *pp;
 369         struct tbl_head *h, *hp;
 370
 371         p = mandoc_calloc(1, sizeof(struct tbl_cell));
 372
 373         if (NULL != (pp = rp->last)) {
 374                 pp->next = p;
 375                 h = pp->head->next;
 376         } else {
 377                 rp->first = p;
 378                 h = tbl->first_head;
 379         }
 380         rp->last = p;
 381
 382         p->pos = pos;
 383         p->vert = vert;
 384
 385         /* Re-use header. */
 386
 387         if (h) {
 388                 p->head = h;
 389                 return(p);
 390         }
 391
 392         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 393         hp->ident = tbl->opts.cols++;
 394         hp->vert = vert;
 395
 396         if (tbl->last_head) {
 397                 hp->prev = tbl->last_head;
 398                 tbl->last_head->next = hp;
 399         } else
 400                 tbl->first_head = hp;
 401         tbl->last_head = hp;
 402
 403         p->head = hp;
 404         return(p);
 405 }