tbl_layout.c

   1 /*      $Id: tbl_layout.c,v 1.30 2014/11/25 21:41:47 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  struct tbl_cell *cell_alloc(struct tbl_node *, struct tbl_row *,
  63                                 enum tbl_cellt, int vert);
  64
  65
  66 static int
  67 mods(struct tbl_node *tbl, struct tbl_cell *cp,
  68                 int ln, const char *p, int *pos)
  69 {
  70         char             buf[5];
  71         int              i;
  72
  73         /* Not all types accept modifiers. */
  74
  75         switch (cp->pos) {
  76         case TBL_CELL_DOWN:
  77                 /* FALLTHROUGH */
  78         case TBL_CELL_HORIZ:
  79                 /* FALLTHROUGH */
  80         case TBL_CELL_DHORIZ:
  81                 return(1);
  82         default:
  83                 break;
  84         }
  85
  86 mod:
  87         /*
  88          * XXX: since, at least for now, modifiers are non-conflicting
  89          * (are separable by value, regardless of position), we let
  90          * modifiers come in any order.  The existing tbl doesn't let
  91          * this happen.
  92          */
  93         switch (p[*pos]) {
  94         case '\0':
  95                 /* FALLTHROUGH */
  96         case ' ':
  97                 /* FALLTHROUGH */
  98         case '\t':
  99                 /* FALLTHROUGH */
 100         case ',':
 101                 /* FALLTHROUGH */
 102         case '.':
 103                 /* FALLTHROUGH */
 104         case '|':
 105                 return(1);
 106         default:
 107                 break;
 108         }
 109
 110         /* Throw away parenthesised expression. */
 111
 112         if ('(' == p[*pos]) {
 113                 (*pos)++;
 114                 while (p[*pos] && ')' != p[*pos])
 115                         (*pos)++;
 116                 if (')' == p[*pos]) {
 117                         (*pos)++;
 118                         goto mod;
 119                 }
 120                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 121                     ln, *pos, NULL);
 122                 return(0);
 123         }
 124
 125         /* Parse numerical spacing from modifier string. */
 126
 127         if (isdigit((unsigned char)p[*pos])) {
 128                 for (i = 0; i < 4; i++) {
 129                         if ( ! isdigit((unsigned char)p[*pos + i]))
 130                                 break;
 131                         buf[i] = p[*pos + i];
 132                 }
 133                 buf[i] = '\0';
 134
 135                 /* No greater than 4 digits. */
 136
 137                 if (4 == i) {
 138                         mandoc_msg(MANDOCERR_TBLLAYOUT,
 139                             tbl->parse, ln, *pos, NULL);
 140                         return(0);
 141                 }
 142
 143                 *pos += i;
 144                 cp->spacing = (size_t)atoi(buf);
 145
 146                 goto mod;
 147                 /* NOTREACHED */
 148         }
 149
 150         /* TODO: GNU has many more extensions. */
 151
 152         switch (tolower((unsigned char)p[(*pos)++])) {
 153         case 'z':
 154                 cp->flags |= TBL_CELL_WIGN;
 155                 goto mod;
 156         case 'u':
 157                 cp->flags |= TBL_CELL_UP;
 158                 goto mod;
 159         case 'e':
 160                 cp->flags |= TBL_CELL_EQUAL;
 161                 goto mod;
 162         case 't':
 163                 cp->flags |= TBL_CELL_TALIGN;
 164                 goto mod;
 165         case 'd':
 166                 cp->flags |= TBL_CELL_BALIGN;
 167                 goto mod;
 168         case 'w':  /* XXX for now, ignore minimal column width */
 169                 goto mod;
 170         case 'x':
 171                 cp->flags |= TBL_CELL_WMAX;
 172                 goto mod;
 173         case 'f':
 174                 break;
 175         case 'r':
 176                 /* FALLTHROUGH */
 177         case 'b':
 178                 /* FALLTHROUGH */
 179         case 'i':
 180                 (*pos)--;
 181                 break;
 182         default:
 183                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 184                     ln, *pos - 1, NULL);
 185                 return(0);
 186         }
 187
 188         switch (tolower((unsigned char)p[(*pos)++])) {
 189         case '3':
 190                 /* FALLTHROUGH */
 191         case 'b':
 192                 cp->flags |= TBL_CELL_BOLD;
 193                 goto mod;
 194         case '2':
 195                 /* FALLTHROUGH */
 196         case 'i':
 197                 cp->flags |= TBL_CELL_ITALIC;
 198                 goto mod;
 199         case '1':
 200                 /* FALLTHROUGH */
 201         case 'r':
 202                 goto mod;
 203         default:
 204                 break;
 205         }
 206         if (isalnum((unsigned char)p[*pos - 1])) {
 207                 mandoc_vmsg(MANDOCERR_FT_BAD, tbl->parse,
 208                     ln, *pos - 1, "TS f%c", p[*pos - 1]);
 209                 goto mod;
 210         }
 211
 212         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 213             ln, *pos - 1, NULL);
 214         return(0);
 215 }
 216
 217 static int
 218 cell(struct tbl_node *tbl, struct tbl_row *rp,
 219                 int ln, const char *p, int *pos)
 220 {
 221         int              vert, i;
 222         enum tbl_cellt   c;
 223
 224         /* Handle vertical lines. */
 225
 226         for (vert = 0; '|' == p[*pos]; ++*pos)
 227                 vert++;
 228         while (' ' == p[*pos])
 229                 (*pos)++;
 230
 231         /* Handle trailing vertical lines */
 232
 233         if ('.' == p[*pos] || '\0' == p[*pos]) {
 234                 rp->vert = vert;
 235                 return(1);
 236         }
 237
 238         /* Parse the column position (`c', `l', `r', ...). */
 239
 240         for (i = 0; i < KEYS_MAX; i++)
 241                 if (tolower((unsigned char)p[*pos]) == keys[i].name)
 242                         break;
 243
 244         if (KEYS_MAX == i) {
 245                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 246                     ln, *pos, NULL);
 247                 return(0);
 248         }
 249
 250         c = keys[i].key;
 251
 252         /*
 253          * If a span cell is found first, raise a warning and abort the
 254          * parse.  If a span cell is found and the last layout element
 255          * isn't a "normal" layout, bail.
 256          *
 257          * FIXME: recover from this somehow?
 258          */
 259
 260         if (TBL_CELL_SPAN == c) {
 261                 if (NULL == rp->first) {
 262                         mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse,
 263                             ln, *pos, NULL);
 264                         return(0);
 265                 } else if (rp->last)
 266                         switch (rp->last->pos) {
 267                         case TBL_CELL_HORIZ:
 268                                 /* FALLTHROUGH */
 269                         case TBL_CELL_DHORIZ:
 270                                 mandoc_msg(MANDOCERR_TBLLAYOUT,
 271                                     tbl->parse, ln, *pos, NULL);
 272                                 return(0);
 273                         default:
 274                                 break;
 275                         }
 276         }
 277
 278         /*
 279          * If a vertical spanner is found, we may not be in the first
 280          * row.
 281          */
 282
 283         if (TBL_CELL_DOWN == c && rp == tbl->first_row) {
 284                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos, NULL);
 285                 return(0);
 286         }
 287
 288         (*pos)++;
 289
 290         /* Disallow adjacent spacers. */
 291
 292         if (vert > 2) {
 293                 mandoc_msg(MANDOCERR_TBLLAYOUT, tbl->parse, ln, *pos - 1, NULL);
 294                 return(0);
 295         }
 296
 297         /* Allocate cell then parse its modifiers. */
 298
 299         return(mods(tbl, cell_alloc(tbl, rp, c, vert), ln, p, pos));
 300 }
 301
 302 int
 303 tbl_layout(struct tbl_node *tbl, int ln, const char *p)
 304 {
 305         struct tbl_row  *rp;
 306         int              pos;
 307
 308         pos = 0;
 309         rp = NULL;
 310
 311         for (;;) {
 312                 /* Skip whitespace before and after each cell. */
 313
 314                 while (isspace((unsigned char)p[pos]))
 315                         pos++;
 316
 317                 switch (p[pos]) {
 318                 case ',':  /* Next row on this input line. */
 319                         pos++;
 320                         rp = NULL;
 321                         continue;
 322                 case '\0':  /* Next row on next input line. */
 323                         return(1);
 324                 case '.':  /* End of layout. */
 325                         pos++;
 326                         tbl->part = TBL_PART_DATA;
 327                         if (tbl->first_row != NULL)
 328                                 return(1);
 329                         mandoc_msg(MANDOCERR_TBLNOLAYOUT,
 330                             tbl->parse, ln, pos, NULL);
 331                         rp = mandoc_calloc(1, sizeof(*rp));
 332                         cell_alloc(tbl, rp, TBL_CELL_LEFT, 0);
 333                         tbl->first_row = tbl->last_row = rp;
 334                         return(1);
 335                 default:  /* Cell. */
 336                         break;
 337                 }
 338
 339                 if (rp == NULL) {  /* First cell on this line. */
 340                         rp = mandoc_calloc(1, sizeof(*rp));
 341                         if (tbl->last_row)
 342                                 tbl->last_row->next = rp;
 343                         else
 344                                 tbl->first_row = rp;
 345                         tbl->last_row = rp;
 346                 }
 347                 if ( ! cell(tbl, rp, ln, p, &pos))
 348                         return(1);
 349         }
 350 }
 351
 352 static struct tbl_cell *
 353 cell_alloc(struct tbl_node *tbl, struct tbl_row *rp, enum tbl_cellt pos,
 354                 int vert)
 355 {
 356         struct tbl_cell *p, *pp;
 357         struct tbl_head *h, *hp;
 358
 359         p = mandoc_calloc(1, sizeof(struct tbl_cell));
 360
 361         if (NULL != (pp = rp->last)) {
 362                 pp->next = p;
 363                 h = pp->head->next;
 364         } else {
 365                 rp->first = p;
 366                 h = tbl->first_head;
 367         }
 368         rp->last = p;
 369
 370         p->pos = pos;
 371         p->vert = vert;
 372
 373         /* Re-use header. */
 374
 375         if (h) {
 376                 p->head = h;
 377                 return(p);
 378         }
 379
 380         hp = mandoc_calloc(1, sizeof(struct tbl_head));
 381         hp->ident = tbl->opts.cols++;
 382         hp->vert = vert;
 383
 384         if (tbl->last_head) {
 385                 hp->prev = tbl->last_head;
 386                 tbl->last_head->next = hp;
 387         } else
 388                 tbl->first_head = hp;
 389         tbl->last_head = hp;
 390
 391         p->head = hp;
 392         return(p);
 393 }