tbl_data.c

   1 /*      $Id: tbl_data.c,v 1.47 2018/11/25 21:17:34 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2011, 2015, 2017, 2018 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 <assert.h>
  23 #include <ctype.h>
  24 #include <stdlib.h>
  25 #include <string.h>
  26 #include <time.h>
  27
  28 #include "mandoc.h"
  29 #include "mandoc_aux.h"
  30 #include "libmandoc.h"
  31 #include "libroff.h"
  32
  33 static  void             getdata(struct tbl_node *, struct tbl_span *,
  34                                 int, const char *, int *);
  35 static  struct tbl_span *newspan(struct tbl_node *, int,
  36                                 struct tbl_row *);
  37
  38
  39 static void
  40 getdata(struct tbl_node *tbl, struct tbl_span *dp,
  41                 int ln, const char *p, int *pos)
  42 {
  43         struct tbl_dat  *dat, *pdat;
  44         struct tbl_cell *cp;
  45         struct tbl_span *pdp;
  46         int              sv;
  47
  48         /*
  49          * Determine the length of the string in the cell
  50          * and advance the parse point to the end of the cell.
  51          */
  52
  53         sv = *pos;
  54         while (p[*pos] != '\0' && p[*pos] != tbl->opts.tab)
  55                 (*pos)++;
  56
  57         /* Advance to the next layout cell, skipping spanners. */
  58
  59         cp = dp->last == NULL ? dp->layout->first : dp->last->layout->next;
  60         while (cp != NULL && cp->pos == TBL_CELL_SPAN)
  61                 cp = cp->next;
  62
  63         /*
  64          * If the current layout row is out of cells, allocate
  65          * a new cell if another row of the table has at least
  66          * this number of columns, or discard the input if we
  67          * are beyond the last column of the table as a whole.
  68          */
  69
  70         if (cp == NULL) {
  71                 if (dp->layout->last->col + 1 < dp->opts->cols) {
  72                         cp = mandoc_calloc(1, sizeof(*cp));
  73                         cp->pos = TBL_CELL_LEFT;
  74                         dp->layout->last->next = cp;
  75                         cp->col = dp->layout->last->col + 1;
  76                         dp->layout->last = cp;
  77                 } else {
  78                         mandoc_msg(MANDOCERR_TBLDATA_EXTRA, tbl->parse,
  79                             ln, sv, p + sv);
  80                         while (p[*pos] != '\0')
  81                                 (*pos)++;
  82                         return;
  83                 }
  84         }
  85
  86         dat = mandoc_malloc(sizeof(*dat));
  87         dat->layout = cp;
  88         dat->next = NULL;
  89         dat->string = NULL;
  90         dat->hspans = 0;
  91         dat->vspans = 0;
  92         dat->block = 0;
  93         dat->pos = TBL_DATA_NONE;
  94
  95         /*
  96          * Increment the number of vertical spans in a data cell above,
  97          * if this cell vertically extends one or more cells above.
  98          * The iteration must be done over data rows,
  99          * not over layout rows, because one layout row
 100          * can be reused for more than one data row.
 101          */
 102
 103         if (cp->pos == TBL_CELL_DOWN ||
 104             (*pos - sv == 2 && p[sv] == '\\' && p[sv + 1] == '^')) {
 105                 pdp = dp;
 106                 while ((pdp = pdp->prev) != NULL) {
 107                         pdat = pdp->first;
 108                         while (pdat != NULL &&
 109                             pdat->layout->col < dat->layout->col)
 110                                 pdat = pdat->next;
 111                         if (pdat == NULL)
 112                                 break;
 113                         if (pdat->layout->pos != TBL_CELL_DOWN &&
 114                             strcmp(pdat->string, "\\^") != 0) {
 115                                 pdat->vspans++;
 116                                 break;
 117                         }
 118                 }
 119         }
 120
 121         /*
 122          * Count the number of horizontal spans to the right of this cell.
 123          * This is purely a matter of the layout, independent of the data.
 124          */
 125
 126         for (cp = cp->next; cp != NULL; cp = cp->next)
 127                 if (cp->pos == TBL_CELL_SPAN)
 128                         dat->hspans++;
 129                 else
 130                         break;
 131
 132         /* Append the new data cell to the data row. */
 133
 134         if (dp->last == NULL)
 135                 dp->first = dat;
 136         else
 137                 dp->last->next = dat;
 138         dp->last = dat;
 139
 140         /*
 141          * Check for a continued-data scope opening.  This consists of a
 142          * trailing `T{' at the end of the line.  Subsequent lines,
 143          * until a standalone `T}', are included in our cell.
 144          */
 145
 146         if (*pos - sv == 2 && p[sv] == 'T' && p[sv + 1] == '{') {
 147                 tbl->part = TBL_PART_CDATA;
 148                 return;
 149         }
 150
 151         dat->string = mandoc_strndup(p + sv, *pos - sv);
 152
 153         if (p[*pos] != '\0')
 154                 (*pos)++;
 155
 156         if ( ! strcmp(dat->string, "_"))
 157                 dat->pos = TBL_DATA_HORIZ;
 158         else if ( ! strcmp(dat->string, "="))
 159                 dat->pos = TBL_DATA_DHORIZ;
 160         else if ( ! strcmp(dat->string, "\\_"))
 161                 dat->pos = TBL_DATA_NHORIZ;
 162         else if ( ! strcmp(dat->string, "\\="))
 163                 dat->pos = TBL_DATA_NDHORIZ;
 164         else
 165                 dat->pos = TBL_DATA_DATA;
 166
 167         if ((dat->layout->pos == TBL_CELL_HORIZ ||
 168             dat->layout->pos == TBL_CELL_DHORIZ ||
 169             dat->layout->pos == TBL_CELL_DOWN) &&
 170             dat->pos == TBL_DATA_DATA && *dat->string != '\0')
 171                 mandoc_msg(MANDOCERR_TBLDATA_SPAN,
 172                     tbl->parse, ln, sv, dat->string);
 173 }
 174
 175 void
 176 tbl_cdata(struct tbl_node *tbl, int ln, const char *p, int pos)
 177 {
 178         struct tbl_dat  *dat;
 179         size_t           sz;
 180
 181         dat = tbl->last_span->last;
 182
 183         if (p[pos] == 'T' && p[pos + 1] == '}') {
 184                 pos += 2;
 185                 if (p[pos] == tbl->opts.tab) {
 186                         tbl->part = TBL_PART_DATA;
 187                         pos++;
 188                         while (p[pos] != '\0')
 189                                 getdata(tbl, tbl->last_span, ln, p, &pos);
 190                         return;
 191                 } else if (p[pos] == '\0') {
 192                         tbl->part = TBL_PART_DATA;
 193                         return;
 194                 }
 195
 196                 /* Fallthrough: T} is part of a word. */
 197         }
 198
 199         dat->pos = TBL_DATA_DATA;
 200         dat->block = 1;
 201
 202         if (dat->string != NULL) {
 203                 sz = strlen(p + pos) + strlen(dat->string) + 2;
 204                 dat->string = mandoc_realloc(dat->string, sz);
 205                 (void)strlcat(dat->string, " ", sz);
 206                 (void)strlcat(dat->string, p + pos, sz);
 207         } else
 208                 dat->string = mandoc_strdup(p + pos);
 209
 210         if (dat->layout->pos == TBL_CELL_DOWN)
 211                 mandoc_msg(MANDOCERR_TBLDATA_SPAN, tbl->parse,
 212                     ln, pos, dat->string);
 213 }
 214
 215 static struct tbl_span *
 216 newspan(struct tbl_node *tbl, int line, struct tbl_row *rp)
 217 {
 218         struct tbl_span *dp;
 219
 220         dp = mandoc_calloc(1, sizeof(*dp));
 221         dp->line = line;
 222         dp->opts = &tbl->opts;
 223         dp->layout = rp;
 224         dp->prev = tbl->last_span;
 225
 226         if (dp->prev == NULL) {
 227                 tbl->first_span = dp;
 228                 tbl->current_span = NULL;
 229         } else
 230                 dp->prev->next = dp;
 231         tbl->last_span = dp;
 232
 233         return dp;
 234 }
 235
 236 void
 237 tbl_data(struct tbl_node *tbl, int ln, const char *p, int pos)
 238 {
 239         struct tbl_row  *rp;
 240         struct tbl_cell *cp;
 241         struct tbl_span *sp;
 242
 243         rp = (sp = tbl->last_span) == NULL ? tbl->first_row :
 244             sp->pos == TBL_SPAN_DATA && sp->layout->next != NULL ?
 245             sp->layout->next : sp->layout;
 246
 247         assert(rp != NULL);
 248
 249         if ( ! strcmp(p, "_")) {
 250                 sp = newspan(tbl, ln, rp);
 251                 sp->pos = TBL_SPAN_HORIZ;
 252                 return;
 253         } else if ( ! strcmp(p, "=")) {
 254                 sp = newspan(tbl, ln, rp);
 255                 sp->pos = TBL_SPAN_DHORIZ;
 256                 return;
 257         }
 258
 259         /*
 260          * If the layout row contains nothing but horizontal lines,
 261          * allocate an empty span for it and assign the current span
 262          * to the next layout row accepting data.
 263          */
 264
 265         while (rp->next != NULL) {
 266                 if (rp->last->col + 1 < tbl->opts.cols)
 267                         break;
 268                 for (cp = rp->first; cp != NULL; cp = cp->next)
 269                         if (cp->pos != TBL_CELL_HORIZ &&
 270                             cp->pos != TBL_CELL_DHORIZ)
 271                                 break;
 272                 if (cp != NULL)
 273                         break;
 274                 sp = newspan(tbl, ln, rp);
 275                 sp->pos = TBL_SPAN_DATA;
 276                 rp = rp->next;
 277         }
 278
 279         /* Process a real data row. */
 280
 281         sp = newspan(tbl, ln, rp);
 282         sp->pos = TBL_SPAN_DATA;
 283         while (p[pos] != '\0')
 284                 getdata(tbl, sp, ln, p, &pos);
 285 }