tbl_data.c

   1 /*      $Id: tbl_data.c,v 1.53 2020/01/11 20:48:18 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2011,2015,2017,2018,2019 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 <stdint.h>
  25 #include <stdio.h>
  26 #include <stdlib.h>
  27 #include <string.h>
  28 #include <time.h>
  29
  30 #include "mandoc_aux.h"
  31 #include "mandoc.h"
  32 #include "tbl.h"
  33 #include "libmandoc.h"
  34 #include "tbl_int.h"
  35
  36 static  void             getdata(struct tbl_node *, struct tbl_span *,
  37                                 int, const char *, int *);
  38 static  struct tbl_span *newspan(struct tbl_node *, int,
  39                                 struct tbl_row *);
  40
  41
  42 static void
  43 getdata(struct tbl_node *tbl, struct tbl_span *dp,
  44                 int ln, const char *p, int *pos)
  45 {
  46         struct tbl_dat  *dat, *pdat;
  47         struct tbl_cell *cp;
  48         struct tbl_span *pdp;
  49         int              sv;
  50
  51         /*
  52          * Determine the length of the string in the cell
  53          * and advance the parse point to the end of the cell.
  54          */
  55
  56         sv = *pos;
  57         while (p[*pos] != '\0' && p[*pos] != tbl->opts.tab)
  58                 (*pos)++;
  59
  60         /* Advance to the next layout cell, skipping spanners. */
  61
  62         cp = dp->last == NULL ? dp->layout->first : dp->last->layout->next;
  63         while (cp != NULL && cp->pos == TBL_CELL_SPAN)
  64                 cp = cp->next;
  65
  66         /*
  67          * If the current layout row is out of cells, allocate
  68          * a new cell if another row of the table has at least
  69          * this number of columns, or discard the input if we
  70          * are beyond the last column of the table as a whole.
  71          */
  72
  73         if (cp == NULL) {
  74                 if (dp->layout->last->col + 1 < dp->opts->cols) {
  75                         cp = mandoc_calloc(1, sizeof(*cp));
  76                         cp->pos = TBL_CELL_LEFT;
  77                         cp->spacing = SIZE_MAX;
  78                         dp->layout->last->next = cp;
  79                         cp->col = dp->layout->last->col + 1;
  80                         dp->layout->last = cp;
  81                 } else {
  82                         mandoc_msg(MANDOCERR_TBLDATA_EXTRA,
  83                             ln, sv, "%s", p + sv);
  84                         while (p[*pos] != '\0')
  85                                 (*pos)++;
  86                         return;
  87                 }
  88         }
  89
  90         dat = mandoc_malloc(sizeof(*dat));
  91         dat->layout = cp;
  92         dat->next = NULL;
  93         dat->string = NULL;
  94         dat->hspans = 0;
  95         dat->vspans = 0;
  96         dat->block = 0;
  97         dat->pos = TBL_DATA_NONE;
  98
  99         /*
 100          * Increment the number of vertical spans in a data cell above,
 101          * if this cell vertically extends one or more cells above.
 102          * The iteration must be done over data rows,
 103          * not over layout rows, because one layout row
 104          * can be reused for more than one data row.
 105          */
 106
 107         if (cp->pos == TBL_CELL_DOWN ||
 108             (*pos - sv == 2 && p[sv] == '\\' && p[sv + 1] == '^')) {
 109                 pdp = dp;
 110                 while ((pdp = pdp->prev) != NULL) {
 111                         pdat = pdp->first;
 112                         while (pdat != NULL &&
 113                             pdat->layout->col < dat->layout->col)
 114                                 pdat = pdat->next;
 115                         if (pdat == NULL)
 116                                 break;
 117                         if (pdat->layout->pos != TBL_CELL_DOWN &&
 118                             strcmp(pdat->string, "\\^") != 0) {
 119                                 pdat->vspans++;
 120                                 break;
 121                         }
 122                 }
 123         }
 124
 125         /*
 126          * Count the number of horizontal spans to the right of this cell.
 127          * This is purely a matter of the layout, independent of the data.
 128          */
 129
 130         for (cp = cp->next; cp != NULL; cp = cp->next)
 131                 if (cp->pos == TBL_CELL_SPAN)
 132                         dat->hspans++;
 133                 else
 134                         break;
 135
 136         /* Append the new data cell to the data row. */
 137
 138         if (dp->last == NULL)
 139                 dp->first = dat;
 140         else
 141                 dp->last->next = dat;
 142         dp->last = dat;
 143
 144         /*
 145          * Check for a continued-data scope opening.  This consists of a
 146          * trailing `T{' at the end of the line.  Subsequent lines,
 147          * until a standalone `T}', are included in our cell.
 148          */
 149
 150         if (*pos - sv == 2 && p[sv] == 'T' && p[sv + 1] == '{') {
 151                 tbl->part = TBL_PART_CDATA;
 152                 return;
 153         }
 154
 155         dat->string = mandoc_strndup(p + sv, *pos - sv);
 156
 157         if (p[*pos] != '\0')
 158                 (*pos)++;
 159
 160         if ( ! strcmp(dat->string, "_"))
 161                 dat->pos = TBL_DATA_HORIZ;
 162         else if ( ! strcmp(dat->string, "="))
 163                 dat->pos = TBL_DATA_DHORIZ;
 164         else if ( ! strcmp(dat->string, "\\_"))
 165                 dat->pos = TBL_DATA_NHORIZ;
 166         else if ( ! strcmp(dat->string, "\\="))
 167                 dat->pos = TBL_DATA_NDHORIZ;
 168         else
 169                 dat->pos = TBL_DATA_DATA;
 170
 171         if ((dat->layout->pos == TBL_CELL_HORIZ ||
 172             dat->layout->pos == TBL_CELL_DHORIZ ||
 173             dat->layout->pos == TBL_CELL_DOWN) &&
 174             dat->pos == TBL_DATA_DATA && *dat->string != '\0')
 175                 mandoc_msg(MANDOCERR_TBLDATA_SPAN,
 176                     ln, sv, "%s", dat->string);
 177 }
 178
 179 void
 180 tbl_cdata(struct tbl_node *tbl, int ln, const char *p, int pos)
 181 {
 182         struct tbl_dat  *dat;
 183         size_t           sz;
 184
 185         dat = tbl->last_span->last;
 186
 187         if (p[pos] == 'T' && p[pos + 1] == '}') {
 188                 pos += 2;
 189                 if (p[pos] == tbl->opts.tab) {
 190                         tbl->part = TBL_PART_DATA;
 191                         pos++;
 192                         while (p[pos] != '\0')
 193                                 getdata(tbl, tbl->last_span, ln, p, &pos);
 194                         return;
 195                 } else if (p[pos] == '\0') {
 196                         tbl->part = TBL_PART_DATA;
 197                         return;
 198                 }
 199
 200                 /* Fallthrough: T} is part of a word. */
 201         }
 202
 203         dat->pos = TBL_DATA_DATA;
 204         dat->block = 1;
 205
 206         if (dat->string != NULL) {
 207                 sz = strlen(p + pos) + strlen(dat->string) + 2;
 208                 dat->string = mandoc_realloc(dat->string, sz);
 209                 (void)strlcat(dat->string, " ", sz);
 210                 (void)strlcat(dat->string, p + pos, sz);
 211         } else
 212                 dat->string = mandoc_strdup(p + pos);
 213
 214         if (dat->layout->pos == TBL_CELL_DOWN)
 215                 mandoc_msg(MANDOCERR_TBLDATA_SPAN,
 216                     ln, pos, "%s", dat->string);
 217 }
 218
 219 static struct tbl_span *
 220 newspan(struct tbl_node *tbl, int line, struct tbl_row *rp)
 221 {
 222         struct tbl_span *dp;
 223
 224         dp = mandoc_calloc(1, sizeof(*dp));
 225         dp->line = line;
 226         dp->opts = &tbl->opts;
 227         dp->layout = rp;
 228         dp->prev = tbl->last_span;
 229
 230         if (dp->prev == NULL) {
 231                 tbl->first_span = dp;
 232                 tbl->current_span = NULL;
 233         } else
 234                 dp->prev->next = dp;
 235         tbl->last_span = dp;
 236
 237         return dp;
 238 }
 239
 240 void
 241 tbl_data(struct tbl_node *tbl, int ln, const char *p, int pos)
 242 {
 243         struct tbl_row  *rp;
 244         struct tbl_cell *cp;
 245         struct tbl_span *sp;
 246
 247         rp = (sp = tbl->last_span) == NULL ? tbl->first_row :
 248             sp->pos == TBL_SPAN_DATA && sp->layout->next != NULL ?
 249             sp->layout->next : sp->layout;
 250
 251         assert(rp != NULL);
 252
 253         if (p[1] == '\0') {
 254                 switch (p[0]) {
 255                 case '.':
 256                         /*
 257                          * Empty request lines must be handled here
 258                          * and cannot be discarded in roff_parseln()
 259                          * because in the layout section, they
 260                          * are significant and end the layout.
 261                          */
 262                         return;
 263                 case '_':
 264                         sp = newspan(tbl, ln, rp);
 265                         sp->pos = TBL_SPAN_HORIZ;
 266                         return;
 267                 case '=':
 268                         sp = newspan(tbl, ln, rp);
 269                         sp->pos = TBL_SPAN_DHORIZ;
 270                         return;
 271                 default:
 272                         break;
 273                 }
 274         }
 275
 276         /*
 277          * If the layout row contains nothing but horizontal lines,
 278          * allocate an empty span for it and assign the current span
 279          * to the next layout row accepting data.
 280          */
 281
 282         while (rp->next != NULL) {
 283                 if (rp->last->col + 1 < tbl->opts.cols)
 284                         break;
 285                 for (cp = rp->first; cp != NULL; cp = cp->next)
 286                         if (cp->pos != TBL_CELL_HORIZ &&
 287                             cp->pos != TBL_CELL_DHORIZ)
 288                                 break;
 289                 if (cp != NULL)
 290                         break;
 291                 sp = newspan(tbl, ln, rp);
 292                 sp->pos = TBL_SPAN_DATA;
 293                 rp = rp->next;
 294         }
 295
 296         /* Process a real data row. */
 297
 298         sp = newspan(tbl, ln, rp);
 299         sp->pos = TBL_SPAN_DATA;
 300         while (p[pos] != '\0')
 301                 getdata(tbl, sp, ln, p, &pos);
 302 }