tbl_data.c

   1 /*      $Id: tbl_data.c,v 1.58 2021/09/10 12:07:21 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2011,2015,2017-2019,2021 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         const char      *ccp;
  50         int              startpos, endpos;
  51
  52         /*
  53          * Determine the length of the string in the cell
  54          * and advance the parse point to the end of the cell.
  55          */
  56
  57         startpos = *pos;
  58         ccp = p + startpos;
  59         while (*ccp != '\0' && *ccp != tbl->opts.tab)
  60                 if (*ccp++ == '\\')
  61                         mandoc_escape(&ccp, NULL, NULL);
  62         *pos = ccp - p;
  63
  64         /* Advance to the next layout cell, skipping spanners. */
  65
  66         cp = dp->last == NULL ? dp->layout->first : dp->last->layout->next;
  67         while (cp != NULL && cp->pos == TBL_CELL_SPAN)
  68                 cp = cp->next;
  69
  70         /*
  71          * If the current layout row is out of cells, allocate
  72          * a new cell if another row of the table has at least
  73          * this number of columns, or discard the input if we
  74          * are beyond the last column of the table as a whole.
  75          */
  76
  77         if (cp == NULL) {
  78                 if (dp->layout->last->col + 1 < dp->opts->cols) {
  79                         cp = mandoc_calloc(1, sizeof(*cp));
  80                         cp->pos = TBL_CELL_LEFT;
  81                         cp->font = ESCAPE_FONTROMAN;
  82                         cp->spacing = SIZE_MAX;
  83                         dp->layout->last->next = cp;
  84                         cp->col = dp->layout->last->col + 1;
  85                         dp->layout->last = cp;
  86                 } else {
  87                         mandoc_msg(MANDOCERR_TBLDATA_EXTRA,
  88                             ln, startpos, "%s", p + startpos);
  89                         while (p[*pos] != '\0')
  90                                 (*pos)++;
  91                         return;
  92                 }
  93         }
  94
  95         dat = mandoc_malloc(sizeof(*dat));
  96         dat->layout = cp;
  97         dat->next = NULL;
  98         dat->string = NULL;
  99         dat->hspans = 0;
 100         dat->vspans = 0;
 101         dat->block = 0;
 102         dat->pos = TBL_DATA_NONE;
 103
 104         /*
 105          * Increment the number of vertical spans in a data cell above,
 106          * if this cell vertically extends one or more cells above.
 107          * The iteration must be done over data rows,
 108          * not over layout rows, because one layout row
 109          * can be reused for more than one data row.
 110          */
 111
 112         if (cp->pos == TBL_CELL_DOWN ||
 113             (*pos - startpos == 2 &&
 114              p[startpos] == '\\' && p[startpos + 1] == '^')) {
 115                 pdp = dp;
 116                 while ((pdp = pdp->prev) != NULL) {
 117                         pdat = pdp->first;
 118                         while (pdat != NULL &&
 119                             pdat->layout->col < dat->layout->col)
 120                                 pdat = pdat->next;
 121                         if (pdat == NULL)
 122                                 break;
 123                         if (pdat->layout->pos != TBL_CELL_DOWN &&
 124                             strcmp(pdat->string, "\\^") != 0) {
 125                                 pdat->vspans++;
 126                                 break;
 127                         }
 128                 }
 129         }
 130
 131         /*
 132          * Count the number of horizontal spans to the right of this cell.
 133          * This is purely a matter of the layout, independent of the data.
 134          */
 135
 136         for (cp = cp->next; cp != NULL; cp = cp->next)
 137                 if (cp->pos == TBL_CELL_SPAN)
 138                         dat->hspans++;
 139                 else
 140                         break;
 141
 142         /* Append the new data cell to the data row. */
 143
 144         if (dp->last == NULL)
 145                 dp->first = dat;
 146         else
 147                 dp->last->next = dat;
 148         dp->last = dat;
 149
 150         /*
 151          * Check for a continued-data scope opening.  This consists of a
 152          * trailing `T{' at the end of the line.  Subsequent lines,
 153          * until a standalone `T}', are included in our cell.
 154          */
 155
 156         if (*pos - startpos == 2 &&
 157             p[startpos] == 'T' && p[startpos + 1] == '{') {
 158                 tbl->part = TBL_PART_CDATA;
 159                 return;
 160         }
 161
 162         endpos = *pos;
 163         if (dp->opts->opts & TBL_OPT_NOSPACE) {
 164                 while (p[startpos] == ' ')
 165                         startpos++;
 166                 while (endpos > startpos && p[endpos - 1] == ' ')
 167                         endpos--;
 168         }
 169         dat->string = mandoc_strndup(p + startpos, endpos - startpos);
 170
 171         if (p[*pos] != '\0')
 172                 (*pos)++;
 173
 174         if ( ! strcmp(dat->string, "_"))
 175                 dat->pos = TBL_DATA_HORIZ;
 176         else if ( ! strcmp(dat->string, "="))
 177                 dat->pos = TBL_DATA_DHORIZ;
 178         else if ( ! strcmp(dat->string, "\\_"))
 179                 dat->pos = TBL_DATA_NHORIZ;
 180         else if ( ! strcmp(dat->string, "\\="))
 181                 dat->pos = TBL_DATA_NDHORIZ;
 182         else
 183                 dat->pos = TBL_DATA_DATA;
 184
 185         if ((dat->layout->pos == TBL_CELL_HORIZ ||
 186             dat->layout->pos == TBL_CELL_DHORIZ ||
 187             dat->layout->pos == TBL_CELL_DOWN) &&
 188             dat->pos == TBL_DATA_DATA && *dat->string != '\0')
 189                 mandoc_msg(MANDOCERR_TBLDATA_SPAN,
 190                     ln, startpos, "%s", dat->string);
 191 }
 192
 193 void
 194 tbl_cdata(struct tbl_node *tbl, int ln, const char *p, int pos)
 195 {
 196         struct tbl_dat  *dat;
 197         size_t           sz;
 198
 199         dat = tbl->last_span->last;
 200
 201         if (p[pos] == 'T' && p[pos + 1] == '}') {
 202                 pos += 2;
 203                 if (tbl->opts.opts & TBL_OPT_NOSPACE)
 204                         while (p[pos] == ' ')
 205                                 pos++;
 206                 if (p[pos] == tbl->opts.tab) {
 207                         tbl->part = TBL_PART_DATA;
 208                         pos++;
 209                         while (p[pos] != '\0')
 210                                 getdata(tbl, tbl->last_span, ln, p, &pos);
 211                         return;
 212                 } else if (p[pos] == '\0') {
 213                         tbl->part = TBL_PART_DATA;
 214                         return;
 215                 }
 216
 217                 /* Fallthrough: T} is part of a word. */
 218         }
 219
 220         dat->pos = TBL_DATA_DATA;
 221         dat->block = 1;
 222
 223         if (dat->string != NULL) {
 224                 sz = strlen(p + pos) + strlen(dat->string) + 2;
 225                 dat->string = mandoc_realloc(dat->string, sz);
 226                 (void)strlcat(dat->string, " ", sz);
 227                 (void)strlcat(dat->string, p + pos, sz);
 228         } else
 229                 dat->string = mandoc_strdup(p + pos);
 230
 231         if (dat->layout->pos == TBL_CELL_DOWN)
 232                 mandoc_msg(MANDOCERR_TBLDATA_SPAN,
 233                     ln, pos, "%s", dat->string);
 234 }
 235
 236 static struct tbl_span *
 237 newspan(struct tbl_node *tbl, int line, struct tbl_row *rp)
 238 {
 239         struct tbl_span *dp;
 240
 241         dp = mandoc_calloc(1, sizeof(*dp));
 242         dp->line = line;
 243         dp->opts = &tbl->opts;
 244         dp->layout = rp;
 245         dp->prev = tbl->last_span;
 246
 247         if (dp->prev == NULL) {
 248                 tbl->first_span = dp;
 249                 tbl->current_span = NULL;
 250         } else
 251                 dp->prev->next = dp;
 252         tbl->last_span = dp;
 253
 254         return dp;
 255 }
 256
 257 void
 258 tbl_data(struct tbl_node *tbl, int ln, const char *p, int pos)
 259 {
 260         struct tbl_row  *rp;
 261         struct tbl_cell *cp;
 262         struct tbl_span *sp;
 263
 264         for (sp = tbl->last_span; sp != NULL; sp = sp->prev)
 265                 if (sp->pos == TBL_SPAN_DATA)
 266                         break;
 267         rp = sp == NULL ? tbl->first_row :
 268             sp->layout->next == NULL ? sp->layout : sp->layout->next;
 269         assert(rp != NULL);
 270
 271         if (p[1] == '\0') {
 272                 switch (p[0]) {
 273                 case '.':
 274                         /*
 275                          * Empty request lines must be handled here
 276                          * and cannot be discarded in roff_parseln()
 277                          * because in the layout section, they
 278                          * are significant and end the layout.
 279                          */
 280                         return;
 281                 case '_':
 282                         sp = newspan(tbl, ln, rp);
 283                         sp->pos = TBL_SPAN_HORIZ;
 284                         return;
 285                 case '=':
 286                         sp = newspan(tbl, ln, rp);
 287                         sp->pos = TBL_SPAN_DHORIZ;
 288                         return;
 289                 default:
 290                         break;
 291                 }
 292         }
 293
 294         /*
 295          * If the layout row contains nothing but horizontal lines,
 296          * allocate an empty span for it and assign the current span
 297          * to the next layout row accepting data.
 298          */
 299
 300         while (rp->next != NULL) {
 301                 if (rp->last->col + 1 < tbl->opts.cols)
 302                         break;
 303                 for (cp = rp->first; cp != NULL; cp = cp->next)
 304                         if (cp->pos != TBL_CELL_HORIZ &&
 305                             cp->pos != TBL_CELL_DHORIZ)
 306                                 break;
 307                 if (cp != NULL)
 308                         break;
 309                 sp = newspan(tbl, ln, rp);
 310                 sp->pos = TBL_SPAN_DATA;
 311                 rp = rp->next;
 312         }
 313
 314         /* Process a real data row. */
 315
 316         sp = newspan(tbl, ln, rp);
 317         sp->pos = TBL_SPAN_DATA;
 318         while (p[pos] != '\0')
 319                 getdata(tbl, sp, ln, p, &pos);
 320 }