out.c

   1 /*      $Id: out.c,v 1.80 2021/08/10 12:55:03 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2011,2014,2015,2017,2018 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 <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 "out.h"
  34
  35 struct  tbl_colgroup {
  36         struct tbl_colgroup     *next;
  37         size_t                   wanted;
  38         int                      startcol;
  39         int                      endcol;
  40 };
  41
  42 static  size_t  tblcalc_data(struct rofftbl *, struct roffcol *,
  43                         const struct tbl_opts *, const struct tbl_dat *,
  44                         size_t);
  45 static  size_t  tblcalc_literal(struct rofftbl *, struct roffcol *,
  46                         const struct tbl_dat *, size_t);
  47 static  size_t  tblcalc_number(struct rofftbl *, struct roffcol *,
  48                         const struct tbl_opts *, const struct tbl_dat *);
  49
  50
  51 /*
  52  * Parse the *src string and store a scaling unit into *dst.
  53  * If the string doesn't specify the unit, use the default.
  54  * If no default is specified, fail.
  55  * Return a pointer to the byte after the last byte used,
  56  * or NULL on total failure.
  57  */
  58 const char *
  59 a2roffsu(const char *src, struct roffsu *dst, enum roffscale def)
  60 {
  61         char            *endptr;
  62
  63         dst->unit = def == SCALE_MAX ? SCALE_BU : def;
  64         dst->scale = strtod(src, &endptr);
  65         if (endptr == src)
  66                 return NULL;
  67
  68         switch (*endptr++) {
  69         case 'c':
  70                 dst->unit = SCALE_CM;
  71                 break;
  72         case 'i':
  73                 dst->unit = SCALE_IN;
  74                 break;
  75         case 'f':
  76                 dst->unit = SCALE_FS;
  77                 break;
  78         case 'M':
  79                 dst->unit = SCALE_MM;
  80                 break;
  81         case 'm':
  82                 dst->unit = SCALE_EM;
  83                 break;
  84         case 'n':
  85                 dst->unit = SCALE_EN;
  86                 break;
  87         case 'P':
  88                 dst->unit = SCALE_PC;
  89                 break;
  90         case 'p':
  91                 dst->unit = SCALE_PT;
  92                 break;
  93         case 'u':
  94                 dst->unit = SCALE_BU;
  95                 break;
  96         case 'v':
  97                 dst->unit = SCALE_VS;
  98                 break;
  99         default:
 100                 endptr--;
 101                 if (SCALE_MAX == def)
 102                         return NULL;
 103                 dst->unit = def;
 104                 break;
 105         }
 106         return endptr;
 107 }
 108
 109 /*
 110  * Calculate the abstract widths and decimal positions of columns in a
 111  * table.  This routine allocates the columns structures then runs over
 112  * all rows and cells in the table.  The function pointers in "tbl" are
 113  * used for the actual width calculations.
 114  */
 115 void
 116 tblcalc(struct rofftbl *tbl, const struct tbl_span *sp_first,
 117     size_t offset, size_t rmargin)
 118 {
 119         struct roffsu            su;
 120         const struct tbl_opts   *opts;
 121         const struct tbl_span   *sp;
 122         const struct tbl_dat    *dp;
 123         struct roffcol          *col;
 124         struct tbl_colgroup     *first_group, **gp, *g;
 125         size_t                  *colwidth;
 126         size_t                   ewidth, min1, min2, wanted, width, xwidth;
 127         int                      done, icol, maxcol, necol, nxcol, quirkcol;
 128
 129         /*
 130          * Allocate the master column specifiers.  These will hold the
 131          * widths and decimal positions for all cells in the column.  It
 132          * must be freed and nullified by the caller.
 133          */
 134
 135         assert(tbl->cols == NULL);
 136         tbl->cols = mandoc_calloc((size_t)sp_first->opts->cols,
 137             sizeof(struct roffcol));
 138         opts = sp_first->opts;
 139
 140         maxcol = -1;
 141         first_group = NULL;
 142         for (sp = sp_first; sp != NULL; sp = sp->next) {
 143                 if (sp->pos != TBL_SPAN_DATA)
 144                         continue;
 145
 146                 /*
 147                  * Account for the data cells in the layout, matching it
 148                  * to data cells in the data section.
 149                  */
 150
 151                 gp = &first_group;
 152                 for (dp = sp->first; dp != NULL; dp = dp->next) {
 153                         icol = dp->layout->col;
 154                         while (maxcol < icol + dp->hspans)
 155                                 tbl->cols[++maxcol].spacing = SIZE_MAX;
 156                         col = tbl->cols + icol;
 157                         col->flags |= dp->layout->flags;
 158                         if (dp->layout->flags & TBL_CELL_WIGN)
 159                                 continue;
 160
 161                         /* Handle explicit width specifications. */
 162
 163                         if (dp->layout->wstr != NULL &&
 164                             dp->layout->width == 0 &&
 165                             a2roffsu(dp->layout->wstr, &su, SCALE_EN)
 166                             != NULL)
 167                                 dp->layout->width =
 168                                     (*tbl->sulen)(&su, tbl->arg);
 169                         if (col->width < dp->layout->width)
 170                                 col->width = dp->layout->width;
 171                         if (dp->layout->spacing != SIZE_MAX &&
 172                             (col->spacing == SIZE_MAX ||
 173                              col->spacing < dp->layout->spacing))
 174                                 col->spacing = dp->layout->spacing;
 175
 176                         /*
 177                          * Calculate an automatic width.
 178                          * Except for spanning cells, apply it.
 179                          */
 180
 181                         width = tblcalc_data(tbl,
 182                             dp->hspans == 0 ? col : NULL,
 183                             opts, dp,
 184                             dp->block == 0 ? 0 :
 185                             dp->layout->width ? dp->layout->width :
 186                             rmargin ? (rmargin + sp->opts->cols / 2)
 187                             / (sp->opts->cols + 1) : 0);
 188                         if (dp->hspans == 0)
 189                                 continue;
 190
 191                         /*
 192                          * Build an ordered, singly linked list
 193                          * of all groups of columns joined by spans,
 194                          * recording the minimum width for each group.
 195                          */
 196
 197                         while (*gp != NULL && ((*gp)->startcol < icol ||
 198                             (*gp)->endcol < icol + dp->hspans))
 199                                 gp = &(*gp)->next;
 200                         if (*gp == NULL || (*gp)->startcol > icol ||
 201                             (*gp)->endcol > icol + dp->hspans) {
 202                                 g = mandoc_malloc(sizeof(*g));
 203                                 g->next = *gp;
 204                                 g->wanted = width;
 205                                 g->startcol = icol;
 206                                 g->endcol = icol + dp->hspans;
 207                                 *gp = g;
 208                         } else if ((*gp)->wanted < width)
 209                                 (*gp)->wanted = width;
 210                 }
 211         }
 212
 213         /*
 214          * The minimum width of columns explicitly specified
 215          * in the layout is 1n.
 216          */
 217
 218         if (maxcol < sp_first->opts->cols - 1)
 219                 maxcol = sp_first->opts->cols - 1;
 220         for (icol = 0; icol <= maxcol; icol++) {
 221                 col = tbl->cols + icol;
 222                 if (col->width < 1)
 223                         col->width = 1;
 224
 225                 /*
 226                  * Column spacings are needed for span width
 227                  * calculations, so set the default values now.
 228                  */
 229
 230                 if (col->spacing == SIZE_MAX || icol == maxcol)
 231                         col->spacing = 3;
 232         }
 233
 234         /*
 235          * Replace the minimum widths with the missing widths,
 236          * and dismiss groups that are already wide enough.
 237          */
 238
 239         gp = &first_group;
 240         while ((g = *gp) != NULL) {
 241                 done = 0;
 242                 for (icol = g->startcol; icol <= g->endcol; icol++) {
 243                         width = tbl->cols[icol].width;
 244                         if (icol < g->endcol)
 245                                 width += tbl->cols[icol].spacing;
 246                         if (g->wanted <= width) {
 247                                 done = 1;
 248                                 break;
 249                         } else
 250                                 (*gp)->wanted -= width;
 251                 }
 252                 if (done) {
 253                         *gp = g->next;
 254                         free(g);
 255                 } else
 256                         gp = &(*gp)->next;
 257         }
 258
 259         colwidth = mandoc_reallocarray(NULL, maxcol + 1, sizeof(*colwidth));
 260         while (first_group != NULL) {
 261
 262                 /*
 263                  * Rebuild the array of the widths of all columns
 264                  * participating in spans that require expansion.
 265                  */
 266
 267                 for (icol = 0; icol <= maxcol; icol++)
 268                         colwidth[icol] = SIZE_MAX;
 269                 for (g = first_group; g != NULL; g = g->next)
 270                         for (icol = g->startcol; icol <= g->endcol; icol++)
 271                                 colwidth[icol] = tbl->cols[icol].width;
 272
 273                 /*
 274                  * Find the smallest and second smallest column width
 275                  * among the columns which may need expamsion.
 276                  */
 277
 278                 min1 = min2 = SIZE_MAX;
 279                 for (icol = 0; icol <= maxcol; icol++) {
 280                         if (min1 > colwidth[icol]) {
 281                                 min2 = min1;
 282                                 min1 = colwidth[icol];
 283                         } else if (min1 < colwidth[icol] &&
 284                             min2 > colwidth[icol])
 285                                 min2 = colwidth[icol];
 286                 }
 287
 288                 /*
 289                  * Find the minimum wanted width
 290                  * for any one of the narrowest columns,
 291                  * and mark the columns wanting that width.
 292                  */
 293
 294                 wanted = min2;
 295                 for (g = first_group; g != NULL; g = g->next) {
 296                         necol = 0;
 297                         for (icol = g->startcol; icol <= g->endcol; icol++)
 298                                 if (tbl->cols[icol].width == min1)
 299                                         necol++;
 300                         if (necol == 0)
 301                                 continue;
 302                         width = min1 + (g->wanted - 1) / necol + 1;
 303                         if (width > min2)
 304                                 width = min2;
 305                         if (wanted > width)
 306                                 wanted = width;
 307                         for (icol = g->startcol; icol <= g->endcol; icol++)
 308                                 if (colwidth[icol] == min1 ||
 309                                     (colwidth[icol] < min2 &&
 310                                      colwidth[icol] > width))
 311                                         colwidth[icol] = width;
 312                 }
 313
 314                 /* Record the effect of the widening on the group list. */
 315
 316                 gp = &first_group;
 317                 while ((g = *gp) != NULL) {
 318                         done = 0;
 319                         for (icol = g->startcol; icol <= g->endcol; icol++) {
 320                                 if (colwidth[icol] != wanted ||
 321                                     tbl->cols[icol].width == wanted)
 322                                         continue;
 323                                 if (g->wanted <= wanted - min1) {
 324                                         done = 1;
 325                                         break;
 326                                 }
 327                                 g->wanted -= wanted - min1;
 328                         }
 329                         if (done) {
 330                                 *gp = g->next;
 331                                 free(g);
 332                         } else
 333                                 gp = &(*gp)->next;
 334                 }
 335
 336                 /* Record the effect of the widening on the columns. */
 337
 338                 for (icol = 0; icol <= maxcol; icol++)
 339                         if (colwidth[icol] == wanted)
 340                                 tbl->cols[icol].width = wanted;
 341         }
 342         free(colwidth);
 343
 344         /*
 345          * Align numbers with text.
 346          * Count columns to equalize and columns to maximize.
 347          * Find maximum width of the columns to equalize.
 348          * Find total width of the columns *not* to maximize.
 349          */
 350
 351         necol = nxcol = 0;
 352         ewidth = xwidth = 0;
 353         for (icol = 0; icol <= maxcol; icol++) {
 354                 col = tbl->cols + icol;
 355                 if (col->width > col->nwidth)
 356                         col->decimal += (col->width - col->nwidth) / 2;
 357                 else
 358                         col->width = col->nwidth;
 359                 if (col->flags & TBL_CELL_EQUAL) {
 360                         necol++;
 361                         if (ewidth < col->width)
 362                                 ewidth = col->width;
 363                 }
 364                 if (col->flags & TBL_CELL_WMAX)
 365                         nxcol++;
 366                 else
 367                         xwidth += col->width;
 368         }
 369
 370         /*
 371          * Equalize columns, if requested for any of them.
 372          * Update total width of the columns not to maximize.
 373          */
 374
 375         if (necol) {
 376                 for (icol = 0; icol <= maxcol; icol++) {
 377                         col = tbl->cols + icol;
 378                         if ( ! (col->flags & TBL_CELL_EQUAL))
 379                                 continue;
 380                         if (col->width == ewidth)
 381                                 continue;
 382                         if (nxcol && rmargin)
 383                                 xwidth += ewidth - col->width;
 384                         col->width = ewidth;
 385                 }
 386         }
 387
 388         /*
 389          * If there are any columns to maximize, find the total
 390          * available width, deducting 3n margins between columns.
 391          * Distribute the available width evenly.
 392          */
 393
 394         if (nxcol && rmargin) {
 395                 xwidth += 3*maxcol +
 396                     (opts->opts & (TBL_OPT_BOX | TBL_OPT_DBOX) ?
 397                      2 : !!opts->lvert + !!opts->rvert);
 398                 if (rmargin <= offset + xwidth)
 399                         return;
 400                 xwidth = rmargin - offset - xwidth;
 401
 402                 /*
 403                  * Emulate a bug in GNU tbl width calculation that
 404                  * manifests itself for large numbers of x-columns.
 405                  * Emulating it for 5 x-columns gives identical
 406                  * behaviour for up to 6 x-columns.
 407                  */
 408
 409                 if (nxcol == 5) {
 410                         quirkcol = xwidth % nxcol + 2;
 411                         if (quirkcol != 3 && quirkcol != 4)
 412                                 quirkcol = -1;
 413                 } else
 414                         quirkcol = -1;
 415
 416                 necol = 0;
 417                 ewidth = 0;
 418                 for (icol = 0; icol <= maxcol; icol++) {
 419                         col = tbl->cols + icol;
 420                         if ( ! (col->flags & TBL_CELL_WMAX))
 421                                 continue;
 422                         col->width = (double)xwidth * ++necol / nxcol
 423                             - ewidth + 0.4995;
 424                         if (necol == quirkcol)
 425                                 col->width--;
 426                         ewidth += col->width;
 427                 }
 428         }
 429 }
 430
 431 static size_t
 432 tblcalc_data(struct rofftbl *tbl, struct roffcol *col,
 433     const struct tbl_opts *opts, const struct tbl_dat *dp, size_t mw)
 434 {
 435         size_t           sz;
 436
 437         /* Branch down into data sub-types. */
 438
 439         switch (dp->layout->pos) {
 440         case TBL_CELL_HORIZ:
 441         case TBL_CELL_DHORIZ:
 442                 sz = (*tbl->len)(1, tbl->arg);
 443                 if (col != NULL && col->width < sz)
 444                         col->width = sz;
 445                 return sz;
 446         case TBL_CELL_LONG:
 447         case TBL_CELL_CENTRE:
 448         case TBL_CELL_LEFT:
 449         case TBL_CELL_RIGHT:
 450                 return tblcalc_literal(tbl, col, dp, mw);
 451         case TBL_CELL_NUMBER:
 452                 return tblcalc_number(tbl, col, opts, dp);
 453         case TBL_CELL_DOWN:
 454                 return 0;
 455         default:
 456                 abort();
 457         }
 458 }
 459
 460 static size_t
 461 tblcalc_literal(struct rofftbl *tbl, struct roffcol *col,
 462     const struct tbl_dat *dp, size_t mw)
 463 {
 464         const char      *str;   /* Beginning of the first line. */
 465         const char      *beg;   /* Beginning of the current line. */
 466         char            *end;   /* End of the current line. */
 467         size_t           lsz;   /* Length of the current line. */
 468         size_t           wsz;   /* Length of the current word. */
 469         size_t           msz;   /* Length of the longest line. */
 470
 471         if (dp->string == NULL || *dp->string == '\0')
 472                 return 0;
 473         str = mw ? mandoc_strdup(dp->string) : dp->string;
 474         msz = lsz = 0;
 475         for (beg = str; beg != NULL && *beg != '\0'; beg = end) {
 476                 end = mw ? strchr(beg, ' ') : NULL;
 477                 if (end != NULL) {
 478                         *end++ = '\0';
 479                         while (*end == ' ')
 480                                 end++;
 481                 }
 482                 wsz = (*tbl->slen)(beg, tbl->arg);
 483                 if (mw && lsz && lsz + 1 + wsz <= mw)
 484                         lsz += 1 + wsz;
 485                 else
 486                         lsz = wsz;
 487                 if (msz < lsz)
 488                         msz = lsz;
 489         }
 490         if (mw)
 491                 free((void *)str);
 492         if (col != NULL && col->width < msz)
 493                 col->width = msz;
 494         return msz;
 495 }
 496
 497 static size_t
 498 tblcalc_number(struct rofftbl *tbl, struct roffcol *col,
 499                 const struct tbl_opts *opts, const struct tbl_dat *dp)
 500 {
 501         const char      *cp, *lastdigit, *lastpoint;
 502         size_t           intsz, totsz;
 503         char             buf[2];
 504
 505         if (dp->string == NULL || *dp->string == '\0')
 506                 return 0;
 507
 508         totsz = (*tbl->slen)(dp->string, tbl->arg);
 509         if (col == NULL)
 510                 return totsz;
 511
 512         /*
 513          * Find the last digit and
 514          * the last decimal point that is adjacent to a digit.
 515          * The alignment indicator "\&" overrides everything.
 516          */
 517
 518         lastdigit = lastpoint = NULL;
 519         for (cp = dp->string; cp[0] != '\0'; cp++) {
 520                 if (cp[0] == '\\' && cp[1] == '&') {
 521                         lastdigit = lastpoint = cp;
 522                         break;
 523                 } else if (cp[0] == opts->decimal &&
 524                     (isdigit((unsigned char)cp[1]) ||
 525                      (cp > dp->string && isdigit((unsigned char)cp[-1]))))
 526                         lastpoint = cp;
 527                 else if (isdigit((unsigned char)cp[0]))
 528                         lastdigit = cp;
 529         }
 530
 531         /* Not a number, treat as a literal string. */
 532
 533         if (lastdigit == NULL) {
 534                 if (col != NULL && col->width < totsz)
 535                         col->width = totsz;
 536                 return totsz;
 537         }
 538
 539         /* Measure the width of the integer part. */
 540
 541         if (lastpoint == NULL)
 542                 lastpoint = lastdigit + 1;
 543         intsz = 0;
 544         buf[1] = '\0';
 545         for (cp = dp->string; cp < lastpoint; cp++) {
 546                 buf[0] = cp[0];
 547                 intsz += (*tbl->slen)(buf, tbl->arg);
 548         }
 549
 550         /*
 551          * If this number has more integer digits than all numbers
 552          * seen on earlier lines, shift them all to the right.
 553          * If it has fewer, shift this number to the right.
 554          */
 555
 556         if (intsz > col->decimal) {
 557                 col->nwidth += intsz - col->decimal;
 558                 col->decimal = intsz;
 559         } else
 560                 totsz += col->decimal - intsz;
 561
 562         /* Update the maximum total width seen so far. */
 563
 564         if (totsz > col->nwidth)
 565                 col->nwidth = totsz;
 566         return totsz;
 567 }