contrib/mdocml/out.c

   1 /*      $Id: out.c,v 1.82 2021/09/07 17:07:58 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2011, 2014, 2015, 2017, 2018, 2019, 2021
   5  *               Ingo Schwarze <schwarze@openbsd.org>
   6  *
   7  * Permission to use, copy, modify, and distribute this software for any
   8  * purpose with or without fee is hereby granted, provided that the above
   9  * copyright notice and this permission notice appear in all copies.
  10  *
  11  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  12  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  13  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  14  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  15  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  16  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  17  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  18  */
  19 #include "config.h"
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <ctype.h>
  25 #include <stdint.h>
  26 #include <stdio.h>
  27 #include <stdlib.h>
  28 #include <string.h>
  29 #include <time.h>
  30
  31 #include "mandoc_aux.h"
  32 #include "mandoc.h"
  33 #include "tbl.h"
  34 #include "out.h"
  35
  36 struct  tbl_colgroup {
  37         struct tbl_colgroup     *next;
  38         size_t                   wanted;
  39         int                      startcol;
  40         int                      endcol;
  41 };
  42
  43 static  size_t  tblcalc_data(struct rofftbl *, struct roffcol *,
  44                         const struct tbl_opts *, const struct tbl_dat *,
  45                         size_t);
  46 static  size_t  tblcalc_literal(struct rofftbl *, struct roffcol *,
  47                         const struct tbl_dat *, size_t);
  48 static  size_t  tblcalc_number(struct rofftbl *, struct roffcol *,
  49                         const struct tbl_opts *, const struct tbl_dat *);
  50
  51
  52 /*
  53  * Parse the *src string and store a scaling unit into *dst.
  54  * If the string doesn't specify the unit, use the default.
  55  * If no default is specified, fail.
  56  * Return a pointer to the byte after the last byte used,
  57  * or NULL on total failure.
  58  */
  59 const char *
  60 a2roffsu(const char *src, struct roffsu *dst, enum roffscale def)
  61 {
  62         char            *endptr;
  63
  64         dst->unit = def == SCALE_MAX ? SCALE_BU : def;
  65         dst->scale = strtod(src, &endptr);
  66         if (endptr == src)
  67                 return NULL;
  68
  69         switch (*endptr++) {
  70         case 'c':
  71                 dst->unit = SCALE_CM;
  72                 break;
  73         case 'i':
  74                 dst->unit = SCALE_IN;
  75                 break;
  76         case 'f':
  77                 dst->unit = SCALE_FS;
  78                 break;
  79         case 'M':
  80                 dst->unit = SCALE_MM;
  81                 break;
  82         case 'm':
  83                 dst->unit = SCALE_EM;
  84                 break;
  85         case 'n':
  86                 dst->unit = SCALE_EN;
  87                 break;
  88         case 'P':
  89                 dst->unit = SCALE_PC;
  90                 break;
  91         case 'p':
  92                 dst->unit = SCALE_PT;
  93                 break;
  94         case 'u':
  95                 dst->unit = SCALE_BU;
  96                 break;
  97         case 'v':
  98                 dst->unit = SCALE_VS;
  99                 break;
 100         default:
 101                 endptr--;
 102                 if (SCALE_MAX == def)
 103                         return NULL;
 104                 dst->unit = def;
 105                 break;
 106         }
 107         return endptr;
 108 }
 109
 110 /*
 111  * Calculate the abstract widths and decimal positions of columns in a
 112  * table.  This routine allocates the columns structures then runs over
 113  * all rows and cells in the table.  The function pointers in "tbl" are
 114  * used for the actual width calculations.
 115  */
 116 void
 117 tblcalc(struct rofftbl *tbl, const struct tbl_span *sp_first,
 118     size_t offset, size_t rmargin)
 119 {
 120         struct roffsu            su;
 121         const struct tbl_opts   *opts;
 122         const struct tbl_span   *sp;
 123         const struct tbl_dat    *dp;
 124         struct roffcol          *col;
 125         struct tbl_colgroup     *first_group, **gp, *g;
 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         while (first_group != NULL) {
 260
 261                 /*
 262                  * Find the smallest and second smallest column width
 263                  * among the columns which may need expamsion.
 264                  */
 265
 266                 min1 = min2 = SIZE_MAX;
 267                 for (icol = 0; icol <= maxcol; icol++) {
 268                         width = tbl->cols[icol].width;
 269                         if (min1 > width) {
 270                                 min2 = min1;
 271                                 min1 = width;
 272                         } else if (min1 < width && min2 > width)
 273                                 min2 = width;
 274                 }
 275
 276                 /*
 277                  * Find the minimum wanted width
 278                  * for any one of the narrowest columns,
 279                  * and mark the columns wanting that width.
 280                  */
 281
 282                 wanted = min2;
 283                 for (g = first_group; g != NULL; g = g->next) {
 284                         necol = 0;
 285                         for (icol = g->startcol; icol <= g->endcol; icol++)
 286                                 if (tbl->cols[icol].width == min1)
 287                                         necol++;
 288                         if (necol == 0)
 289                                 continue;
 290                         width = min1 + (g->wanted - 1) / necol + 1;
 291                         if (width > min2)
 292                                 width = min2;
 293                         if (wanted > width)
 294                                 wanted = width;
 295                 }
 296
 297                 /* Record the effect of the widening. */
 298
 299                 gp = &first_group;
 300                 while ((g = *gp) != NULL) {
 301                         done = 0;
 302                         for (icol = g->startcol; icol <= g->endcol; icol++) {
 303                                 if (tbl->cols[icol].width != min1)
 304                                         continue;
 305                                 if (g->wanted <= wanted - min1) {
 306                                         tbl->cols[icol].width += g->wanted;
 307                                         done = 1;
 308                                         break;
 309                                 }
 310                                 tbl->cols[icol].width = wanted;
 311                                 g->wanted -= wanted - min1;
 312                         }
 313                         if (done) {
 314                                 *gp = g->next;
 315                                 free(g);
 316                         } else
 317                                 gp = &(*gp)->next;
 318                 }
 319         }
 320
 321         /*
 322          * Align numbers with text.
 323          * Count columns to equalize and columns to maximize.
 324          * Find maximum width of the columns to equalize.
 325          * Find total width of the columns *not* to maximize.
 326          */
 327
 328         necol = nxcol = 0;
 329         ewidth = xwidth = 0;
 330         for (icol = 0; icol <= maxcol; icol++) {
 331                 col = tbl->cols + icol;
 332                 if (col->width > col->nwidth)
 333                         col->decimal += (col->width - col->nwidth) / 2;
 334                 if (col->flags & TBL_CELL_EQUAL) {
 335                         necol++;
 336                         if (ewidth < col->width)
 337                                 ewidth = col->width;
 338                 }
 339                 if (col->flags & TBL_CELL_WMAX)
 340                         nxcol++;
 341                 else
 342                         xwidth += col->width;
 343         }
 344
 345         /*
 346          * Equalize columns, if requested for any of them.
 347          * Update total width of the columns not to maximize.
 348          */
 349
 350         if (necol) {
 351                 for (icol = 0; icol <= maxcol; icol++) {
 352                         col = tbl->cols + icol;
 353                         if ( ! (col->flags & TBL_CELL_EQUAL))
 354                                 continue;
 355                         if (col->width == ewidth)
 356                                 continue;
 357                         if (nxcol && rmargin)
 358                                 xwidth += ewidth - col->width;
 359                         col->width = ewidth;
 360                 }
 361         }
 362
 363         /*
 364          * If there are any columns to maximize, find the total
 365          * available width, deducting 3n margins between columns.
 366          * Distribute the available width evenly.
 367          */
 368
 369         if (nxcol && rmargin) {
 370                 xwidth += 3*maxcol +
 371                     (opts->opts & (TBL_OPT_BOX | TBL_OPT_DBOX) ?
 372                      2 : !!opts->lvert + !!opts->rvert);
 373                 if (rmargin <= offset + xwidth)
 374                         return;
 375                 xwidth = rmargin - offset - xwidth;
 376
 377                 /*
 378                  * Emulate a bug in GNU tbl width calculation that
 379                  * manifests itself for large numbers of x-columns.
 380                  * Emulating it for 5 x-columns gives identical
 381                  * behaviour for up to 6 x-columns.
 382                  */
 383
 384                 if (nxcol == 5) {
 385                         quirkcol = xwidth % nxcol + 2;
 386                         if (quirkcol != 3 && quirkcol != 4)
 387                                 quirkcol = -1;
 388                 } else
 389                         quirkcol = -1;
 390
 391                 necol = 0;
 392                 ewidth = 0;
 393                 for (icol = 0; icol <= maxcol; icol++) {
 394                         col = tbl->cols + icol;
 395                         if ( ! (col->flags & TBL_CELL_WMAX))
 396                                 continue;
 397                         col->width = (double)xwidth * ++necol / nxcol
 398                             - ewidth + 0.4995;
 399                         if (necol == quirkcol)
 400                                 col->width--;
 401                         ewidth += col->width;
 402                 }
 403         }
 404 }
 405
 406 static size_t
 407 tblcalc_data(struct rofftbl *tbl, struct roffcol *col,
 408     const struct tbl_opts *opts, const struct tbl_dat *dp, size_t mw)
 409 {
 410         size_t           sz;
 411
 412         /* Branch down into data sub-types. */
 413
 414         switch (dp->layout->pos) {
 415         case TBL_CELL_HORIZ:
 416         case TBL_CELL_DHORIZ:
 417                 sz = (*tbl->len)(1, tbl->arg);
 418                 if (col != NULL && col->width < sz)
 419                         col->width = sz;
 420                 return sz;
 421         case TBL_CELL_LONG:
 422         case TBL_CELL_CENTRE:
 423         case TBL_CELL_LEFT:
 424         case TBL_CELL_RIGHT:
 425                 return tblcalc_literal(tbl, col, dp, mw);
 426         case TBL_CELL_NUMBER:
 427                 return tblcalc_number(tbl, col, opts, dp);
 428         case TBL_CELL_DOWN:
 429                 return 0;
 430         default:
 431                 abort();
 432         }
 433 }
 434
 435 static size_t
 436 tblcalc_literal(struct rofftbl *tbl, struct roffcol *col,
 437     const struct tbl_dat *dp, size_t mw)
 438 {
 439         const char      *str;   /* Beginning of the first line. */
 440         const char      *beg;   /* Beginning of the current line. */
 441         char            *end;   /* End of the current line. */
 442         size_t           lsz;   /* Length of the current line. */
 443         size_t           wsz;   /* Length of the current word. */
 444         size_t           msz;   /* Length of the longest line. */
 445
 446         if (dp->string == NULL || *dp->string == '\0')
 447                 return 0;
 448         str = mw ? mandoc_strdup(dp->string) : dp->string;
 449         msz = lsz = 0;
 450         for (beg = str; beg != NULL && *beg != '\0'; beg = end) {
 451                 end = mw ? strchr(beg, ' ') : NULL;
 452                 if (end != NULL) {
 453                         *end++ = '\0';
 454                         while (*end == ' ')
 455                                 end++;
 456                 }
 457                 wsz = (*tbl->slen)(beg, tbl->arg);
 458                 if (mw && lsz && lsz + 1 + wsz <= mw)
 459                         lsz += 1 + wsz;
 460                 else
 461                         lsz = wsz;
 462                 if (msz < lsz)
 463                         msz = lsz;
 464         }
 465         if (mw)
 466                 free((void *)str);
 467         if (col != NULL && col->width < msz)
 468                 col->width = msz;
 469         return msz;
 470 }
 471
 472 static size_t
 473 tblcalc_number(struct rofftbl *tbl, struct roffcol *col,
 474                 const struct tbl_opts *opts, const struct tbl_dat *dp)
 475 {
 476         const char      *cp, *lastdigit, *lastpoint;
 477         size_t           intsz, totsz;
 478         char             buf[2];
 479
 480         if (dp->string == NULL || *dp->string == '\0')
 481                 return 0;
 482
 483         totsz = (*tbl->slen)(dp->string, tbl->arg);
 484         if (col == NULL)
 485                 return totsz;
 486
 487         /*
 488          * Find the last digit and
 489          * the last decimal point that is adjacent to a digit.
 490          * The alignment indicator "\&" overrides everything.
 491          */
 492
 493         lastdigit = lastpoint = NULL;
 494         for (cp = dp->string; cp[0] != '\0'; cp++) {
 495                 if (cp[0] == '\\' && cp[1] == '&') {
 496                         lastdigit = lastpoint = cp;
 497                         break;
 498                 } else if (cp[0] == opts->decimal &&
 499                     (isdigit((unsigned char)cp[1]) ||
 500                      (cp > dp->string && isdigit((unsigned char)cp[-1]))))
 501                         lastpoint = cp;
 502                 else if (isdigit((unsigned char)cp[0]))
 503                         lastdigit = cp;
 504         }
 505
 506         /* Not a number, treat as a literal string. */
 507
 508         if (lastdigit == NULL) {
 509                 if (col != NULL && col->width < totsz)
 510                         col->width = totsz;
 511                 return totsz;
 512         }
 513
 514         /* Measure the width of the integer part. */
 515
 516         if (lastpoint == NULL)
 517                 lastpoint = lastdigit + 1;
 518         intsz = 0;
 519         buf[1] = '\0';
 520         for (cp = dp->string; cp < lastpoint; cp++) {
 521                 buf[0] = cp[0];
 522                 intsz += (*tbl->slen)(buf, tbl->arg);
 523         }
 524
 525         /*
 526          * If this number has more integer digits than all numbers
 527          * seen on earlier lines, shift them all to the right.
 528          * If it has fewer, shift this number to the right.
 529          */
 530
 531         if (intsz > col->decimal) {
 532                 col->nwidth += intsz - col->decimal;
 533                 col->decimal = intsz;
 534         } else
 535                 totsz += col->decimal - intsz;
 536
 537         /* Update the maximum total width seen so far. */
 538
 539         if (totsz > col->nwidth)
 540                 col->nwidth = totsz;
 541         if (col->nwidth > col->width)
 542                 col->width = col->nwidth;
 543         return totsz;
 544 }