term.c

   1 /*      $Id: term.c,v 1.171 2010/09/15 14:36:16 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010 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 #ifdef HAVE_CONFIG_H
  19 #include "config.h"
  20 #endif
  21
  22 #include <sys/types.h>
  23
  24 #include <assert.h>
  25 #include <ctype.h>
  26 #include <stdint.h>
  27 #include <stdio.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30
  31 #include "mandoc.h"
  32 #include "chars.h"
  33 #include "out.h"
  34 #include "term.h"
  35 #include "main.h"
  36
  37 static  void              spec(struct termp *, enum roffdeco,
  38                                 const char *, size_t);
  39 static  void              res(struct termp *, const char *, size_t);
  40 static  void              bufferc(struct termp *, char);
  41 static  void              adjbuf(struct termp *p, size_t);
  42 static  void              encode(struct termp *, const char *, size_t);
  43
  44
  45 void
  46 term_free(struct termp *p)
  47 {
  48
  49         if (p->buf)
  50                 free(p->buf);
  51         if (p->symtab)
  52                 chars_free(p->symtab);
  53
  54         free(p);
  55 }
  56
  57
  58 void
  59 term_begin(struct termp *p, term_margin head,
  60                 term_margin foot, const void *arg)
  61 {
  62
  63         p->headf = head;
  64         p->footf = foot;
  65         p->argf = arg;
  66         (*p->begin)(p);
  67 }
  68
  69
  70 void
  71 term_end(struct termp *p)
  72 {
  73
  74         (*p->end)(p);
  75 }
  76
  77
  78 struct termp *
  79 term_alloc(enum termenc enc)
  80 {
  81         struct termp    *p;
  82
  83         p = calloc(1, sizeof(struct termp));
  84         if (NULL == p) {
  85                 perror(NULL);
  86                 exit((int)MANDOCLEVEL_SYSERR);
  87         }
  88
  89         p->enc = enc;
  90         return(p);
  91 }
  92
  93
  94 /*
  95  * Flush a line of text.  A "line" is loosely defined as being something
  96  * that should be followed by a newline, regardless of whether it's
  97  * broken apart by newlines getting there.  A line can also be a
  98  * fragment of a columnar list (`Bl -tag' or `Bl -column'), which does
  99  * not have a trailing newline.
 100  *
 101  * The following flags may be specified:
 102  *
 103  *  - TERMP_NOLPAD: when beginning to write the line, don't left-pad the
 104  *    offset value.  This is useful when doing columnar lists where the
 105  *    prior column has right-padded.
 106  *
 107  *  - TERMP_NOBREAK: this is the most important and is used when making
 108  *    columns.  In short: don't print a newline and instead pad to the
 109  *    right margin.  Used in conjunction with TERMP_NOLPAD.
 110  *
 111  *  - TERMP_TWOSPACE: when padding, make sure there are at least two
 112  *    space characters of padding.  Otherwise, rather break the line.
 113  *
 114  *  - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and
 115  *    the line is overrun, and don't pad-right if it's underrun.
 116  *
 117  *  - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when
 118  *    overruning, instead save the position and continue at that point
 119  *    when the next invocation.
 120  *
 121  *  In-line line breaking:
 122  *
 123  *  If TERMP_NOBREAK is specified and the line overruns the right
 124  *  margin, it will break and pad-right to the right margin after
 125  *  writing.  If maxrmargin is violated, it will break and continue
 126  *  writing from the right-margin, which will lead to the above scenario
 127  *  upon exit.  Otherwise, the line will break at the right margin.
 128  */
 129 void
 130 term_flushln(struct termp *p)
 131 {
 132         int              i;     /* current input position in p->buf */
 133         size_t           vis;   /* current visual position on output */
 134         size_t           vbl;   /* number of blanks to prepend to output */
 135         size_t           vend;  /* end of word visual position on output */
 136         size_t           bp;    /* visual right border position */
 137         int              j;     /* temporary loop index for p->buf */
 138         int              jhy;   /* last hyph before overflow w/r/t j */
 139         size_t           maxvis; /* output position of visible boundary */
 140         size_t           mmax; /* used in calculating bp */
 141
 142         /*
 143          * First, establish the maximum columns of "visible" content.
 144          * This is usually the difference between the right-margin and
 145          * an indentation, but can be, for tagged lists or columns, a
 146          * small set of values.
 147          */
 148
 149         assert(p->offset < p->rmargin);
 150
 151         maxvis = (int)(p->rmargin - p->offset) - p->overstep < 0 ?
 152                 /* LINTED */
 153                 0 : p->rmargin - p->offset - p->overstep;
 154         mmax = (int)(p->maxrmargin - p->offset) - p->overstep < 0 ?
 155                 /* LINTED */
 156                 0 : p->maxrmargin - p->offset - p->overstep;
 157
 158         bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
 159
 160         /*
 161          * Indent the first line of a paragraph.
 162          */
 163         vbl = p->flags & TERMP_NOLPAD ? 0 : p->offset;
 164
 165         vis = vend = i = 0;
 166
 167         while (i < (int)p->col) {
 168                 /*
 169                  * Handle literal tab characters: collapse all
 170                  * subsequent tabs into a single huge set of spaces.
 171                  */
 172                 while (i < (int)p->col && '\t' == p->buf[i]) {
 173                         vend = (vis / p->tabwidth + 1) * p->tabwidth;
 174                         vbl += vend - vis;
 175                         vis = vend;
 176                         i++;
 177                 }
 178
 179                 /*
 180                  * Count up visible word characters.  Control sequences
 181                  * (starting with the CSI) aren't counted.  A space
 182                  * generates a non-printing word, which is valid (the
 183                  * space is printed according to regular spacing rules).
 184                  */
 185
 186                 /* LINTED */
 187                 for (j = i, jhy = 0; j < (int)p->col; j++) {
 188                         if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j])
 189                                 break;
 190
 191                         /* Back over the the last printed character. */
 192                         if (8 == p->buf[j]) {
 193                                 assert(j);
 194                                 vend -= (*p->width)(p, p->buf[j - 1]);
 195                                 continue;
 196                         }
 197
 198                         /* Regular word. */
 199                         /* Break at the hyphen point if we overrun. */
 200                         if (vend > vis && vend < bp &&
 201                                         ASCII_HYPH == p->buf[j])
 202                                 jhy = j;
 203
 204                         vend += (*p->width)(p, p->buf[j]);
 205                 }
 206
 207                 /*
 208                  * Find out whether we would exceed the right margin.
 209                  * If so, break to the next line.
 210                  */
 211                 if (vend > bp && 0 == jhy && vis > 0) {
 212                         vend -= vis;
 213                         (*p->endline)(p);
 214                         if (TERMP_NOBREAK & p->flags) {
 215                                 p->viscol = p->rmargin;
 216                                 (*p->advance)(p, p->rmargin);
 217                                 vend += p->rmargin - p->offset;
 218                         } else {
 219                                 p->viscol = 0;
 220                                 vbl = p->offset;
 221                         }
 222
 223                         /* Remove the p->overstep width. */
 224
 225                         bp += (int)/* LINTED */
 226                                 p->overstep;
 227                         p->overstep = 0;
 228                 }
 229
 230                 /* Write out the [remaining] word. */
 231                 for ( ; i < (int)p->col; i++) {
 232                         if (vend > bp && jhy > 0 && i > jhy)
 233                                 break;
 234                         if ('\t' == p->buf[i])
 235                                 break;
 236                         if (' ' == p->buf[i]) {
 237                                 j = i;
 238                                 while (' ' == p->buf[i])
 239                                         i++;
 240                                 vbl += (i - j) * (*p->width)(p, ' ');
 241                                 break;
 242                         }
 243                         if (ASCII_NBRSP == p->buf[i]) {
 244                                 vbl += (*p->width)(p, ' ');
 245                                 continue;
 246                         }
 247
 248                         /*
 249                          * Now we definitely know there will be
 250                          * printable characters to output,
 251                          * so write preceding white space now.
 252                          */
 253                         if (vbl) {
 254                                 (*p->advance)(p, vbl);
 255                                 p->viscol += vbl;
 256                                 vbl = 0;
 257                         }
 258
 259                         if (ASCII_HYPH == p->buf[i]) {
 260                                 (*p->letter)(p, '-');
 261                                 p->viscol += (*p->width)(p, '-');
 262                         } else {
 263                                 (*p->letter)(p, p->buf[i]);
 264                                 p->viscol += (*p->width)(p, p->buf[i]);
 265                         }
 266                 }
 267                 vend += vbl;
 268                 vis = vend;
 269         }
 270
 271         /*
 272          * If there was trailing white space, it was not printed;
 273          * so reset the cursor position accordingly.
 274          */
 275         vis -= vbl;
 276
 277         p->col = 0;
 278         p->overstep = 0;
 279
 280         if ( ! (TERMP_NOBREAK & p->flags)) {
 281                 p->viscol = 0;
 282                 (*p->endline)(p);
 283                 return;
 284         }
 285
 286         if (TERMP_HANG & p->flags) {
 287                 /* We need one blank after the tag. */
 288                 p->overstep = /* LINTED */
 289                         vis - maxvis + (*p->width)(p, ' ');
 290
 291                 /*
 292                  * Behave exactly the same way as groff:
 293                  * If we have overstepped the margin, temporarily move
 294                  * it to the right and flag the rest of the line to be
 295                  * shorter.
 296                  * If we landed right at the margin, be happy.
 297                  * If we are one step before the margin, temporarily
 298                  * move it one step LEFT and flag the rest of the line
 299                  * to be longer.
 300                  */
 301                 if (p->overstep >= -1) {
 302                         assert((int)maxvis + p->overstep >= 0);
 303                         /* LINTED */
 304                         maxvis += p->overstep;
 305                 } else
 306                         p->overstep = 0;
 307
 308         } else if (TERMP_DANGLE & p->flags)
 309                 return;
 310
 311         /* Right-pad. */
 312         if (maxvis > vis + /* LINTED */
 313                         ((TERMP_TWOSPACE & p->flags) ?
 314                          (*p->width)(p, ' ') : 0)) {
 315                 p->viscol += maxvis - vis;
 316                 (*p->advance)(p, maxvis - vis);
 317                 vis += (maxvis - vis);
 318         } else {        /* ...or newline break. */
 319                 (*p->endline)(p);
 320                 p->viscol = p->rmargin;
 321                 (*p->advance)(p, p->rmargin);
 322         }
 323 }
 324
 325
 326 /*
 327  * A newline only breaks an existing line; it won't assert vertical
 328  * space.  All data in the output buffer is flushed prior to the newline
 329  * assertion.
 330  */
 331 void
 332 term_newln(struct termp *p)
 333 {
 334
 335         p->flags |= TERMP_NOSPACE;
 336         if (0 == p->col && 0 == p->viscol) {
 337                 p->flags &= ~TERMP_NOLPAD;
 338                 return;
 339         }
 340         term_flushln(p);
 341         p->flags &= ~TERMP_NOLPAD;
 342 }
 343
 344
 345 /*
 346  * Asserts a vertical space (a full, empty line-break between lines).
 347  * Note that if used twice, this will cause two blank spaces and so on.
 348  * All data in the output buffer is flushed prior to the newline
 349  * assertion.
 350  */
 351 void
 352 term_vspace(struct termp *p)
 353 {
 354
 355         term_newln(p);
 356         p->viscol = 0;
 357         (*p->endline)(p);
 358 }
 359
 360
 361 static void
 362 spec(struct termp *p, enum roffdeco d, const char *word, size_t len)
 363 {
 364         const char      *rhs;
 365         size_t           sz;
 366
 367         rhs = chars_spec2str(p->symtab, word, len, &sz);
 368         if (rhs)
 369                 encode(p, rhs, sz);
 370         else if (DECO_SSPECIAL == d)
 371                 encode(p, word, len);
 372 }
 373
 374
 375 static void
 376 res(struct termp *p, const char *word, size_t len)
 377 {
 378         const char      *rhs;
 379         size_t           sz;
 380
 381         rhs = chars_res2str(p->symtab, word, len, &sz);
 382         if (rhs)
 383                 encode(p, rhs, sz);
 384 }
 385
 386
 387 void
 388 term_fontlast(struct termp *p)
 389 {
 390         enum termfont    f;
 391
 392         f = p->fontl;
 393         p->fontl = p->fontq[p->fonti];
 394         p->fontq[p->fonti] = f;
 395 }
 396
 397
 398 void
 399 term_fontrepl(struct termp *p, enum termfont f)
 400 {
 401
 402         p->fontl = p->fontq[p->fonti];
 403         p->fontq[p->fonti] = f;
 404 }
 405
 406
 407 void
 408 term_fontpush(struct termp *p, enum termfont f)
 409 {
 410
 411         assert(p->fonti + 1 < 10);
 412         p->fontl = p->fontq[p->fonti];
 413         p->fontq[++p->fonti] = f;
 414 }
 415
 416
 417 const void *
 418 term_fontq(struct termp *p)
 419 {
 420
 421         return(&p->fontq[p->fonti]);
 422 }
 423
 424
 425 enum termfont
 426 term_fonttop(struct termp *p)
 427 {
 428
 429         return(p->fontq[p->fonti]);
 430 }
 431
 432
 433 void
 434 term_fontpopq(struct termp *p, const void *key)
 435 {
 436
 437         while (p->fonti >= 0 && key != &p->fontq[p->fonti])
 438                 p->fonti--;
 439         assert(p->fonti >= 0);
 440 }
 441
 442
 443 void
 444 term_fontpop(struct termp *p)
 445 {
 446
 447         assert(p->fonti);
 448         p->fonti--;
 449 }
 450
 451
 452 /*
 453  * Handle pwords, partial words, which may be either a single word or a
 454  * phrase that cannot be broken down (such as a literal string).  This
 455  * handles word styling.
 456  */
 457 void
 458 term_word(struct termp *p, const char *word)
 459 {
 460         const char      *sv, *seq;
 461         size_t           ssz;
 462         enum roffdeco    deco;
 463
 464         sv = word;
 465
 466         if (word[0] && '\0' == word[1])
 467                 switch (word[0]) {
 468                 case('.'):
 469                         /* FALLTHROUGH */
 470                 case(','):
 471                         /* FALLTHROUGH */
 472                 case(';'):
 473                         /* FALLTHROUGH */
 474                 case(':'):
 475                         /* FALLTHROUGH */
 476                 case('?'):
 477                         /* FALLTHROUGH */
 478                 case('!'):
 479                         /* FALLTHROUGH */
 480                 case(')'):
 481                         /* FALLTHROUGH */
 482                 case(']'):
 483                         if ( ! (TERMP_IGNDELIM & p->flags))
 484                                 p->flags |= TERMP_NOSPACE;
 485                         break;
 486                 default:
 487                         break;
 488                 }
 489
 490         if ( ! (TERMP_NOSPACE & p->flags)) {
 491                 if ( ! (TERMP_KEEP & p->flags)) {
 492                         if (TERMP_PREKEEP & p->flags)
 493                                 p->flags |= TERMP_KEEP;
 494                         bufferc(p, ' ');
 495                         if (TERMP_SENTENCE & p->flags)
 496                                 bufferc(p, ' ');
 497                 } else
 498                         bufferc(p, ASCII_NBRSP);
 499         }
 500
 501         if ( ! (p->flags & TERMP_NONOSPACE))
 502                 p->flags &= ~TERMP_NOSPACE;
 503         else
 504                 p->flags |= TERMP_NOSPACE;
 505
 506         p->flags &= ~TERMP_SENTENCE;
 507
 508         while (*word) {
 509                 if ((ssz = strcspn(word, "\\")) > 0)
 510                         encode(p, word, ssz);
 511
 512                 word += ssz;
 513                 if ('\\' != *word)
 514                         continue;
 515
 516                 seq = ++word;
 517                 word += a2roffdeco(&deco, &seq, &ssz);
 518
 519                 switch (deco) {
 520                 case (DECO_RESERVED):
 521                         res(p, seq, ssz);
 522                         break;
 523                 case (DECO_SPECIAL):
 524                         /* FALLTHROUGH */
 525                 case (DECO_SSPECIAL):
 526                         spec(p, deco, seq, ssz);
 527                         break;
 528                 case (DECO_BOLD):
 529                         term_fontrepl(p, TERMFONT_BOLD);
 530                         break;
 531                 case (DECO_ITALIC):
 532                         term_fontrepl(p, TERMFONT_UNDER);
 533                         break;
 534                 case (DECO_ROMAN):
 535                         term_fontrepl(p, TERMFONT_NONE);
 536                         break;
 537                 case (DECO_PREVIOUS):
 538                         term_fontlast(p);
 539                         break;
 540                 default:
 541                         break;
 542                 }
 543
 544                 if (DECO_NOSPACE == deco && '\0' == *word)
 545                         p->flags |= TERMP_NOSPACE;
 546         }
 547
 548         /*
 549          * Note that we don't process the pipe: the parser sees it as
 550          * punctuation, but we don't in terms of typography.
 551          */
 552         if (sv[0] && '\0' == sv[1])
 553                 switch (sv[0]) {
 554                 case('('):
 555                         /* FALLTHROUGH */
 556                 case('['):
 557                         p->flags |= TERMP_NOSPACE;
 558                         break;
 559                 default:
 560                         break;
 561                 }
 562 }
 563
 564
 565 static void
 566 adjbuf(struct termp *p, size_t sz)
 567 {
 568
 569         if (0 == p->maxcols)
 570                 p->maxcols = 1024;
 571         while (sz >= p->maxcols)
 572                 p->maxcols <<= 2;
 573
 574         p->buf = realloc(p->buf, p->maxcols);
 575         if (NULL == p->buf) {
 576                 perror(NULL);
 577                 exit((int)MANDOCLEVEL_SYSERR);
 578         }
 579 }
 580
 581
 582 static void
 583 bufferc(struct termp *p, char c)
 584 {
 585
 586         if (p->col + 1 >= p->maxcols)
 587                 adjbuf(p, p->col + 1);
 588
 589         p->buf[(int)p->col++] = c;
 590 }
 591
 592
 593 static void
 594 encode(struct termp *p, const char *word, size_t sz)
 595 {
 596         enum termfont     f;
 597         int               i;
 598
 599         /*
 600          * Encode and buffer a string of characters.  If the current
 601          * font mode is unset, buffer directly, else encode then buffer
 602          * character by character.
 603          */
 604
 605         if (TERMFONT_NONE == (f = term_fonttop(p))) {
 606                 if (p->col + sz >= p->maxcols)
 607                         adjbuf(p, p->col + sz);
 608                 memcpy(&p->buf[(int)p->col], word, sz);
 609                 p->col += sz;
 610                 return;
 611         }
 612
 613         /* Pre-buffer, assuming worst-case. */
 614
 615         if (p->col + 1 + (sz * 3) >= p->maxcols)
 616                 adjbuf(p, p->col + 1 + (sz * 3));
 617
 618         for (i = 0; i < (int)sz; i++) {
 619                 if ( ! isgraph((u_char)word[i])) {
 620                         p->buf[(int)p->col++] = word[i];
 621                         continue;
 622                 }
 623
 624                 if (TERMFONT_UNDER == f)
 625                         p->buf[(int)p->col++] = '_';
 626                 else
 627                         p->buf[(int)p->col++] = word[i];
 628
 629                 p->buf[(int)p->col++] = 8;
 630                 p->buf[(int)p->col++] = word[i];
 631         }
 632 }
 633
 634
 635 size_t
 636 term_len(const struct termp *p, size_t sz)
 637 {
 638
 639         return((*p->width)(p, ' ') * sz);
 640 }
 641
 642
 643 size_t
 644 term_strlen(const struct termp *p, const char *cp)
 645 {
 646         size_t           sz, ssz, rsz, i;
 647         enum roffdeco    d;
 648         const char      *seq, *rhs;
 649
 650         for (sz = 0; '\0' != *cp; )
 651                 /*
 652                  * Account for escaped sequences within string length
 653                  * calculations.  This follows the logic in term_word()
 654                  * as we must calculate the width of produced strings.
 655                  */
 656                 if ('\\' == *cp) {
 657                         seq = ++cp;
 658                         cp += a2roffdeco(&d, &seq, &ssz);
 659
 660                         switch (d) {
 661                         case (DECO_RESERVED):
 662                                 rhs = chars_res2str
 663                                         (p->symtab, seq, ssz, &rsz);
 664                                 break;
 665                         case (DECO_SPECIAL):
 666                                 /* FALLTHROUGH */
 667                         case (DECO_SSPECIAL):
 668                                 rhs = chars_spec2str
 669                                         (p->symtab, seq, ssz, &rsz);
 670
 671                                 /* Allow for one-char escapes. */
 672                                 if (DECO_SSPECIAL != d || rhs)
 673                                         break;
 674
 675                                 rhs = seq;
 676                                 rsz = ssz;
 677                                 break;
 678                         default:
 679                                 rhs = NULL;
 680                                 break;
 681                         }
 682
 683                         if (rhs)
 684                                 for (i = 0; i < rsz; i++)
 685                                         sz += (*p->width)(p, *rhs++);
 686                 } else
 687                         sz += (*p->width)(p, *cp++);
 688
 689         return(sz);
 690 }
 691
 692
 693 /* ARGSUSED */
 694 size_t
 695 term_vspan(const struct termp *p, const struct roffsu *su)
 696 {
 697         double           r;
 698
 699         switch (su->unit) {
 700         case (SCALE_CM):
 701                 r = su->scale * 2;
 702                 break;
 703         case (SCALE_IN):
 704                 r = su->scale * 6;
 705                 break;
 706         case (SCALE_PC):
 707                 r = su->scale;
 708                 break;
 709         case (SCALE_PT):
 710                 r = su->scale / 8;
 711                 break;
 712         case (SCALE_MM):
 713                 r = su->scale / 1000;
 714                 break;
 715         case (SCALE_VS):
 716                 r = su->scale;
 717                 break;
 718         default:
 719                 r = su->scale - 1;
 720                 break;
 721         }
 722
 723         if (r < 0.0)
 724                 r = 0.0;
 725         return(/* LINTED */(size_t)
 726                         r);
 727 }
 728
 729
 730 size_t
 731 term_hspan(const struct termp *p, const struct roffsu *su)
 732 {
 733         double           v;
 734
 735         v = ((*p->hspan)(p, su));
 736         if (v < 0.0)
 737                 v = 0.0;
 738         return((size_t) /* LINTED */
 739                         v);
 740 }