term.c

   1 /*      $Id: term.c,v 1.188 2011/05/14 18:15:20 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2011 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 "out.h"
  33 #include "term.h"
  34 #include "main.h"
  35
  36 static  void              spec(struct termp *, const char *, size_t);
  37 static  void              res(struct termp *, const char *, size_t);
  38 static  void              bufferc(struct termp *, char);
  39 static  void              adjbuf(struct termp *p, int);
  40 static  void              encode(struct termp *, const char *, size_t);
  41
  42
  43 void
  44 term_free(struct termp *p)
  45 {
  46
  47         if (p->buf)
  48                 free(p->buf);
  49         if (p->symtab)
  50                 mchars_free(p->symtab);
  51
  52         free(p);
  53 }
  54
  55
  56 void
  57 term_begin(struct termp *p, term_margin head,
  58                 term_margin foot, const void *arg)
  59 {
  60
  61         p->headf = head;
  62         p->footf = foot;
  63         p->argf = arg;
  64         (*p->begin)(p);
  65 }
  66
  67
  68 void
  69 term_end(struct termp *p)
  70 {
  71
  72         (*p->end)(p);
  73 }
  74
  75
  76 struct termp *
  77 term_alloc(enum termenc enc)
  78 {
  79         struct termp    *p;
  80
  81         p = mandoc_calloc(1, sizeof(struct termp));
  82         p->enc = enc;
  83         return(p);
  84 }
  85
  86
  87 /*
  88  * Flush a line of text.  A "line" is loosely defined as being something
  89  * that should be followed by a newline, regardless of whether it's
  90  * broken apart by newlines getting there.  A line can also be a
  91  * fragment of a columnar list (`Bl -tag' or `Bl -column'), which does
  92  * not have a trailing newline.
  93  *
  94  * The following flags may be specified:
  95  *
  96  *  - TERMP_NOLPAD: when beginning to write the line, don't left-pad the
  97  *    offset value.  This is useful when doing columnar lists where the
  98  *    prior column has right-padded.
  99  *
 100  *  - TERMP_NOBREAK: this is the most important and is used when making
 101  *    columns.  In short: don't print a newline and instead pad to the
 102  *    right margin.  Used in conjunction with TERMP_NOLPAD.
 103  *
 104  *  - TERMP_TWOSPACE: when padding, make sure there are at least two
 105  *    space characters of padding.  Otherwise, rather break the line.
 106  *
 107  *  - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and
 108  *    the line is overrun, and don't pad-right if it's underrun.
 109  *
 110  *  - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when
 111  *    overruning, instead save the position and continue at that point
 112  *    when the next invocation.
 113  *
 114  *  In-line line breaking:
 115  *
 116  *  If TERMP_NOBREAK is specified and the line overruns the right
 117  *  margin, it will break and pad-right to the right margin after
 118  *  writing.  If maxrmargin is violated, it will break and continue
 119  *  writing from the right-margin, which will lead to the above scenario
 120  *  upon exit.  Otherwise, the line will break at the right margin.
 121  */
 122 void
 123 term_flushln(struct termp *p)
 124 {
 125         int              i;     /* current input position in p->buf */
 126         size_t           vis;   /* current visual position on output */
 127         size_t           vbl;   /* number of blanks to prepend to output */
 128         size_t           vend;  /* end of word visual position on output */
 129         size_t           bp;    /* visual right border position */
 130         size_t           dv;    /* temporary for visual pos calculations */
 131         int              j;     /* temporary loop index for p->buf */
 132         int              jhy;   /* last hyph before overflow w/r/t j */
 133         size_t           maxvis; /* output position of visible boundary */
 134         size_t           mmax; /* used in calculating bp */
 135
 136         /*
 137          * First, establish the maximum columns of "visible" content.
 138          * This is usually the difference between the right-margin and
 139          * an indentation, but can be, for tagged lists or columns, a
 140          * small set of values.
 141          */
 142         assert  (p->rmargin >= p->offset);
 143         dv     = p->rmargin - p->offset;
 144         maxvis = (int)dv > p->overstep ? dv - (size_t)p->overstep : 0;
 145         dv     = p->maxrmargin - p->offset;
 146         mmax   = (int)dv > p->overstep ? dv - (size_t)p->overstep : 0;
 147
 148         bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
 149
 150         /*
 151          * Indent the first line of a paragraph.
 152          */
 153         vbl = p->flags & TERMP_NOLPAD ? (size_t)0 : p->offset;
 154
 155         vis = vend = 0;
 156         i = 0;
 157
 158         while (i < p->col) {
 159                 /*
 160                  * Handle literal tab characters: collapse all
 161                  * subsequent tabs into a single huge set of spaces.
 162                  */
 163                 while (i < p->col && '\t' == p->buf[i]) {
 164                         vend = (vis / p->tabwidth + 1) * p->tabwidth;
 165                         vbl += vend - vis;
 166                         vis = vend;
 167                         i++;
 168                 }
 169
 170                 /*
 171                  * Count up visible word characters.  Control sequences
 172                  * (starting with the CSI) aren't counted.  A space
 173                  * generates a non-printing word, which is valid (the
 174                  * space is printed according to regular spacing rules).
 175                  */
 176
 177                 for (j = i, jhy = 0; j < p->col; j++) {
 178                         if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j])
 179                                 break;
 180
 181                         /* Back over the the last printed character. */
 182                         if (8 == p->buf[j]) {
 183                                 assert(j);
 184                                 vend -= (*p->width)(p, p->buf[j - 1]);
 185                                 continue;
 186                         }
 187
 188                         /* Regular word. */
 189                         /* Break at the hyphen point if we overrun. */
 190                         if (vend > vis && vend < bp &&
 191                                         ASCII_HYPH == p->buf[j])
 192                                 jhy = j;
 193
 194                         vend += (*p->width)(p, p->buf[j]);
 195                 }
 196
 197                 /*
 198                  * Find out whether we would exceed the right margin.
 199                  * If so, break to the next line.
 200                  */
 201                 if (vend > bp && 0 == jhy && vis > 0) {
 202                         vend -= vis;
 203                         (*p->endline)(p);
 204                         if (TERMP_NOBREAK & p->flags) {
 205                                 p->viscol = p->rmargin;
 206                                 (*p->advance)(p, p->rmargin);
 207                                 vend += p->rmargin - p->offset;
 208                         } else {
 209                                 p->viscol = 0;
 210                                 vbl = p->offset;
 211                         }
 212
 213                         /* Remove the p->overstep width. */
 214
 215                         bp += (size_t)p->overstep;
 216                         p->overstep = 0;
 217                 }
 218
 219                 /* Write out the [remaining] word. */
 220                 for ( ; i < p->col; i++) {
 221                         if (vend > bp && jhy > 0 && i > jhy)
 222                                 break;
 223                         if ('\t' == p->buf[i])
 224                                 break;
 225                         if (' ' == p->buf[i]) {
 226                                 j = i;
 227                                 while (' ' == p->buf[i])
 228                                         i++;
 229                                 dv = (size_t)(i - j) * (*p->width)(p, ' ');
 230                                 vbl += dv;
 231                                 vend += dv;
 232                                 break;
 233                         }
 234                         if (ASCII_NBRSP == p->buf[i]) {
 235                                 vbl += (*p->width)(p, ' ');
 236                                 continue;
 237                         }
 238
 239                         /*
 240                          * Now we definitely know there will be
 241                          * printable characters to output,
 242                          * so write preceding white space now.
 243                          */
 244                         if (vbl) {
 245                                 (*p->advance)(p, vbl);
 246                                 p->viscol += vbl;
 247                                 vbl = 0;
 248                         }
 249
 250                         if (ASCII_HYPH == p->buf[i]) {
 251                                 (*p->letter)(p, '-');
 252                                 p->viscol += (*p->width)(p, '-');
 253                         } else {
 254                                 (*p->letter)(p, p->buf[i]);
 255                                 p->viscol += (*p->width)(p, p->buf[i]);
 256                         }
 257                 }
 258                 vis = vend;
 259         }
 260
 261         /*
 262          * If there was trailing white space, it was not printed;
 263          * so reset the cursor position accordingly.
 264          */
 265         vis -= vbl;
 266
 267         p->col = 0;
 268         p->overstep = 0;
 269
 270         if ( ! (TERMP_NOBREAK & p->flags)) {
 271                 p->viscol = 0;
 272                 (*p->endline)(p);
 273                 return;
 274         }
 275
 276         if (TERMP_HANG & p->flags) {
 277                 /* We need one blank after the tag. */
 278                 p->overstep = (int)(vis - maxvis + (*p->width)(p, ' '));
 279
 280                 /*
 281                  * Behave exactly the same way as groff:
 282                  * If we have overstepped the margin, temporarily move
 283                  * it to the right and flag the rest of the line to be
 284                  * shorter.
 285                  * If we landed right at the margin, be happy.
 286                  * If we are one step before the margin, temporarily
 287                  * move it one step LEFT and flag the rest of the line
 288                  * to be longer.
 289                  */
 290                 if (p->overstep >= -1) {
 291                         assert((int)maxvis + p->overstep >= 0);
 292                         maxvis += (size_t)p->overstep;
 293                 } else
 294                         p->overstep = 0;
 295
 296         } else if (TERMP_DANGLE & p->flags)
 297                 return;
 298
 299         /* Right-pad. */
 300         if (maxvis > vis +
 301             ((TERMP_TWOSPACE & p->flags) ? (*p->width)(p, ' ') : 0)) {
 302                 p->viscol += maxvis - vis;
 303                 (*p->advance)(p, maxvis - vis);
 304                 vis += (maxvis - vis);
 305         } else {        /* ...or newline break. */
 306                 (*p->endline)(p);
 307                 p->viscol = p->rmargin;
 308                 (*p->advance)(p, p->rmargin);
 309         }
 310 }
 311
 312
 313 /*
 314  * A newline only breaks an existing line; it won't assert vertical
 315  * space.  All data in the output buffer is flushed prior to the newline
 316  * assertion.
 317  */
 318 void
 319 term_newln(struct termp *p)
 320 {
 321
 322         p->flags |= TERMP_NOSPACE;
 323         if (0 == p->col && 0 == p->viscol) {
 324                 p->flags &= ~TERMP_NOLPAD;
 325                 return;
 326         }
 327         term_flushln(p);
 328         p->flags &= ~TERMP_NOLPAD;
 329 }
 330
 331
 332 /*
 333  * Asserts a vertical space (a full, empty line-break between lines).
 334  * Note that if used twice, this will cause two blank spaces and so on.
 335  * All data in the output buffer is flushed prior to the newline
 336  * assertion.
 337  */
 338 void
 339 term_vspace(struct termp *p)
 340 {
 341
 342         term_newln(p);
 343         p->viscol = 0;
 344         (*p->endline)(p);
 345 }
 346
 347
 348 static void
 349 numbered(struct termp *p, const char *word, size_t len)
 350 {
 351         char             c;
 352
 353         if ('\0' != (c = mchars_num2char(word, len)))
 354                 encode(p, &c, 1);
 355 }
 356
 357
 358 static void
 359 spec(struct termp *p, const char *word, size_t len)
 360 {
 361         const char      *rhs;
 362         size_t           sz;
 363
 364         rhs = mchars_spec2str(p->symtab, word, len, &sz);
 365         if (rhs)
 366                 encode(p, rhs, sz);
 367         else if (1 == len)
 368                 encode(p, word, len);
 369 }
 370
 371
 372 static void
 373 res(struct termp *p, const char *word, size_t len)
 374 {
 375         const char      *rhs;
 376         size_t           sz;
 377
 378         rhs = mchars_res2str(p->symtab, word, len, &sz);
 379         if (rhs)
 380                 encode(p, rhs, sz);
 381 }
 382
 383
 384 void
 385 term_fontlast(struct termp *p)
 386 {
 387         enum termfont    f;
 388
 389         f = p->fontl;
 390         p->fontl = p->fontq[p->fonti];
 391         p->fontq[p->fonti] = f;
 392 }
 393
 394
 395 void
 396 term_fontrepl(struct termp *p, enum termfont f)
 397 {
 398
 399         p->fontl = p->fontq[p->fonti];
 400         p->fontq[p->fonti] = f;
 401 }
 402
 403
 404 void
 405 term_fontpush(struct termp *p, enum termfont f)
 406 {
 407
 408         assert(p->fonti + 1 < 10);
 409         p->fontl = p->fontq[p->fonti];
 410         p->fontq[++p->fonti] = f;
 411 }
 412
 413
 414 const void *
 415 term_fontq(struct termp *p)
 416 {
 417
 418         return(&p->fontq[p->fonti]);
 419 }
 420
 421
 422 enum termfont
 423 term_fonttop(struct termp *p)
 424 {
 425
 426         return(p->fontq[p->fonti]);
 427 }
 428
 429
 430 void
 431 term_fontpopq(struct termp *p, const void *key)
 432 {
 433
 434         while (p->fonti >= 0 && key != &p->fontq[p->fonti])
 435                 p->fonti--;
 436         assert(p->fonti >= 0);
 437 }
 438
 439
 440 void
 441 term_fontpop(struct termp *p)
 442 {
 443
 444         assert(p->fonti);
 445         p->fonti--;
 446 }
 447
 448
 449 /*
 450  * Handle pwords, partial words, which may be either a single word or a
 451  * phrase that cannot be broken down (such as a literal string).  This
 452  * handles word styling.
 453  */
 454 void
 455 term_word(struct termp *p, const char *word)
 456 {
 457         const char      *seq;
 458         int              sz;
 459         size_t           ssz;
 460         enum mandoc_esc  esc;
 461
 462         if ( ! (TERMP_NOSPACE & p->flags)) {
 463                 if ( ! (TERMP_KEEP & p->flags)) {
 464                         if (TERMP_PREKEEP & p->flags)
 465                                 p->flags |= TERMP_KEEP;
 466                         bufferc(p, ' ');
 467                         if (TERMP_SENTENCE & p->flags)
 468                                 bufferc(p, ' ');
 469                 } else
 470                         bufferc(p, ASCII_NBRSP);
 471         }
 472
 473         if ( ! (p->flags & TERMP_NONOSPACE))
 474                 p->flags &= ~TERMP_NOSPACE;
 475         else
 476                 p->flags |= TERMP_NOSPACE;
 477
 478         p->flags &= ~(TERMP_SENTENCE | TERMP_IGNDELIM);
 479
 480         while ('\0' != *word) {
 481                 if ((ssz = strcspn(word, "\\")) > 0)
 482                         encode(p, word, ssz);
 483
 484                 word += (int)ssz;
 485                 if ('\\' != *word)
 486                         continue;
 487
 488                 word++;
 489                 esc = mandoc_escape(&word, &seq, &sz);
 490                 if (ESCAPE_ERROR == esc)
 491                         break;
 492
 493                 switch (esc) {
 494                 case (ESCAPE_NUMBERED):
 495                         numbered(p, seq, sz);
 496                         break;
 497                 case (ESCAPE_PREDEF):
 498                         res(p, seq, sz);
 499                         break;
 500                 case (ESCAPE_SPECIAL):
 501                         spec(p, seq, sz);
 502                         break;
 503                 case (ESCAPE_FONTBOLD):
 504                         term_fontrepl(p, TERMFONT_BOLD);
 505                         break;
 506                 case (ESCAPE_FONTITALIC):
 507                         term_fontrepl(p, TERMFONT_UNDER);
 508                         break;
 509                 case (ESCAPE_FONTROMAN):
 510                         term_fontrepl(p, TERMFONT_NONE);
 511                         break;
 512                 case (ESCAPE_FONTPREV):
 513                         term_fontlast(p);
 514                         break;
 515                 case (ESCAPE_NOSPACE):
 516                         if ('\0' == *word)
 517                                 p->flags |= TERMP_NOSPACE;
 518                         break;
 519                 default:
 520                         break;
 521                 }
 522         }
 523 }
 524
 525
 526 static void
 527 adjbuf(struct termp *p, int sz)
 528 {
 529
 530         if (0 == p->maxcols)
 531                 p->maxcols = 1024;
 532         while (sz >= p->maxcols)
 533                 p->maxcols <<= 2;
 534
 535         p->buf = mandoc_realloc
 536                 (p->buf, sizeof(int) * (size_t)p->maxcols);
 537 }
 538
 539
 540 static void
 541 bufferc(struct termp *p, char c)
 542 {
 543
 544         if (p->col + 1 >= p->maxcols)
 545                 adjbuf(p, p->col + 1);
 546
 547         p->buf[p->col++] = c;
 548 }
 549
 550
 551 static void
 552 encode(struct termp *p, const char *word, size_t sz)
 553 {
 554         enum termfont     f;
 555         int               i, len;
 556
 557         /* LINTED */
 558         len = sz;
 559
 560         /*
 561          * Encode and buffer a string of characters.  If the current
 562          * font mode is unset, buffer directly, else encode then buffer
 563          * character by character.
 564          */
 565
 566         if (TERMFONT_NONE == (f = term_fonttop(p))) {
 567                 if (p->col + len >= p->maxcols)
 568                         adjbuf(p, p->col + len);
 569                 for (i = 0; i < len; i++)
 570                         p->buf[p->col++] = word[i];
 571                 return;
 572         }
 573
 574         /* Pre-buffer, assuming worst-case. */
 575
 576         if (p->col + 1 + (len * 3) >= p->maxcols)
 577                 adjbuf(p, p->col + 1 + (len * 3));
 578
 579         for (i = 0; i < len; i++) {
 580                 if ( ! isgraph((unsigned char)word[i])) {
 581                         p->buf[p->col++] = word[i];
 582                         continue;
 583                 }
 584
 585                 if (TERMFONT_UNDER == f)
 586                         p->buf[p->col++] = '_';
 587                 else
 588                         p->buf[p->col++] = word[i];
 589
 590                 p->buf[p->col++] = 8;
 591                 p->buf[p->col++] = word[i];
 592         }
 593 }
 594
 595
 596 size_t
 597 term_len(const struct termp *p, size_t sz)
 598 {
 599
 600         return((*p->width)(p, ' ') * sz);
 601 }
 602
 603
 604 size_t
 605 term_strlen(const struct termp *p, const char *cp)
 606 {
 607         size_t           sz, rsz, i;
 608         int              ssz;
 609         enum mandoc_esc  esc;
 610         const char      *seq, *rhs;
 611
 612         /*
 613          * Account for escaped sequences within string length
 614          * calculations.  This follows the logic in term_word() as we
 615          * must calculate the width of produced strings.
 616          */
 617
 618         sz = 0;
 619         while ('\0' != *cp)
 620                 switch (*cp) {
 621                 case ('\\'):
 622                         ++cp;
 623                         esc = mandoc_escape(&cp, &seq, &ssz);
 624                         if (ESCAPE_ERROR == esc)
 625                                 return(sz);
 626
 627                         switch (esc) {
 628                         case (ESCAPE_PREDEF):
 629                                 rhs = mchars_res2str
 630                                         (p->symtab, seq, ssz, &rsz);
 631                                 break;
 632                         case (ESCAPE_SPECIAL):
 633                                 rhs = mchars_spec2str
 634                                         (p->symtab, seq, ssz, &rsz);
 635
 636                                 if (ssz != 1 || rhs)
 637                                         break;
 638
 639                                 rhs = seq;
 640                                 rsz = ssz;
 641                                 break;
 642                         default:
 643                                 rhs = NULL;
 644                                 break;
 645                         }
 646
 647                         if (NULL == rhs)
 648                                 break;
 649
 650                         for (i = 0; i < rsz; i++)
 651                                 sz += (*p->width)(p, *rhs++);
 652                         break;
 653                 case (ASCII_NBRSP):
 654                         sz += (*p->width)(p, ' ');
 655                         cp++;
 656                         break;
 657                 case (ASCII_HYPH):
 658                         sz += (*p->width)(p, '-');
 659                         cp++;
 660                         break;
 661                 default:
 662                         sz += (*p->width)(p, *cp++);
 663                         break;
 664                 }
 665
 666         return(sz);
 667 }
 668
 669
 670 /* ARGSUSED */
 671 size_t
 672 term_vspan(const struct termp *p, const struct roffsu *su)
 673 {
 674         double           r;
 675
 676         switch (su->unit) {
 677         case (SCALE_CM):
 678                 r = su->scale * 2;
 679                 break;
 680         case (SCALE_IN):
 681                 r = su->scale * 6;
 682                 break;
 683         case (SCALE_PC):
 684                 r = su->scale;
 685                 break;
 686         case (SCALE_PT):
 687                 r = su->scale / 8;
 688                 break;
 689         case (SCALE_MM):
 690                 r = su->scale / 1000;
 691                 break;
 692         case (SCALE_VS):
 693                 r = su->scale;
 694                 break;
 695         default:
 696                 r = su->scale - 1;
 697                 break;
 698         }
 699
 700         if (r < 0.0)
 701                 r = 0.0;
 702         return(/* LINTED */(size_t)
 703                         r);
 704 }
 705
 706
 707 size_t
 708 term_hspan(const struct termp *p, const struct roffsu *su)
 709 {
 710         double           v;
 711
 712         v = ((*p->hspan)(p, su));
 713         if (v < 0.0)
 714                 v = 0.0;
 715         return((size_t) /* LINTED */
 716                         v);
 717 }