term.c

   1 /*      $Id: term.c,v 1.128 2010/01/01 17:14:30 kristaps Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #ifdef HAVE_CONFIG_H
  18 #include "config.h"
  19 #endif
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <ctype.h>
  25 #include <stdio.h>
  26 #include <stdlib.h>
  27 #include <string.h>
  28 #include <time.h>
  29
  30 #include "chars.h"
  31 #include "out.h"
  32 #include "term.h"
  33 #include "man.h"
  34 #include "mdoc.h"
  35 #include "main.h"
  36
  37 static  struct termp     *term_alloc(enum termenc);
  38 static  void              term_free(struct termp *);
  39 static  void              spec(struct termp *, const char *, size_t);
  40 static  void              res(struct termp *, const char *, size_t);
  41 static  void              buffera(struct termp *, const char *, size_t);
  42 static  void              bufferc(struct termp *, char);
  43 static  void              adjbuf(struct termp *p, size_t);
  44 static  void              encode(struct termp *, const char *, size_t);
  45
  46
  47 void *
  48 ascii_alloc(void)
  49 {
  50
  51         return(term_alloc(TERMENC_ASCII));
  52 }
  53
  54
  55 void
  56 terminal_free(void *arg)
  57 {
  58
  59         term_free((struct termp *)arg);
  60 }
  61
  62
  63 static void
  64 term_free(struct termp *p)
  65 {
  66
  67         if (p->buf)
  68                 free(p->buf);
  69         if (p->symtab)
  70                 chars_free(p->symtab);
  71
  72         free(p);
  73 }
  74
  75
  76 static struct termp *
  77 term_alloc(enum termenc enc)
  78 {
  79         struct termp *p;
  80
  81         p = calloc(1, sizeof(struct termp));
  82         if (NULL == p) {
  83                 perror(NULL);
  84                 exit(EXIT_FAILURE);
  85         }
  86         p->maxrmargin = 78;
  87         p->enc = enc;
  88         return(p);
  89 }
  90
  91
  92 /*
  93  * Flush a line of text.  A "line" is loosely defined as being something
  94  * that should be followed by a newline, regardless of whether it's
  95  * broken apart by newlines getting there.  A line can also be a
  96  * fragment of a columnar list.
  97  *
  98  * Specifically, a line is whatever's in p->buf of length p->col, which
  99  * is zeroed after this function returns.
 100  *
 101  * The usage of termp:flags is as follows:
 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           vsz;   /* visual characters to write to output */
 136         size_t           bp;    /* visual right border position */
 137         int              j;     /* temporary loop index */
 138         size_t           maxvis, mmax;
 139         static int       overstep = 0;
 140
 141         /*
 142          * First, establish the maximum columns of "visible" content.
 143          * This is usually the difference between the right-margin and
 144          * an indentation, but can be, for tagged lists or columns, a
 145          * small set of values.
 146          */
 147
 148         assert(p->offset < p->rmargin);
 149
 150         maxvis = (int)(p->rmargin - p->offset) - overstep < 0 ?
 151                 /* LINTED */
 152                 0 : p->rmargin - p->offset - overstep;
 153         mmax = (int)(p->maxrmargin - p->offset) - overstep < 0 ?
 154                 /* LINTED */
 155                 0 : p->maxrmargin - p->offset - overstep;
 156
 157         bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
 158
 159         /*
 160          * FIXME: if bp is zero, we still output the first word before
 161          * breaking the line.
 162          */
 163
 164         vis = 0;
 165
 166         /*
 167          * If in the standard case (left-justified), then begin with our
 168          * indentation, otherwise (columns, etc.) just start spitting
 169          * out text.
 170          */
 171
 172         if ( ! (p->flags & TERMP_NOLPAD))
 173                 /* LINTED */
 174                 for (j = 0; j < (int)p->offset; j++)
 175                         putchar(' ');
 176
 177         for (i = 0; i < (int)p->col; i++) {
 178                 /*
 179                  * Count up visible word characters.  Control sequences
 180                  * (starting with the CSI) aren't counted.  A space
 181                  * generates a non-printing word, which is valid (the
 182                  * space is printed according to regular spacing rules).
 183                  */
 184
 185                 /* LINTED */
 186                 for (j = i, vsz = 0; j < (int)p->col; j++) {
 187                         if (j && ' ' == p->buf[j])
 188                                 break;
 189                         else if (8 == p->buf[j])
 190                                 vsz--;
 191                         else
 192                                 vsz++;
 193                 }
 194
 195                 /*
 196                  * Choose the number of blanks to prepend: no blank at the
 197                  * beginning of a line, one between words -- but do not
 198                  * actually write them yet.
 199                  */
 200                 vbl = (size_t)(0 == vis ? 0 : 1);
 201
 202                 /*
 203                  * Find out whether we would exceed the right margin.
 204                  * If so, break to the next line.  (TODO: hyphenate)
 205                  * Otherwise, write the chosen number of blanks now.
 206                  */
 207                 if (vis && vis + vbl + vsz > bp) {
 208                         putchar('\n');
 209                         if (TERMP_NOBREAK & p->flags) {
 210                                 for (j = 0; j < (int)p->rmargin; j++)
 211                                         putchar(' ');
 212                                 vis = p->rmargin - p->offset;
 213                         } else {
 214                                 for (j = 0; j < (int)p->offset; j++)
 215                                         putchar(' ');
 216                                 vis = 0;
 217                         }
 218                         /* Remove the overstep width. */
 219                         bp += (int)/* LINTED */
 220                                 overstep;
 221                         overstep = 0;
 222                 } else {
 223                         for (j = 0; j < (int)vbl; j++)
 224                                 putchar(' ');
 225                         vis += vbl;
 226                 }
 227
 228                 /*
 229                  * Finally, write out the word.
 230                  */
 231                 for ( ; i < (int)p->col; i++) {
 232                         if (' ' == p->buf[i])
 233                                 break;
 234
 235                         /* The unit sep. is a non-breaking space. */
 236                         if (31 == p->buf[i])
 237                                 putchar(' ');
 238                         else
 239                                 putchar(p->buf[i]);
 240                 }
 241                 vis += vsz;
 242         }
 243
 244         p->col = 0;
 245         overstep = 0;
 246
 247         if ( ! (TERMP_NOBREAK & p->flags)) {
 248                 putchar('\n');
 249                 return;
 250         }
 251
 252         if (TERMP_HANG & p->flags) {
 253                 /* We need one blank after the tag. */
 254                 overstep = /* LINTED */
 255                         vis - maxvis + 1;
 256
 257                 /*
 258                  * Behave exactly the same way as groff:
 259                  * If we have overstepped the margin, temporarily move
 260                  * it to the right and flag the rest of the line to be
 261                  * shorter.
 262                  * If we landed right at the margin, be happy.
 263                  * If we are one step before the margin, temporarily
 264                  * move it one step LEFT and flag the rest of the line
 265                  * to be longer.
 266                  */
 267                 if (overstep >= -1) {
 268                         assert((int)maxvis + overstep >= 0);
 269                         /* LINTED */
 270                         maxvis += overstep;
 271                 } else
 272                         overstep = 0;
 273
 274         } else if (TERMP_DANGLE & p->flags)
 275                 return;
 276
 277         /* Right-pad. */
 278         if (maxvis > vis + /* LINTED */
 279                         ((TERMP_TWOSPACE & p->flags) ? 1 : 0))
 280                 for ( ; vis < maxvis; vis++)
 281                         putchar(' ');
 282         else {  /* ...or newline break. */
 283                 putchar('\n');
 284                 for (i = 0; i < (int)p->rmargin; i++)
 285                         putchar(' ');
 286         }
 287 }
 288
 289
 290 /*
 291  * A newline only breaks an existing line; it won't assert vertical
 292  * space.  All data in the output buffer is flushed prior to the newline
 293  * assertion.
 294  */
 295 void
 296 term_newln(struct termp *p)
 297 {
 298
 299         p->flags |= TERMP_NOSPACE;
 300         if (0 == p->col) {
 301                 p->flags &= ~TERMP_NOLPAD;
 302                 return;
 303         }
 304         term_flushln(p);
 305         p->flags &= ~TERMP_NOLPAD;
 306 }
 307
 308
 309 /*
 310  * Asserts a vertical space (a full, empty line-break between lines).
 311  * Note that if used twice, this will cause two blank spaces and so on.
 312  * All data in the output buffer is flushed prior to the newline
 313  * assertion.
 314  */
 315 void
 316 term_vspace(struct termp *p)
 317 {
 318
 319         term_newln(p);
 320         putchar('\n');
 321 }
 322
 323
 324 static void
 325 spec(struct termp *p, const char *word, size_t len)
 326 {
 327         const char      *rhs;
 328         size_t           sz;
 329
 330         rhs = chars_a2ascii(p->symtab, word, len, &sz);
 331         if (rhs)
 332                 encode(p, rhs, sz);
 333 }
 334
 335
 336 static void
 337 res(struct termp *p, const char *word, size_t len)
 338 {
 339         const char      *rhs;
 340         size_t           sz;
 341
 342         rhs = chars_a2res(p->symtab, word, len, &sz);
 343         if (rhs)
 344                 encode(p, rhs, sz);
 345 }
 346
 347
 348 void
 349 term_fontlast(struct termp *p)
 350 {
 351         enum termfont    f;
 352
 353         f = p->fontl;
 354         p->fontl = p->fontq[p->fonti];
 355         p->fontq[p->fonti] = f;
 356 }
 357
 358
 359 void
 360 term_fontrepl(struct termp *p, enum termfont f)
 361 {
 362
 363         p->fontl = p->fontq[p->fonti];
 364         p->fontq[p->fonti] = f;
 365 }
 366
 367
 368 void
 369 term_fontpush(struct termp *p, enum termfont f)
 370 {
 371
 372         assert(p->fonti + 1 < 10);
 373         p->fontl = p->fontq[p->fonti];
 374         p->fontq[++p->fonti] = f;
 375 }
 376
 377
 378 const void *
 379 term_fontq(struct termp *p)
 380 {
 381
 382         return(&p->fontq[p->fonti]);
 383 }
 384
 385
 386 enum termfont
 387 term_fonttop(struct termp *p)
 388 {
 389
 390         return(p->fontq[p->fonti]);
 391 }
 392
 393
 394 void
 395 term_fontpopq(struct termp *p, const void *key)
 396 {
 397
 398         while (p->fonti >= 0 && key != &p->fontq[p->fonti])
 399                 p->fonti--;
 400         assert(p->fonti >= 0);
 401 }
 402
 403
 404 void
 405 term_fontpop(struct termp *p)
 406 {
 407
 408         assert(p->fonti);
 409         p->fonti--;
 410 }
 411
 412
 413 /*
 414  * Handle pwords, partial words, which may be either a single word or a
 415  * phrase that cannot be broken down (such as a literal string).  This
 416  * handles word styling.
 417  */
 418 void
 419 term_word(struct termp *p, const char *word)
 420 {
 421         const char      *sv, *seq;
 422         int              sz;
 423         size_t           ssz;
 424         enum roffdeco    deco;
 425
 426         sv = word;
 427
 428         if (word[0] && '\0' == word[1])
 429                 switch (word[0]) {
 430                 case('.'):
 431                         /* FALLTHROUGH */
 432                 case(','):
 433                         /* FALLTHROUGH */
 434                 case(';'):
 435                         /* FALLTHROUGH */
 436                 case(':'):
 437                         /* FALLTHROUGH */
 438                 case('?'):
 439                         /* FALLTHROUGH */
 440                 case('!'):
 441                         /* FALLTHROUGH */
 442                 case(')'):
 443                         /* FALLTHROUGH */
 444                 case(']'):
 445                         /* FALLTHROUGH */
 446                 case('}'):
 447                         if ( ! (TERMP_IGNDELIM & p->flags))
 448                                 p->flags |= TERMP_NOSPACE;
 449                         break;
 450                 default:
 451                         break;
 452                 }
 453
 454         if ( ! (TERMP_NOSPACE & p->flags))
 455                 bufferc(p, ' ');
 456
 457         if ( ! (p->flags & TERMP_NONOSPACE))
 458                 p->flags &= ~TERMP_NOSPACE;
 459
 460         /* FIXME: use strcspn. */
 461
 462         while (*word) {
 463                 if ('\\' != *word) {
 464                         encode(p, word, 1);
 465                         word++;
 466                         continue;
 467                 }
 468
 469                 seq = ++word;
 470                 sz = a2roffdeco(&deco, &seq, &ssz);
 471
 472                 switch (deco) {
 473                 case (DECO_RESERVED):
 474                         res(p, seq, ssz);
 475                         break;
 476                 case (DECO_SPECIAL):
 477                         spec(p, seq, ssz);
 478                         break;
 479                 case (DECO_BOLD):
 480                         term_fontrepl(p, TERMFONT_BOLD);
 481                         break;
 482                 case (DECO_ITALIC):
 483                         term_fontrepl(p, TERMFONT_UNDER);
 484                         break;
 485                 case (DECO_ROMAN):
 486                         term_fontrepl(p, TERMFONT_NONE);
 487                         break;
 488                 case (DECO_PREVIOUS):
 489                         term_fontlast(p);
 490                         break;
 491                 default:
 492                         break;
 493                 }
 494
 495                 word += sz;
 496                 if (DECO_NOSPACE == deco && '\0' == *word)
 497                         p->flags |= TERMP_NOSPACE;
 498         }
 499
 500         if (sv[0] && 0 == sv[1])
 501                 switch (sv[0]) {
 502                 case('('):
 503                         /* FALLTHROUGH */
 504                 case('['):
 505                         /* FALLTHROUGH */
 506                 case('{'):
 507                         p->flags |= TERMP_NOSPACE;
 508                         break;
 509                 default:
 510                         break;
 511                 }
 512 }
 513
 514
 515 static void
 516 adjbuf(struct termp *p, size_t sz)
 517 {
 518
 519         if (0 == p->maxcols)
 520                 p->maxcols = 1024;
 521         while (sz >= p->maxcols)
 522                 p->maxcols <<= 2;
 523
 524         p->buf = realloc(p->buf, p->maxcols);
 525         if (NULL == p->buf) {
 526                 perror(NULL);
 527                 exit(EXIT_FAILURE);
 528         }
 529 }
 530
 531
 532 static void
 533 buffera(struct termp *p, const char *word, size_t sz)
 534 {
 535
 536         if (p->col + sz >= p->maxcols)
 537                 adjbuf(p, p->col + sz);
 538
 539         memcpy(&p->buf[(int)p->col], word, sz);
 540         p->col += sz;
 541 }
 542
 543
 544 static void
 545 bufferc(struct termp *p, char c)
 546 {
 547
 548         if (p->col + 1 >= p->maxcols)
 549                 adjbuf(p, p->col + 1);
 550
 551         p->buf[(int)p->col++] = c;
 552 }
 553
 554
 555 static void
 556 encode(struct termp *p, const char *word, size_t sz)
 557 {
 558         enum termfont     f;
 559         int               i;
 560
 561         /*
 562          * Encode and buffer a string of characters.  If the current
 563          * font mode is unset, buffer directly, else encode then buffer
 564          * character by character.
 565          */
 566
 567         if (TERMFONT_NONE == (f = term_fonttop(p))) {
 568                 buffera(p, word, sz);
 569                 return;
 570         }
 571
 572         for (i = 0; i < (int)sz; i++) {
 573                 if ( ! isgraph((u_char)word[i])) {
 574                         bufferc(p, word[i]);
 575                         continue;
 576                 }
 577
 578                 if (TERMFONT_UNDER == f)
 579                         bufferc(p, '_');
 580                 else
 581                         bufferc(p, word[i]);
 582
 583                 bufferc(p, 8);
 584                 bufferc(p, word[i]);
 585         }
 586 }
 587
 588
 589 size_t
 590 term_vspan(const struct roffsu *su)
 591 {
 592         double           r;
 593
 594         switch (su->unit) {
 595         case (SCALE_CM):
 596                 r = su->scale * 2;
 597                 break;
 598         case (SCALE_IN):
 599                 r = su->scale * 6;
 600                 break;
 601         case (SCALE_PC):
 602                 r = su->scale;
 603                 break;
 604         case (SCALE_PT):
 605                 r = su->scale / 8;
 606                 break;
 607         case (SCALE_MM):
 608                 r = su->scale / 1000;
 609                 break;
 610         case (SCALE_VS):
 611                 r = su->scale;
 612                 break;
 613         default:
 614                 r = su->scale - 1;
 615                 break;
 616         }
 617
 618         if (r < 0.0)
 619                 r = 0.0;
 620         return(/* LINTED */(size_t)
 621                         r);
 622 }
 623
 624
 625 size_t
 626 term_hspan(const struct roffsu *su)
 627 {
 628         double           r;
 629
 630         /* XXX: CM, IN, and PT are approximations. */
 631
 632         switch (su->unit) {
 633         case (SCALE_CM):
 634                 r = 4 * su->scale;
 635                 break;
 636         case (SCALE_IN):
 637                 /* XXX: this is an approximation. */
 638                 r = 10 * su->scale;
 639                 break;
 640         case (SCALE_PC):
 641                 r = (10 * su->scale) / 6;
 642                 break;
 643         case (SCALE_PT):
 644                 r = (10 * su->scale) / 72;
 645                 break;
 646         case (SCALE_MM):
 647                 r = su->scale / 1000; /* FIXME: double-check. */
 648                 break;
 649         case (SCALE_VS):
 650                 r = su->scale * 2 - 1; /* FIXME: double-check. */
 651                 break;
 652         default:
 653                 r = su->scale;
 654                 break;
 655         }
 656
 657         if (r < 0.0)
 658                 r = 0.0;
 659         return((size_t)/* LINTED */
 660                         r);
 661 }
 662
 663