term.c

   1 /*      $Id: term.c,v 1.125 2009/11/12 05:50:12 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 #include <assert.h>
  18 #include <ctype.h>
  19 #include <stdio.h>
  20 #include <stdlib.h>
  21 #include <string.h>
  22 #include <time.h>
  23
  24 #include "chars.h"
  25 #include "out.h"
  26 #include "term.h"
  27 #include "man.h"
  28 #include "mdoc.h"
  29 #include "main.h"
  30
  31 static  struct termp     *term_alloc(enum termenc);
  32 static  void              term_free(struct termp *);
  33 static  void              spec(struct termp *, const char *, size_t);
  34 static  void              res(struct termp *, const char *, size_t);
  35 static  void              buffera(struct termp *, const char *, size_t);
  36 static  void              bufferc(struct termp *, char);
  37 static  void              adjbuf(struct termp *p, size_t);
  38 static  void              encode(struct termp *, const char *, size_t);
  39
  40
  41 void *
  42 ascii_alloc(void)
  43 {
  44
  45         return(term_alloc(TERMENC_ASCII));
  46 }
  47
  48
  49 void
  50 terminal_free(void *arg)
  51 {
  52
  53         term_free((struct termp *)arg);
  54 }
  55
  56
  57 static void
  58 term_free(struct termp *p)
  59 {
  60
  61         if (p->buf)
  62                 free(p->buf);
  63         if (p->symtab)
  64                 chars_free(p->symtab);
  65
  66         free(p);
  67 }
  68
  69
  70 static struct termp *
  71 term_alloc(enum termenc enc)
  72 {
  73         struct termp *p;
  74
  75         p = calloc(1, sizeof(struct termp));
  76         if (NULL == p) {
  77                 perror(NULL);
  78                 exit(EXIT_FAILURE);
  79         }
  80         p->maxrmargin = 78;
  81         p->enc = enc;
  82         return(p);
  83 }
  84
  85
  86 /*
  87  * Flush a line of text.  A "line" is loosely defined as being something
  88  * that should be followed by a newline, regardless of whether it's
  89  * broken apart by newlines getting there.  A line can also be a
  90  * fragment of a columnar list.
  91  *
  92  * Specifically, a line is whatever's in p->buf of length p->col, which
  93  * is zeroed after this function returns.
  94  *
  95  * The usage of termp:flags is as follows:
  96  *
  97  *  - TERMP_NOLPAD: when beginning to write the line, don't left-pad the
  98  *    offset value.  This is useful when doing columnar lists where the
  99  *    prior column has right-padded.
 100  *
 101  *  - TERMP_NOBREAK: this is the most important and is used when making
 102  *    columns.  In short: don't print a newline and instead pad to the
 103  *    right margin.  Used in conjunction with TERMP_NOLPAD.
 104  *
 105  *  - TERMP_TWOSPACE: when padding, make sure there are at least two
 106  *    space characters of padding.  Otherwise, rather break the line.
 107  *
 108  *  - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and
 109  *    the line is overrun, and don't pad-right if it's underrun.
 110  *
 111  *  - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when
 112  *    overruning, instead save the position and continue at that point
 113  *    when the next invocation.
 114  *
 115  *  In-line line breaking:
 116  *
 117  *  If TERMP_NOBREAK is specified and the line overruns the right
 118  *  margin, it will break and pad-right to the right margin after
 119  *  writing.  If maxrmargin is violated, it will break and continue
 120  *  writing from the right-margin, which will lead to the above scenario
 121  *  upon exit.  Otherwise, the line will break at the right margin.
 122  */
 123 void
 124 term_flushln(struct termp *p)
 125 {
 126         int              i;     /* current input position in p->buf */
 127         size_t           vis;   /* current visual position on output */
 128         size_t           vbl;   /* number of blanks to prepend to output */
 129         size_t           vsz;   /* visual characters to write to output */
 130         size_t           bp;    /* visual right border position */
 131         int              j;     /* temporary loop index */
 132         size_t           maxvis, mmax;
 133         static int       overstep = 0;
 134
 135         /*
 136          * First, establish the maximum columns of "visible" content.
 137          * This is usually the difference between the right-margin and
 138          * an indentation, but can be, for tagged lists or columns, a
 139          * small set of values.
 140          */
 141
 142         assert(p->offset < p->rmargin);
 143
 144         maxvis = (int)(p->rmargin - p->offset) - overstep < 0 ?
 145                 /* LINTED */
 146                 0 : p->rmargin - p->offset - overstep;
 147         mmax = (int)(p->maxrmargin - p->offset) - overstep < 0 ?
 148                 /* LINTED */
 149                 0 : p->maxrmargin - p->offset - overstep;
 150
 151         bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
 152
 153         /*
 154          * FIXME: if bp is zero, we still output the first word before
 155          * breaking the line.
 156          */
 157
 158         vis = 0;
 159
 160         /*
 161          * If in the standard case (left-justified), then begin with our
 162          * indentation, otherwise (columns, etc.) just start spitting
 163          * out text.
 164          */
 165
 166         if ( ! (p->flags & TERMP_NOLPAD))
 167                 /* LINTED */
 168                 for (j = 0; j < (int)p->offset; j++)
 169                         putchar(' ');
 170
 171         for (i = 0; i < (int)p->col; i++) {
 172                 /*
 173                  * Count up visible word characters.  Control sequences
 174                  * (starting with the CSI) aren't counted.  A space
 175                  * generates a non-printing word, which is valid (the
 176                  * space is printed according to regular spacing rules).
 177                  */
 178
 179                 /* LINTED */
 180                 for (j = i, vsz = 0; j < (int)p->col; j++) {
 181                         if (j && ' ' == p->buf[j])
 182                                 break;
 183                         else if (8 == p->buf[j])
 184                                 vsz--;
 185                         else
 186                                 vsz++;
 187                 }
 188
 189                 /*
 190                  * Choose the number of blanks to prepend: no blank at the
 191                  * beginning of a line, one between words -- but do not
 192                  * actually write them yet.
 193                  */
 194                 vbl = (size_t)(0 == vis ? 0 : 1);
 195
 196                 /*
 197                  * Find out whether we would exceed the right margin.
 198                  * If so, break to the next line.  (TODO: hyphenate)
 199                  * Otherwise, write the chosen number of blanks now.
 200                  */
 201                 if (vis && vis + vbl + vsz > bp) {
 202                         putchar('\n');
 203                         if (TERMP_NOBREAK & p->flags) {
 204                                 for (j = 0; j < (int)p->rmargin; j++)
 205                                         putchar(' ');
 206                                 vis = p->rmargin - p->offset;
 207                         } else {
 208                                 for (j = 0; j < (int)p->offset; j++)
 209                                         putchar(' ');
 210                                 vis = 0;
 211                         }
 212                         /* Remove the overstep width. */
 213                         bp += (int)/* LINTED */
 214                                 overstep;
 215                         overstep = 0;
 216                 } else {
 217                         for (j = 0; j < (int)vbl; j++)
 218                                 putchar(' ');
 219                         vis += vbl;
 220                 }
 221
 222                 /*
 223                  * Finally, write out the word.
 224                  */
 225                 for ( ; i < (int)p->col; i++) {
 226                         if (' ' == p->buf[i])
 227                                 break;
 228
 229                         /* The unit sep. is a non-breaking space. */
 230                         if (31 == p->buf[i])
 231                                 putchar(' ');
 232                         else
 233                                 putchar(p->buf[i]);
 234                 }
 235                 vis += vsz;
 236         }
 237
 238         p->col = 0;
 239         overstep = 0;
 240
 241         if ( ! (TERMP_NOBREAK & p->flags)) {
 242                 putchar('\n');
 243                 return;
 244         }
 245
 246         if (TERMP_HANG & p->flags) {
 247                 /* We need one blank after the tag. */
 248                 overstep = /* LINTED */
 249                         vis - maxvis + 1;
 250
 251                 /*
 252                  * Behave exactly the same way as groff:
 253                  * If we have overstepped the margin, temporarily move
 254                  * it to the right and flag the rest of the line to be
 255                  * shorter.
 256                  * If we landed right at the margin, be happy.
 257                  * If we are one step before the margin, temporarily
 258                  * move it one step LEFT and flag the rest of the line
 259                  * to be longer.
 260                  */
 261                 if (overstep >= -1) {
 262                         assert((int)maxvis + overstep >= 0);
 263                         /* LINTED */
 264                         maxvis += overstep;
 265                 } else
 266                         overstep = 0;
 267
 268         } else if (TERMP_DANGLE & p->flags)
 269                 return;
 270
 271         /* Right-pad. */
 272         if (maxvis > vis + /* LINTED */
 273                         ((TERMP_TWOSPACE & p->flags) ? 1 : 0))
 274                 for ( ; vis < maxvis; vis++)
 275                         putchar(' ');
 276         else {  /* ...or newline break. */
 277                 putchar('\n');
 278                 for (i = 0; i < (int)p->rmargin; i++)
 279                         putchar(' ');
 280         }
 281 }
 282
 283
 284 /*
 285  * A newline only breaks an existing line; it won't assert vertical
 286  * space.  All data in the output buffer is flushed prior to the newline
 287  * assertion.
 288  */
 289 void
 290 term_newln(struct termp *p)
 291 {
 292
 293         p->flags |= TERMP_NOSPACE;
 294         if (0 == p->col) {
 295                 p->flags &= ~TERMP_NOLPAD;
 296                 return;
 297         }
 298         term_flushln(p);
 299         p->flags &= ~TERMP_NOLPAD;
 300 }
 301
 302
 303 /*
 304  * Asserts a vertical space (a full, empty line-break between lines).
 305  * Note that if used twice, this will cause two blank spaces and so on.
 306  * All data in the output buffer is flushed prior to the newline
 307  * assertion.
 308  */
 309 void
 310 term_vspace(struct termp *p)
 311 {
 312
 313         term_newln(p);
 314         putchar('\n');
 315 }
 316
 317
 318 static void
 319 spec(struct termp *p, const char *word, size_t len)
 320 {
 321         const char      *rhs;
 322         size_t           sz;
 323
 324         rhs = chars_a2ascii(p->symtab, word, len, &sz);
 325         if (rhs)
 326                 encode(p, rhs, sz);
 327 }
 328
 329
 330 static void
 331 res(struct termp *p, const char *word, size_t len)
 332 {
 333         const char      *rhs;
 334         size_t           sz;
 335
 336         rhs = chars_a2res(p->symtab, word, len, &sz);
 337         if (rhs)
 338                 encode(p, rhs, sz);
 339 }
 340
 341
 342 void
 343 term_fontlast(struct termp *p)
 344 {
 345         enum termfont    f;
 346
 347         f = p->fontl;
 348         p->fontl = p->fontq[p->fonti];
 349         p->fontq[p->fonti] = f;
 350 }
 351
 352
 353 void
 354 term_fontrepl(struct termp *p, enum termfont f)
 355 {
 356
 357         p->fontl = p->fontq[p->fonti];
 358         p->fontq[p->fonti] = f;
 359 }
 360
 361
 362 void
 363 term_fontpush(struct termp *p, enum termfont f)
 364 {
 365
 366         assert(p->fonti + 1 < 10);
 367         p->fontl = p->fontq[p->fonti];
 368         p->fontq[++p->fonti] = f;
 369 }
 370
 371
 372 const void *
 373 term_fontq(struct termp *p)
 374 {
 375
 376         return(&p->fontq[p->fonti]);
 377 }
 378
 379
 380 enum termfont
 381 term_fonttop(struct termp *p)
 382 {
 383
 384         return(p->fontq[p->fonti]);
 385 }
 386
 387
 388 void
 389 term_fontpopq(struct termp *p, const void *key)
 390 {
 391
 392         while (p->fonti >= 0 && key != &p->fontq[p->fonti])
 393                 p->fonti--;
 394         assert(p->fonti >= 0);
 395 }
 396
 397
 398 void
 399 term_fontpop(struct termp *p)
 400 {
 401
 402         assert(p->fonti);
 403         p->fonti--;
 404 }
 405
 406
 407 /*
 408  * Handle pwords, partial words, which may be either a single word or a
 409  * phrase that cannot be broken down (such as a literal string).  This
 410  * handles word styling.
 411  */
 412 void
 413 term_word(struct termp *p, const char *word)
 414 {
 415         const char      *sv, *seq;
 416         int              sz;
 417         size_t           ssz;
 418         enum roffdeco    deco;
 419
 420         sv = word;
 421
 422         if (word[0] && '\0' == word[1])
 423                 switch (word[0]) {
 424                 case('.'):
 425                         /* FALLTHROUGH */
 426                 case(','):
 427                         /* FALLTHROUGH */
 428                 case(';'):
 429                         /* FALLTHROUGH */
 430                 case(':'):
 431                         /* FALLTHROUGH */
 432                 case('?'):
 433                         /* FALLTHROUGH */
 434                 case('!'):
 435                         /* FALLTHROUGH */
 436                 case(')'):
 437                         /* FALLTHROUGH */
 438                 case(']'):
 439                         /* FALLTHROUGH */
 440                 case('}'):
 441                         if ( ! (TERMP_IGNDELIM & p->flags))
 442                                 p->flags |= TERMP_NOSPACE;
 443                         break;
 444                 default:
 445                         break;
 446                 }
 447
 448         if ( ! (TERMP_NOSPACE & p->flags))
 449                 bufferc(p, ' ');
 450
 451         if ( ! (p->flags & TERMP_NONOSPACE))
 452                 p->flags &= ~TERMP_NOSPACE;
 453
 454         /* FIXME: use strcspn. */
 455
 456         while (*word) {
 457                 if ('\\' != *word) {
 458                         encode(p, word, 1);
 459                         word++;
 460                         continue;
 461                 }
 462
 463                 seq = ++word;
 464                 sz = a2roffdeco(&deco, &seq, &ssz);
 465
 466                 switch (deco) {
 467                 case (DECO_RESERVED):
 468                         res(p, seq, ssz);
 469                         break;
 470                 case (DECO_SPECIAL):
 471                         spec(p, seq, ssz);
 472                         break;
 473                 case (DECO_BOLD):
 474                         term_fontrepl(p, TERMFONT_BOLD);
 475                         break;
 476                 case (DECO_ITALIC):
 477                         term_fontrepl(p, TERMFONT_UNDER);
 478                         break;
 479                 case (DECO_ROMAN):
 480                         term_fontrepl(p, TERMFONT_NONE);
 481                         break;
 482                 case (DECO_PREVIOUS):
 483                         term_fontlast(p);
 484                         break;
 485                 default:
 486                         break;
 487                 }
 488                 word += sz;
 489         }
 490
 491         if (sv[0] && 0 == sv[1])
 492                 switch (sv[0]) {
 493                 case('('):
 494                         /* FALLTHROUGH */
 495                 case('['):
 496                         /* FALLTHROUGH */
 497                 case('{'):
 498                         p->flags |= TERMP_NOSPACE;
 499                         break;
 500                 default:
 501                         break;
 502                 }
 503 }
 504
 505
 506 static void
 507 adjbuf(struct termp *p, size_t sz)
 508 {
 509
 510         if (0 == p->maxcols)
 511                 p->maxcols = 1024;
 512         while (sz >= p->maxcols)
 513                 p->maxcols <<= 2;
 514
 515         p->buf = realloc(p->buf, p->maxcols);
 516         if (NULL == p->buf) {
 517                 perror(NULL);
 518                 exit(EXIT_FAILURE);
 519         }
 520 }
 521
 522
 523 static void
 524 buffera(struct termp *p, const char *word, size_t sz)
 525 {
 526
 527         if (p->col + sz >= p->maxcols)
 528                 adjbuf(p, p->col + sz);
 529
 530         memcpy(&p->buf[p->col], word, sz);
 531         p->col += sz;
 532 }
 533
 534
 535 static void
 536 bufferc(struct termp *p, char c)
 537 {
 538
 539         if (p->col + 1 >= p->maxcols)
 540                 adjbuf(p, p->col + 1);
 541
 542         p->buf[p->col++] = c;
 543 }
 544
 545
 546 static void
 547 encode(struct termp *p, const char *word, size_t sz)
 548 {
 549         enum termfont     f;
 550         int               i;
 551
 552         /*
 553          * Encode and buffer a string of characters.  If the current
 554          * font mode is unset, buffer directly, else encode then buffer
 555          * character by character.
 556          */
 557
 558         if (TERMFONT_NONE == (f = term_fonttop(p))) {
 559                 buffera(p, word, sz);
 560                 return;
 561         }
 562
 563         for (i = 0; i < (int)sz; i++) {
 564                 if ( ! isgraph((u_char)word[i])) {
 565                         bufferc(p, word[i]);
 566                         continue;
 567                 }
 568
 569                 if (TERMFONT_UNDER == f)
 570                         bufferc(p, '_');
 571                 else
 572                         bufferc(p, word[i]);
 573
 574                 bufferc(p, 8);
 575                 bufferc(p, word[i]);
 576         }
 577 }
 578
 579
 580 size_t
 581 term_vspan(const struct roffsu *su)
 582 {
 583         double           r;
 584
 585         switch (su->unit) {
 586         case (SCALE_CM):
 587                 r = su->scale * 2;
 588                 break;
 589         case (SCALE_IN):
 590                 r = su->scale * 6;
 591                 break;
 592         case (SCALE_PC):
 593                 r = su->scale;
 594                 break;
 595         case (SCALE_PT):
 596                 r = su->scale / 8;
 597                 break;
 598         case (SCALE_MM):
 599                 r = su->scale / 1000;
 600                 break;
 601         case (SCALE_VS):
 602                 r = su->scale;
 603                 break;
 604         default:
 605                 r = su->scale - 1;
 606                 break;
 607         }
 608
 609         if (r < 0.0)
 610                 r = 0.0;
 611         return(/* LINTED */(size_t)
 612                         r);
 613 }
 614
 615
 616 size_t
 617 term_hspan(const struct roffsu *su)
 618 {
 619         double           r;
 620
 621         /* XXX: CM, IN, and PT are approximations. */
 622
 623         switch (su->unit) {
 624         case (SCALE_CM):
 625                 r = 4 * su->scale;
 626                 break;
 627         case (SCALE_IN):
 628                 /* XXX: this is an approximation. */
 629                 r = 10 * su->scale;
 630                 break;
 631         case (SCALE_PC):
 632                 r = (10 * su->scale) / 6;
 633                 break;
 634         case (SCALE_PT):
 635                 r = (10 * su->scale) / 72;
 636                 break;
 637         case (SCALE_MM):
 638                 r = su->scale / 1000; /* FIXME: double-check. */
 639                 break;
 640         case (SCALE_VS):
 641                 r = su->scale * 2 - 1; /* FIXME: double-check. */
 642                 break;
 643         default:
 644                 r = su->scale;
 645                 break;
 646         }
 647
 648         if (r < 0.0)
 649                 r = 0.0;
 650         return((size_t)/* LINTED */
 651                         r);
 652 }
 653
 654