term.c

   1 /*      $Id: term.c,v 1.89 2009/07/16 13:17:51 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 <err.h>
  19 #include <stdio.h>
  20 #include <stdlib.h>
  21 #include <string.h>
  22
  23 #include "term.h"
  24 #include "man.h"
  25 #include "mdoc.h"
  26
  27 extern  int               man_run(struct termp *,
  28                                 const struct man *);
  29 extern  int               mdoc_run(struct termp *,
  30                                 const struct mdoc *);
  31
  32 static  struct termp     *term_alloc(enum termenc);
  33 static  void              term_free(struct termp *);
  34 static  void              term_pescape(struct termp *, const char **);
  35 static  void              term_nescape(struct termp *,
  36                                 const char *, size_t);
  37 static  void              term_chara(struct termp *, char);
  38 static  void              term_encodea(struct termp *, char);
  39 static  int               term_isopendelim(const char *);
  40 static  int               term_isclosedelim(const char *);
  41
  42
  43 void *
  44 ascii_alloc(void)
  45 {
  46
  47         return(term_alloc(TERMENC_ASCII));
  48 }
  49
  50
  51 int
  52 terminal_man(void *arg, const struct man *man)
  53 {
  54         struct termp    *p;
  55
  56         p = (struct termp *)arg;
  57         if (NULL == p->symtab)
  58                 p->symtab = term_ascii2htab();
  59
  60         return(man_run(p, man));
  61 }
  62
  63
  64 int
  65 terminal_mdoc(void *arg, const struct mdoc *mdoc)
  66 {
  67         struct termp    *p;
  68
  69         p = (struct termp *)arg;
  70         if (NULL == p->symtab)
  71                 p->symtab = term_ascii2htab();
  72
  73         return(mdoc_run(p, mdoc));
  74 }
  75
  76
  77 void
  78 terminal_free(void *arg)
  79 {
  80
  81         term_free((struct termp *)arg);
  82 }
  83
  84
  85 static void
  86 term_free(struct termp *p)
  87 {
  88
  89         if (p->buf)
  90                 free(p->buf);
  91         if (TERMENC_ASCII == p->enc && p->symtab)
  92                 term_asciifree(p->symtab);
  93
  94         free(p);
  95 }
  96
  97
  98 static struct termp *
  99 term_alloc(enum termenc enc)
 100 {
 101         struct termp *p;
 102
 103         if (NULL == (p = malloc(sizeof(struct termp))))
 104                 err(1, "malloc");
 105         bzero(p, sizeof(struct termp));
 106         p->maxrmargin = 78;
 107         p->enc = enc;
 108         return(p);
 109 }
 110
 111
 112 static int
 113 term_isclosedelim(const char *p)
 114 {
 115
 116         if ( ! (*p && 0 == *(p + 1)))
 117                 return(0);
 118
 119         switch (*p) {
 120         case('.'):
 121                 /* FALLTHROUGH */
 122         case(','):
 123                 /* FALLTHROUGH */
 124         case(';'):
 125                 /* FALLTHROUGH */
 126         case(':'):
 127                 /* FALLTHROUGH */
 128         case('?'):
 129                 /* FALLTHROUGH */
 130         case('!'):
 131                 /* FALLTHROUGH */
 132         case(')'):
 133                 /* FALLTHROUGH */
 134         case(']'):
 135                 /* FALLTHROUGH */
 136         case('}'):
 137                 return(1);
 138         default:
 139                 break;
 140         }
 141
 142         return(0);
 143 }
 144
 145
 146 static int
 147 term_isopendelim(const char *p)
 148 {
 149
 150         if ( ! (*p && 0 == *(p + 1)))
 151                 return(0);
 152
 153         switch (*p) {
 154         case('('):
 155                 /* FALLTHROUGH */
 156         case('['):
 157                 /* FALLTHROUGH */
 158         case('{'):
 159                 return(1);
 160         default:
 161                 break;
 162         }
 163
 164         return(0);
 165 }
 166
 167
 168 /*
 169  * Flush a line of text.  A "line" is loosely defined as being something
 170  * that should be followed by a newline, regardless of whether it's
 171  * broken apart by newlines getting there.  A line can also be a
 172  * fragment of a columnar list.
 173  *
 174  * Specifically, a line is whatever's in p->buf of length p->col, which
 175  * is zeroed after this function returns.
 176  *
 177  * The usage of termp:flags is as follows:
 178  *
 179  *  - TERMP_NOLPAD: when beginning to write the line, don't left-pad the
 180  *    offset value.  This is useful when doing columnar lists where the
 181  *    prior column has right-padded.
 182  *
 183  *  - TERMP_NOBREAK: this is the most important and is used when making
 184  *    columns.  In short: don't print a newline and instead pad to the
 185  *    right margin.  Used in conjunction with TERMP_NOLPAD.
 186  *
 187  *  - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and
 188  *    the line is overrun, and don't pad-right if it's underrun.
 189  *
 190  *  - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when
 191  *    overruning, instead save the position and continue at that point
 192  *    when the next invocation.
 193  *
 194  *  In-line line breaking:
 195  *
 196  *  If TERMP_NOBREAK is specified and the line overruns the right
 197  *  margin, it will break and pad-right to the right margin after
 198  *  writing.  If maxrmargin is violated, it will break and continue
 199  *  writing from the right-margin, which will lead to the above
 200  *  scenario upon exit.
 201  *
 202  *  Otherwise, the line will break at the right margin.  Extremely long
 203  *  lines will cause the system to emit a warning (TODO: hyphenate, if
 204  *  possible).
 205  *
 206  *  FIXME: newline breaks occur (in groff) also occur when a single
 207  *  space follows a NOBREAK!
 208  */
 209 void
 210 term_flushln(struct termp *p)
 211 {
 212         int              i, j;
 213         size_t           vbl, vsz, vis, maxvis, mmax, bp;
 214         static int       sv = -1;
 215
 216         /*
 217          * First, establish the maximum columns of "visible" content.
 218          * This is usually the difference between the right-margin and
 219          * an indentation, but can be, for tagged lists or columns, a
 220          * small set of values.
 221          */
 222
 223         assert(p->offset < p->rmargin);
 224         maxvis = p->rmargin - p->offset;
 225         mmax = p->maxrmargin - p->offset;
 226         bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
 227         vis = 0;
 228
 229         if (sv >= 0) {
 230                 vis = (size_t)sv;
 231                 sv = -1;
 232         }
 233
 234         /*
 235          * If in the standard case (left-justified), then begin with our
 236          * indentation, otherwise (columns, etc.) just start spitting
 237          * out text.
 238          */
 239
 240         if ( ! (p->flags & TERMP_NOLPAD))
 241                 /* LINTED */
 242                 for (j = 0; j < (int)p->offset; j++)
 243                         putchar(' ');
 244
 245         for (i = 0; i < (int)p->col; i++) {
 246                 /*
 247                  * Count up visible word characters.  Control sequences
 248                  * (starting with the CSI) aren't counted.  A space
 249                  * generates a non-printing word, which is valid (the
 250                  * space is printed according to regular spacing rules).
 251                  */
 252
 253                 /* LINTED */
 254                 for (j = i, vsz = 0; j < (int)p->col; j++) {
 255                         if (' ' == p->buf[j])
 256                                 break;
 257                         else if (8 == p->buf[j])
 258                                 vsz--;
 259                         else
 260                                 vsz++;
 261                 }
 262
 263                 /*
 264                  * Choose the number of blanks to prepend: no blank at the
 265                  * beginning of a line, one between words -- but do not
 266                  * actually write them yet.
 267                  */
 268                 vbl = (size_t)(0 == vis ? 0 : 1);
 269
 270                 /*
 271                  * Find out whether we would exceed the right margin.
 272                  * If so, break to the next line.  (TODO: hyphenate)
 273                  * Otherwise, write the chosen number of blanks now.
 274                  */
 275                 if (vis && vis + vbl + vsz > bp) {
 276                         putchar('\n');
 277                         if (TERMP_NOBREAK & p->flags) {
 278                                 for (j = 0; j < (int)p->rmargin; j++)
 279                                         putchar(' ');
 280                                 vis = p->rmargin - p->offset;
 281                         } else {
 282                                 for (j = 0; j < (int)p->offset; j++)
 283                                         putchar(' ');
 284                                 vis = 0;
 285                         }
 286                 } else {
 287                         for (j = 0; j < (int)vbl; j++)
 288                                 putchar(' ');
 289                         vis += vbl;
 290                 }
 291
 292                 /*
 293                  * Finally, write out the word.
 294                  */
 295                 for ( ; i < (int)p->col; i++) {
 296                         if (' ' == p->buf[i])
 297                                 break;
 298                         putchar(p->buf[i]);
 299                 }
 300                 vis += vsz;
 301         }
 302
 303         /*
 304          * If we've overstepped our maximum visible no-break space, then
 305          * cause a newline and offset at the right margin.
 306          */
 307
 308         if ((TERMP_NOBREAK & p->flags) && vis >= maxvis) {
 309                 if ( ! (TERMP_DANGLE & p->flags) &&
 310                                 ! (TERMP_HANG & p->flags)) {
 311                         putchar('\n');
 312                         for (i = 0; i < (int)p->rmargin; i++)
 313                                 putchar(' ');
 314                 }
 315                 if (TERMP_HANG & p->flags)
 316                         sv = (int)(vis - maxvis);
 317                 p->col = 0;
 318                 return;
 319         }
 320
 321         /*
 322          * If we're not to right-marginalise it (newline), then instead
 323          * pad to the right margin and stay off.
 324          */
 325
 326         if (p->flags & TERMP_NOBREAK) {
 327                 if ( ! (TERMP_DANGLE & p->flags))
 328                         for ( ; vis < maxvis; vis++)
 329                                 putchar(' ');
 330         } else
 331                 putchar('\n');
 332
 333         p->col = 0;
 334 }
 335
 336
 337 /*
 338  * A newline only breaks an existing line; it won't assert vertical
 339  * space.  All data in the output buffer is flushed prior to the newline
 340  * assertion.
 341  */
 342 void
 343 term_newln(struct termp *p)
 344 {
 345
 346         p->flags |= TERMP_NOSPACE;
 347         if (0 == p->col) {
 348                 p->flags &= ~TERMP_NOLPAD;
 349                 return;
 350         }
 351         term_flushln(p);
 352         p->flags &= ~TERMP_NOLPAD;
 353 }
 354
 355
 356 /*
 357  * Asserts a vertical space (a full, empty line-break between lines).
 358  * Note that if used twice, this will cause two blank spaces and so on.
 359  * All data in the output buffer is flushed prior to the newline
 360  * assertion.
 361  */
 362 void
 363 term_vspace(struct termp *p)
 364 {
 365
 366         term_newln(p);
 367         putchar('\n');
 368 }
 369
 370
 371 /*
 372  * Determine the symbol indicated by an escape sequences, that is, one
 373  * starting with a backslash.  Once done, we pass this value into the
 374  * output buffer by way of the symbol table.
 375  */
 376 static void
 377 term_nescape(struct termp *p, const char *word, size_t len)
 378 {
 379         const char      *rhs;
 380         size_t           sz;
 381         int              i;
 382
 383         rhs = term_a2ascii(p->symtab, word, len, &sz);
 384
 385         if (rhs)
 386                 for (i = 0; i < (int)sz; i++)
 387                         term_encodea(p, rhs[i]);
 388 }
 389
 390
 391 /*
 392  * Handle an escape sequence: determine its length and pass it to the
 393  * escape-symbol look table.  Note that we assume mdoc(3) has validated
 394  * the escape sequence (we assert upon badly-formed escape sequences).
 395  */
 396 static void
 397 term_pescape(struct termp *p, const char **word)
 398 {
 399         int              j;
 400         const char      *wp;
 401
 402         wp = *word;
 403
 404         if (0 == *(++wp)) {
 405                 *word = wp;
 406                 return;
 407         }
 408
 409         if ('(' == *wp) {
 410                 wp++;
 411                 if (0 == *wp || 0 == *(wp + 1)) {
 412                         *word = 0 == *wp ? wp : wp + 1;
 413                         return;
 414                 }
 415
 416                 term_nescape(p, wp, 2);
 417                 *word = ++wp;
 418                 return;
 419
 420         } else if ('*' == *wp) {
 421                 if (0 == *(++wp)) {
 422                         *word = wp;
 423                         return;
 424                 }
 425
 426                 switch (*wp) {
 427                 case ('('):
 428                         wp++;
 429                         if (0 == *wp || 0 == *(wp + 1)) {
 430                                 *word = 0 == *wp ? wp : wp + 1;
 431                                 return;
 432                         }
 433
 434                         term_nescape(p, wp, 2);
 435                         *word = ++wp;
 436                         return;
 437                 case ('['):
 438                         break;
 439                 default:
 440                         term_nescape(p, wp, 1);
 441                         *word = wp;
 442                         return;
 443                 }
 444
 445         } else if ('f' == *wp) {
 446                 if (0 == *(++wp)) {
 447                         *word = wp;
 448                         return;
 449                 }
 450
 451                 switch (*wp) {
 452                 case ('B'):
 453                         p->flags |= TERMP_BOLD;
 454                         break;
 455                 case ('I'):
 456                         p->flags |= TERMP_UNDER;
 457                         break;
 458                 case ('P'):
 459                         /* FALLTHROUGH */
 460                 case ('R'):
 461                         p->flags &= ~TERMP_STYLE;
 462                         break;
 463                 default:
 464                         break;
 465                 }
 466
 467                 *word = wp;
 468                 return;
 469
 470         } else if ('[' != *wp) {
 471                 term_nescape(p, wp, 1);
 472                 *word = wp;
 473                 return;
 474         }
 475
 476         wp++;
 477         for (j = 0; *wp && ']' != *wp; wp++, j++)
 478                 /* Loop... */ ;
 479
 480         if (0 == *wp) {
 481                 *word = wp;
 482                 return;
 483         }
 484
 485         term_nescape(p, wp - j, (size_t)j);
 486         *word = wp;
 487 }
 488
 489
 490 /*
 491  * Handle pwords, partial words, which may be either a single word or a
 492  * phrase that cannot be broken down (such as a literal string).  This
 493  * handles word styling.
 494  */
 495 void
 496 term_word(struct termp *p, const char *word)
 497 {
 498         const char       *sv;
 499
 500         if (term_isclosedelim(word))
 501                 if ( ! (TERMP_IGNDELIM & p->flags))
 502                         p->flags |= TERMP_NOSPACE;
 503
 504         if ( ! (TERMP_NOSPACE & p->flags))
 505                 term_chara(p, ' ');
 506
 507         if ( ! (p->flags & TERMP_NONOSPACE))
 508                 p->flags &= ~TERMP_NOSPACE;
 509
 510         /*
 511          * If ANSI (word-length styling), then apply our style now,
 512          * before the word.
 513          */
 514
 515         for (sv = word; *word; word++)
 516                 if ('\\' != *word)
 517                         term_encodea(p, *word);
 518                 else
 519                         term_pescape(p, &word);
 520
 521         if (term_isopendelim(sv))
 522                 p->flags |= TERMP_NOSPACE;
 523 }
 524
 525
 526 /*
 527  * Insert a single character into the line-buffer.  If the buffer's
 528  * space is exceeded, then allocate more space by doubling the buffer
 529  * size.
 530  */
 531 static void
 532 term_chara(struct termp *p, char c)
 533 {
 534         size_t           s;
 535
 536         if (p->col + 1 >= p->maxcols) {
 537                 if (0 == p->maxcols)
 538                         p->maxcols = 256;
 539                 s = p->maxcols * 2;
 540                 p->buf = realloc(p->buf, s);
 541                 if (NULL == p->buf)
 542                         err(1, "realloc");
 543                 p->maxcols = s;
 544         }
 545         p->buf[(int)(p->col)++] = c;
 546 }
 547
 548
 549 static void
 550 term_encodea(struct termp *p, char c)
 551 {
 552
 553         if (' ' != c && TERMP_STYLE & p->flags) {
 554                 if (TERMP_BOLD & p->flags) {
 555                         term_chara(p, c);
 556                         term_chara(p, 8);
 557                 }
 558                 if (TERMP_UNDER & p->flags) {
 559                         term_chara(p, '_');
 560                         term_chara(p, 8);
 561                 }
 562         }
 563         term_chara(p, c);
 564 }