mdw@git.distorted.org.uk Git - sgt/halibut/blob - bk_text.c

   1 /*
   2  * text backend for Halibut
   3  */
   4
   5 #include <stdio.h>
   6 #include <stdlib.h>
   7 #include <assert.h>
   8 #include "halibut.h"
   9
  10 typedef enum { LEFT, LEFTPLUS, CENTRE } alignment;
  11 typedef struct {
  12     alignment align;
  13     int just_numbers;
  14     wchar_t underline;
  15     wchar_t *number_suffix;
  16 } alignstruct;
  17
  18 typedef struct {
  19     int indent, indent_code;
  20     int listindentbefore, listindentafter;
  21     int width;
  22     alignstruct atitle, achapter, *asect;
  23     int nasect;
  24     int include_version_id;
  25     int indent_preambles;
  26     word bullet;
  27 } textconfig;
  28
  29 static int text_convert(wchar_t *, char **);
  30
  31 static void text_heading(FILE *, word *, word *, word *, alignstruct, int,int);
  32 static void text_rule(FILE *, int, int);
  33 static void text_para(FILE *, word *, char *, word *, int, int, int);
  34 static void text_codepara(FILE *, word *, int, int);
  35 static void text_versionid(FILE *, word *);
  36
  37 static alignment utoalign(wchar_t *p) {
  38     if (!ustricmp(p, L"centre") || !ustricmp(p, L"center"))
  39         return CENTRE;
  40     if (!ustricmp(p, L"leftplus"))
  41         return LEFTPLUS;
  42     return LEFT;
  43 }
  44
  45 static textconfig text_configure(paragraph *source) {
  46     textconfig ret;
  47
  48     /*
  49      * Non-negotiables.
  50      */
  51     ret.bullet.next = NULL;
  52     ret.bullet.alt = NULL;
  53     ret.bullet.type = word_Normal;
  54     ret.atitle.just_numbers = FALSE;   /* ignored */
  55
  56     /*
  57      * Defaults.
  58      */
  59     ret.indent = 7;
  60     ret.indent_code = 2;
  61     ret.listindentbefore = 1;
  62     ret.listindentafter = 3;
  63     ret.width = 68;
  64     ret.atitle.align = CENTRE;
  65     ret.atitle.underline = L'=';
  66     ret.achapter.align = LEFT;
  67     ret.achapter.just_numbers = FALSE;
  68     ret.achapter.number_suffix = ustrdup(L": ");
  69     ret.achapter.underline = L'-';
  70     ret.nasect = 1;
  71     ret.asect = mknewa(alignstruct, ret.nasect);
  72     ret.asect[0].align = LEFTPLUS;
  73     ret.asect[0].just_numbers = TRUE;
  74     ret.asect[0].number_suffix = ustrdup(L" ");
  75     ret.asect[0].underline = L'\0';
  76     ret.include_version_id = TRUE;
  77     ret.indent_preambles = FALSE;
  78     ret.bullet.text = ustrdup(L"-");
  79
  80     for (; source; source = source->next) {
  81         if (source->type == para_Config) {
  82             if (!ustricmp(source->keyword, L"text-indent")) {
  83                 ret.indent = utoi(uadv(source->keyword));
  84             } else if (!ustricmp(source->keyword, L"text-indent-code")) {
  85                 ret.indent_code = utoi(uadv(source->keyword));
  86             } else if (!ustricmp(source->keyword, L"text-width")) {
  87                 ret.width = utoi(uadv(source->keyword));
  88             } else if (!ustricmp(source->keyword, L"text-list-indent")) {
  89                 ret.listindentbefore = utoi(uadv(source->keyword));
  90             } else if (!ustricmp(source->keyword, L"text-listitem-indent")) {
  91                 ret.listindentafter = utoi(uadv(source->keyword));
  92             } else if (!ustricmp(source->keyword, L"text-chapter-align")) {
  93                 ret.achapter.align = utoalign(uadv(source->keyword));
  94             } else if (!ustricmp(source->keyword, L"text-chapter-underline")) {
  95                 ret.achapter.underline = *uadv(source->keyword);
  96             } else if (!ustricmp(source->keyword, L"text-chapter-numeric")) {
  97                 ret.achapter.just_numbers = utob(uadv(source->keyword));
  98             } else if (!ustricmp(source->keyword, L"text-chapter-suffix")) {
  99                 ret.achapter.number_suffix = uadv(source->keyword);
 100             } else if (!ustricmp(source->keyword, L"text-section-align")) {
 101                 wchar_t *p = uadv(source->keyword);
 102                 int n = 0;
 103                 if (uisdigit(*p)) {
 104                     n = utoi(p);
 105                     p = uadv(p);
 106                 }
 107                 if (n >= ret.nasect) {
 108                     int i;
 109                     ret.asect = resize(ret.asect, n+1);
 110                     for (i = ret.nasect; i <= n; i++)
 111                         ret.asect[i] = ret.asect[ret.nasect-1];
 112                     ret.nasect = n+1;
 113                 }
 114                 ret.asect[n].align = utoalign(p);
 115             } else if (!ustricmp(source->keyword, L"text-section-underline")) {
 116                 wchar_t *p = uadv(source->keyword);
 117                 int n = 0;
 118                 if (uisdigit(*p)) {
 119                     n = utoi(p);
 120                     p = uadv(p);
 121                 }
 122                 if (n >= ret.nasect) {
 123                     int i;
 124                     ret.asect = resize(ret.asect, n+1);
 125                     for (i = ret.nasect; i <= n; i++)
 126                         ret.asect[i] = ret.asect[ret.nasect-1];
 127                     ret.nasect = n+1;
 128                 }
 129                 ret.asect[n].underline = *p;
 130             } else if (!ustricmp(source->keyword, L"text-section-numeric")) {
 131                 wchar_t *p = uadv(source->keyword);
 132                 int n = 0;
 133                 if (uisdigit(*p)) {
 134                     n = utoi(p);
 135                     p = uadv(p);
 136                 }
 137                 if (n >= ret.nasect) {
 138                     int i;
 139                     ret.asect = resize(ret.asect, n+1);
 140                     for (i = ret.nasect; i <= n; i++)
 141                         ret.asect[i] = ret.asect[ret.nasect-1];
 142                     ret.nasect = n+1;
 143                 }
 144                 ret.asect[n].just_numbers = utob(p);
 145             } else if (!ustricmp(source->keyword, L"text-section-suffix")) {
 146                 wchar_t *p = uadv(source->keyword);
 147                 int n = 0;
 148                 if (uisdigit(*p)) {
 149                     n = utoi(p);
 150                     p = uadv(p);
 151                 }
 152                 if (n >= ret.nasect) {
 153                     int i;
 154                     ret.asect = resize(ret.asect, n+1);
 155                     for (i = ret.nasect; i <= n; i++)
 156                         ret.asect[i] = ret.asect[ret.nasect-1];
 157                     ret.nasect = n+1;
 158                 }
 159                 ret.asect[n].number_suffix = p;
 160             } else if (!ustricmp(source->keyword, L"text-title-align")) {
 161                 ret.atitle.align = utoalign(uadv(source->keyword));
 162             } else if (!ustricmp(source->keyword, L"text-title-underline")) {
 163                 ret.atitle.underline = *uadv(source->keyword);
 164             } else if (!ustricmp(source->keyword, L"text-versionid")) {
 165                 ret.include_version_id = utob(uadv(source->keyword));
 166             } else if (!ustricmp(source->keyword, L"text-indent-preamble")) {
 167                 ret.indent_preambles = utob(uadv(source->keyword));
 168             } else if (!ustricmp(source->keyword, L"text-bullet")) {
 169                 ret.bullet.text = uadv(source->keyword);
 170             }
 171         }
 172     }
 173
 174     return ret;
 175 }
 176
 177 void text_backend(paragraph *sourceform, keywordlist *keywords,
 178                   indexdata *idx) {
 179     paragraph *p;
 180     textconfig conf;
 181     word *prefix, *body, *wp;
 182     word spaceword;
 183     FILE *fp;
 184     char *prefixextra;
 185     int indentb, indenta;
 186
 187     IGNORE(keywords);                  /* we don't happen to need this */
 188     IGNORE(idx);                       /* or this */
 189
 190     conf = text_configure(sourceform);
 191
 192     /*
 193      * Determine the output file name, and open the output file
 194      *
 195      * FIXME: want configurable output file names here. For the
 196      * moment, we'll just call it `output.txt'.
 197      */
 198     fp = fopen("output.txt", "w");
 199     if (!fp) {
 200         error(err_cantopenw, "output.txt");
 201         return;
 202     }
 203
 204     /* Do the title */
 205     for (p = sourceform; p; p = p->next)
 206         if (p->type == para_Title)
 207             text_heading(fp, NULL, NULL, p->words,
 208                          conf.atitle, conf.indent, conf.width);
 209
 210     /* Do the preamble and copyright */
 211     for (p = sourceform; p; p = p->next)
 212         if (p->type == para_Preamble)
 213             text_para(fp, NULL, NULL, p->words,
 214                       conf.indent_preambles ? conf.indent : 0, 0,
 215                       conf.width + (conf.indent_preambles ? 0 : conf.indent));
 216     for (p = sourceform; p; p = p->next)
 217         if (p->type == para_Copyright)
 218             text_para(fp, NULL, NULL, p->words,
 219                       conf.indent_preambles ? conf.indent : 0, 0,
 220                       conf.width + (conf.indent_preambles ? 0 : conf.indent));
 221
 222     /* Do the main document */
 223     for (p = sourceform; p; p = p->next) switch (p->type) {
 224
 225         /*
 226          * Things we ignore because we've already processed them or
 227          * aren't going to touch them in this pass.
 228          */
 229       case para_IM:
 230       case para_BR:
 231       case para_Biblio:                /* only touch BiblioCited */
 232       case para_VersionID:
 233       case para_Copyright:
 234       case para_Preamble:
 235       case para_NoCite:
 236       case para_Title:
 237         break;
 238
 239         /*
 240          * Chapter titles.
 241          */
 242       case para_Chapter:
 243       case para_Appendix:
 244       case para_UnnumberedChapter:
 245         text_heading(fp, p->kwtext, p->kwtext2, p->words,
 246                      conf.achapter, conf.indent, conf.width);
 247         break;
 248
 249       case para_Heading:
 250       case para_Subsect:
 251         text_heading(fp, p->kwtext, p->kwtext2, p->words,
 252                      conf.asect[p->aux>=conf.nasect ? conf.nasect-1 : p->aux],
 253                      conf.indent, conf.width);
 254         break;
 255
 256       case para_Rule:
 257         text_rule(fp, conf.indent, conf.width);
 258         break;
 259
 260       case para_Normal:
 261       case para_BiblioCited:
 262       case para_Bullet:
 263       case para_NumberedList:
 264         if (p->type == para_Bullet) {
 265             prefix = &conf.bullet;
 266             prefixextra = NULL;
 267             indentb = conf.listindentbefore;
 268             indenta = conf.listindentafter;
 269         } else if (p->type == para_NumberedList) {
 270             prefix = p->kwtext;
 271             prefixextra = ".";         /* FIXME: configurability */
 272             indentb = conf.listindentbefore;
 273             indenta = conf.listindentafter;
 274         } else {
 275             prefix = NULL;
 276             prefixextra = NULL;
 277             indentb = indenta = 0;
 278         }
 279         if (p->type == para_BiblioCited) {
 280             body = dup_word_list(p->kwtext);
 281             for (wp = body; wp->next; wp = wp->next);
 282             wp->next = &spaceword;
 283             spaceword.next = p->words;
 284             spaceword.alt = NULL;
 285             spaceword.type = word_WhiteSpace;
 286             spaceword.text = NULL;
 287         } else {
 288             wp = NULL;
 289             body = p->words;
 290         }
 291         text_para(fp, prefix, prefixextra, body,
 292                   conf.indent + indentb, indenta,
 293                   conf.width - indentb - indenta);
 294         if (wp) {
 295             wp->next = NULL;
 296             free_word_list(body);
 297         }
 298         break;
 299
 300       case para_Code:
 301         text_codepara(fp, p->words, conf.indent + conf.indent_code, conf.width - 2 * conf.indent_code);
 302         break;
 303     }
 304
 305     /* Do the version ID */
 306     if (conf.include_version_id) {
 307         for (p = sourceform; p; p = p->next)
 308             if (p->type == para_VersionID)
 309                 text_versionid(fp, p->words);
 310     }
 311
 312     /*
 313      * Tidy up
 314      */
 315     fclose(fp);
 316     sfree(conf.bullet.text);
 317 }
 318
 319 /*
 320  * Convert a wide string into a string of chars. If `result' is
 321  * non-NULL, mallocs the resulting string and stores a pointer to
 322  * it in `*result'. If `result' is NULL, merely checks whether all
 323  * characters in the string are feasible for the output character
 324  * set.
 325  *
 326  * Return is nonzero if all characters are OK. If not all
 327  * characters are OK but `result' is non-NULL, a result _will_
 328  * still be generated!
 329  */
 330 static int text_convert(wchar_t *s, char **result) {
 331     /*
 332      * FIXME. Currently this is ISO8859-1 only.
 333      */
 334     int doing = (result != 0);
 335     int ok = TRUE;
 336     char *p = NULL;
 337     int plen = 0, psize = 0;
 338
 339     for (; *s; s++) {
 340         wchar_t c = *s;
 341         char outc;
 342
 343         if ((c >= 32 && c <= 126) ||
 344             (c >= 160 && c <= 255)) {
 345             /* Char is OK. */
 346             outc = (char)c;
 347         } else {
 348             /* Char is not OK. */
 349             ok = FALSE;
 350             outc = 0xBF;               /* approximate the good old DEC `uh?' */
 351         }
 352         if (doing) {
 353             if (plen >= psize) {
 354                 psize = plen + 256;
 355                 p = resize(p, psize);
 356             }
 357             p[plen++] = outc;
 358         }
 359     }
 360     if (doing) {
 361         p = resize(p, plen+1);
 362         p[plen] = '\0';
 363         *result = p;
 364     }
 365     return ok;
 366 }
 367
 368 static void text_rdaddwc(rdstringc *rs, word *text, word *end) {
 369     char *c;
 370
 371     for (; text && text != end; text = text->next) switch (text->type) {
 372       case word_HyperLink:
 373       case word_HyperEnd:
 374       case word_UpperXref:
 375       case word_LowerXref:
 376       case word_XrefEnd:
 377       case word_IndexRef:
 378         break;
 379
 380       case word_Normal:
 381       case word_Emph:
 382       case word_Code:
 383       case word_WeakCode:
 384       case word_WhiteSpace:
 385       case word_EmphSpace:
 386       case word_CodeSpace:
 387       case word_WkCodeSpace:
 388       case word_Quote:
 389       case word_EmphQuote:
 390       case word_CodeQuote:
 391       case word_WkCodeQuote:
 392         assert(text->type != word_CodeQuote &&
 393                text->type != word_WkCodeQuote);
 394         if (towordstyle(text->type) == word_Emph &&
 395             (attraux(text->aux) == attr_First ||
 396              attraux(text->aux) == attr_Only))
 397             rdaddc(rs, '_');           /* FIXME: configurability */
 398         else if (towordstyle(text->type) == word_Code &&
 399                  (attraux(text->aux) == attr_First ||
 400                   attraux(text->aux) == attr_Only))
 401             rdaddc(rs, '`');           /* FIXME: configurability */
 402         if (removeattr(text->type) == word_Normal) {
 403             if (text_convert(text->text, &c))
 404                 rdaddsc(rs, c);
 405             else
 406                 text_rdaddwc(rs, text->alt, NULL);
 407             sfree(c);
 408         } else if (removeattr(text->type) == word_WhiteSpace) {
 409             rdaddc(rs, ' ');
 410         } else if (removeattr(text->type) == word_Quote) {
 411             rdaddc(rs, quoteaux(text->aux) == quote_Open ? '`' : '\'');
 412                                        /* FIXME: configurability */
 413         }
 414         if (towordstyle(text->type) == word_Emph &&
 415             (attraux(text->aux) == attr_Last ||
 416              attraux(text->aux) == attr_Only))
 417             rdaddc(rs, '_');           /* FIXME: configurability */
 418         else if (towordstyle(text->type) == word_Code &&
 419                  (attraux(text->aux) == attr_Last ||
 420                   attraux(text->aux) == attr_Only))
 421             rdaddc(rs, '\'');          /* FIXME: configurability */
 422         break;
 423     }
 424 }
 425
 426 static int text_width(word *);
 427
 428 static int text_width_list(word *text) {
 429     int w = 0;
 430     while (text) {
 431         w += text_width(text);
 432         text = text->next;
 433     }
 434     return w;
 435 }
 436
 437 static int text_width(word *text) {
 438     switch (text->type) {
 439       case word_HyperLink:
 440       case word_HyperEnd:
 441       case word_UpperXref:
 442       case word_LowerXref:
 443       case word_XrefEnd:
 444       case word_IndexRef:
 445         return 0;
 446
 447       case word_Normal:
 448       case word_Emph:
 449       case word_Code:
 450       case word_WeakCode:
 451         return (((text->type == word_Emph ||
 452                   text->type == word_Code)
 453                  ? (attraux(text->aux) == attr_Only ? 2 :
 454                     attraux(text->aux) == attr_Always ? 0 : 1)
 455                  : 0) +
 456                 (text_convert(text->text, NULL) ?
 457                  ustrlen(text->text) :
 458                  text_width_list(text->alt)));
 459
 460       case word_WhiteSpace:
 461       case word_EmphSpace:
 462       case word_CodeSpace:
 463       case word_WkCodeSpace:
 464       case word_Quote:
 465       case word_EmphQuote:
 466       case word_CodeQuote:
 467       case word_WkCodeQuote:
 468         assert(text->type != word_CodeQuote &&
 469                text->type != word_WkCodeQuote);
 470         return (((towordstyle(text->type) == word_Emph ||
 471                   towordstyle(text->type) == word_Code)
 472                  ? (attraux(text->aux) == attr_Only ? 2 :
 473                     attraux(text->aux) == attr_Always ? 0 : 1)
 474                  : 0) + 1);
 475     }
 476     return 0;                          /* should never happen */
 477 }
 478
 479 static void text_heading(FILE *fp, word *tprefix, word *nprefix, word *text,
 480                          alignstruct align, int indent, int width) {
 481     rdstringc t = { 0, 0, NULL };
 482     int margin, length;
 483     int firstlinewidth, wrapwidth;
 484     wrappedline *wrapping, *p;
 485
 486     if (align.just_numbers && nprefix) {
 487         char *c;
 488         text_rdaddwc(&t, nprefix, NULL);
 489         if (text_convert(align.number_suffix, &c)) {
 490             rdaddsc(&t, c);
 491             sfree(c);
 492         }
 493     } else if (!align.just_numbers && tprefix) {
 494         char *c;
 495         text_rdaddwc(&t, tprefix, NULL);
 496         if (text_convert(align.number_suffix, &c)) {
 497             rdaddsc(&t, c);
 498             sfree(c);
 499         }
 500     }
 501     margin = length = (t.text ? strlen(t.text) : 0);
 502
 503     if (align.align == LEFTPLUS) {
 504         margin = indent - margin;
 505         if (margin < 0) margin = 0;
 506         firstlinewidth = indent + width - margin - length;
 507         wrapwidth = width;
 508     } else if (align.align == LEFT || align.align == CENTRE) {
 509         margin = 0;
 510         firstlinewidth = indent + width - length;
 511         wrapwidth = indent + width;
 512     }
 513
 514     wrapping = wrap_para(text, firstlinewidth, wrapwidth, text_width);
 515     for (p = wrapping; p; p = p->next) {
 516         text_rdaddwc(&t, p->begin, p->end);
 517         length = (t.text ? strlen(t.text) : 0);
 518         if (align.align == CENTRE) {
 519             margin = (indent + width - length)/2;
 520             if (margin < 0) margin = 0;
 521         }
 522         fprintf(fp, "%*s%s\n", margin, "", t.text);
 523         if (align.underline != L'\0') {
 524             char *u, uc;
 525             wchar_t uw[2];
 526             uw[0] = align.underline; uw[1] = L'\0';
 527             text_convert(uw, &u);
 528             uc = u[0];
 529             sfree(u);
 530             fprintf(fp, "%*s", margin, "");
 531             while (length--)
 532                 putc(uc, fp);
 533             putc('\n', fp);
 534         }
 535         if (align.align == LEFTPLUS)
 536             margin = indent;
 537         else
 538             margin = 0;
 539         sfree(t.text);
 540         t = empty_rdstringc;
 541     }
 542     wrap_free(wrapping);
 543     putc('\n', fp);
 544
 545     sfree(t.text);
 546 }
 547
 548 static void text_rule(FILE *fp, int indent, int width) {
 549     while (indent--) putc(' ', fp);
 550     while (width--) putc('-', fp);     /* FIXME: configurability! */
 551     putc('\n', fp);
 552     putc('\n', fp);
 553 }
 554
 555 static void text_para(FILE *fp, word *prefix, char *prefixextra, word *text,
 556                       int indent, int extraindent, int width) {
 557     wrappedline *wrapping, *p;
 558     rdstringc pfx = { 0, 0, NULL };
 559     int e;
 560     int firstlinewidth = width;
 561
 562     if (prefix) {
 563         text_rdaddwc(&pfx, prefix, NULL);
 564         if (prefixextra)
 565             rdaddsc(&pfx, prefixextra);
 566         fprintf(fp, "%*s%s", indent, "", pfx.text);
 567         /* If the prefix is too long, shorten the first line to fit. */
 568         e = extraindent - strlen(pfx.text);
 569         if (e < 0) {
 570             firstlinewidth += e;       /* this decreases it, since e < 0 */
 571             if (firstlinewidth < 0) {
 572                 e = indent + extraindent;
 573                 firstlinewidth = width;
 574                 fprintf(fp, "\n");
 575             } else
 576                 e = 0;
 577         }
 578         sfree(pfx.text);
 579     } else
 580         e = indent + extraindent;
 581
 582     wrapping = wrap_para(text, firstlinewidth, width, text_width);
 583     for (p = wrapping; p; p = p->next) {
 584         rdstringc t = { 0, 0, NULL };
 585         text_rdaddwc(&t, p->begin, p->end);
 586         fprintf(fp, "%*s%s\n", e, "", t.text);
 587         e = indent + extraindent;
 588         sfree(t.text);
 589     }
 590     wrap_free(wrapping);
 591     putc('\n', fp);
 592 }
 593
 594 static void text_codepara(FILE *fp, word *text, int indent, int width) {
 595     for (; text; text = text->next) if (text->type == word_WeakCode) {
 596         char *c;
 597         text_convert(text->text, &c);
 598         if (strlen(c) > (size_t)width) {
 599             /* FIXME: warn */
 600         }
 601         fprintf(fp, "%*s%s\n", indent, "", c);
 602         sfree(c);
 603     }
 604
 605     putc('\n', fp);
 606 }
 607
 608 static void text_versionid(FILE *fp, word *text) {
 609     rdstringc t = { 0, 0, NULL };
 610
 611     rdaddc(&t, '[');                   /* FIXME: configurability */
 612     text_rdaddwc(&t, text, NULL);
 613     rdaddc(&t, ']');                   /* FIXME: configurability */
 614
 615     fprintf(fp, "%s\n", t.text);
 616     sfree(t.text);
 617 }