[sgt/halibut] / bk_text.c

/*
 * text backend for Halibut
 */

#include <stdio.h>
#include <stdlib.h>
#include <assert.h>
#include "halibut.h"

typedef enum { LEFT, LEFTPLUS, CENTRE } alignment;
typedef struct {
    alignment align;
    int just_numbers;
    wchar_t underline;
} alignstruct;

typedef struct {
    int indent, indent_code;
    int listindentbefore, listindentafter;
    int width;
    alignstruct atitle, achapter, *asect;
    int nasect;
    int include_version_id;
    int indent_preambles;
    word bullet;
} textconfig;

static int text_convert(wchar_t *, char **);

static void text_heading(FILE *, word *, word *, word *, alignstruct, int,int);
static void text_rule(FILE *, int, int);
static void text_para(FILE *, word *, char *, word *, int, int, int);
static void text_codepara(FILE *, word *, int, int);
static void text_versionid(FILE *, word *);

static alignment utoalign(wchar_t *p) {
    if (!ustricmp(p, L"centre") || !ustricmp(p, L"center"))
	return CENTRE;
    if (!ustricmp(p, L"leftplus"))
	return LEFTPLUS;
    return LEFT;
}

static textconfig text_configure(paragraph *source) {
    textconfig ret;

    /*
     * Non-negotiables.
     */
    ret.bullet.next = NULL;
    ret.bullet.alt = NULL;
    ret.bullet.type = word_Normal;
    ret.atitle.just_numbers = FALSE;   /* ignored */

    /*
     * Defaults.
     */
    ret.indent = 7;
    ret.indent_code = 2;
    ret.listindentbefore = 1;
    ret.listindentafter = 3;
    ret.width = 68;
    ret.atitle.align = CENTRE;
    ret.atitle.underline = L'=';
    ret.achapter.align = LEFT;
    ret.achapter.just_numbers = FALSE;
    ret.achapter.underline = L'-';
    ret.nasect = 1;
    ret.asect = mknewa(alignstruct, ret.nasect);
    ret.asect[0].align = LEFTPLUS;
    ret.asect[0].just_numbers = TRUE;
    ret.asect[0].underline = L'\0';
    ret.include_version_id = TRUE;
    ret.indent_preambles = FALSE;
    ret.bullet.text = ustrdup(L"-");

    for (; source; source = source->next) {
	if (source->type == para_Config) {
	    if (!ustricmp(source->keyword, L"text-indent")) {
		ret.indent = utoi(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-indent-code")) {
		ret.indent_code = utoi(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-width")) {
		ret.width = utoi(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-list-indent")) {
		ret.listindentbefore = utoi(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-listitem-indent")) {
		ret.listindentafter = utoi(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-chapter-align")) {
		ret.achapter.align = utoalign(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-chapter-underline")) {
		ret.achapter.underline = *uadv(source->keyword);
	    } else if (!ustricmp(source->keyword, L"text-chapter-numeric")) {
		ret.achapter.just_numbers = utob(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-section-align")) {
		wchar_t *p = uadv(source->keyword);
		int n = 0;
		if (uisdigit(*p)) {
		    n = utoi(p);
		    p = uadv(p);
		}
		if (n >= ret.nasect) {
		    int i;
		    ret.asect = resize(ret.asect, n+1);
		    for (i = ret.nasect; i <= n; i++)
			ret.asect[i] = ret.asect[ret.nasect-1];
		    ret.nasect = n+1;
		}
		ret.asect[n].align = utoalign(p);
	    } else if (!ustricmp(source->keyword, L"text-section-underline")) {
		wchar_t *p = uadv(source->keyword);
		int n = 0;
		if (uisdigit(*p)) {
		    n = utoi(p);
		    p = uadv(p);
		}
		if (n >= ret.nasect) {
		    int i;
		    ret.asect = resize(ret.asect, n+1);
		    for (i = ret.nasect; i <= n; i++)
			ret.asect[i] = ret.asect[ret.nasect-1];
		    ret.nasect = n+1;
		}
		ret.asect[n].underline = *p;
	    } else if (!ustricmp(source->keyword, L"text-section-numeric")) {
		wchar_t *p = uadv(source->keyword);
		int n = 0;
		if (uisdigit(*p)) {
		    n = utoi(p);
		    p = uadv(p);
		}
		if (n >= ret.nasect) {
		    int i;
		    ret.asect = resize(ret.asect, n+1);
		    for (i = ret.nasect; i <= n; i++)
			ret.asect[i] = ret.asect[ret.nasect-1];
		    ret.nasect = n+1;
		}
		ret.asect[n].just_numbers = utob(p);
	    } else if (!ustricmp(source->keyword, L"text-title-align")) {
		ret.atitle.align = utoalign(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-title-underline")) {
		ret.atitle.underline = *uadv(source->keyword);
	    } else if (!ustricmp(source->keyword, L"text-versionid")) {
		ret.include_version_id = utob(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-indent-preamble")) {
		ret.indent_preambles = utob(uadv(source->keyword));
	    } else if (!ustricmp(source->keyword, L"text-bullet")) {
		ret.bullet.text = uadv(source->keyword);
	    }
	}
    }

    return ret;
}

void text_backend(paragraph *sourceform, keywordlist *keywords,
		  indexdata *idx) {
    paragraph *p;
    textconfig conf;
    word *prefix, *body, *wp;
    word spaceword;
    FILE *fp;
    char *prefixextra;
    int indentb, indenta;

    IGNORE(keywords);		       /* we don't happen to need this */
    IGNORE(idx);		       /* or this */

    conf = text_configure(sourceform);

    /*
     * Determine the output file name, and open the output file
     *
     * FIXME: want configurable output file names here. For the
     * moment, we'll just call it `output.txt'.
     */
    fp = fopen("output.txt", "w");
    if (!fp) {
	error(err_cantopenw, "output.txt");
	return;
    }

    /* Do the title */
    for (p = sourceform; p; p = p->next)
	if (p->type == para_Title)
	    text_heading(fp, NULL, NULL, p->words,
			 conf.atitle, conf.indent, conf.width);

    /* Do the preamble and copyright */
    for (p = sourceform; p; p = p->next)
	if (p->type == para_Preamble)
	    text_para(fp, NULL, NULL, p->words,
		      conf.indent_preambles ? conf.indent : 0, 0,
		      conf.width + (conf.indent_preambles ? 0 : conf.indent));
    for (p = sourceform; p; p = p->next)
	if (p->type == para_Copyright)
	    text_para(fp, NULL, NULL, p->words,
		      conf.indent_preambles ? conf.indent : 0, 0,
		      conf.width + (conf.indent_preambles ? 0 : conf.indent));

    /* Do the main document */
    for (p = sourceform; p; p = p->next) switch (p->type) {

	/*
	 * Things we ignore because we've already processed them or
	 * aren't going to touch them in this pass.
	 */
      case para_IM:
      case para_BR:
      case para_Biblio:		       /* only touch BiblioCited */
      case para_VersionID:
      case para_Copyright:
      case para_Preamble:
      case para_NoCite:
      case para_Title:
	break;

	/*
	 * Chapter titles.
	 */
      case para_Chapter:
      case para_Appendix:
      case para_UnnumberedChapter:
	text_heading(fp, p->kwtext, p->kwtext2, p->words,
		     conf.achapter, conf.indent, conf.width);
	break;

      case para_Heading:
      case para_Subsect:
	text_heading(fp, p->kwtext, p->kwtext2, p->words,
		     conf.asect[p->aux>=conf.nasect ? conf.nasect-1 : p->aux],
		     conf.indent, conf.width);
	break;

      case para_Rule:
	text_rule(fp, conf.indent, conf.width);
	break;

      case para_Normal:
      case para_BiblioCited:
      case para_Bullet:
      case para_NumberedList:
	if (p->type == para_Bullet) {
	    prefix = &conf.bullet;
	    prefixextra = NULL;
	    indentb = conf.listindentbefore;
	    indenta = conf.listindentafter;
	} else if (p->type == para_NumberedList) {
	    prefix = p->kwtext;
	    prefixextra = ".";	       /* FIXME: configurability */
	    indentb = conf.listindentbefore;
	    indenta = conf.listindentafter;
	} else {
	    prefix = NULL;
	    prefixextra = NULL;
	    indentb = indenta = 0;
	}
	if (p->type == para_BiblioCited) {
	    body = dup_word_list(p->kwtext);
	    for (wp = body; wp->next; wp = wp->next);
	    wp->next = &spaceword;
	    spaceword.next = p->words;
	    spaceword.alt = NULL;
	    spaceword.type = word_WhiteSpace;
	    spaceword.text = NULL;
	} else {
	    wp = NULL;
	    body = p->words;
	}
	text_para(fp, prefix, prefixextra, body,
		  conf.indent + indentb, indenta,
		  conf.width - indentb - indenta);
	if (wp) {
	    wp->next = NULL;
	    free_word_list(body);
	}
	break;

      case para_Code:
	text_codepara(fp, p->words, conf.indent + conf.indent_code, conf.width - 2 * conf.indent_code);
	break;
    }

    /* Do the version ID */
    if (conf.include_version_id) {
	for (p = sourceform; p; p = p->next)
	    if (p->type == para_VersionID)
 		text_versionid(fp, p->words);
    }

    /*
     * Tidy up
     */
    fclose(fp);
    sfree(conf.bullet.text);
}

/*
 * Convert a wide string into a string of chars. If `result' is
 * non-NULL, mallocs the resulting string and stores a pointer to
 * it in `*result'. If `result' is NULL, merely checks whether all
 * characters in the string are feasible for the output character
 * set.
 *
 * Return is nonzero if all characters are OK. If not all
 * characters are OK but `result' is non-NULL, a result _will_
 * still be generated!
 */
static int text_convert(wchar_t *s, char **result) {
    /*
     * FIXME. Currently this is ISO8859-1 only.
     */
    int doing = (result != 0);
    int ok = TRUE;
    char *p = NULL;
    int plen = 0, psize = 0;

    for (; *s; s++) {
	wchar_t c = *s;
	char outc;

	if ((c >= 32 && c <= 126) ||
	    (c >= 160 && c <= 255)) {
	    /* Char is OK. */
	    outc = (char)c;
	} else {
	    /* Char is not OK. */
	    ok = FALSE;
	    outc = 0xBF;	       /* approximate the good old DEC `uh?' */
	}
	if (doing) {
	    if (plen >= psize) {
		psize = plen + 256;
		p = resize(p, psize);
	    }
	    p[plen++] = outc;
	}
    }
    if (doing) {
	p = resize(p, plen+1);
	p[plen] = '\0';
	*result = p;
    }
    return ok;
}

static void text_rdaddwc(rdstringc *rs, word *text, word *end) {
    char *c;

    for (; text && text != end; text = text->next) switch (text->type) {
      case word_HyperLink:
      case word_HyperEnd:
      case word_UpperXref:
      case word_LowerXref:
      case word_XrefEnd:
      case word_IndexRef:
	break;

      case word_Normal:
      case word_Emph:
      case word_Code:
      case word_WeakCode:
      case word_WhiteSpace:
      case word_EmphSpace:
      case word_CodeSpace:
      case word_WkCodeSpace:
      case word_Quote:
      case word_EmphQuote:
      case word_CodeQuote:
      case word_WkCodeQuote:
	assert(text->type != word_CodeQuote &&
	       text->type != word_WkCodeQuote);
	if (towordstyle(text->type) == word_Emph &&
	    (attraux(text->aux) == attr_First ||
	     attraux(text->aux) == attr_Only))
	    rdaddc(rs, '_');	       /* FIXME: configurability */
	else if (towordstyle(text->type) == word_Code &&
		 (attraux(text->aux) == attr_First ||
		  attraux(text->aux) == attr_Only))
	    rdaddc(rs, '`');	       /* FIXME: configurability */
	if (removeattr(text->type) == word_Normal) {
	    if (text_convert(text->text, &c))
		rdaddsc(rs, c);
	    else
		text_rdaddwc(rs, text->alt, NULL);
	    sfree(c);
	} else if (removeattr(text->type) == word_WhiteSpace) {
	    rdaddc(rs, ' ');
	} else if (removeattr(text->type) == word_Quote) {
	    rdaddc(rs, quoteaux(text->aux) == quote_Open ? '`' : '\'');
				       /* FIXME: configurability */
	}
	if (towordstyle(text->type) == word_Emph &&
	    (attraux(text->aux) == attr_Last ||
	     attraux(text->aux) == attr_Only))
	    rdaddc(rs, '_');	       /* FIXME: configurability */
	else if (towordstyle(text->type) == word_Code &&
		 (attraux(text->aux) == attr_Last ||
		  attraux(text->aux) == attr_Only))
	    rdaddc(rs, '\'');	       /* FIXME: configurability */
	break;
    }
}

static int text_width(word *);

static int text_width_list(word *text) {
    int w = 0;
    while (text) {
	w += text_width(text);
	text = text->next;
    }
    return w;
}

static int text_width(word *text) {
    switch (text->type) {
      case word_HyperLink:
      case word_HyperEnd:
      case word_UpperXref:
      case word_LowerXref:
      case word_XrefEnd:
      case word_IndexRef:
	return 0;

      case word_Normal:
      case word_Emph:
      case word_Code:
      case word_WeakCode:
	return (((text->type == word_Emph ||
		  text->type == word_Code)
		 ? (attraux(text->aux) == attr_Only ? 2 :
		    attraux(text->aux) == attr_Always ? 0 : 1)
		 : 0) +
		(text_convert(text->text, NULL) ?
		 ustrlen(text->text) :
		 text_width_list(text->alt)));

      case word_WhiteSpace:
      case word_EmphSpace:
      case word_CodeSpace:
      case word_WkCodeSpace:
      case word_Quote:
      case word_EmphQuote:
      case word_CodeQuote:
      case word_WkCodeQuote:
	assert(text->type != word_CodeQuote &&
	       text->type != word_WkCodeQuote);
	return (((towordstyle(text->type) == word_Emph ||
		  towordstyle(text->type) == word_Code)
		 ? (attraux(text->aux) == attr_Only ? 2 :
		    attraux(text->aux) == attr_Always ? 0 : 1)
		 : 0) + 1);
    }
    return 0;			       /* should never happen */
}

static void text_heading(FILE *fp, word *tprefix, word *nprefix, word *text,
			 alignstruct align, int indent, int width) {
    rdstringc t = { 0, 0, NULL };
    int margin, length;
    int firstlinewidth, wrapwidth;
    wrappedline *wrapping, *p;

    if (align.just_numbers && nprefix) {
	text_rdaddwc(&t, nprefix, NULL);
	rdaddc(&t, ' ');	       /* FIXME: as below */
    } else if (!align.just_numbers && tprefix) {
	text_rdaddwc(&t, tprefix, NULL);
	rdaddsc(&t, ": ");	       /* FIXME: configurability */
    }
    margin = length = (t.text ? strlen(t.text) : 0);

    if (align.align == LEFTPLUS) {
	margin = indent - margin;
	if (margin < 0) margin = 0;
	firstlinewidth = indent + width - margin - length;
	wrapwidth = width;
    } else if (align.align == LEFT || align.align == CENTRE) {
	margin = 0;
	firstlinewidth = indent + width - length;
	wrapwidth = indent + width;
    }

    wrapping = wrap_para(text, firstlinewidth, wrapwidth, text_width);
    for (p = wrapping; p; p = p->next) {
	text_rdaddwc(&t, p->begin, p->end);
	length = (t.text ? strlen(t.text) : 0);
	if (align.align == CENTRE) {
	    margin = (indent + width - length)/2;
	    if (margin < 0) margin = 0;
	}
	fprintf(fp, "%*s%s\n", margin, "", t.text);
	if (align.underline != L'\0') {
	    char *u, uc;
	    wchar_t uw[2];
	    uw[0] = align.underline; uw[1] = L'\0';
	    text_convert(uw, &u);
	    uc = u[0];
	    sfree(u);
	    fprintf(fp, "%*s", margin, "");
	    while (length--)
		putc(uc, fp);
	    putc('\n', fp);
	}
	if (align.align == LEFTPLUS)
	    margin = indent;
	else
	    margin = 0;
	sfree(t.text);
	t = empty_rdstringc;
    }
    wrap_free(wrapping);
    putc('\n', fp);

    sfree(t.text);
}

static void text_rule(FILE *fp, int indent, int width) {
    while (indent--) putc(' ', fp);
    while (width--) putc('-', fp);     /* FIXME: configurability! */
    putc('\n', fp);
    putc('\n', fp);
}

static void text_para(FILE *fp, word *prefix, char *prefixextra, word *text,
		      int indent, int extraindent, int width) {
    wrappedline *wrapping, *p;
    rdstringc pfx = { 0, 0, NULL };
    int e;
    int firstlinewidth = width;

    if (prefix) {
	text_rdaddwc(&pfx, prefix, NULL);
	if (prefixextra)
	    rdaddsc(&pfx, prefixextra);
	fprintf(fp, "%*s%s", indent, "", pfx.text);
	/* If the prefix is too long, shorten the first line to fit. */
	e = extraindent - strlen(pfx.text);
	if (e < 0) {
	    firstlinewidth += e;       /* this decreases it, since e < 0 */
	    if (firstlinewidth < 0) {
		e = indent + extraindent;
		firstlinewidth = width;
		fprintf(fp, "\n");
	    } else
		e = 0;
	}
	sfree(pfx.text);
    } else
	e = indent + extraindent;

    wrapping = wrap_para(text, firstlinewidth, width, text_width);
    for (p = wrapping; p; p = p->next) {
	rdstringc t = { 0, 0, NULL };
	text_rdaddwc(&t, p->begin, p->end);
	fprintf(fp, "%*s%s\n", e, "", t.text);
	e = indent + extraindent;
	sfree(t.text);
    }
    wrap_free(wrapping);
    putc('\n', fp);
}

static void text_codepara(FILE *fp, word *text, int indent, int width) {
    for (; text; text = text->next) if (text->type == word_WeakCode) {
	char *c;
	text_convert(text->text, &c);
	if (strlen(c) > (size_t)width) {
	    /* FIXME: warn */
	}
	fprintf(fp, "%*s%s\n", indent, "", c);
	sfree(c);
    }

    putc('\n', fp);
}

static void text_versionid(FILE *fp, word *text) {
    rdstringc t = { 0, 0, NULL };

    rdaddc(&t, '[');		       /* FIXME: configurability */
    text_rdaddwc(&t, text, NULL);
    rdaddc(&t, ']');		       /* FIXME: configurability */

    fprintf(fp, "%s\n", t.text);
    sfree(t.text);
}
Commit	Line	Data
d7482997	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	} alignstruct;
	16
	17	typedef struct {
	18	int indent, indent_code;
	19	int listindentbefore, listindentafter;
	20	int width;
	21	alignstruct atitle, achapter, *asect;
	22	int nasect;
	23	int include_version_id;
	24	int indent_preambles;
	25	word bullet;
	26	} textconfig;
	27
	28	static int text_convert(wchar_t , char *);
	29
	30	static void text_heading(FILE , word , word , word , alignstruct, int,int);
	31	static void text_rule(FILE *, int, int);
	32	static void text_para(FILE , word , char , word , int, int, int);
	33	static void text_codepara(FILE , word , int, int);
	34	static void text_versionid(FILE , word );
	35
	36	static alignment utoalign(wchar_t *p) {
	37	if (!ustricmp(p, L"centre") \|\| !ustricmp(p, L"center"))
	38	return CENTRE;
	39	if (!ustricmp(p, L"leftplus"))
	40	return LEFTPLUS;
	41	return LEFT;
	42	}
	43
	44	static textconfig text_configure(paragraph *source) {
	45	textconfig ret;
	46
	47	/*
	48	* Non-negotiables.
	49	*/
	50	ret.bullet.next = NULL;
	51	ret.bullet.alt = NULL;
	52	ret.bullet.type = word_Normal;
	53	ret.atitle.just_numbers = FALSE; /* ignored */
	54
	55	/*
	56	* Defaults.
	57	*/
	58	ret.indent = 7;
	59	ret.indent_code = 2;
	60	ret.listindentbefore = 1;
	61	ret.listindentafter = 3;
	62	ret.width = 68;
	63	ret.atitle.align = CENTRE;
	64	ret.atitle.underline = L'=';
65	ret.achapter.align = LEFT;
66	ret.achapter.just_numbers = FALSE;
67	ret.achapter.underline = L'-';
68	ret.nasect = 1;
69	ret.asect = mknewa(alignstruct, ret.nasect);
70	ret.asect[0].align = LEFTPLUS;
71	ret.asect[0].just_numbers = TRUE;
72	ret.asect[0].underline = L'\0';
73	ret.include_version_id = TRUE;
74	ret.indent_preambles = FALSE;
75	ret.bullet.text = ustrdup(L"-");
76
77	for (; source; source = source->next) {
78	if (source->type == para_Config) {
79	if (!ustricmp(source->keyword, L"text-indent")) {
80	ret.indent = utoi(uadv(source->keyword));
81	} else if (!ustricmp(source->keyword, L"text-indent-code")) {
82	ret.indent_code = utoi(uadv(source->keyword));
83	} else if (!ustricmp(source->keyword, L"text-width")) {
84	ret.width = utoi(uadv(source->keyword));
85	} else if (!ustricmp(source->keyword, L"text-list-indent")) {
86	ret.listindentbefore = utoi(uadv(source->keyword));
87	} else if (!ustricmp(source->keyword, L"text-listitem-indent")) {
88	ret.listindentafter = utoi(uadv(source->keyword));
89	} else if (!ustricmp(source->keyword, L"text-chapter-align")) {
90	ret.achapter.align = utoalign(uadv(source->keyword));
91	} else if (!ustricmp(source->keyword, L"text-chapter-underline")) {
92	ret.achapter.underline = *uadv(source->keyword);
93	} else if (!ustricmp(source->keyword, L"text-chapter-numeric")) {
c83c6495	94	ret.achapter.just_numbers = utob(uadv(source->keyword));
d7482997	95	} else if (!ustricmp(source->keyword, L"text-section-align")) {
	96	wchar_t *p = uadv(source->keyword);
	97	int n = 0;
	98	if (uisdigit(*p)) {
	99	n = utoi(p);
	100	p = uadv(p);
	101	}
	102	if (n >= ret.nasect) {
	103	int i;
	104	ret.asect = resize(ret.asect, n+1);
	105	for (i = ret.nasect; i <= n; i++)
	106	ret.asect[i] = ret.asect[ret.nasect-1];
	107	ret.nasect = n+1;
	108	}
	109	ret.asect[n].align = utoalign(p);
	110	} else if (!ustricmp(source->keyword, L"text-section-underline")) {
	111	wchar_t *p = uadv(source->keyword);
	112	int n = 0;
	113	if (uisdigit(*p)) {
	114	n = utoi(p);
	115	p = uadv(p);
	116	}
	117	if (n >= ret.nasect) {
	118	int i;
	119	ret.asect = resize(ret.asect, n+1);
	120	for (i = ret.nasect; i <= n; i++)
	121	ret.asect[i] = ret.asect[ret.nasect-1];
	122	ret.nasect = n+1;
	123	}
	124	ret.asect[n].underline = *p;
	125	} else if (!ustricmp(source->keyword, L"text-section-numeric")) {
	126	wchar_t *p = uadv(source->keyword);
	127	int n = 0;
	128	if (uisdigit(*p)) {
	129	n = utoi(p);
	130	p = uadv(p);
	131	}
	132	if (n >= ret.nasect) {
	133	int i;
	134	ret.asect = resize(ret.asect, n+1);
	135	for (i = ret.nasect; i <= n; i++)
	136	ret.asect[i] = ret.asect[ret.nasect-1];
	137	ret.nasect = n+1;
	138	}
	139	ret.asect[n].just_numbers = utob(p);
	140	} else if (!ustricmp(source->keyword, L"text-title-align")) {
	141	ret.atitle.align = utoalign(uadv(source->keyword));
	142	} else if (!ustricmp(source->keyword, L"text-title-underline")) {
	143	ret.atitle.underline = *uadv(source->keyword);
	144	} else if (!ustricmp(source->keyword, L"text-versionid")) {
	145	ret.include_version_id = utob(uadv(source->keyword));
	146	} else if (!ustricmp(source->keyword, L"text-indent-preamble")) {
	147	ret.indent_preambles = utob(uadv(source->keyword));
	148	} else if (!ustricmp(source->keyword, L"text-bullet")) {
	149	ret.bullet.text = uadv(source->keyword);
	150	}
	151	}
	152	}
	153
	154	return ret;
	155	}
	156
	157	void text_backend(paragraph sourceform, keywordlist keywords,
	158	indexdata *idx) {
159	paragraph *p;
160	textconfig conf;
161	word prefix, body, *wp;
162	word spaceword;
163	FILE *fp;
164	char *prefixextra;
165	int indentb, indenta;
166
167	IGNORE(keywords); /* we don't happen to need this */
168	IGNORE(idx); /* or this */
169
170	conf = text_configure(sourceform);
171
172	/*
173	* Determine the output file name, and open the output file
174	*
175	* FIXME: want configurable output file names here. For the
176	* moment, we'll just call it `output.txt'.
177	*/
178	fp = fopen("output.txt", "w");
179	if (!fp) {
180	error(err_cantopenw, "output.txt");
181	return;
182	}
183
184	/* Do the title */
185	for (p = sourceform; p; p = p->next)
186	if (p->type == para_Title)
187	text_heading(fp, NULL, NULL, p->words,
188	conf.atitle, conf.indent, conf.width);
189
190	/* Do the preamble and copyright */
191	for (p = sourceform; p; p = p->next)
192	if (p->type == para_Preamble)
193	text_para(fp, NULL, NULL, p->words,
194	conf.indent_preambles ? conf.indent : 0, 0,
195	conf.width + (conf.indent_preambles ? 0 : conf.indent));
196	for (p = sourceform; p; p = p->next)
197	if (p->type == para_Copyright)
198	text_para(fp, NULL, NULL, p->words,
199	conf.indent_preambles ? conf.indent : 0, 0,
200	conf.width + (conf.indent_preambles ? 0 : conf.indent));
201
202	/* Do the main document */
203	for (p = sourceform; p; p = p->next) switch (p->type) {
204
205	/*
206	* Things we ignore because we've already processed them or
207	* aren't going to touch them in this pass.
208	*/
209	case para_IM:
210	case para_BR:
211	case para_Biblio: /* only touch BiblioCited */
212	case para_VersionID:
213	case para_Copyright:
214	case para_Preamble:
215	case para_NoCite:
216	case para_Title:
217	break;
218
219	/*
220	* Chapter titles.
221	*/
222	case para_Chapter:
223	case para_Appendix:
224	case para_UnnumberedChapter:
225	text_heading(fp, p->kwtext, p->kwtext2, p->words,
226	conf.achapter, conf.indent, conf.width);
227	break;
228
229	case para_Heading:
230	case para_Subsect:
231	text_heading(fp, p->kwtext, p->kwtext2, p->words,
232	conf.asect[p->aux>=conf.nasect ? conf.nasect-1 : p->aux],
233	conf.indent, conf.width);
234	break;
235
236	case para_Rule:
237	text_rule(fp, conf.indent, conf.width);
238	break;
239
240	case para_Normal:
241	case para_BiblioCited:
242	case para_Bullet:
243	case para_NumberedList:
244	if (p->type == para_Bullet) {
245	prefix = &conf.bullet;
246	prefixextra = NULL;
247	indentb = conf.listindentbefore;
248	indenta = conf.listindentafter;
249	} else if (p->type == para_NumberedList) {
250	prefix = p->kwtext;
251	prefixextra = "."; /* FIXME: configurability */
252	indentb = conf.listindentbefore;
253	indenta = conf.listindentafter;
254	} else {
255	prefix = NULL;
256	prefixextra = NULL;
257	indentb = indenta = 0;
258	}
259	if (p->type == para_BiblioCited) {
260	body = dup_word_list(p->kwtext);
261	for (wp = body; wp->next; wp = wp->next);
262	wp->next = &spaceword;
263	spaceword.next = p->words;
264	spaceword.alt = NULL;
265	spaceword.type = word_WhiteSpace;
266	spaceword.text = NULL;
267	} else {
268	wp = NULL;
269	body = p->words;
270	}
271	text_para(fp, prefix, prefixextra, body,
c83c6495	272	conf.indent + indentb, indenta,
c83c6495	273	conf.width - indentb - indenta);
d7482997	274	if (wp) {
	275	wp->next = NULL;
	276	free_word_list(body);
	277	}
	278	break;
	279
	280	case para_Code:
	281	text_codepara(fp, p->words, conf.indent + conf.indent_code, conf.width - 2 * conf.indent_code);
	282	break;
	283	}
	284
	285	/* Do the version ID */
	286	if (conf.include_version_id) {
	287	for (p = sourceform; p; p = p->next)
	288	if (p->type == para_VersionID)
	289	text_versionid(fp, p->words);
	290	}
	291
	292	/*
	293	* Tidy up
	294	*/
	295	fclose(fp);
	296	sfree(conf.bullet.text);
	297	}
	298
	299	/*
	300	* Convert a wide string into a string of chars. If `result' is
	301	* non-NULL, mallocs the resulting string and stores a pointer to
	302	* it in `*result'. If `result' is NULL, merely checks whether all
	303	* characters in the string are feasible for the output character
	304	* set.
	305	*
	306	* Return is nonzero if all characters are OK. If not all
	307	* characters are OK but `result' is non-NULL, a result _will_
	308	* still be generated!
	309	*/
	310	static int text_convert(wchar_t s, char *result) {
	311	/*
	312	* FIXME. Currently this is ISO8859-1 only.
	313	*/
	314	int doing = (result != 0);
	315	int ok = TRUE;
	316	char *p = NULL;
	317	int plen = 0, psize = 0;
	318
	319	for (; *s; s++) {
	320	wchar_t c = *s;
	321	char outc;
	322
	323	if ((c >= 32 && c <= 126) \|\|
	324	(c >= 160 && c <= 255)) {
	325	/* Char is OK. */
	326	outc = (char)c;
	327	} else {
	328	/* Char is not OK. */
	329	ok = FALSE;
	330	outc = 0xBF; /* approximate the good old DEC `uh?' */
	331	}
	332	if (doing) {
	333	if (plen >= psize) {
	334	psize = plen + 256;
	335	p = resize(p, psize);
	336	}
	337	p[plen++] = outc;
338	}
339	}
340	if (doing) {
341	p = resize(p, plen+1);
342	p[plen] = '\0';
343	*result = p;
344	}
345	return ok;
346	}
347
348	static void text_rdaddwc(rdstringc rs, word text, word *end) {
349	char *c;
350
351	for (; text && text != end; text = text->next) switch (text->type) {
352	case word_HyperLink:
353	case word_HyperEnd:
354	case word_UpperXref:
355	case word_LowerXref:
356	case word_XrefEnd:
357	case word_IndexRef:
358	break;
359
360	case word_Normal:
361	case word_Emph:
362	case word_Code:
363	case word_WeakCode:
364	case word_WhiteSpace:
365	case word_EmphSpace:
366	case word_CodeSpace:
367	case word_WkCodeSpace:
368	case word_Quote:
369	case word_EmphQuote:
370	case word_CodeQuote:
371	case word_WkCodeQuote:
372	assert(text->type != word_CodeQuote &&
373	text->type != word_WkCodeQuote);
374	if (towordstyle(text->type) == word_Emph &&
375	(attraux(text->aux) == attr_First \|\|
376	attraux(text->aux) == attr_Only))
377	rdaddc(rs, '_'); /* FIXME: configurability */
378	else if (towordstyle(text->type) == word_Code &&
379	(attraux(text->aux) == attr_First \|\|
380	attraux(text->aux) == attr_Only))
381	rdaddc(rs, '`'); /* FIXME: configurability */
382	if (removeattr(text->type) == word_Normal) {
383	if (text_convert(text->text, &c))
384	rdaddsc(rs, c);
385	else
386	text_rdaddwc(rs, text->alt, NULL);
387	sfree(c);
388	} else if (removeattr(text->type) == word_WhiteSpace) {
389	rdaddc(rs, ' ');
390	} else if (removeattr(text->type) == word_Quote) {
391	rdaddc(rs, quoteaux(text->aux) == quote_Open ? '`' : '\'');
392	/* FIXME: configurability */
393	}
394	if (towordstyle(text->type) == word_Emph &&
395	(attraux(text->aux) == attr_Last \|\|
396	attraux(text->aux) == attr_Only))
397	rdaddc(rs, '_'); /* FIXME: configurability */
398	else if (towordstyle(text->type) == word_Code &&
399	(attraux(text->aux) == attr_Last \|\|
400	attraux(text->aux) == attr_Only))
401	rdaddc(rs, '\''); /* FIXME: configurability */
402	break;
403	}
404	}
405
406	static int text_width(word *);
407
408	static int text_width_list(word *text) {
409	int w = 0;
410	while (text) {
411	w += text_width(text);
412	text = text->next;
413	}
414	return w;
415	}
416
417	static int text_width(word *text) {
418	switch (text->type) {
419	case word_HyperLink:
420	case word_HyperEnd:
421	case word_UpperXref:
422	case word_LowerXref:
423	case word_XrefEnd:
424	case word_IndexRef:
425	return 0;
426
427	case word_Normal:
428	case word_Emph:
429	case word_Code:
430	case word_WeakCode:
431	return (((text->type == word_Emph \|\|
432	text->type == word_Code)
433	? (attraux(text->aux) == attr_Only ? 2 :
434	attraux(text->aux) == attr_Always ? 0 : 1)
435	: 0) +
436	(text_convert(text->text, NULL) ?
437	ustrlen(text->text) :
438	text_width_list(text->alt)));
439
440	case word_WhiteSpace:
441	case word_EmphSpace:
442	case word_CodeSpace:
443	case word_WkCodeSpace:
444	case word_Quote:
445	case word_EmphQuote:
446	case word_CodeQuote:
447	case word_WkCodeQuote:
448	assert(text->type != word_CodeQuote &&
449	text->type != word_WkCodeQuote);
450	return (((towordstyle(text->type) == word_Emph \|\|
451	towordstyle(text->type) == word_Code)
452	? (attraux(text->aux) == attr_Only ? 2 :
453	attraux(text->aux) == attr_Always ? 0 : 1)
454	: 0) + 1);
455	}
456	return 0; /* should never happen */
457	}
458
459	static void text_heading(FILE fp, word tprefix, word nprefix, word text,
460	alignstruct align, int indent, int width) {
461	rdstringc t = { 0, 0, NULL };
462	int margin, length;
463	int firstlinewidth, wrapwidth;
464	wrappedline wrapping, p;
465
466	if (align.just_numbers && nprefix) {
467	text_rdaddwc(&t, nprefix, NULL);
468	rdaddc(&t, ' '); /* FIXME: as below */
469	} else if (!align.just_numbers && tprefix) {
470	text_rdaddwc(&t, tprefix, NULL);
471	rdaddsc(&t, ": "); /* FIXME: configurability */
472	}
473	margin = length = (t.text ? strlen(t.text) : 0);
474
475	if (align.align == LEFTPLUS) {
476	margin = indent - margin;
477	if (margin < 0) margin = 0;
478	firstlinewidth = indent + width - margin - length;
479	wrapwidth = width;
480	} else if (align.align == LEFT \|\| align.align == CENTRE) {
481	margin = 0;
482	firstlinewidth = indent + width - length;
483	wrapwidth = indent + width;
484	}
485
486	wrapping = wrap_para(text, firstlinewidth, wrapwidth, text_width);
487	for (p = wrapping; p; p = p->next) {
488	text_rdaddwc(&t, p->begin, p->end);
489	length = (t.text ? strlen(t.text) : 0);
490	if (align.align == CENTRE) {
491	margin = (indent + width - length)/2;
492	if (margin < 0) margin = 0;
493	}
494	fprintf(fp, "%*s%s\n", margin, "", t.text);
495	if (align.underline != L'\0') {
496	char *u, uc;
497	wchar_t uw[2];
498	uw[0] = align.underline; uw[1] = L'\0';
499	text_convert(uw, &u);
500	uc = u[0];
501	sfree(u);
502	fprintf(fp, "%*s", margin, "");
503	while (length--)
504	putc(uc, fp);
505	putc('\n', fp);
506	}
507	if (align.align == LEFTPLUS)
508	margin = indent;
509	else
510	margin = 0;
511	sfree(t.text);
512	t = empty_rdstringc;
513	}
514	wrap_free(wrapping);
515	putc('\n', fp);
516
517	sfree(t.text);
518	}
519
520	static void text_rule(FILE *fp, int indent, int width) {
521	while (indent--) putc(' ', fp);
522	while (width--) putc('-', fp); /* FIXME: configurability! */
523	putc('\n', fp);
524	putc('\n', fp);
525	}
526
527	static void text_para(FILE fp, word prefix, char prefixextra, word text,
528	int indent, int extraindent, int width) {
529	wrappedline wrapping, p;
530	rdstringc pfx = { 0, 0, NULL };
531	int e;
532	int firstlinewidth = width;
533
534	if (prefix) {
535	text_rdaddwc(&pfx, prefix, NULL);
536	if (prefixextra)
537	rdaddsc(&pfx, prefixextra);
538	fprintf(fp, "%*s%s", indent, "", pfx.text);
c83c6495	539	/* If the prefix is too long, shorten the first line to fit. */
d7482997	540	e = extraindent - strlen(pfx.text);
d7482997	541	if (e < 0) {
c83c6495	542	firstlinewidth += e; /* this decreases it, since e < 0 */
d7482997	543	if (firstlinewidth < 0) {
	544	e = indent + extraindent;
	545	firstlinewidth = width;
	546	fprintf(fp, "\n");
c83c6495	547	} else
c83c6495	548	e = 0;
d7482997	549	}
	550	sfree(pfx.text);
	551	} else
	552	e = indent + extraindent;
	553
	554	wrapping = wrap_para(text, firstlinewidth, width, text_width);
	555	for (p = wrapping; p; p = p->next) {
	556	rdstringc t = { 0, 0, NULL };
	557	text_rdaddwc(&t, p->begin, p->end);
	558	fprintf(fp, "%*s%s\n", e, "", t.text);
	559	e = indent + extraindent;
	560	sfree(t.text);
	561	}
	562	wrap_free(wrapping);
	563	putc('\n', fp);
	564	}
	565
	566	static void text_codepara(FILE fp, word text, int indent, int width) {
	567	for (; text; text = text->next) if (text->type == word_WeakCode) {
	568	char *c;
	569	text_convert(text->text, &c);
	570	if (strlen(c) > (size_t)width) {
	571	/* FIXME: warn */
	572	}
	573	fprintf(fp, "%*s%s\n", indent, "", c);
	574	sfree(c);
	575	}
	576
	577	putc('\n', fp);
	578	}
	579
	580	static void text_versionid(FILE fp, word text) {
	581	rdstringc t = { 0, 0, NULL };
	582
	583	rdaddc(&t, '['); /* FIXME: configurability */
	584	text_rdaddwc(&t, text, NULL);
	585	rdaddc(&t, ']'); /* FIXME: configurability */
	586
	587	fprintf(fp, "%s\n", t.text);
	588	sfree(t.text);
	589	}