[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;
    wchar_t *number_suffix;
} 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;
    char *filename;
} 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.number_suffix = L": ";
    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].number_suffix = L" ";
    ret.asect[0].underline = L'\0';
    ret.include_version_id = TRUE;
    ret.indent_preambles = FALSE;
    ret.bullet.text = L"-";
    ret.filename = dupstr("output.txt");

    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-filename")) {
		sfree(ret.filename);
		ret.filename = utoa_dup(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-chapter-suffix")) {
		ret.achapter.number_suffix = 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-section-suffix")) {
		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].number_suffix = 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;
}

paragraph *text_config_filename(char *filename)
{
    paragraph *p;
    wchar_t *ufilename, *up;
    int len;

    p = mknew(paragraph);
    memset(p, 0, sizeof(*p));
    p->type = para_Config;
    p->next = NULL;
    p->fpos.filename = "<command line>";
    p->fpos.line = p->fpos.col = -1;

    ufilename = ufroma_dup(filename);
    len = ustrlen(ufilename) + 2 + lenof(L"text-filename");
    p->keyword = mknewa(wchar_t, len);
    up = p->keyword;
    ustrcpy(up, L"text-filename");
    up = uadv(up);
    ustrcpy(up, ufilename);
    up = uadv(up);
    *up = L'\0';
    assert(up - p->keyword < len);
    sfree(ufilename);

    return p;
}

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

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

    conf = text_configure(sourceform);

    /*
     * Open the output file.
     */
    fp = fopen(conf.filename, "w");
    if (!fp) {
	error(err_cantopenw, conf.filename);
	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);

    nestindent = conf.listindentbefore + conf.listindentafter;
    nesting = (conf.indent_preambles ? 0 : -conf.indent);

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

      case para_QuotePush:
	nesting += 2;
	break;
      case para_QuotePop:
	nesting -= 2;
	assert(nesting >= 0);
	break;

      case para_LcontPush:
	nesting += nestindent;
	break;
      case para_LcontPop:
	nesting -= nestindent;
	assert(nesting >= 0);
	break;

	/*
	 * 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_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);
	nesting = 0;
	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 + nesting, conf.width - nesting);
	break;

      case para_Normal:
      case para_Copyright:
      case para_DescribedThing:
      case para_Description:
      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 if (p->type == para_Description) {
	    prefix = NULL;
	    prefixextra = NULL;
	    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 + nesting + indentb, indenta,
		  conf.width - nesting - indentb - indenta);
	if (wp) {
	    wp->next = NULL;
	    free_word_list(body);
	}
	break;

      case para_Code:
	text_codepara(fp, p->words,
		      conf.indent + nesting + conf.indent_code,
		      conf.width - nesting - 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.asect);
    sfree(conf.filename);
}

/*
 * 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) {
	char *c;
	text_rdaddwc(&t, nprefix, NULL);
	if (text_convert(align.number_suffix, &c)) {
	    rdaddsc(&t, c);
	    sfree(c);
	}
    } else if (!align.just_numbers && tprefix) {
	char *c;
	text_rdaddwc(&t, tprefix, NULL);
	if (text_convert(align.number_suffix, &c)) {
	    rdaddsc(&t, c);
	    sfree(c);
	}
    }
    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;
63223c78	15	wchar_t *number_suffix;
d7482997	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;
50d6b4bd	27	char *filename;
d7482997	28	} textconfig;
	29
	30	static int text_convert(wchar_t , char *);
	31
	32	static void text_heading(FILE , word , word , word , alignstruct, int,int);
	33	static void text_rule(FILE *, int, int);
	34	static void text_para(FILE , word , char , word , int, int, int);
	35	static void text_codepara(FILE , word , int, int);
	36	static void text_versionid(FILE , word );
	37
	38	static alignment utoalign(wchar_t *p) {
	39	if (!ustricmp(p, L"centre") \|\| !ustricmp(p, L"center"))
	40	return CENTRE;
	41	if (!ustricmp(p, L"leftplus"))
	42	return LEFTPLUS;
	43	return LEFT;
	44	}
	45
	46	static textconfig text_configure(paragraph *source) {
	47	textconfig ret;
	48
	49	/*
	50	* Non-negotiables.
	51	*/
	52	ret.bullet.next = NULL;
	53	ret.bullet.alt = NULL;
	54	ret.bullet.type = word_Normal;
	55	ret.atitle.just_numbers = FALSE; /* ignored */
	56
	57	/*
	58	* Defaults.
	59	*/
	60	ret.indent = 7;
	61	ret.indent_code = 2;
	62	ret.listindentbefore = 1;
	63	ret.listindentafter = 3;
	64	ret.width = 68;
	65	ret.atitle.align = CENTRE;
	66	ret.atitle.underline = L'=';
	67	ret.achapter.align = LEFT;
	68	ret.achapter.just_numbers = FALSE;
e5e6bf9d	69	ret.achapter.number_suffix = L": ";
d7482997	70	ret.achapter.underline = L'-';
	71	ret.nasect = 1;
	72	ret.asect = mknewa(alignstruct, ret.nasect);
	73	ret.asect[0].align = LEFTPLUS;
	74	ret.asect[0].just_numbers = TRUE;
e5e6bf9d	75	ret.asect[0].number_suffix = L" ";
d7482997	76	ret.asect[0].underline = L'\0';
	77	ret.include_version_id = TRUE;
	78	ret.indent_preambles = FALSE;
e5e6bf9d	79	ret.bullet.text = L"-";
50d6b4bd	80	ret.filename = dupstr("output.txt");
d7482997	81
	82	for (; source; source = source->next) {
	83	if (source->type == para_Config) {
	84	if (!ustricmp(source->keyword, L"text-indent")) {
	85	ret.indent = utoi(uadv(source->keyword));
50d6b4bd	86	} else if (!ustricmp(source->keyword, L"text-filename")) {
	87	sfree(ret.filename);
	88	ret.filename = utoa_dup(uadv(source->keyword));
d7482997	89	} else if (!ustricmp(source->keyword, L"text-indent-code")) {
	90	ret.indent_code = utoi(uadv(source->keyword));
	91	} else if (!ustricmp(source->keyword, L"text-width")) {
	92	ret.width = utoi(uadv(source->keyword));
	93	} else if (!ustricmp(source->keyword, L"text-list-indent")) {
	94	ret.listindentbefore = utoi(uadv(source->keyword));
	95	} else if (!ustricmp(source->keyword, L"text-listitem-indent")) {
	96	ret.listindentafter = utoi(uadv(source->keyword));
	97	} else if (!ustricmp(source->keyword, L"text-chapter-align")) {
	98	ret.achapter.align = utoalign(uadv(source->keyword));
	99	} else if (!ustricmp(source->keyword, L"text-chapter-underline")) {
	100	ret.achapter.underline = *uadv(source->keyword);
	101	} else if (!ustricmp(source->keyword, L"text-chapter-numeric")) {
c83c6495	102	ret.achapter.just_numbers = utob(uadv(source->keyword));
63223c78	103	} else if (!ustricmp(source->keyword, L"text-chapter-suffix")) {
e5e6bf9d	104	ret.achapter.number_suffix = uadv(source->keyword);
d7482997	105	} else if (!ustricmp(source->keyword, L"text-section-align")) {
	106	wchar_t *p = uadv(source->keyword);
	107	int n = 0;
	108	if (uisdigit(*p)) {
	109	n = utoi(p);
	110	p = uadv(p);
	111	}
	112	if (n >= ret.nasect) {
	113	int i;
	114	ret.asect = resize(ret.asect, n+1);
	115	for (i = ret.nasect; i <= n; i++)
	116	ret.asect[i] = ret.asect[ret.nasect-1];
	117	ret.nasect = n+1;
	118	}
	119	ret.asect[n].align = utoalign(p);
	120	} else if (!ustricmp(source->keyword, L"text-section-underline")) {
	121	wchar_t *p = uadv(source->keyword);
	122	int n = 0;
	123	if (uisdigit(*p)) {
	124	n = utoi(p);
	125	p = uadv(p);
	126	}
	127	if (n >= ret.nasect) {
	128	int i;
	129	ret.asect = resize(ret.asect, n+1);
	130	for (i = ret.nasect; i <= n; i++)
	131	ret.asect[i] = ret.asect[ret.nasect-1];
	132	ret.nasect = n+1;
	133	}
	134	ret.asect[n].underline = *p;
	135	} else if (!ustricmp(source->keyword, L"text-section-numeric")) {
	136	wchar_t *p = uadv(source->keyword);
	137	int n = 0;
	138	if (uisdigit(*p)) {
	139	n = utoi(p);
	140	p = uadv(p);
	141	}
	142	if (n >= ret.nasect) {
	143	int i;
	144	ret.asect = resize(ret.asect, n+1);
	145	for (i = ret.nasect; i <= n; i++)
	146	ret.asect[i] = ret.asect[ret.nasect-1];
	147	ret.nasect = n+1;
	148	}
	149	ret.asect[n].just_numbers = utob(p);
63223c78	150	} else if (!ustricmp(source->keyword, L"text-section-suffix")) {
	151	wchar_t *p = uadv(source->keyword);
	152	int n = 0;
	153	if (uisdigit(*p)) {
	154	n = utoi(p);
	155	p = uadv(p);
	156	}
	157	if (n >= ret.nasect) {
	158	int i;
	159	ret.asect = resize(ret.asect, n+1);
e5e6bf9d	160	for (i = ret.nasect; i <= n; i++) {
63223c78	161	ret.asect[i] = ret.asect[ret.nasect-1];
e5e6bf9d	162	}
63223c78	163	ret.nasect = n+1;
63223c78	164	}
e5e6bf9d	165	ret.asect[n].number_suffix = p;
d7482997	166	} else if (!ustricmp(source->keyword, L"text-title-align")) {
	167	ret.atitle.align = utoalign(uadv(source->keyword));
	168	} else if (!ustricmp(source->keyword, L"text-title-underline")) {
	169	ret.atitle.underline = *uadv(source->keyword);
	170	} else if (!ustricmp(source->keyword, L"text-versionid")) {
	171	ret.include_version_id = utob(uadv(source->keyword));
	172	} else if (!ustricmp(source->keyword, L"text-indent-preamble")) {
	173	ret.indent_preambles = utob(uadv(source->keyword));
	174	} else if (!ustricmp(source->keyword, L"text-bullet")) {
	175	ret.bullet.text = uadv(source->keyword);
	176	}
	177	}
	178	}
	179
	180	return ret;
	181	}
	182
ba9c1487	183	paragraph text_config_filename(char filename)
	184	{
	185	paragraph *p;
	186	wchar_t ufilename, up;
	187	int len;
	188
	189	p = mknew(paragraph);
	190	memset(p, 0, sizeof(*p));
	191	p->type = para_Config;
	192	p->next = NULL;
	193	p->fpos.filename = "<command line>";
	194	p->fpos.line = p->fpos.col = -1;
	195
	196	ufilename = ufroma_dup(filename);
	197	len = ustrlen(ufilename) + 2 + lenof(L"text-filename");
	198	p->keyword = mknewa(wchar_t, len);
	199	up = p->keyword;
	200	ustrcpy(up, L"text-filename");
	201	up = uadv(up);
	202	ustrcpy(up, ufilename);
	203	up = uadv(up);
	204	*up = L'\0';
	205	assert(up - p->keyword < len);
	206	sfree(ufilename);
	207
	208	return p;
	209	}
	210
d7482997	211	void text_backend(paragraph sourceform, keywordlist keywords,
e62b3302	212	indexdata *idx) {
d7482997	213	paragraph *p;
	214	textconfig conf;
	215	word prefix, body, *wp;
	216	word spaceword;
	217	FILE *fp;
	218	char *prefixextra;
7136a6c7	219	int nesting, nestindent;
d7482997	220	int indentb, indenta;
	221
	222	IGNORE(keywords); /* we don't happen to need this */
	223	IGNORE(idx); /* or this */
	224
	225	conf = text_configure(sourceform);
	226
	227	/*
50d6b4bd	228	* Open the output file.
d7482997	229	*/
50d6b4bd	230	fp = fopen(conf.filename, "w");
d7482997	231	if (!fp) {
50d6b4bd	232	error(err_cantopenw, conf.filename);
d7482997	233	return;
	234	}
	235
	236	/* Do the title */
	237	for (p = sourceform; p; p = p->next)
	238	if (p->type == para_Title)
	239	text_heading(fp, NULL, NULL, p->words,
	240	conf.atitle, conf.indent, conf.width);
	241
7136a6c7	242	nestindent = conf.listindentbefore + conf.listindentafter;
8902e0ed	243	nesting = (conf.indent_preambles ? 0 : -conf.indent);
7136a6c7	244
d7482997	245	/* Do the main document */
	246	for (p = sourceform; p; p = p->next) switch (p->type) {
	247
2614b01d	248	case para_QuotePush:
	249	nesting += 2;
	250	break;
	251	case para_QuotePop:
	252	nesting -= 2;
	253	assert(nesting >= 0);
	254	break;
	255
7136a6c7	256	case para_LcontPush:
2614b01d	257	nesting += nestindent;
7136a6c7	258	break;
7136a6c7	259	case para_LcontPop:
2614b01d	260	nesting -= nestindent;
2614b01d	261	assert(nesting >= 0);
7136a6c7	262	break;
7136a6c7	263
d7482997	264	/*
	265	* Things we ignore because we've already processed them or
	266	* aren't going to touch them in this pass.
	267	*/
	268	case para_IM:
	269	case para_BR:
	270	case para_Biblio: /* only touch BiblioCited */
	271	case para_VersionID:
d7482997	272	case para_NoCite:
	273	case para_Title:
	274	break;
	275
	276	/*
	277	* Chapter titles.
	278	*/
	279	case para_Chapter:
	280	case para_Appendix:
	281	case para_UnnumberedChapter:
	282	text_heading(fp, p->kwtext, p->kwtext2, p->words,
	283	conf.achapter, conf.indent, conf.width);
8902e0ed	284	nesting = 0;
d7482997	285	break;
	286
	287	case para_Heading:
	288	case para_Subsect:
	289	text_heading(fp, p->kwtext, p->kwtext2, p->words,
	290	conf.asect[p->aux>=conf.nasect ? conf.nasect-1 : p->aux],
	291	conf.indent, conf.width);
	292	break;
	293
	294	case para_Rule:
2614b01d	295	text_rule(fp, conf.indent + nesting, conf.width - nesting);
d7482997	296	break;
	297
	298	case para_Normal:
9057a0a8	299	case para_Copyright:
7136a6c7	300	case para_DescribedThing:
7136a6c7	301	case para_Description:
d7482997	302	case para_BiblioCited:
	303	case para_Bullet:
	304	case para_NumberedList:
	305	if (p->type == para_Bullet) {
	306	prefix = &conf.bullet;
	307	prefixextra = NULL;
	308	indentb = conf.listindentbefore;
	309	indenta = conf.listindentafter;
	310	} else if (p->type == para_NumberedList) {
	311	prefix = p->kwtext;
	312	prefixextra = "."; /* FIXME: configurability */
	313	indentb = conf.listindentbefore;
	314	indenta = conf.listindentafter;
7136a6c7	315	} else if (p->type == para_Description) {
	316	prefix = NULL;
	317	prefixextra = NULL;
	318	indentb = conf.listindentbefore;
	319	indenta = conf.listindentafter;
d7482997	320	} else {
	321	prefix = NULL;
	322	prefixextra = NULL;
	323	indentb = indenta = 0;
	324	}
	325	if (p->type == para_BiblioCited) {
	326	body = dup_word_list(p->kwtext);
	327	for (wp = body; wp->next; wp = wp->next);
	328	wp->next = &spaceword;
	329	spaceword.next = p->words;
	330	spaceword.alt = NULL;
	331	spaceword.type = word_WhiteSpace;
	332	spaceword.text = NULL;
	333	} else {
	334	wp = NULL;
	335	body = p->words;
	336	}
	337	text_para(fp, prefix, prefixextra, body,
2614b01d	338	conf.indent + nesting + indentb, indenta,
2614b01d	339	conf.width - nesting - indentb - indenta);
d7482997	340	if (wp) {
	341	wp->next = NULL;
	342	free_word_list(body);
	343	}
	344	break;
	345
	346	case para_Code:
7136a6c7	347	text_codepara(fp, p->words,
2614b01d	348	conf.indent + nesting + conf.indent_code,
2614b01d	349	conf.width - nesting - 2 * conf.indent_code);
d7482997	350	break;
	351	}
	352
	353	/* Do the version ID */
	354	if (conf.include_version_id) {
	355	for (p = sourceform; p; p = p->next)
	356	if (p->type == para_VersionID)
	357	text_versionid(fp, p->words);
	358	}
	359
	360	/*
	361	* Tidy up
	362	*/
	363	fclose(fp);
e5e6bf9d	364	sfree(conf.asect);
50d6b4bd	365	sfree(conf.filename);
d7482997	366	}
	367
	368	/*
	369	* Convert a wide string into a string of chars. If `result' is
	370	* non-NULL, mallocs the resulting string and stores a pointer to
	371	* it in `*result'. If `result' is NULL, merely checks whether all
	372	* characters in the string are feasible for the output character
	373	* set.
	374	*
	375	* Return is nonzero if all characters are OK. If not all
	376	* characters are OK but `result' is non-NULL, a result _will_
	377	* still be generated!
	378	*/
	379	static int text_convert(wchar_t s, char *result) {
	380	/*
	381	* FIXME. Currently this is ISO8859-1 only.
	382	*/
	383	int doing = (result != 0);
	384	int ok = TRUE;
	385	char *p = NULL;
	386	int plen = 0, psize = 0;
	387
	388	for (; *s; s++) {
	389	wchar_t c = *s;
	390	char outc;
	391
	392	if ((c >= 32 && c <= 126) \|\|
	393	(c >= 160 && c <= 255)) {
	394	/* Char is OK. */
	395	outc = (char)c;
	396	} else {
	397	/* Char is not OK. */
	398	ok = FALSE;
	399	outc = 0xBF; /* approximate the good old DEC `uh?' */
	400	}
	401	if (doing) {
	402	if (plen >= psize) {
	403	psize = plen + 256;
	404	p = resize(p, psize);
	405	}
	406	p[plen++] = outc;
	407	}
	408	}
	409	if (doing) {
	410	p = resize(p, plen+1);
	411	p[plen] = '\0';
	412	*result = p;
	413	}
	414	return ok;
	415	}
	416
	417	static void text_rdaddwc(rdstringc rs, word text, word *end) {
	418	char *c;
	419
	420	for (; text && text != end; text = text->next) switch (text->type) {
	421	case word_HyperLink:
	422	case word_HyperEnd:
	423	case word_UpperXref:
	424	case word_LowerXref:
	425	case word_XrefEnd:
	426	case word_IndexRef:
	427	break;
	428
	429	case word_Normal:
430	case word_Emph:
431	case word_Code:
432	case word_WeakCode:
433	case word_WhiteSpace:
434	case word_EmphSpace:
435	case word_CodeSpace:
436	case word_WkCodeSpace:
437	case word_Quote:
438	case word_EmphQuote:
439	case word_CodeQuote:
440	case word_WkCodeQuote:
441	assert(text->type != word_CodeQuote &&
442	text->type != word_WkCodeQuote);
443	if (towordstyle(text->type) == word_Emph &&
444	(attraux(text->aux) == attr_First \|\|
445	attraux(text->aux) == attr_Only))
446	rdaddc(rs, '_'); /* FIXME: configurability */
447	else if (towordstyle(text->type) == word_Code &&
448	(attraux(text->aux) == attr_First \|\|
449	attraux(text->aux) == attr_Only))
450	rdaddc(rs, '`'); /* FIXME: configurability */
451	if (removeattr(text->type) == word_Normal) {
452	if (text_convert(text->text, &c))
453	rdaddsc(rs, c);
454	else
455	text_rdaddwc(rs, text->alt, NULL);
456	sfree(c);
457	} else if (removeattr(text->type) == word_WhiteSpace) {
458	rdaddc(rs, ' ');
459	} else if (removeattr(text->type) == word_Quote) {
460	rdaddc(rs, quoteaux(text->aux) == quote_Open ? '`' : '\'');
461	/* FIXME: configurability */
462	}
463	if (towordstyle(text->type) == word_Emph &&
464	(attraux(text->aux) == attr_Last \|\|
465	attraux(text->aux) == attr_Only))
466	rdaddc(rs, '_'); /* FIXME: configurability */
467	else if (towordstyle(text->type) == word_Code &&
468	(attraux(text->aux) == attr_Last \|\|
469	attraux(text->aux) == attr_Only))
470	rdaddc(rs, '\''); /* FIXME: configurability */
471	break;
472	}
473	}
474
475	static int text_width(word *);
476
477	static int text_width_list(word *text) {
478	int w = 0;
479	while (text) {
480	w += text_width(text);
481	text = text->next;
482	}
483	return w;
484	}
485
486	static int text_width(word *text) {
487	switch (text->type) {
488	case word_HyperLink:
489	case word_HyperEnd:
490	case word_UpperXref:
491	case word_LowerXref:
492	case word_XrefEnd:
493	case word_IndexRef:
494	return 0;
495
496	case word_Normal:
497	case word_Emph:
498	case word_Code:
499	case word_WeakCode:
500	return (((text->type == word_Emph \|\|
501	text->type == word_Code)
502	? (attraux(text->aux) == attr_Only ? 2 :
503	attraux(text->aux) == attr_Always ? 0 : 1)
504	: 0) +
505	(text_convert(text->text, NULL) ?
506	ustrlen(text->text) :
507	text_width_list(text->alt)));
508
509	case word_WhiteSpace:
510	case word_EmphSpace:
511	case word_CodeSpace:
512	case word_WkCodeSpace:
513	case word_Quote:
514	case word_EmphQuote:
515	case word_CodeQuote:
516	case word_WkCodeQuote:
517	assert(text->type != word_CodeQuote &&
518	text->type != word_WkCodeQuote);
519	return (((towordstyle(text->type) == word_Emph \|\|
520	towordstyle(text->type) == word_Code)
521	? (attraux(text->aux) == attr_Only ? 2 :
522	attraux(text->aux) == attr_Always ? 0 : 1)
523	: 0) + 1);
524	}
525	return 0; /* should never happen */
526	}
527
528	static void text_heading(FILE fp, word tprefix, word nprefix, word text,
529	alignstruct align, int indent, int width) {
530	rdstringc t = { 0, 0, NULL };
531	int margin, length;
532	int firstlinewidth, wrapwidth;
533	wrappedline wrapping, p;
534
535	if (align.just_numbers && nprefix) {
63223c78	536	char *c;
d7482997	537	text_rdaddwc(&t, nprefix, NULL);
63223c78	538	if (text_convert(align.number_suffix, &c)) {
	539	rdaddsc(&t, c);
	540	sfree(c);
	541	}
d7482997	542	} else if (!align.just_numbers && tprefix) {
63223c78	543	char *c;
d7482997	544	text_rdaddwc(&t, tprefix, NULL);
63223c78	545	if (text_convert(align.number_suffix, &c)) {
	546	rdaddsc(&t, c);
	547	sfree(c);
	548	}
d7482997	549	}
	550	margin = length = (t.text ? strlen(t.text) : 0);
	551
	552	if (align.align == LEFTPLUS) {
	553	margin = indent - margin;
	554	if (margin < 0) margin = 0;
	555	firstlinewidth = indent + width - margin - length;
	556	wrapwidth = width;
	557	} else if (align.align == LEFT \|\| align.align == CENTRE) {
	558	margin = 0;
	559	firstlinewidth = indent + width - length;
	560	wrapwidth = indent + width;
	561	}
	562
	563	wrapping = wrap_para(text, firstlinewidth, wrapwidth, text_width);
	564	for (p = wrapping; p; p = p->next) {
	565	text_rdaddwc(&t, p->begin, p->end);
	566	length = (t.text ? strlen(t.text) : 0);
	567	if (align.align == CENTRE) {
	568	margin = (indent + width - length)/2;
	569	if (margin < 0) margin = 0;
	570	}
	571	fprintf(fp, "%*s%s\n", margin, "", t.text);
	572	if (align.underline != L'\0') {
	573	char *u, uc;
	574	wchar_t uw[2];
	575	uw[0] = align.underline; uw[1] = L'\0';
	576	text_convert(uw, &u);
	577	uc = u[0];
	578	sfree(u);
	579	fprintf(fp, "%*s", margin, "");
	580	while (length--)
	581	putc(uc, fp);
	582	putc('\n', fp);
	583	}
	584	if (align.align == LEFTPLUS)
	585	margin = indent;
	586	else
	587	margin = 0;
	588	sfree(t.text);
	589	t = empty_rdstringc;
	590	}
	591	wrap_free(wrapping);
	592	putc('\n', fp);
	593
	594	sfree(t.text);
	595	}
	596
	597	static void text_rule(FILE *fp, int indent, int width) {
	598	while (indent--) putc(' ', fp);
	599	while (width--) putc('-', fp); /* FIXME: configurability! */
	600	putc('\n', fp);
	601	putc('\n', fp);
	602	}
	603
	604	static void text_para(FILE fp, word prefix, char prefixextra, word text,
	605	int indent, int extraindent, int width) {
	606	wrappedline wrapping, p;
	607	rdstringc pfx = { 0, 0, NULL };
	608	int e;
	609	int firstlinewidth = width;
	610
	611	if (prefix) {
	612	text_rdaddwc(&pfx, prefix, NULL);
613	if (prefixextra)
614	rdaddsc(&pfx, prefixextra);
615	fprintf(fp, "%*s%s", indent, "", pfx.text);
c83c6495	616	/* If the prefix is too long, shorten the first line to fit. */
d7482997	617	e = extraindent - strlen(pfx.text);
d7482997	618	if (e < 0) {
c83c6495	619	firstlinewidth += e; /* this decreases it, since e < 0 */
d7482997	620	if (firstlinewidth < 0) {
	621	e = indent + extraindent;
	622	firstlinewidth = width;
	623	fprintf(fp, "\n");
c83c6495	624	} else
c83c6495	625	e = 0;
d7482997	626	}
	627	sfree(pfx.text);
	628	} else
	629	e = indent + extraindent;
	630
	631	wrapping = wrap_para(text, firstlinewidth, width, text_width);
	632	for (p = wrapping; p; p = p->next) {
	633	rdstringc t = { 0, 0, NULL };
	634	text_rdaddwc(&t, p->begin, p->end);
	635	fprintf(fp, "%*s%s\n", e, "", t.text);
	636	e = indent + extraindent;
	637	sfree(t.text);
	638	}
	639	wrap_free(wrapping);
	640	putc('\n', fp);
	641	}
	642
	643	static void text_codepara(FILE fp, word text, int indent, int width) {
	644	for (; text; text = text->next) if (text->type == word_WeakCode) {
	645	char *c;
	646	text_convert(text->text, &c);
	647	if (strlen(c) > (size_t)width) {
	648	/* FIXME: warn */
	649	}
	650	fprintf(fp, "%*s%s\n", indent, "", c);
	651	sfree(c);
	652	}
	653
	654	putc('\n', fp);
	655	}
	656
	657	static void text_versionid(FILE fp, word text) {
	658	rdstringc t = { 0, 0, NULL };
	659
	660	rdaddc(&t, '['); /* FIXME: configurability */
	661	text_rdaddwc(&t, text, NULL);
	662	rdaddc(&t, ']'); /* FIXME: configurability */
	663
	664	fprintf(fp, "%s\n", t.text);
	665	sfree(t.text);
	666	}