[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;
    int charset;
    word bullet;
    char *filename;
} textconfig;

typedef struct {
    FILE *fp;
    int charset;
    charset_state state;
} textfile;

static void text_heading(textfile *, word *, word *, word *, alignstruct,
			 int,int);
static void text_rule(textfile *, int, int);
static void text_para(textfile *, word *, wchar_t *, word *, int, int, int);
static void text_codepara(textfile *, word *, int, int);
static void text_versionid(textfile *, word *);

static void text_output(textfile *, const wchar_t *);
static void text_output_many(textfile *, int, wchar_t);

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");
    ret.charset = CS_ASCII;

    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-charset")) {
		char *csname = utoa_dup(uadv(source->keyword), CS_ASCII);
		ret.charset = charset_from_localenc(csname);
		sfree(csname);
	    } else if (!ustricmp(source->keyword, L"text-filename")) {
		sfree(ret.filename);
		ret.filename = dupstr(adv(source->origkeyword));
	    } 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)
{
    return cmdline_cfg_simple("text-filename", filename, NULL);
}

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

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

    conf = text_configure(sourceform);

    /*
     * Open the output file.
     */
    tf.fp = fopen(conf.filename, "w");
    if (!tf.fp) {
	error(err_cantopenw, conf.filename);
	return;
    }
    tf.charset = conf.charset;
    tf.state = charset_init_state;

    /* Do the title */
    for (p = sourceform; p; p = p->next)
	if (p->type == para_Title)
	    text_heading(&tf, 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(&tf, p->kwtext, p->kwtext2, p->words,
		     conf.achapter, conf.indent, conf.width);
	nesting = 0;
	break;

      case para_Heading:
      case para_Subsect:
	text_heading(&tf, 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(&tf, 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 = L".";	       /* 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(&tf, 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(&tf, 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(&tf, p->words);
    }

    /*
     * Tidy up
     */
    text_output(&tf, NULL);	       /* end charset conversion */
    fclose(tf.fp);
    sfree(conf.asect);
    sfree(conf.filename);
}

static void text_output(textfile *tf, const wchar_t *s)
{
    char buf[256];
    int ret, len;
    const wchar_t **sp;

    if (!s) {
	sp = NULL;
	len = 1;
    } else {
	sp = &s;
	len = ustrlen(s);
    }

    while (len > 0) {
	ret = charset_from_unicode(sp, &len, buf, lenof(buf),
				   tf->charset, &tf->state, NULL);
	if (!sp)
	    len = 0;
	fwrite(buf, 1, ret, tf->fp);
    }
}

static void text_output_many(textfile *tf, int n, wchar_t c)
{
    wchar_t s[2];
    s[0] = c;
    s[1] = L'\0';
    while (n--)
	text_output(tf, s);
}

static void text_rdaddw(int charset, rdstring *rs, word *text, word *end) {
    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))
	    rdadd(rs, L'_');	       /* FIXME: configurability */
	else if (towordstyle(text->type) == word_Code &&
		 (attraux(text->aux) == attr_First ||
		  attraux(text->aux) == attr_Only))
	    rdadd(rs, L'`');	       /* FIXME: configurability */
	if (removeattr(text->type) == word_Normal) {
	    if (cvt_ok(charset, text->text) || !text->alt)
		rdadds(rs, text->text);
	    else
		text_rdaddw(charset, rs, text->alt, NULL);
	} else if (removeattr(text->type) == word_WhiteSpace) {
	    rdadd(rs, L' ');
	} else if (removeattr(text->type) == word_Quote) {
	    rdadd(rs, quoteaux(text->aux) == quote_Open ? L'`' : L'\'');
				       /* FIXME: configurability */
	}
	if (towordstyle(text->type) == word_Emph &&
	    (attraux(text->aux) == attr_Last ||
	     attraux(text->aux) == attr_Only))
	    rdadd(rs, L'_');	       /* FIXME: configurability */
	else if (towordstyle(text->type) == word_Code &&
		 (attraux(text->aux) == attr_Last ||
		  attraux(text->aux) == attr_Only))
	    rdadd(rs, L'\'');	       /* FIXME: configurability */
	break;
    }
}

static int text_width(void *, word *);

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

static int text_width(void *ctx, word *text) {
    int charset = * (int *) ctx;

    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) +
		(cvt_ok(charset, text->text) || !text->alt ?
		 ustrwid(text->text, charset) :
		 text_width_list(ctx, 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(textfile *tf, word *tprefix, word *nprefix,
			 word *text, alignstruct align,
			 int indent, int width) {
    rdstring t = { 0, 0, NULL };
    int margin, length;
    int firstlinewidth, wrapwidth;
    wrappedline *wrapping, *p;

    if (align.just_numbers && nprefix) {
	text_rdaddw(tf->charset, &t, nprefix, NULL);
	rdadds(&t, align.number_suffix);
    } else if (!align.just_numbers && tprefix) {
	text_rdaddw(tf->charset, &t, tprefix, NULL);
	rdadds(&t, align.number_suffix);
    }
    margin = length = t.pos;

    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, &tf->charset, 0);
    for (p = wrapping; p; p = p->next) {
	text_rdaddw(tf->charset, &t, p->begin, p->end);
	length = t.pos;
	if (align.align == CENTRE) {
	    margin = (indent + width - length)/2;
	    if (margin < 0) margin = 0;
	}
	text_output_many(tf, margin, L' ');
	text_output(tf, t.text);
	text_output(tf, L"\n");
	if (align.underline != L'\0') {
	    text_output_many(tf, margin, L' ');
	    text_output_many(tf, length, align.underline);
	    text_output(tf, L"\n");
	}
	if (align.align == LEFTPLUS)
	    margin = indent;
	else
	    margin = 0;
	sfree(t.text);
	t = empty_rdstring;
    }
    wrap_free(wrapping);
    text_output(tf, L"\n");

    sfree(t.text);
}

static void text_rule(textfile *tf, int indent, int width) {
    text_output_many(tf, indent, L' ');
    text_output_many(tf, width, L'-');     /* FIXME: configurability! */
    text_output_many(tf, 2, L'\n');
}

static void text_para(textfile *tf, word *prefix, wchar_t *prefixextra,
		      word *text, int indent, int extraindent, int width) {
    wrappedline *wrapping, *p;
    rdstring pfx = { 0, 0, NULL };
    int e;
    int firstlinewidth = width;

    if (prefix) {
	text_rdaddw(tf->charset, &pfx, prefix, NULL);
	if (prefixextra)
	    rdadds(&pfx, prefixextra);
	text_output_many(tf, indent, L' ');
	text_output(tf, pfx.text);
	/* If the prefix is too long, shorten the first line to fit. */
	e = extraindent - pfx.pos;
	if (e < 0) {
	    firstlinewidth += e;       /* this decreases it, since e < 0 */
	    if (firstlinewidth < 0) {
		e = indent + extraindent;
		firstlinewidth = width;
		text_output(tf, L"\n");
	    } else
		e = 0;
	}
	sfree(pfx.text);
    } else
	e = indent + extraindent;

    wrapping = wrap_para(text, firstlinewidth, width,
			 text_width, &tf->charset, 0);
    for (p = wrapping; p; p = p->next) {
	rdstring t = { 0, 0, NULL };
	text_rdaddw(tf->charset, &t, p->begin, p->end);
	text_output_many(tf, e, L' ');
	text_output(tf, t.text);
	text_output(tf, L"\n");
	e = indent + extraindent;
	sfree(t.text);
    }
    wrap_free(wrapping);
    text_output(tf, L"\n");
}

static void text_codepara(textfile *tf, word *text, int indent, int width) {
    for (; text; text = text->next) if (text->type == word_WeakCode) {
	if (ustrlen(text->text) > width) {
	    /* FIXME: warn */
	}
	text_output_many(tf, indent, L' ');
	text_output(tf, text->text);
	text_output(tf, L"\n");
    }

    text_output(tf, L"\n");
}

static void text_versionid(textfile *tf, word *text) {
    rdstring t = { 0, 0, NULL };

    rdadd(&t, L'[');		       /* FIXME: configurability */
    text_rdaddw(tf->charset, &t, text, NULL);
    rdadd(&t, L']');		       /* FIXME: configurability */
    rdadd(&t, L'\n');

    text_output(tf, 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;
2ac8ceac	26	int charset;
d7482997	27	word bullet;
50d6b4bd	28	char *filename;
d7482997	29	} textconfig;
d7482997	30
2ac8ceac	31	typedef struct {
	32	FILE *fp;
	33	int charset;
	34	charset_state state;
	35	} textfile;
	36
	37	static void text_heading(textfile , word , word , word , alignstruct,
	38	int,int);
	39	static void text_rule(textfile *, int, int);
	40	static void text_para(textfile , word , wchar_t , word , int, int, int);
	41	static void text_codepara(textfile , word , int, int);
	42	static void text_versionid(textfile , word );
d7482997	43
2ac8ceac	44	static void text_output(textfile , const wchar_t );
2ac8ceac	45	static void text_output_many(textfile *, int, wchar_t);
d7482997	46
	47	static alignment utoalign(wchar_t *p) {
	48	if (!ustricmp(p, L"centre") \|\| !ustricmp(p, L"center"))
	49	return CENTRE;
	50	if (!ustricmp(p, L"leftplus"))
	51	return LEFTPLUS;
	52	return LEFT;
	53	}
	54
	55	static textconfig text_configure(paragraph *source) {
	56	textconfig ret;
	57
	58	/*
	59	* Non-negotiables.
	60	*/
	61	ret.bullet.next = NULL;
	62	ret.bullet.alt = NULL;
	63	ret.bullet.type = word_Normal;
	64	ret.atitle.just_numbers = FALSE; /* ignored */
	65
	66	/*
	67	* Defaults.
	68	*/
	69	ret.indent = 7;
	70	ret.indent_code = 2;
	71	ret.listindentbefore = 1;
	72	ret.listindentafter = 3;
	73	ret.width = 68;
	74	ret.atitle.align = CENTRE;
	75	ret.atitle.underline = L'=';
	76	ret.achapter.align = LEFT;
	77	ret.achapter.just_numbers = FALSE;
e5e6bf9d	78	ret.achapter.number_suffix = L": ";
d7482997	79	ret.achapter.underline = L'-';
	80	ret.nasect = 1;
	81	ret.asect = mknewa(alignstruct, ret.nasect);
	82	ret.asect[0].align = LEFTPLUS;
	83	ret.asect[0].just_numbers = TRUE;
e5e6bf9d	84	ret.asect[0].number_suffix = L" ";
d7482997	85	ret.asect[0].underline = L'\0';
	86	ret.include_version_id = TRUE;
	87	ret.indent_preambles = FALSE;
e5e6bf9d	88	ret.bullet.text = L"-";
50d6b4bd	89	ret.filename = dupstr("output.txt");
2ac8ceac	90	ret.charset = CS_ASCII;
d7482997	91
	92	for (; source; source = source->next) {
	93	if (source->type == para_Config) {
	94	if (!ustricmp(source->keyword, L"text-indent")) {
	95	ret.indent = utoi(uadv(source->keyword));
2ac8ceac	96	} else if (!ustricmp(source->keyword, L"text-charset")) {
	97	char *csname = utoa_dup(uadv(source->keyword), CS_ASCII);
	98	ret.charset = charset_from_localenc(csname);
	99	sfree(csname);
50d6b4bd	100	} else if (!ustricmp(source->keyword, L"text-filename")) {
50d6b4bd	101	sfree(ret.filename);
e4ea58f8	102	ret.filename = dupstr(adv(source->origkeyword));
d7482997	103	} else if (!ustricmp(source->keyword, L"text-indent-code")) {
	104	ret.indent_code = utoi(uadv(source->keyword));
	105	} else if (!ustricmp(source->keyword, L"text-width")) {
	106	ret.width = utoi(uadv(source->keyword));
	107	} else if (!ustricmp(source->keyword, L"text-list-indent")) {
	108	ret.listindentbefore = utoi(uadv(source->keyword));
	109	} else if (!ustricmp(source->keyword, L"text-listitem-indent")) {
	110	ret.listindentafter = utoi(uadv(source->keyword));
	111	} else if (!ustricmp(source->keyword, L"text-chapter-align")) {
	112	ret.achapter.align = utoalign(uadv(source->keyword));
	113	} else if (!ustricmp(source->keyword, L"text-chapter-underline")) {
	114	ret.achapter.underline = *uadv(source->keyword);
	115	} else if (!ustricmp(source->keyword, L"text-chapter-numeric")) {
c83c6495	116	ret.achapter.just_numbers = utob(uadv(source->keyword));
63223c78	117	} else if (!ustricmp(source->keyword, L"text-chapter-suffix")) {
e5e6bf9d	118	ret.achapter.number_suffix = uadv(source->keyword);
d7482997	119	} else if (!ustricmp(source->keyword, L"text-section-align")) {
	120	wchar_t *p = uadv(source->keyword);
	121	int n = 0;
	122	if (uisdigit(*p)) {
	123	n = utoi(p);
	124	p = uadv(p);
	125	}
	126	if (n >= ret.nasect) {
	127	int i;
	128	ret.asect = resize(ret.asect, n+1);
	129	for (i = ret.nasect; i <= n; i++)
	130	ret.asect[i] = ret.asect[ret.nasect-1];
	131	ret.nasect = n+1;
	132	}
	133	ret.asect[n].align = utoalign(p);
	134	} else if (!ustricmp(source->keyword, L"text-section-underline")) {
	135	wchar_t *p = uadv(source->keyword);
	136	int n = 0;
	137	if (uisdigit(*p)) {
	138	n = utoi(p);
	139	p = uadv(p);
	140	}
	141	if (n >= ret.nasect) {
	142	int i;
	143	ret.asect = resize(ret.asect, n+1);
	144	for (i = ret.nasect; i <= n; i++)
	145	ret.asect[i] = ret.asect[ret.nasect-1];
	146	ret.nasect = n+1;
	147	}
	148	ret.asect[n].underline = *p;
	149	} else if (!ustricmp(source->keyword, L"text-section-numeric")) {
	150	wchar_t *p = uadv(source->keyword);
	151	int n = 0;
	152	if (uisdigit(*p)) {
	153	n = utoi(p);
	154	p = uadv(p);
	155	}
	156	if (n >= ret.nasect) {
	157	int i;
	158	ret.asect = resize(ret.asect, n+1);
	159	for (i = ret.nasect; i <= n; i++)
	160	ret.asect[i] = ret.asect[ret.nasect-1];
	161	ret.nasect = n+1;
	162	}
	163	ret.asect[n].just_numbers = utob(p);
63223c78	164	} else if (!ustricmp(source->keyword, L"text-section-suffix")) {
	165	wchar_t *p = uadv(source->keyword);
	166	int n = 0;
	167	if (uisdigit(*p)) {
	168	n = utoi(p);
	169	p = uadv(p);
	170	}
	171	if (n >= ret.nasect) {
	172	int i;
	173	ret.asect = resize(ret.asect, n+1);
e5e6bf9d	174	for (i = ret.nasect; i <= n; i++) {
63223c78	175	ret.asect[i] = ret.asect[ret.nasect-1];
e5e6bf9d	176	}
63223c78	177	ret.nasect = n+1;
63223c78	178	}
e5e6bf9d	179	ret.asect[n].number_suffix = p;
d7482997	180	} else if (!ustricmp(source->keyword, L"text-title-align")) {
	181	ret.atitle.align = utoalign(uadv(source->keyword));
	182	} else if (!ustricmp(source->keyword, L"text-title-underline")) {
	183	ret.atitle.underline = *uadv(source->keyword);
	184	} else if (!ustricmp(source->keyword, L"text-versionid")) {
	185	ret.include_version_id = utob(uadv(source->keyword));
	186	} else if (!ustricmp(source->keyword, L"text-indent-preamble")) {
	187	ret.indent_preambles = utob(uadv(source->keyword));
	188	} else if (!ustricmp(source->keyword, L"text-bullet")) {
	189	ret.bullet.text = uadv(source->keyword);
	190	}
	191	}
	192	}
	193
	194	return ret;
	195	}
	196
ba9c1487	197	paragraph text_config_filename(char filename)
ba9c1487	198	{
e4ea58f8	199	return cmdline_cfg_simple("text-filename", filename, NULL);
ba9c1487	200	}
ba9c1487	201
d7482997	202	void text_backend(paragraph sourceform, keywordlist keywords,
43341922	203	indexdata idx, void unused) {
d7482997	204	paragraph *p;
	205	textconfig conf;
	206	word prefix, body, *wp;
	207	word spaceword;
2ac8ceac	208	textfile tf;
2ac8ceac	209	wchar_t *prefixextra;
7136a6c7	210	int nesting, nestindent;
d7482997	211	int indentb, indenta;
d7482997	212
43341922	213	IGNORE(unused);
d7482997	214	IGNORE(keywords); /* we don't happen to need this */
	215	IGNORE(idx); /* or this */
	216
	217	conf = text_configure(sourceform);
	218
	219	/*
50d6b4bd	220	* Open the output file.
d7482997	221	*/
2ac8ceac	222	tf.fp = fopen(conf.filename, "w");
2ac8ceac	223	if (!tf.fp) {
50d6b4bd	224	error(err_cantopenw, conf.filename);
d7482997	225	return;
d7482997	226	}
2ac8ceac	227	tf.charset = conf.charset;
2ac8ceac	228	tf.state = charset_init_state;
d7482997	229
	230	/* Do the title */
	231	for (p = sourceform; p; p = p->next)
	232	if (p->type == para_Title)
2ac8ceac	233	text_heading(&tf, NULL, NULL, p->words,
d7482997	234	conf.atitle, conf.indent, conf.width);
d7482997	235
7136a6c7	236	nestindent = conf.listindentbefore + conf.listindentafter;
8902e0ed	237	nesting = (conf.indent_preambles ? 0 : -conf.indent);
7136a6c7	238
d7482997	239	/* Do the main document */
	240	for (p = sourceform; p; p = p->next) switch (p->type) {
	241
2614b01d	242	case para_QuotePush:
	243	nesting += 2;
	244	break;
	245	case para_QuotePop:
	246	nesting -= 2;
	247	assert(nesting >= 0);
	248	break;
	249
7136a6c7	250	case para_LcontPush:
2614b01d	251	nesting += nestindent;
7136a6c7	252	break;
7136a6c7	253	case para_LcontPop:
2614b01d	254	nesting -= nestindent;
2614b01d	255	assert(nesting >= 0);
7136a6c7	256	break;
7136a6c7	257
d7482997	258	/*
	259	* Things we ignore because we've already processed them or
	260	* aren't going to touch them in this pass.
	261	*/
	262	case para_IM:
	263	case para_BR:
	264	case para_Biblio: /* only touch BiblioCited */
	265	case para_VersionID:
d7482997	266	case para_NoCite:
	267	case para_Title:
	268	break;
	269
	270	/*
	271	* Chapter titles.
	272	*/
	273	case para_Chapter:
	274	case para_Appendix:
	275	case para_UnnumberedChapter:
2ac8ceac	276	text_heading(&tf, p->kwtext, p->kwtext2, p->words,
d7482997	277	conf.achapter, conf.indent, conf.width);
8902e0ed	278	nesting = 0;
d7482997	279	break;
	280
	281	case para_Heading:
	282	case para_Subsect:
2ac8ceac	283	text_heading(&tf, p->kwtext, p->kwtext2, p->words,
d7482997	284	conf.asect[p->aux>=conf.nasect ? conf.nasect-1 : p->aux],
	285	conf.indent, conf.width);
	286	break;
	287
	288	case para_Rule:
2ac8ceac	289	text_rule(&tf, conf.indent + nesting, conf.width - nesting);
d7482997	290	break;
	291
	292	case para_Normal:
9057a0a8	293	case para_Copyright:
7136a6c7	294	case para_DescribedThing:
7136a6c7	295	case para_Description:
d7482997	296	case para_BiblioCited:
	297	case para_Bullet:
	298	case para_NumberedList:
	299	if (p->type == para_Bullet) {
	300	prefix = &conf.bullet;
	301	prefixextra = NULL;
	302	indentb = conf.listindentbefore;
	303	indenta = conf.listindentafter;
	304	} else if (p->type == para_NumberedList) {
	305	prefix = p->kwtext;
2ac8ceac	306	prefixextra = L"."; /* FIXME: configurability */
d7482997	307	indentb = conf.listindentbefore;
d7482997	308	indenta = conf.listindentafter;
7136a6c7	309	} else if (p->type == para_Description) {
	310	prefix = NULL;
	311	prefixextra = NULL;
	312	indentb = conf.listindentbefore;
	313	indenta = conf.listindentafter;
d7482997	314	} else {
	315	prefix = NULL;
	316	prefixextra = NULL;
	317	indentb = indenta = 0;
	318	}
	319	if (p->type == para_BiblioCited) {
	320	body = dup_word_list(p->kwtext);
	321	for (wp = body; wp->next; wp = wp->next);
	322	wp->next = &spaceword;
	323	spaceword.next = p->words;
	324	spaceword.alt = NULL;
	325	spaceword.type = word_WhiteSpace;
	326	spaceword.text = NULL;
	327	} else {
	328	wp = NULL;
	329	body = p->words;
	330	}
2ac8ceac	331	text_para(&tf, prefix, prefixextra, body,
2614b01d	332	conf.indent + nesting + indentb, indenta,
2614b01d	333	conf.width - nesting - indentb - indenta);
d7482997	334	if (wp) {
	335	wp->next = NULL;
	336	free_word_list(body);
	337	}
	338	break;
	339
	340	case para_Code:
2ac8ceac	341	text_codepara(&tf, p->words,
2614b01d	342	conf.indent + nesting + conf.indent_code,
2614b01d	343	conf.width - nesting - 2 * conf.indent_code);
d7482997	344	break;
	345	}
	346
	347	/* Do the version ID */
	348	if (conf.include_version_id) {
	349	for (p = sourceform; p; p = p->next)
	350	if (p->type == para_VersionID)
2ac8ceac	351	text_versionid(&tf, p->words);
d7482997	352	}
	353
	354	/*
	355	* Tidy up
	356	*/
2ac8ceac	357	text_output(&tf, NULL); /* end charset conversion */
2ac8ceac	358	fclose(tf.fp);
e5e6bf9d	359	sfree(conf.asect);
50d6b4bd	360	sfree(conf.filename);
d7482997	361	}
d7482997	362
2ac8ceac	363	static void text_output(textfile tf, const wchar_t s)
	364	{
	365	char buf[256];
	366	int ret, len;
	367	const wchar_t **sp;
	368
	369	if (!s) {
	370	sp = NULL;
	371	len = 1;
	372	} else {
	373	sp = &s;
	374	len = ustrlen(s);
	375	}
	376
	377	while (len > 0) {
	378	ret = charset_from_unicode(sp, &len, buf, lenof(buf),
	379	tf->charset, &tf->state, NULL);
	380	if (!sp)
	381	len = 0;
	382	fwrite(buf, 1, ret, tf->fp);
d7482997	383	}
d7482997	384	}
d7482997	385
2ac8ceac	386	static void text_output_many(textfile *tf, int n, wchar_t c)
	387	{
	388	wchar_t s[2];
	389	s[0] = c;
	390	s[1] = L'\0';
	391	while (n--)
	392	text_output(tf, s);
	393	}
d7482997	394
2ac8ceac	395	static void text_rdaddw(int charset, rdstring rs, word text, word *end) {
d7482997	396	for (; text && text != end; text = text->next) switch (text->type) {
	397	case word_HyperLink:
	398	case word_HyperEnd:
	399	case word_UpperXref:
	400	case word_LowerXref:
	401	case word_XrefEnd:
	402	case word_IndexRef:
	403	break;
	404
	405	case word_Normal:
	406	case word_Emph:
	407	case word_Code:
	408	case word_WeakCode:
	409	case word_WhiteSpace:
	410	case word_EmphSpace:
	411	case word_CodeSpace:
	412	case word_WkCodeSpace:
	413	case word_Quote:
	414	case word_EmphQuote:
	415	case word_CodeQuote:
	416	case word_WkCodeQuote:
	417	assert(text->type != word_CodeQuote &&
	418	text->type != word_WkCodeQuote);
	419	if (towordstyle(text->type) == word_Emph &&
	420	(attraux(text->aux) == attr_First \|\|
	421	attraux(text->aux) == attr_Only))
2ac8ceac	422	rdadd(rs, L'_'); /* FIXME: configurability */
d7482997	423	else if (towordstyle(text->type) == word_Code &&
	424	(attraux(text->aux) == attr_First \|\|
	425	attraux(text->aux) == attr_Only))
2ac8ceac	426	rdadd(rs, L'`'); /* FIXME: configurability */
d7482997	427	if (removeattr(text->type) == word_Normal) {
91f93b94	428	if (cvt_ok(charset, text->text) \|\| !text->alt)
2ac8ceac	429	rdadds(rs, text->text);
d7482997	430	else
2ac8ceac	431	text_rdaddw(charset, rs, text->alt, NULL);
d7482997	432	} else if (removeattr(text->type) == word_WhiteSpace) {
2ac8ceac	433	rdadd(rs, L' ');
d7482997	434	} else if (removeattr(text->type) == word_Quote) {
2ac8ceac	435	rdadd(rs, quoteaux(text->aux) == quote_Open ? L'`' : L'\'');
d7482997	436	/* FIXME: configurability */
	437	}
	438	if (towordstyle(text->type) == word_Emph &&
	439	(attraux(text->aux) == attr_Last \|\|
	440	attraux(text->aux) == attr_Only))
2ac8ceac	441	rdadd(rs, L'_'); /* FIXME: configurability */
d7482997	442	else if (towordstyle(text->type) == word_Code &&
	443	(attraux(text->aux) == attr_Last \|\|
	444	attraux(text->aux) == attr_Only))
2ac8ceac	445	rdadd(rs, L'\''); /* FIXME: configurability */
d7482997	446	break;
	447	}
	448	}
	449
43341922	450	static int text_width(void , word );
d7482997	451
43341922	452	static int text_width_list(void ctx, word text) {
d7482997	453	int w = 0;
d7482997	454	while (text) {
43341922	455	w += text_width(ctx, text);
d7482997	456	text = text->next;
	457	}
	458	return w;
	459	}
	460
43341922	461	static int text_width(void ctx, word text) {
2ac8ceac	462	int charset = * (int *) ctx;
43341922	463
d7482997	464	switch (text->type) {
	465	case word_HyperLink:
	466	case word_HyperEnd:
	467	case word_UpperXref:
	468	case word_LowerXref:
	469	case word_XrefEnd:
	470	case word_IndexRef:
	471	return 0;
	472
	473	case word_Normal:
	474	case word_Emph:
	475	case word_Code:
	476	case word_WeakCode:
	477	return (((text->type == word_Emph \|\|
	478	text->type == word_Code)
	479	? (attraux(text->aux) == attr_Only ? 2 :
	480	attraux(text->aux) == attr_Always ? 0 : 1)
	481	: 0) +
91f93b94	482	(cvt_ok(charset, text->text) \|\| !text->alt ?
e5cd393f	483	ustrwid(text->text, charset) :
43341922	484	text_width_list(ctx, text->alt)));
d7482997	485
	486	case word_WhiteSpace:
	487	case word_EmphSpace:
	488	case word_CodeSpace:
	489	case word_WkCodeSpace:
	490	case word_Quote:
	491	case word_EmphQuote:
	492	case word_CodeQuote:
	493	case word_WkCodeQuote:
	494	assert(text->type != word_CodeQuote &&
	495	text->type != word_WkCodeQuote);
	496	return (((towordstyle(text->type) == word_Emph \|\|
	497	towordstyle(text->type) == word_Code)
	498	? (attraux(text->aux) == attr_Only ? 2 :
	499	attraux(text->aux) == attr_Always ? 0 : 1)
	500	: 0) + 1);
	501	}
	502	return 0; /* should never happen */
	503	}
	504
2ac8ceac	505	static void text_heading(textfile tf, word tprefix, word *nprefix,
	506	word *text, alignstruct align,
	507	int indent, int width) {
	508	rdstring t = { 0, 0, NULL };
d7482997	509	int margin, length;
	510	int firstlinewidth, wrapwidth;
	511	wrappedline wrapping, p;
	512
	513	if (align.just_numbers && nprefix) {
2ac8ceac	514	text_rdaddw(tf->charset, &t, nprefix, NULL);
2ac8ceac	515	rdadds(&t, align.number_suffix);
d7482997	516	} else if (!align.just_numbers && tprefix) {
2ac8ceac	517	text_rdaddw(tf->charset, &t, tprefix, NULL);
2ac8ceac	518	rdadds(&t, align.number_suffix);
d7482997	519	}
2ac8ceac	520	margin = length = t.pos;
d7482997	521
	522	if (align.align == LEFTPLUS) {
	523	margin = indent - margin;
	524	if (margin < 0) margin = 0;
	525	firstlinewidth = indent + width - margin - length;
	526	wrapwidth = width;
	527	} else if (align.align == LEFT \|\| align.align == CENTRE) {
	528	margin = 0;
	529	firstlinewidth = indent + width - length;
	530	wrapwidth = indent + width;
	531	}
	532
2ac8ceac	533	wrapping = wrap_para(text, firstlinewidth, wrapwidth,
2ac8ceac	534	text_width, &tf->charset, 0);
d7482997	535	for (p = wrapping; p; p = p->next) {
2ac8ceac	536	text_rdaddw(tf->charset, &t, p->begin, p->end);
2ac8ceac	537	length = t.pos;
d7482997	538	if (align.align == CENTRE) {
	539	margin = (indent + width - length)/2;
	540	if (margin < 0) margin = 0;
	541	}
2ac8ceac	542	text_output_many(tf, margin, L' ');
	543	text_output(tf, t.text);
	544	text_output(tf, L"\n");
d7482997	545	if (align.underline != L'\0') {
2ac8ceac	546	text_output_many(tf, margin, L' ');
	547	text_output_many(tf, length, align.underline);
	548	text_output(tf, L"\n");
d7482997	549	}
	550	if (align.align == LEFTPLUS)
	551	margin = indent;
	552	else
	553	margin = 0;
	554	sfree(t.text);
2ac8ceac	555	t = empty_rdstring;
d7482997	556	}
d7482997	557	wrap_free(wrapping);
2ac8ceac	558	text_output(tf, L"\n");
d7482997	559
	560	sfree(t.text);
	561	}
	562
2ac8ceac	563	static void text_rule(textfile *tf, int indent, int width) {
	564	text_output_many(tf, indent, L' ');
	565	text_output_many(tf, width, L'-'); /* FIXME: configurability! */
	566	text_output_many(tf, 2, L'\n');
d7482997	567	}
d7482997	568
2ac8ceac	569	static void text_para(textfile tf, word prefix, wchar_t *prefixextra,
2ac8ceac	570	word *text, int indent, int extraindent, int width) {
d7482997	571	wrappedline wrapping, p;
2ac8ceac	572	rdstring pfx = { 0, 0, NULL };
d7482997	573	int e;
	574	int firstlinewidth = width;
	575
	576	if (prefix) {
2ac8ceac	577	text_rdaddw(tf->charset, &pfx, prefix, NULL);
d7482997	578	if (prefixextra)
2ac8ceac	579	rdadds(&pfx, prefixextra);
	580	text_output_many(tf, indent, L' ');
	581	text_output(tf, pfx.text);
c83c6495	582	/* If the prefix is too long, shorten the first line to fit. */
2ac8ceac	583	e = extraindent - pfx.pos;
d7482997	584	if (e < 0) {
c83c6495	585	firstlinewidth += e; /* this decreases it, since e < 0 */
d7482997	586	if (firstlinewidth < 0) {
	587	e = indent + extraindent;
	588	firstlinewidth = width;
2ac8ceac	589	text_output(tf, L"\n");
c83c6495	590	} else
c83c6495	591	e = 0;
d7482997	592	}
	593	sfree(pfx.text);
	594	} else
	595	e = indent + extraindent;
	596
2ac8ceac	597	wrapping = wrap_para(text, firstlinewidth, width,
2ac8ceac	598	text_width, &tf->charset, 0);
d7482997	599	for (p = wrapping; p; p = p->next) {
2ac8ceac	600	rdstring t = { 0, 0, NULL };
	601	text_rdaddw(tf->charset, &t, p->begin, p->end);
	602	text_output_many(tf, e, L' ');
	603	text_output(tf, t.text);
	604	text_output(tf, L"\n");
d7482997	605	e = indent + extraindent;
	606	sfree(t.text);
	607	}
	608	wrap_free(wrapping);
2ac8ceac	609	text_output(tf, L"\n");
d7482997	610	}
d7482997	611
2ac8ceac	612	static void text_codepara(textfile tf, word text, int indent, int width) {
d7482997	613	for (; text; text = text->next) if (text->type == word_WeakCode) {
2ac8ceac	614	if (ustrlen(text->text) > width) {
d7482997	615	/* FIXME: warn */
d7482997	616	}
2ac8ceac	617	text_output_many(tf, indent, L' ');
	618	text_output(tf, text->text);
	619	text_output(tf, L"\n");
d7482997	620	}
d7482997	621
2ac8ceac	622	text_output(tf, L"\n");
d7482997	623	}
d7482997	624
2ac8ceac	625	static void text_versionid(textfile tf, word text) {
2ac8ceac	626	rdstring t = { 0, 0, NULL };
d7482997	627
2ac8ceac	628	rdadd(&t, L'['); /* FIXME: configurability */
	629	text_rdaddw(tf->charset, &t, text, NULL);
	630	rdadd(&t, L']'); /* FIXME: configurability */
	631	rdadd(&t, L'\n');
d7482997	632
2ac8ceac	633	text_output(tf, t.text);
d7482997	634	sfree(t.text);
d7482997	635	}