sgml.c

elinks下lynx是最重要的二个文本浏览器, 在linux下非常实用, lynx比elinks早的多, 目前好像停止开发, 这是lynx源代码

#include <AttrList.h>
static int current_is_class = 0;
#endif

/*	Handle Attribute
**	----------------
*/
/* PUBLIC CONST char * SGML_default = "";   ?? */

PRIVATE void handle_attribute_name ARGS2(
	HTStream *,	context,
	CONST char *,	s)
{
    HTTag * tag = context->current_tag;
    attr * attributes = tag->attributes;
    int high, low, i, diff;

#ifdef USE_PRETTYSRC
    if (psrc_view) {
	attr_is_href = FALSE;
	attr_is_name = FALSE;
    }
#endif
    /*
    **	Ignore unknown tag. - KW
    */
    if (tag == context->unknown_tag) {
#ifdef USE_PRETTYSRC
	if (psrc_view)
	    context->current_attribute_number = 1; /* anything !=INVALID */
#endif
	return;
    }

    /*
    **	Binary search for attribute name.
    */
    for (low = 0, high = tag->number_of_attributes;
	 high > low;
	 diff < 0 ? (low = i+1) : (high = i)) {
	i = (low + (high-low)/2);
	diff = my_casecomp(attributes[i].name, s);
	if (diff == 0) {		/* success: found it */
	    context->current_attribute_number = i;
#ifdef USE_PRETTYSRC
	    if (psrc_view) {
		attr_is_name = (BOOL) (attributes[i].type == HTMLA_ANAME);
		attr_is_href = (BOOL) (attributes[i].type == HTMLA_HREF);
	    } else
#endif
	    {
	    context->present[i] = YES;
	    Clear_extra(context->value[i]);
#ifdef USE_COLOR_STYLE
#   ifdef USE_PRETTYSRC
	    current_is_class = IS_C(attributes[i]);
#   else
	    current_is_class = (!strcasecomp("class", s));
#   endif
	    CTRACE((tfp, "SGML: found attribute %s, %d\n", s, current_is_class));
#endif
	    }
	    return;
	} /* if */

    } /* for */

    CTRACE((tfp, "SGML: Unknown attribute %s for tag %s\n",
		s, context->current_tag->name));
    context->current_attribute_number = INVALID;	/* Invalid */
}


/*	Handle attribute value
**	----------------------
*/
PRIVATE void handle_attribute_value ARGS2(
	HTStream *,	context,
	CONST char *,	s)
{
    if (context->current_attribute_number != INVALID) {
	StrAllocCopy_extra(context->value[context->current_attribute_number], s);
#ifdef USE_COLOR_STYLE
	if (current_is_class)
	{
	    strncpy (class_string, s, TEMPSTRINGSIZE);
	    CTRACE((tfp, "SGML: class is '%s'\n", s));
	}
	else
	{
	    CTRACE((tfp, "SGML: attribute value is '%s'\n", s));
	}
#endif
    } else {
	CTRACE((tfp, "SGML: Attribute value %s ***ignored\n", s));
    }
    context->current_attribute_number = INVALID; /* can't have two assignments! */
}


/*
**  Translate some Unicodes to Lynx special codes and output them.
**  Special codes - ones those output depend on parsing.
**
**  Additional issue, like handling bidirectional text if necessary
**  may be called from here:  zwnj (8204), zwj (8205), lrm (8206), rlm (8207)
**  - currently they are ignored in SGML.c and LYCharUtils.c
**  but also in UCdomap.c because they are non printable...
**
*/
PRIVATE BOOL put_special_unicodes ARGS2(
	HTStream *,	context,
	UCode_t,	code)
{
    /* (Tgf_nolyspcl) */
    if (context->no_lynx_specialcodes) {
	/*
	**  We were asked by a "DTD" flag to not generate lynx specials. - kw
	*/
	return NO;
    }

    if (code == CH_NBSP) {  /* S/390 -- gil -- 0657 */
	/*
	**  Use Lynx special character for nbsp.
	*/
#ifdef USE_PRETTYSRC
	if (!psrc_view)
#endif
	PUTC(HT_NON_BREAK_SPACE);
    } else  if (code == CH_SHY) {
	/*
	**  Use Lynx special character for shy.
	*/
#ifdef USE_PRETTYSRC
	if (!psrc_view)
#endif
	PUTC(LY_SOFT_HYPHEN);
    } else if (code == 8194 || code == 8201) {
	/*
	**  Use Lynx special character for ensp or thinsp.
	**
	**  Originally, Lynx use space '32' as word delimiter and omits this
	**  space at end of line if word is wrapped to the next line.  There
	**  are several other spaces in the Unicode repertoire and we should
	**  teach Lynx to understand them, not only as regular characters but
	**  in the context of line wrapping.  Unfortunately, if we use
	**  HT_EN_SPACE we override the chartrans tables for those spaces
	**  with a single '32' for all (but do line wrapping more fancy).
	**
	**  We may treat emsp as one or two ensp (below).
	*/
#ifdef USE_PRETTYSRC
	if (!psrc_view)
#endif
	PUTC(HT_EN_SPACE);
    } else if (code == 8195) {
	/*
	**  Use Lynx special character for emsp.
	*/
#ifdef USE_PRETTYSRC
	if (!psrc_view) {
#endif
	/* PUTC(HT_EN_SPACE);  let's stay with a single space :) */
	PUTC(HT_EN_SPACE);
#ifdef USE_PRETTYSRC
	}
#endif
    } else {
	/*
	**  Return NO if nothing done.
	*/
	return NO;
    }
    /*
    **	We have handled it.
    */
    return YES;
}

#ifdef USE_PRETTYSRC
PRIVATE void put_pretty_entity ARGS2(HTStream *, context, int, term)
{
    PSRCSTART(entity);
    PUTC('&');
    PUTS(entity_string);
    if (term)
	PUTC((char)term);
    PSRCSTOP(entity);
}

PRIVATE void put_pretty_number ARGS1(HTStream *, context)
{
    PSRCSTART(entity);
    PUTS( (context->isHex ? "&#x" : "&#") );
    PUTS(entity_string);
    PUTC(';');
    PSRCSTOP(entity);
}
#endif /* USE_PRETTYSRC */

/*	Handle entity
**	-------------
**
** On entry,
**	s	contains the entity name zero terminated
** Bugs:
**	If the entity name is unknown, the terminator is treated as
**	a printable non-special character in all cases, even if it is '<'
** Bug-fix:
**	Modified SGML_character() so we only come here with terminator
**	as '\0' and check a FoundEntity flag. -- Foteos Macrides
**
** Modified more (for use with Lynx character translation code):
*/
PRIVATE char replace_buf [64];	      /* buffer for replacement strings */
PRIVATE BOOL FoundEntity = FALSE;

PRIVATE void handle_entity ARGS2(
	HTStream *,	context,
	char,		term)
{
    UCode_t code;
    long uck = -1;
    CONST char *s = context->string->data;

    /*
    **	Handle all entities normally. - FM
    */
    FoundEntity = FALSE;
    if ((code = HTMLGetEntityUCValue(s)) != 0) {
	/*
	**  We got a Unicode value for the entity name.
	**  Check for special Unicodes. - FM
	*/
	if (put_special_unicodes(context, code)) {
#ifdef USE_PRETTYSRC
	    if (psrc_view) {
		put_pretty_entity(context, term);
	    }
#endif
	    FoundEntity = TRUE;
	    return;
	}
	/*
	**  Seek a translation from the chartrans tables.
	*/
	if ((uck = UCTransUniChar(code, context->outUCLYhndl)) >= 32 &&
/* =============== work in ASCII below here ===============  S/390 -- gil -- 0672 */
	    uck < 256 &&
	    (uck < 127 ||
	     uck >= LYlowest_eightbit[context->outUCLYhndl])) {
#ifdef USE_PRETTYSRC
	    if (psrc_view) {
		put_pretty_entity(context, term);
	    } else
#endif
	    PUTC(FROMASCII((char)uck));
	    FoundEntity = TRUE;
	    return;
	} else if ((uck == -4 ||
		    (context->T.repl_translated_C0 &&
		     uck > 0 && uck < 32)) &&
		   /*
		   **  Not found; look for replacement string.
		   */
		   (uck = UCTransUniCharStr(replace_buf, 60, code,
					    context->outUCLYhndl, 0) >= 0)) {
#ifdef USE_PRETTYSRC
	    if (psrc_view) {
		put_pretty_entity(context, term);
	    } else
#endif
	    PUTS(replace_buf);
	    FoundEntity = TRUE;
	    return;
	}
	/*
	**  If we're displaying UTF-8, try that now. - FM
	*/
#ifndef USE_PRETTYSRC
	if (context->T.output_utf8 && PUTUTF8(code)) {
	    FoundEntity = TRUE;
	    return;
	}
#else
	if (context->T.output_utf8 && (psrc_view ?
	      (UCPutUtf8_charstring((HTStream *)context->target,
	      (putc_func_t*)(fake_put_character), code)): PUTUTF8(code) ) ) {

	    if (psrc_view) {
		put_pretty_entity(context, term);
	    }

	    FoundEntity = TRUE;
	    return;
	}
#endif
	/*
	**  If it's safe ASCII, use it. - FM
	*/
	if (code >= 32 && code < 127) {
#ifdef USE_PRETTYSRC
	    if (psrc_view) {
		put_pretty_entity(context, term);
	    } else
#endif

	    PUTC(FROMASCII((char)code));
	    FoundEntity = TRUE;
	    return;
	}
/* =============== work in ASCII above here ===============  S/390 -- gil -- 0682 */
	/*
	**  Ignore zwnj (8204) and zwj (8205), if we get to here.
	**  Note that zwnj may have been handled as <WBR>
	**  by the calling function. - FM
	*/
	if (!strcmp(s, "zwnj") ||
	    !strcmp(s, "zwj")) {
	    CTRACE((tfp, "handle_entity: Ignoring '%s'.\n", s));
#ifdef USE_PRETTYSRC
	    if (psrc_view) {
		put_pretty_entity(context, term);
	    }
#endif
	    FoundEntity = TRUE;
	    return;
	}
	/*
	**  Ignore lrm (8206), and rln (8207), if we get to here. - FM
	*/
	if (!strcmp(s, "lrm") ||
	    !strcmp(s, "rlm")) {
	    CTRACE((tfp, "handle_entity: Ignoring '%s'.\n", s));
#ifdef USE_PRETTYSRC
	    if (psrc_view) {
		put_pretty_entity(context, term);
	    }
#endif
	    FoundEntity = TRUE;
	    return;
	}
    }

    /*
    **	If entity string not found, display as text.
    */
#ifdef USE_PRETTYSRC
    if (psrc_view)
	PSRCSTART(badseq);
#endif
    CTRACE((tfp, "SGML: Unknown entity '%s' %ld %ld\n", s, (long)code, uck)); /* S/390 -- gil -- 0695 */
    PUTC('&');
    PUTS(s);
    if (term != '\0')
	PUTC(term);
#ifdef USE_PRETTYSRC
    if (psrc_view)
	PSRCSTOP(badseq);
#endif
}


/*	Handle comment
**	--------------
*/
PRIVATE void handle_comment ARGS1(
	HTStream *,		context)
{
    CONST char *s = context->string->data;

    CTRACE((tfp, "SGML Comment:\n<%s>\n", s));

    if (context->csi == NULL &&
	strncmp(s, "!--#", 4) == 0 &&
	LYCheckForCSI(context->node_anchor, (char **)&context->url) == TRUE) {
	LYDoCSI(context->url, s, (char **)&context->csi);
    } else {
	LYCommentHacks(context->node_anchor, context->string->data);
    }

    return;
}


/*	Handle identifier
**	-----------------
*/
PRIVATE void handle_identifier ARGS1(
	HTStream *,		context)
{
    CONST char *s = context->string->data;

    CTRACE((tfp, "SGML Identifier:\n<%s>\n", s));

    return;
}


/*	Handle doctype
**	--------------
*/
PRIVATE void handle_doctype ARGS1(
	HTStream *,		context)
{
    CONST char *s = context->string->data;

    CTRACE((tfp, "SGML Doctype:\n<%s>\n", s));

    return;
}

PRIVATE void SGML_write PARAMS((
	HTStream *		me,
	CONST char *		s,
	int			l));

/*	Handle marked
**	-------------
*/
PRIVATE void handle_marked ARGS1(
	HTStream *,		context)
{
    CONST char *s = context->string->data;

    CTRACE((tfp, "SGML Marked Section:\n<%s>\n", s));

    if (!strncmp(context->string->data, "![INCLUDE[", 10)) {
	context->string->data[context->string->size - 3] = '\0';
	StrAllocCat(context->include, context->string->data + 10);
	/* @@@ This needs to take charset into account! @@@
	   the wrong assumptions will be made about the data's
	   charset once it is in include - kw */

    } else if (!strncmp(context->string->data, "![CDATA[", 8)) {
	(*context->actions->_write)(context->target,
		   context->string->data + 8,
		   context->string->size - 11);

    }
    return;
}


/*	Handle sgmlent
**	--------------
*/
PRIVATE void handle_sgmlent ARGS1(
	HTStream *,		context)
{
    CONST char *s = context->string->data;

    CTRACE((tfp, "SGML Entity Declaration:\n<%s>\n", s));

    return;
}


/*	Handle sgmlent
**	--------------
*/
PRIVATE void handle_sgmlele ARGS1(
	HTStream *,		context)
{
    CONST char *s = context->string->data;

    CTRACE((tfp, "SGML Element Declaration:\n<%s>\n", s));

    return;
}