htmltree.c.svn-base

这是一个用于解析xml文件的类库。使用这个类库

/*
 * HTMLtree.c : implementation of access function for an HTML tree.
 *
 * See Copyright for the status of this software.
 *
 * [e-mail]
 */


#define IN_LIBXML
#include "libxml.h"
#ifdef LIBXML_HTML_ENABLED

#include <string.h> /* for memset() only ! */

#ifdef HAVE_CTYPE_H
#include <ctype.h>
#endif
#ifdef HAVE_STDLIB_H
#include <stdlib.h>
#endif

#include <libxml/xmlmemory.h>
#include <libxml/HTMLparser.h>
#include <libxml/HTMLtree.h>
#include <libxml/entities.h>
#include <libxml/valid.h>
#include <libxml/xmlerror.h>
#include <libxml/parserInternals.h>
#include <libxml/globals.h>
#include <libxml/uri.h>

/************************************************************************
 *									*
 *   		Getting/Setting encoding meta tags			*
 *									*
 ************************************************************************/

/**
 * htmlGetMetaEncoding:
 * @doc:  the document
 * 
 * Encoding definition lookup in the Meta tags
 *
 * Returns the current encoding as flagged in the HTML source
 */
const xmlChar *
htmlGetMetaEncoding(htmlDocPtr doc) {
    htmlNodePtr cur;
    const xmlChar *content;
    const xmlChar *encoding;

    if (doc == NULL)
	return(NULL);
    cur = doc->children;

    /*
     * Search the html
     */
    while (cur != NULL) {
	if ((cur->type == XML_ELEMENT_NODE) && (cur->name != NULL)) {
	    if (xmlStrEqual(cur->name, BAD_CAST"html"))
		break;
	    if (xmlStrEqual(cur->name, BAD_CAST"head"))
		goto found_head;
	    if (xmlStrEqual(cur->name, BAD_CAST"meta"))
		goto found_meta;
	}
	cur = cur->next;
    }
    if (cur == NULL)
	return(NULL);
    cur = cur->children;

    /*
     * Search the head
     */
    while (cur != NULL) {
	if ((cur->type == XML_ELEMENT_NODE) && (cur->name != NULL)) {
	    if (xmlStrEqual(cur->name, BAD_CAST"head"))
		break;
	    if (xmlStrEqual(cur->name, BAD_CAST"meta"))
		goto found_meta;
	}
	cur = cur->next;
    }
    if (cur == NULL)
	return(NULL);
found_head:
    cur = cur->children;

    /*
     * Search the meta elements
     */
found_meta:
    while (cur != NULL) {
	if ((cur->type == XML_ELEMENT_NODE) && (cur->name != NULL)) {
	    if (xmlStrEqual(cur->name, BAD_CAST"meta")) {
		xmlAttrPtr attr = cur->properties;
		int http;
		const xmlChar *value;

		content = NULL;
		http = 0;
		while (attr != NULL) {
		    if ((attr->children != NULL) &&
		        (attr->children->type == XML_TEXT_NODE) &&
		        (attr->children->next == NULL)) {
			value = attr->children->content;
			if ((!xmlStrcasecmp(attr->name, BAD_CAST"http-equiv"))
			 && (!xmlStrcasecmp(value, BAD_CAST"Content-Type")))
			    http = 1;
			else if ((value != NULL)
			 && (!xmlStrcasecmp(attr->name, BAD_CAST"content")))
			    content = value;
			if ((http != 0) && (content != NULL))
			    goto found_content;
		    }
		    attr = attr->next;
		}
	    }
	}
	cur = cur->next;
    }
    return(NULL);

found_content:
    encoding = xmlStrstr(content, BAD_CAST"charset=");
    if (encoding == NULL) 
	encoding = xmlStrstr(content, BAD_CAST"Charset=");
    if (encoding == NULL) 
	encoding = xmlStrstr(content, BAD_CAST"CHARSET=");
    if (encoding != NULL) {
	encoding += 8;
    } else {
	encoding = xmlStrstr(content, BAD_CAST"charset =");
	if (encoding == NULL) 
	    encoding = xmlStrstr(content, BAD_CAST"Charset =");
	if (encoding == NULL) 
	    encoding = xmlStrstr(content, BAD_CAST"CHARSET =");
	if (encoding != NULL)
	    encoding += 9;
    }
    if (encoding != NULL) {
	while ((*encoding == ' ') || (*encoding == '\t')) encoding++;
    }
    return(encoding);
}

/**
 * htmlSetMetaEncoding:
 * @doc:  the document
 * @encoding:  the encoding string
 * 
 * Sets the current encoding in the Meta tags
 * NOTE: this will not change the document content encoding, just
 * the META flag associated.
 *
 * Returns 0 in case of success and -1 in case of error
 */
int
htmlSetMetaEncoding(htmlDocPtr doc, const xmlChar *encoding) {
    htmlNodePtr cur, meta;
    const xmlChar *content;
    char newcontent[100];


    if (doc == NULL)
	return(-1);

    if (encoding != NULL) {
	_snprintf(newcontent, sizeof(newcontent), "text/html; charset=%s",
                encoding);
	newcontent[sizeof(newcontent) - 1] = 0;
    }

    cur = doc->children;

    /*
     * Search the html
     */
    while (cur != NULL) {
	if ((cur->type == XML_ELEMENT_NODE) && (cur->name != NULL)) {
	    if (xmlStrcasecmp(cur->name, BAD_CAST"html") == 0)
		break;
	    if (xmlStrcasecmp(cur->name, BAD_CAST"head") == 0)
		goto found_head;
	    if (xmlStrcasecmp(cur->name, BAD_CAST"meta") == 0)
		goto found_meta;
	}
	cur = cur->next;
    }
    if (cur == NULL)
	return(-1);
    cur = cur->children;

    /*
     * Search the head
     */
    while (cur != NULL) {
	if ((cur->type == XML_ELEMENT_NODE) && (cur->name != NULL)) {
	    if (xmlStrcasecmp(cur->name, BAD_CAST"head") == 0)
		break;
	    if (xmlStrcasecmp(cur->name, BAD_CAST"meta") == 0)
		goto found_meta;
	}
	cur = cur->next;
    }
    if (cur == NULL)
	return(-1);
found_head:
    if (cur->children == NULL) {
	if (encoding == NULL)
	    return(0);
	meta = xmlNewDocNode(doc, NULL, BAD_CAST"meta", NULL);
	xmlAddChild(cur, meta);
	xmlNewProp(meta, BAD_CAST"http-equiv", BAD_CAST"Content-Type");
	xmlNewProp(meta, BAD_CAST"content", BAD_CAST newcontent);
	return(0);
    }
    cur = cur->children;

found_meta:
    if (encoding != NULL) {
	/*
	 * Create a new Meta element with the right attributes
	 */

	meta = xmlNewDocNode(doc, NULL, BAD_CAST"meta", NULL);
	xmlAddPrevSibling(cur, meta);
	xmlNewProp(meta, BAD_CAST"http-equiv", BAD_CAST"Content-Type");
	xmlNewProp(meta, BAD_CAST"content", BAD_CAST newcontent);
    }

    /*
     * Search and destroy all the remaining the meta elements carrying
     * encoding informations
     */
    while (cur != NULL) {
	if ((cur->type == XML_ELEMENT_NODE) && (cur->name != NULL)) {
	    if (xmlStrcasecmp(cur->name, BAD_CAST"meta") == 0) {
		xmlAttrPtr attr = cur->properties;
		int http;
		const xmlChar *value;

		content = NULL;
		http = 0;
		while (attr != NULL) {
		    if ((attr->children != NULL) &&
		        (attr->children->type == XML_TEXT_NODE) &&
		        (attr->children->next == NULL)) {
			value = attr->children->content;
			if ((!xmlStrcasecmp(attr->name, BAD_CAST"http-equiv"))
			 && (!xmlStrcasecmp(value, BAD_CAST"Content-Type")))
			    http = 1;
			else 
                        {
                           if ((value != NULL) && 
				(!xmlStrcasecmp(attr->name, BAD_CAST"content")))
			      content = value;
                        }
		        if ((http != 0) && (content != NULL))
			    break;
		    }
		    attr = attr->next;
		}
		if ((http != 0) && (content != NULL)) {
		    meta = cur;
		    cur = cur->next;
		    xmlUnlinkNode(meta);
                    xmlFreeNode(meta);
		    continue;
		}

	    }
	}
	cur = cur->next;
    }
    return(0);
}

/**
 * booleanHTMLAttrs:
 *
 * These are the HTML attributes which will be output
 * in minimized form, i.e. <option selected="selected"> will be
 * output as <option selected>, as per XSLT 1.0 16.2 "HTML Output Method"
 *
 */
static const char* htmlBooleanAttrs[] = {
  "checked", "compact", "declare", "defer", "disabled", "ismap",
  "multiple", "nohref", "noresize", "noshade", "nowrap", "readonly",
  "selected", NULL
};


/**
 * htmlIsBooleanAttr:
 * @name:  the name of the attribute to check
 *
 * Determine if a given attribute is a boolean attribute.
 * 
 * returns: false if the attribute is not boolean, true otherwise.
 */
int
htmlIsBooleanAttr(const xmlChar *name)
{
    int i = 0;

    while (htmlBooleanAttrs[i] != NULL) {
        if (xmlStrcasecmp((const xmlChar *)htmlBooleanAttrs[i], name) == 0)
            return 1;
        i++;
    }
    return 0;
}

#ifdef LIBXML_OUTPUT_ENABLED
/************************************************************************
 *									*
 * 			Output error handlers				*
 *									*
 ************************************************************************/
/**
 * htmlSaveErrMemory:
 * @extra:  extra informations
 *
 * Handle an out of memory condition
 */
static void
htmlSaveErrMemory(const char *extra)
{
    __xmlSimpleError(XML_FROM_OUTPUT, XML_ERR_NO_MEMORY, NULL, NULL, extra);
}

/**
 * htmlSaveErr:
 * @code:  the error number
 * @node:  the location of the error.
 * @extra:  extra informations
 *
 * Handle an out of memory condition
 */
static void
htmlSaveErr(int code, xmlNodePtr node, const char *extra)
{
    const char *msg = NULL;

    switch(code) {
        case XML_SAVE_NOT_UTF8:
	    msg = "string is not in UTF-8";
	    break;
	case XML_SAVE_CHAR_INVALID:
	    msg = "invalid character value";
	    break;
	case XML_SAVE_UNKNOWN_ENCODING:
	    msg = "unknown encoding %s";
	    break;
	case XML_SAVE_NO_DOCTYPE:
	    msg = "HTML has no DOCTYPE";
	    break;
	default:
	    msg = "unexpected error number";
    }
    __xmlSimpleError(XML_FROM_OUTPUT, code, node, msg, extra);
}

/************************************************************************
 *									*
 *   		Dumping HTML tree content to a simple buffer		*
 *									*
 ************************************************************************/

static int
htmlNodeDumpFormat(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur,
	           int format);

/**
 * htmlNodeDumpFormat:
 * @buf:  the HTML buffer output
 * @doc:  the document
 * @cur:  the current node
 * @format:  should formatting spaces been added
 *
 * Dump an HTML node, recursive behaviour,children are printed too.
 *
 * Returns the number of byte written or -1 in case of error
 */
static int
htmlNodeDumpFormat(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur,
	           int format) {
    unsigned int use;
    int ret;
    xmlOutputBufferPtr outbuf;

    if (cur == NULL) {
	return (-1);
    }
    if (buf == NULL) {
	return (-1);
    }
    outbuf = (xmlOutputBufferPtr) xmlMalloc(sizeof(xmlOutputBuffer));
    if (outbuf == NULL) {
        htmlSaveErrMemory("allocating HTML output buffer");
	return (-1);
    }
    memset(outbuf, 0, (size_t) sizeof(xmlOutputBuffer));
    outbuf->buffer = buf;
    outbuf->encoder = NULL;
    outbuf->writecallback = NULL;
    outbuf->closecallback = NULL;
    outbuf->context = NULL;
    outbuf->written = 0;

    use = buf->use;
    htmlNodeDumpFormatOutput(outbuf, doc, cur, NULL, format);
    xmlFree(outbuf);
    ret = buf->use - use;
    return (ret);
}

/**
 * htmlNodeDump:
 * @buf:  the HTML buffer output
 * @doc:  the document
 * @cur:  the current node
 *
 * Dump an HTML node, recursive behaviour,children are printed too,
 * and formatting returns are added.
 *
 * Returns the number of byte written or -1 in case of error
 */
int
htmlNodeDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur) {
    xmlInitParser();

    return(htmlNodeDumpFormat(buf, doc, cur, 1));
}

/**
 * htmlNodeDumpFileFormat:
 * @out:  the FILE pointer
 * @doc:  the document
 * @cur:  the current node
 * @encoding: the document encoding
 * @format:  should formatting spaces been added
 *
 * Dump an HTML node, recursive behaviour,children are printed too.
 *
 * TODO: if encoding == NULL try to save in the doc encoding
 *
 * returns: the number of byte written or -1 in case of failure.
 */
int
htmlNodeDumpFileFormat(FILE *out, xmlDocPtr doc,
	               xmlNodePtr cur, const char *encoding, int format) {
    xmlOutputBufferPtr buf;
    xmlCharEncodingHandlerPtr handler = NULL;
    int ret;

    xmlInitParser();

    if (encoding != NULL) {
	xmlCharEncoding enc;

	enc = xmlParseCharEncoding(encoding);
	if (enc != XML_CHAR_ENCODING_UTF8) {
	    handler = xmlFindCharEncodingHandler(encoding);
	    if (handler == NULL)
		return(-1);
	}
    }

    /*
     * Fallback to HTML or ASCII when the encoding is unspecified
     */
    if (handler == NULL)
	handler = xmlFindCharEncodingHandler("HTML");
    if (handler == NULL)
	handler = xmlFindCharEncodingHandler("ascii");

    /* 
     * save the content to a temp buffer.
     */
    buf = xmlOutputBufferCreateFile(out, handler);
    if (buf == NULL) return(0);

    htmlNodeDumpFormatOutput(buf, doc, cur, encoding, format);

    ret = xmlOutputBufferClose(buf);
    return(ret);
}

/**
 * htmlNodeDumpFile:
 * @out:  the FILE pointer
 * @doc:  the document
 * @cur:  the current node
 *
 * Dump an HTML node, recursive behaviour,children are printed too,
 * and formatting returns are added.
 */
void
htmlNodeDumpFile(FILE *out, xmlDocPtr doc, xmlNodePtr cur) {
    htmlNodeDumpFileFormat(out, doc, cur, NULL, 1);
}

/**
 * htmlDocDumpMemory:
 * @cur:  the document
 * @mem:  OUT: the memory pointer
 * @size:  OUT: the memory length
 *
 * Dump an HTML document in memory and return the xmlChar * and it's size.
 * It's up to the caller to free the memory.
 */
void
htmlDocDumpMemory(xmlDocPtr cur, xmlChar**mem, int *size) {
    xmlOutputBufferPtr buf;
    xmlCharEncodingHandlerPtr handler = NULL;
    const char *encoding;

    xmlInitParser();

    if (cur == NULL) {
	*mem = NULL;
	*size = 0;
	return;
    }

    encoding = (const char *) htmlGetMetaEncoding(cur);

    if (encoding != NULL) {
	xmlCharEncoding enc;

	enc = xmlParseCharEncoding(encoding);
	if (enc != cur->charset) {
	    if (cur->charset != XML_CHAR_ENCODING_UTF8) {
		/*
		 * Not supported yet
		 */
		*mem = NULL;
		*size = 0;
		return;
	    }

	    handler = xmlFindCharEncodingHandler(encoding);
	    if (handler == NULL) {
		*mem = NULL;
		*size = 0;
		return;
	    }
	}
    }

    /*
     * Fallback to HTML or ASCII when the encoding is unspecified
     */
    if (handler == NULL)
	handler = xmlFindCharEncodingHandler("HTML");
    if (handler == NULL)
	handler = xmlFindCharEncodingHandler("ascii");

    buf = xmlAllocOutputBuffer(handler);
    if (buf == NULL) {
	*mem = NULL;
	*size = 0;
	return;
    }

    htmlDocContentDumpOutput(buf, cur, NULL);
    xmlOutputBufferFlush(buf);
    if (buf->conv != NULL) {
	*size = buf->conv->use;
	*mem = xmlStrndup(buf->conv->content, *size);
    } else {
	*size = buf->buffer->use;
	*mem = xmlStrndup(buf->buffer->content, *size);
    }
    (void)xmlOutputBufferClose(buf);
}


/************************************************************************
 *									*
 *   		Dumping HTML tree content to an I/O output buffer	*
 *									*
 ************************************************************************/