parser.c

xml开源解析代码.版本为libxml2-2.6.29,可支持GB3212.网络消息发送XML时很有用.

 * Returns the value parsed (as an int), 0 in case of error, str will be
 *         updated to the current value of the index
 */
static int
xmlParseStringCharRef(xmlParserCtxtPtr ctxt, const xmlChar **str) {
    const xmlChar *ptr;
    xmlChar cur;
    unsigned int val = 0;
    unsigned int outofrange = 0;

    if ((str == NULL) || (*str == NULL)) return(0);
    ptr = *str;
    cur = *ptr;
    if ((cur == '&') && (ptr[1] == '#') && (ptr[2] == 'x')) {
	ptr += 3;
	cur = *ptr;
	while (cur != ';') { /* Non input consuming loop */
	    if ((cur >= '0') && (cur <= '9')) 
	        val = val * 16 + (cur - '0');
	    else if ((cur >= 'a') && (cur <= 'f'))
	        val = val * 16 + (cur - 'a') + 10;
	    else if ((cur >= 'A') && (cur <= 'F'))
	        val = val * 16 + (cur - 'A') + 10;
	    else {
		xmlFatalErr(ctxt, XML_ERR_INVALID_HEX_CHARREF, NULL);
		val = 0;
		break;
	    }
	    if (val > 0x10FFFF)
	        outofrange = val;

	    ptr++;
	    cur = *ptr;
	}
	if (cur == ';')
	    ptr++;
    } else if  ((cur == '&') && (ptr[1] == '#')){
	ptr += 2;
	cur = *ptr;
	while (cur != ';') { /* Non input consuming loops */
	    if ((cur >= '0') && (cur <= '9')) 
	        val = val * 10 + (cur - '0');
	    else {
		xmlFatalErr(ctxt, XML_ERR_INVALID_DEC_CHARREF, NULL);
		val = 0;
		break;
	    }
	    if (val > 0x10FFFF)
	        outofrange = val;

	    ptr++;
	    cur = *ptr;
	}
	if (cur == ';')
	    ptr++;
    } else {
	xmlFatalErr(ctxt, XML_ERR_INVALID_CHARREF, NULL);
	return(0);
    }
    *str = ptr;

    /*
     * [ WFC: Legal Character ]
     * Characters referred to using character references must match the
     * production for Char. 
     */
    if ((IS_CHAR(val) && (outofrange == 0))) {
        return(val);
    } else {
        xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
			  "xmlParseStringCharRef: invalid xmlChar value %d\n",
			  val);
    }
    return(0);
}

/**
 * xmlNewBlanksWrapperInputStream:
 * @ctxt:  an XML parser context
 * @entity:  an Entity pointer
 *
 * Create a new input stream for wrapping
 * blanks around a PEReference
 *
 * Returns the new input stream or NULL
 */
 
static void deallocblankswrapper (xmlChar *str) {xmlFree(str);}
 
static xmlParserInputPtr
xmlNewBlanksWrapperInputStream(xmlParserCtxtPtr ctxt, xmlEntityPtr entity) {
    xmlParserInputPtr input;
    xmlChar *buffer;
    size_t length;
    if (entity == NULL) {
	xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
	            "xmlNewBlanksWrapperInputStream entity\n");
	return(NULL);
    }
    if (xmlParserDebugEntities)
	xmlGenericError(xmlGenericErrorContext,
		"new blanks wrapper for entity: %s\n", entity->name);
    input = xmlNewInputStream(ctxt);
    if (input == NULL) {
	return(NULL);
    }
    length = xmlStrlen(entity->name) + 5;
    buffer = xmlMallocAtomic(length);
    if (buffer == NULL) {
	xmlErrMemory(ctxt, NULL);
    	return(NULL);
    }
    buffer [0] = ' ';
    buffer [1] = '%';
    buffer [length-3] = ';';
    buffer [length-2] = ' ';
    buffer [length-1] = 0;
    memcpy(buffer + 2, entity->name, length - 5);
    input->free = deallocblankswrapper;
    input->base = buffer;
    input->cur = buffer;
    input->length = length;
    input->end = &buffer[length];
    return(input);
}

/**
 * xmlParserHandlePEReference:
 * @ctxt:  the parser context
 * 
 * [69] PEReference ::= '%' Name ';'
 *
 * [ WFC: No Recursion ]
 * A parsed entity must not contain a recursive
 * reference to itself, either directly or indirectly. 
 *
 * [ WFC: Entity Declared ]
 * In a document without any DTD, a document with only an internal DTD
 * subset which contains no parameter entity references, or a document
 * with "standalone='yes'", ...  ... The declaration of a parameter
 * entity must precede any reference to it...
 *
 * [ VC: Entity Declared ]
 * In a document with an external subset or external parameter entities
 * with "standalone='no'", ...  ... The declaration of a parameter entity
 * must precede any reference to it...
 *
 * [ WFC: In DTD ]
 * Parameter-entity references may only appear in the DTD.
 * NOTE: misleading but this is handled.
 *
 * A PEReference may have been detected in the current input stream
 * the handling is done accordingly to 
 *      http://www.w3.org/TR/REC-xml#entproc
 * i.e. 
 *   - Included in literal in entity values
 *   - Included as Parameter Entity reference within DTDs
 */
void
xmlParserHandlePEReference(xmlParserCtxtPtr ctxt) {
    const xmlChar *name;
    xmlEntityPtr entity = NULL;
    xmlParserInputPtr input;

    if (RAW != '%') return;
    switch(ctxt->instate) {
	case XML_PARSER_CDATA_SECTION:
	    return;
        case XML_PARSER_COMMENT:
	    return;
	case XML_PARSER_START_TAG:
	    return;
	case XML_PARSER_END_TAG:
	    return;
        case XML_PARSER_EOF:
	    xmlFatalErr(ctxt, XML_ERR_PEREF_AT_EOF, NULL);
	    return;
        case XML_PARSER_PROLOG:
	case XML_PARSER_START:
	case XML_PARSER_MISC:
	    xmlFatalErr(ctxt, XML_ERR_PEREF_IN_PROLOG, NULL);
	    return;
	case XML_PARSER_ENTITY_DECL:
        case XML_PARSER_CONTENT:
        case XML_PARSER_ATTRIBUTE_VALUE:
        case XML_PARSER_PI:
	case XML_PARSER_SYSTEM_LITERAL:
	case XML_PARSER_PUBLIC_LITERAL:
	    /* we just ignore it there */
	    return;
        case XML_PARSER_EPILOG:
	    xmlFatalErr(ctxt, XML_ERR_PEREF_IN_EPILOG, NULL);
	    return;
	case XML_PARSER_ENTITY_VALUE:
	    /*
	     * NOTE: in the case of entity values, we don't do the
	     *       substitution here since we need the literal
	     *       entity value to be able to save the internal
	     *       subset of the document.
	     *       This will be handled by xmlStringDecodeEntities
	     */
	    return;
        case XML_PARSER_DTD:
	    /*
	     * [WFC: Well-Formedness Constraint: PEs in Internal Subset]
	     * In the internal DTD subset, parameter-entity references
	     * can occur only where markup declarations can occur, not
	     * within markup declarations.
	     * In that case this is handled in xmlParseMarkupDecl
	     */
	    if ((ctxt->external == 0) && (ctxt->inputNr == 1))
		return;
	    if (IS_BLANK_CH(NXT(1)) || NXT(1) == 0)
		return;
            break;
        case XML_PARSER_IGNORE:
            return;
    }

    NEXT;
    name = xmlParseName(ctxt);
    if (xmlParserDebugEntities)
	xmlGenericError(xmlGenericErrorContext,
		"PEReference: %s\n", name);
    if (name == NULL) {
	xmlFatalErr(ctxt, XML_ERR_PEREF_NO_NAME, NULL);
    } else {
	if (RAW == ';') {
	    NEXT;
	    if ((ctxt->sax != NULL) && (ctxt->sax->getParameterEntity != NULL))
		entity = ctxt->sax->getParameterEntity(ctxt->userData, name);
	    if (entity == NULL) {
	        
		/*
		 * [ WFC: Entity Declared ]
		 * In a document without any DTD, a document with only an
		 * internal DTD subset which contains no parameter entity
		 * references, or a document with "standalone='yes'", ...
		 * ... The declaration of a parameter entity must precede
		 * any reference to it...
		 */
		if ((ctxt->standalone == 1) ||
		    ((ctxt->hasExternalSubset == 0) &&
		     (ctxt->hasPErefs == 0))) {
		    xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
			 "PEReference: %%%s; not found\n", name);
	        } else {
		    /*
		     * [ VC: Entity Declared ]
		     * In a document with an external subset or external
		     * parameter entities with "standalone='no'", ...
		     * ... The declaration of a parameter entity must precede
		     * any reference to it...
		     */
		    if ((ctxt->validate) && (ctxt->vctxt.error != NULL)) {
		        xmlValidityError(ctxt, XML_WAR_UNDECLARED_ENTITY,
			                 "PEReference: %%%s; not found\n",
				         name);
		    } else 
		        xmlWarningMsg(ctxt, XML_WAR_UNDECLARED_ENTITY,
			              "PEReference: %%%s; not found\n",
				      name, NULL);
		    ctxt->valid = 0;
		}
	    } else if (ctxt->input->free != deallocblankswrapper) {
		    input = xmlNewBlanksWrapperInputStream(ctxt, entity);
		    xmlPushInput(ctxt, input);
	    } else {
	        if ((entity->etype == XML_INTERNAL_PARAMETER_ENTITY) ||
		    (entity->etype == XML_EXTERNAL_PARAMETER_ENTITY)) {
		    xmlChar start[4];
		    xmlCharEncoding enc;

		    /*
		     * handle the extra spaces added before and after
		     * c.f. http://www.w3.org/TR/REC-xml#as-PE
		     * this is done independently.
		     */
		    input = xmlNewEntityInputStream(ctxt, entity);
		    xmlPushInput(ctxt, input);

		    /* 
		     * Get the 4 first bytes and decode the charset
		     * if enc != XML_CHAR_ENCODING_NONE
		     * plug some encoding conversion routines.
		     * Note that, since we may have some non-UTF8
		     * encoding (like UTF16, bug 135229), the 'length'
		     * is not known, but we can calculate based upon
		     * the amount of data in the buffer.
		     */
		    GROW
		    if ((ctxt->input->end - ctxt->input->cur)>=4) {
			start[0] = RAW;
			start[1] = NXT(1);
			start[2] = NXT(2);
			start[3] = NXT(3);
			enc = xmlDetectCharEncoding(start, 4);
			if (enc != XML_CHAR_ENCODING_NONE) {
			    xmlSwitchEncoding(ctxt, enc);
			}
		    }

		    if ((entity->etype == XML_EXTERNAL_PARAMETER_ENTITY) &&
			(CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l' )) &&
			(IS_BLANK_CH(NXT(5)))) {
			xmlParseTextDecl(ctxt);
		    }
		} else {
		    xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_IS_PARAMETER,
			     "PEReference: %s is not a parameter entity\n",
				      name);
		}
	    }
	} else {
	    xmlFatalErr(ctxt, XML_ERR_PEREF_SEMICOL_MISSING, NULL);
	}
    }
}

/*
 * Macro used to grow the current buffer.
 */
#define growBuffer(buffer) {						\
    xmlChar *tmp;							\
    buffer##_size *= 2;							\
    tmp = (xmlChar *)							\
    		xmlRealloc(buffer, buffer##_size * sizeof(xmlChar));	\
    if (tmp == NULL) goto mem_error;					\
    buffer = tmp;							\
}

/**
 * xmlStringLenDecodeEntities:
 * @ctxt:  the parser context
 * @str:  the input string
 * @len: the string length
 * @what:  combination of XML_SUBSTITUTE_REF and XML_SUBSTITUTE_PEREF
 * @end:  an end marker xmlChar, 0 if none
 * @end2:  an end marker xmlChar, 0 if none
 * @end3:  an end marker xmlChar, 0 if none
 * 
 * Takes a entity string content and process to do the adequate substitutions.
 *
 * [67] Reference ::= EntityRef | CharRef
 *
 * [69] PEReference ::= '%' Name ';'
 *
 * Returns A newly allocated string with the substitution done. The caller
 *      must deallocate it !
 */
xmlChar *
xmlStringLenDecodeEntities(xmlParserCtxtPtr ctxt, const xmlChar *str, int len,
		      int what, xmlChar end, xmlChar  end2, xmlChar end3) {
    xmlChar *buffer = NULL;
    int buffer_size = 0;

    xmlChar *current = NULL;
    const xmlChar *last;
    xmlEntityPtr ent;
    int c,l;
    int nbchars = 0;

    if ((ctxt == NULL) || (str == NULL) || (len < 0))
	return(NULL);
    last = str + len;

    if (ctxt->depth > 40) {
	xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
	return(NULL);
    }

    /*
     * allocate a translation buffer.
     */
    buffer_size = XML_PARSER_BIG_BUFFER_SIZE;
    buffer = (xmlChar *) xmlMallocAtomic(buffer_size * sizeof(xmlChar));
    if (buffer == NULL) goto mem_error;

    /*
     * OK loop until we reach one of the ending char or a size limit.
     * we are operating on already parsed values.
     */
    if (str < last)
	c = CUR_SCHAR(str, l);
    else
        c = 0;
    while ((c != 0) && (c != end) && /* non input consuming loop */
	   (c != end2) && (c != end3)) {

	if (c == 0) break;
        if ((c == '&') && (str[1] == '#')) {
	    int val = xmlParseStringCharRef(ctxt, &str);
	    if (val != 0) {
		COPY_BUF(0,buffer,nbchars,val);
	    }
	    if (nbchars > buffer_size - XML_PARSER_BUFFER_SIZE) {
	        growBuffer(buffer);
	    }
	} else if ((c == '&') && (what & XML_SUBSTITUTE_REF)) {
	    if (xmlParserDebugEntities)
		xmlGenericError(xmlGenericErrorContext,
			"String decoding Entity Reference: %.30s\n",
			str);
	    ent = xmlParseStringEntityRef(ctxt, &str);
	    if ((ent != NULL) &&
		(ent->etype == XML_INTERNAL_PREDEFINED_ENTITY)) {
		if (ent->content != NULL) {
		    COPY_BUF(0,buffer,nbchars,ent->content[0]);
		    if (nbchars > buffer_size - XML_PARSER_BUFFER_SIZE) {
			growBuffer(buffer);
		    }
		} else {
		    xmlFatalErrMsg(ctxt, XML_ERR_INTERNAL_ERROR,
			    "predefined entity has no content\n");
		}
	    } else if ((ent != NULL) && (ent->content != NULL)) {
		xmlChar *rep;

		ctxt->depth++;
		rep = xmlStringDecodeEntities(ctxt, ent->content, what,
			                      0, 0, 0);
		ctxt->depth--;
		if (rep != NULL) {
		    current = rep;
		    while (*current != 0) { /* non input consuming loop */
			buffer[nbchars++] = *current++;
			if (nbchars >
		            buffer_size - XML_PARSER_BUFFER_SIZE) {
			    growBuffer(buffer);
			}
		    }
		    xmlFree(rep);
		}
	    } else if (ent != NULL) {
		int i = xmlStrlen(ent->name);
		const xmlChar *cur = ent->name;

		buffer[nbchars++] = '&';
		if (nbchars > buffer_size - i - XML_PARSER_BUFFER_SIZE) {
		    growBuffer(buffer);
		}
		for (;i > 0;i--)
		    buffer[nbchars++] = *cur++;
		buffer[nbchars++] = ';';
	    }
	} else if (c == '%' && (what & XML_SUBSTITUTE_PEREF)) {
	    if (xmlParserDebugEntities)
		xmlGenericError(xmlGenericErrorContext,
			"String decoding PE Reference: %.30s\n", str);
	    ent = xmlParseStringPEReference(ctxt, &str);
	    if (ent != NULL) {
		xmlChar *rep;

		ctxt