parser.c

SIP 1.5.0源代码

        case XML_PARSER_ATTRIBUTE_VALUE:
        case XML_PARSER_PI:
	case XML_PARSER_SYSTEM_LITERAL:
	    /* we just ignore it there */
	    return;
        case XML_PARSER_EPILOG:
	    ctxt->errNo = XML_ERR_PEREF_IN_EPILOG;
	    if ((ctxt->sax != NULL) && (ctxt->sax->error != NULL))
	        ctxt->sax->error(ctxt->userData, "PEReference in epilog!\n");
	    ctxt->wellFormed = 0;
	    ctxt->disableSAX = 1;
	    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 xmlDecodeEntities
	     */
	    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;
    }

    NEXT;
    name = xmlParseName(ctxt);
    if (name == NULL) {
        ctxt->errNo = XML_ERR_PEREF_NO_NAME;
	if ((ctxt->sax != NULL) && (ctxt->sax->error != NULL))
	    ctxt->sax->error(ctxt->userData, "xmlHandlePEReference: no name\n");
	ctxt->wellFormed = 0;
	ctxt->disableSAX = 1;
    } 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))) {
		    if ((ctxt->sax != NULL) && (ctxt->sax->error != NULL))
			ctxt->sax->error(ctxt->userData,
			 "PEReference: %%%s; not found\n", name);
		    ctxt->wellFormed = 0;
		    ctxt->disableSAX = 1;
	        } 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->sax != NULL) && (ctxt->sax->warning != NULL))
			ctxt->sax->warning(ctxt->userData,
			 "PEReference: %%%s; not found\n", name);
		    ctxt->valid = 0;
		}
	    } else {
	        if ((entity->etype == XML_INTERNAL_PARAMETER_ENTITY) ||
		    (entity->etype == XML_EXTERNAL_PARAMETER_ENTITY)) {
		    /*
		     * TODO !!! handle the extra spaces added before and after
		     * c.f. http://www.w3.org/TR/REC-xml#as-PE
		     */
		    input = xmlNewEntityInputStream(ctxt, entity);
		    xmlPushInput(ctxt, input);
		    if ((entity->etype == XML_EXTERNAL_PARAMETER_ENTITY) &&
			(RAW == '<') && (NXT(1) == '?') &&
			(NXT(2) == 'x') && (NXT(3) == 'm') &&
			(NXT(4) == 'l') && (IS_BLANK(NXT(5)))) {
			xmlParseTextDecl(ctxt);
		    }
		    if (ctxt->token == 0)
			ctxt->token = ' ';
		} else {
		    if ((ctxt->sax != NULL) && (ctxt->sax->error != NULL))
			ctxt->sax->error(ctxt->userData,
			 "xmlHandlePEReference: %s is not a parameter entity\n",
			                 name);
		    ctxt->wellFormed = 0;
		    ctxt->disableSAX = 1;
		}
	    }
	} else {
	    ctxt->errNo = XML_ERR_PEREF_SEMICOL_MISSING;
	    if ((ctxt->sax != NULL) && (ctxt->sax->error != NULL))
		ctxt->sax->error(ctxt->userData,
				 "xmlHandlePEReference: expecting ';'\n");
	    ctxt->wellFormed = 0;
	    ctxt->disableSAX = 1;
	}
	xmlFree(name);
    }
}

/*
 * Macro used to grow the current buffer.
 */
#define growBuffer(buffer) {						\
    buffer##_size *= 2;							\
    buffer = (xmlChar *)						\
    		xmlRealloc(buffer, buffer##_size * sizeof(xmlChar));	\
    if (buffer == NULL) {						\
	perror("realloc failed");					\
	return(NULL);							\
    }									\
}

/**
 * xmlDecodeEntities:
 * @ctxt:  the parser context
 * @what:  combination of XML_SUBSTITUTE_REF and XML_SUBSTITUTE_PEREF
 * @len:  the len to decode (in bytes !), -1 for no size limit
 * @end:  an end marker xmlChar, 0 if none
 * @end2:  an end marker xmlChar, 0 if none
 * @end3:  an end marker xmlChar, 0 if none
 * 
 * [67] Reference ::= EntityRef | CharRef
 *
 * [69] PEReference ::= '%' Name ';'
 *
 * Returns A newly allocated string with the substitution done. The caller
 *      must deallocate it !
 */
xmlChar *
xmlDecodeEntities(xmlParserCtxtPtr ctxt, int len, int what,
                  xmlChar end, xmlChar  end2, xmlChar end3) {
    xmlChar *buffer = NULL;
    int buffer_size = 0;
    int nbchars = 0;

    xmlChar *current = NULL;
    xmlEntityPtr ent;
    unsigned int max = (unsigned int) len;
    int c,l;

    if (ctxt->depth > 40) {
	if ((ctxt->sax != NULL) && (ctxt->sax->error != NULL))
	    ctxt->sax->error(ctxt->userData,
		"Detected entity reference loop\n");
	ctxt->wellFormed = 0;
	ctxt->disableSAX = 1;
	ctxt->errNo = XML_ERR_ENTITY_LOOP;
	return(NULL);
    }

    /*
     * allocate a translation buffer.
     */
    buffer_size = XML_PARSER_BIG_BUFFER_SIZE;
    buffer = (xmlChar *) xmlMalloc(buffer_size * sizeof(xmlChar));
    if (buffer == NULL) {
	perror("xmlDecodeEntities: malloc failed");
	return(NULL);
    }

    /*
     * Ok loop until we reach one of the ending char or a size limit.
     */
    c = CUR_CHAR(l);
    while ((nbchars < max) && (c != end) &&
           (c != end2) && (c != end3)) {

	if (c == 0) break;
        if (((c == '&') && (ctxt->token != '&')) && (NXT(1) == '#')) {
	    int val = xmlParseCharRef(ctxt);
	    COPY_BUF(0,buffer,nbchars,val);
	    NEXTL(l);
	} else if ((c == '&') && (ctxt->token != '&') &&
		   (what & XML_SUBSTITUTE_REF)) {
	    ent = xmlParseEntityRef(ctxt);
	    if ((ent != NULL) && 
		(ctxt->replaceEntities != 0)) {
		current = ent->content;
		while (*current != 0) {
		    buffer[nbchars++] = *current++;
		    if (nbchars > buffer_size - XML_PARSER_BUFFER_SIZE) {
			growBuffer(buffer);
		    }
		}
	    } else if (ent != NULL) {
		const xmlChar *cur = ent->name;

		buffer[nbchars++] = '&';
		if (nbchars > buffer_size - XML_PARSER_BUFFER_SIZE) {
		    growBuffer(buffer);
		}
		while (*cur != 0) {
		    buffer[nbchars++] = *cur++;
		}
		buffer[nbchars++] = ';';
	    }
	} else if (c == '%' && (what & XML_SUBSTITUTE_PEREF)) {
	    /*
	     * a PEReference induce to switch the entity flow,
	     * we break here to flush the current set of chars
	     * parsed if any. We will be called back later.
	     */
	    if (nbchars != 0) break;

	    xmlParsePEReference(ctxt);

	    /*
	     * Pop-up of finished entities.
	     */
	    while ((RAW == 0) && (ctxt->inputNr > 1))
		xmlPopInput(ctxt);

	    break;
	} else {
	    COPY_BUF(l,buffer,nbchars,c);
	    NEXTL(l);
	    if (nbchars > buffer_size - XML_PARSER_BUFFER_SIZE) {
	      growBuffer(buffer);
	    }
	}
	c = CUR_CHAR(l);
    }
    buffer[nbchars++] = 0;
    return(buffer);
}

/**
 * xmlStringDecodeEntities:
 * @ctxt:  the parser context
 * @str:  the input string
 * @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
 * 
 * [67] Reference ::= EntityRef | CharRef
 *
 * [69] PEReference ::= '%' Name ';'
 *
 * Returns A newly allocated string with the substitution done. The caller
 *      must deallocate it !
 */
xmlChar *
xmlStringDecodeEntities(xmlParserCtxtPtr ctxt, const xmlChar *str, int what,
		        xmlChar end, xmlChar  end2, xmlChar end3) {
    xmlChar *buffer = NULL;
    int buffer_size = 0;

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

    if (ctxt->depth > 40) {
	if ((ctxt->sax != NULL) && (ctxt->sax->error != NULL))
	    ctxt->sax->error(ctxt->userData,
		"Detected entity reference loop\n");
	ctxt->wellFormed = 0;
	ctxt->disableSAX = 1;
	ctxt->errNo = XML_ERR_ENTITY_LOOP;
	return(NULL);
    }

    /*
     * allocate a translation buffer.
     */
    buffer_size = XML_PARSER_BIG_BUFFER_SIZE;
    buffer = (xmlChar *) xmlMalloc(buffer_size * sizeof(xmlChar));
    if (buffer == NULL) {
	perror("xmlDecodeEntities: malloc failed");
	return(NULL);
    }

    /*
     * Ok loop until we reach one of the ending char or a size limit.
     */
    c = CUR_SCHAR(str, l);
    while ((c != 0) && (c != end) && (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);
	    }
	} else if ((c == '&') && (what & XML_SUBSTITUTE_REF)) {
	    ent = xmlParseStringEntityRef(ctxt, &str);
	    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) {
			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)) {
	    ent = xmlParseStringPEReference(ctxt, &str);
	    if (ent != NULL) {
		xmlChar *rep;

		ctxt->depth++;
		rep = xmlStringDecodeEntities(ctxt, ent->content, what,
			                      0, 0, 0);
		ctxt->depth--;
		if (rep != NULL) {
		    current = rep;
		    while (*current != 0) {
			buffer[nbchars++] = *current++;
			if (nbchars >
		            buffer_size - XML_PARSER_BUFFER_SIZE) {
			    growBuffer(buffer);
			}
		    }
		    xmlFree(rep);
		}
	    }
	} else {
	    COPY_BUF(l,buffer,nbchars,c);
	    str += l;
	    if (nbchars > buffer_size - XML_PARSER_BUFFER_SIZE) {
	      growBuffer(buffer);
	    }
	}
	c = CUR_SCHAR(str, l);
    }
    buffer[nbchars++] = 0;
    return(buffer);
}


/************************************************************************
 *									*
 *		Commodity functions to handle encodings			*
 *									*
 ************************************************************************/

/*
 * xmlCheckLanguageID
 * @lang:  pointer to the string value
 *
 * Checks that the value conforms to the LanguageID production:
 *
 * [33] LanguageID ::= Langcode ('-' Subcode)*
 * [34] Langcode ::= ISO639Code |  IanaCode |  UserCode
 * [35] ISO639Code ::= ([a-z] | [A-Z]) ([a-z] | [A-Z])
 * [36] IanaCode ::= ('i' | 'I') '-' ([a-z] | [A-Z])+
 * [37] UserCode ::= ('x' | 'X') '-' ([a-z] | [A-Z])+
 * [38] Subcode ::= ([a-z] | [A-Z])+
 *
 * Returns 1 if correct 0 otherwise
 **/
int
xmlCheckLanguageID(const xmlChar *lang) {
    const xmlChar *cur = lang;

    if (cur == NULL)
	return(0);
    if (((cur[0] == 'i') && (cur[1] == '-')) ||
	((cur[0] == 'I') && (cur[1] == '-'))) {
	/*
	 * IANA code
	 */
	cur += 2;
        while (((cur[0] >= 'A') && (cur[0] <= 'Z')) ||
	       ((cur[0] >= 'a') && (cur[0] <= 'z')))
	    cur++;
    } else if (((cur[0] == 'x') && (cur[1] == '-')) ||
	       ((cur[0] == 'X') && (cur[1] == '-'))) {
	/*
	 * User code
	 */
	cur += 2;
        while (((cur[0] >= 'A') && (cur[0] <= 'Z')) ||
	       ((cur[0] >= 'a') && (cur[0] <= 'z')))
	    cur++;
    } else if (((cur[0] >= 'A') && (cur[0] <= 'Z')) ||
	       ((cur[0] >= 'a') && (cur[0] <= 'z'))) {
	/*
	 * ISO639
	 */
	cur++;
        if (((cur[0] >= 'A') && (cur[0] <= 'Z')) ||
	    ((cur[0] >= 'a') && (cur[0] <= 'z')))
	    cur++;
	else
	    return(0);
    } else
	return(0);
    while (cur[0] != 0) {
	if (cur[0] != '-')
	    return(0);
	cur++;
        if (((cur[0] >= 'A') && (cur[0] <= 'Z')) ||
	    ((cur[0] >= 'a') && (cur[0] <= 'z')))
	    cur++;
	else
	    return(0);
        while (((cur[0] >= 'A') && (cur[0] <= 'Z')) ||
	       ((cur[0] >= 'a') && (cur[0] <= 'z')))
	    cur++;
    }
    return(1);
}

/**
 * xmlSwitchEncoding:
 * @ctxt:  the parser context
 * @enc:  the encoding value (number)
 *
 * change the input functions when discovering the character encoding
 * of a given entity.
 */
void
xmlSwitchEncoding(xmlParserCtxtPtr ctxt, xmlCharEncoding enc)
{
    xmlCharEncodingHandlerPtr handler;

    handler = xmlGetCharEncodingHandler(enc);
    if (handler != NULL) {
        if (ctxt->input != NU