Upgrade to latest version of xmldom

print-window-tiddler
Jermolene 2017-02-07 18:20:30 +00:00
rodzic d9fd722e50
commit f4a015f120
7 zmienionych plików z 557 dodań i 160 usunięć

Wyświetl plik

Wyświetl plik

@ -2,32 +2,31 @@ function DOMParser(options){
this.options = options ||{locator:{}};
}
DOMParser.prototype.parseFromString = function(source,mimeType){
DOMParser.prototype.parseFromString = function(source,mimeType){
var options = this.options;
var sax = new XMLReader();
var domBuilder = options.domBuilder || new DOMHandler();//contentHandler and LexicalHandler
var errorHandler = options.errorHandler;
var locator = options.locator;
var defaultNSMap = options.xmlns||{};
var entityMap = {'lt':'<','gt':'>','amp':'&','quot':'"','apos':"'"}
var isHTML = /\/x?html?$/.test(mimeType);//mimeType.toLowerCase().indexOf('html') > -1;
var entityMap = isHTML?require('./entities').entityMap:{'lt':'<','gt':'>','amp':'&','quot':'"','apos':"'"};
if(locator){
domBuilder.setDocumentLocator(locator)
}
sax.errorHandler = buildErrorHandler(errorHandler,domBuilder,locator);
sax.domBuilder = options.domBuilder || domBuilder;
if(/\/x?html?$/.test(mimeType)){
entityMap.nbsp = '\xa0';
entityMap.copy = '\xa9';
if(isHTML){
defaultNSMap['']= 'http://www.w3.org/1999/xhtml';
}
defaultNSMap.xml = defaultNSMap.xml || 'http://www.w3.org/XML/1998/namespace';
if(source){
sax.parse(source,defaultNSMap,entityMap);
}else{
sax.errorHandler.error("invalid document source");
sax.errorHandler.error("invalid doc source");
}
return domBuilder.document;
return domBuilder.doc;
}
function buildErrorHandler(errorImpl,domBuilder,locator){
if(!errorImpl){
@ -41,27 +40,20 @@ function buildErrorHandler(errorImpl,domBuilder,locator){
locator = locator||{}
function build(key){
var fn = errorImpl[key];
if(!fn){
if(isCallback){
fn = errorImpl.length == 2?function(msg){errorImpl(key,msg)}:errorImpl;
}else{
var i=arguments.length;
while(--i){
if(fn = errorImpl[arguments[i]]){
break;
}
}
}
if(!fn && isCallback){
fn = errorImpl.length == 2?function(msg){errorImpl(key,msg)}:errorImpl;
}
errorHandler[key] = fn && function(msg){
fn(msg+_locator(locator));
fn('[xmldom '+key+']\t'+msg+_locator(locator));
}||function(){};
}
build('warning','warn');
build('error','warn','warning');
build('fatalError','warn','warning','error');
build('warning');
build('error');
build('fatalError');
return errorHandler;
}
//console.log('#\n\n\n\n\n\n\n####')
/**
* +ContentHandler+ErrorHandler
* +LexicalHandler+EntityResolver2
@ -84,13 +76,13 @@ function position(locator,node){
*/
DOMHandler.prototype = {
startDocument : function() {
this.document = new DOMImplementation().createDocument(null, null, null);
this.doc = new DOMImplementation().createDocument(null, null, null);
if (this.locator) {
this.document.documentURI = this.locator.systemId;
this.doc.documentURI = this.locator.systemId;
}
},
startElement:function(namespaceURI, localName, qName, attrs) {
var doc = this.document;
var doc = this.doc;
var el = doc.createElementNS(namespaceURI, qName||localName);
var len = attrs.length;
appendElement(this, el);
@ -102,24 +94,22 @@ DOMHandler.prototype = {
var value = attrs.getValue(i);
var qName = attrs.getQName(i);
var attr = doc.createAttributeNS(namespaceURI, qName);
if( attr.getOffset){
position(attr.getOffset(1),attr)
}
this.locator &&position(attrs.getLocator(i),attr);
attr.value = attr.nodeValue = value;
el.setAttributeNode(attr)
}
},
endElement:function(namespaceURI, localName, qName) {
var current = this.currentElement
var tagName = current.tagName;
this.currentElement = current.parentNode;
var tagName = current.tagName;
this.currentElement = current.parentNode;
},
startPrefixMapping:function(prefix, uri) {
},
endPrefixMapping:function(prefix) {
},
processingInstruction:function(target, data) {
var ins = this.document.createProcessingInstruction(target, data);
var ins = this.doc.createProcessingInstruction(target, data);
this.locator && position(this.locator,ins)
appendElement(this, ins);
},
@ -128,13 +118,17 @@ DOMHandler.prototype = {
characters:function(chars, start, length) {
chars = _toString.apply(this,arguments)
//console.log(chars)
if(this.currentElement && chars){
if(chars){
if (this.cdata) {
var charNode = this.document.createCDATASection(chars);
this.currentElement.appendChild(charNode);
var charNode = this.doc.createCDATASection(chars);
} else {
var charNode = this.document.createTextNode(chars);
var charNode = this.doc.createTextNode(chars);
}
if(this.currentElement){
this.currentElement.appendChild(charNode);
}else if(/^\s*$/.test(chars)){
this.doc.appendChild(charNode);
//process xml
}
this.locator && position(this.locator,charNode)
}
@ -142,7 +136,7 @@ DOMHandler.prototype = {
skippedEntity:function(name) {
},
endDocument:function() {
this.document.normalize();
this.doc.normalize();
},
setDocumentLocator:function (locator) {
if(this.locator = locator){// && !('lineNumber' in locator)){
@ -152,7 +146,7 @@ DOMHandler.prototype = {
//LexicalHandler
comment:function(chars, start, length) {
chars = _toString.apply(this,arguments)
var comm = this.document.createComment(chars);
var comm = this.doc.createComment(chars);
this.locator && position(this.locator,comm)
appendElement(this, comm);
},
@ -166,7 +160,7 @@ DOMHandler.prototype = {
},
startDTD:function(name, publicId, systemId) {
var impl = this.document.implementation;
var impl = this.doc.implementation;
if (impl && impl.createDocumentType) {
var dt = impl.createDocumentType(name, publicId, systemId);
this.locator && position(this.locator,dt)
@ -178,13 +172,13 @@ DOMHandler.prototype = {
* @link http://www.saxproject.org/apidoc/org/xml/sax/ErrorHandler.html
*/
warning:function(error) {
console.warn(error,_locator(this.locator));
console.warn('[xmldom warning]\t'+error,_locator(this.locator));
},
error:function(error) {
console.error(error,_locator(this.locator));
console.error('[xmldom error]\t'+error,_locator(this.locator));
},
fatalError:function(error) {
console.error(error,_locator(this.locator));
console.error('[xmldom fatalError]\t'+error,_locator(this.locator));
throw error;
}
}
@ -242,15 +236,15 @@ function _toString(chars,start,length){
/* Private static helpers treated below as private instance methods, so don't need to add these to the public API; we might use a Relator to also get rid of non-standard public properties */
function appendElement (hander,node) {
if (!hander.currentElement) {
hander.document.appendChild(node);
hander.doc.appendChild(node);
} else {
hander.currentElement.appendChild(node);
}
}//appendChild and setAttributeNS are preformance key
if(typeof require == 'function'){
//if(typeof require == 'function'){
var XMLReader = require('./sax').XMLReader;
var DOMImplementation = exports.DOMImplementation = require('./dom').DOMImplementation;
exports.XMLSerializer = require('./dom').XMLSerializer ;
exports.DOMParser = DOMParser;
}
//}

Wyświetl plik

@ -109,6 +109,12 @@ NodeList.prototype = {
*/
item: function(index) {
return this[index] || null;
},
toString:function(isHTML,nodeFilter){
for(var buf = [], i = 0;i<this.length;i++){
serializeToString(this[i],buf,isHTML,nodeFilter);
}
return buf.join('');
}
};
function LiveNodeList(node,refresh){
@ -164,6 +170,7 @@ function _addNamedNode(el,list,newAttr,oldAttr){
}
}
function _removeNamedNode(el,list,attr){
//console.log('remove attr:'+attr)
var i = _findNodeIndex(list,attr);
if(i>=0){
var lastIndex = list.length-1
@ -179,7 +186,7 @@ function _removeNamedNode(el,list,attr){
}
}
}else{
throw DOMException(NOT_FOUND_ERR,new Error())
throw DOMException(NOT_FOUND_ERR,new Error(el.tagName+'@'+attr))
}
}
NamedNodeMap.prototype = {
@ -189,9 +196,11 @@ NamedNodeMap.prototype = {
// if(key.indexOf(':')>0 || key == 'xmlns'){
// return null;
// }
//console.log()
var i = this.length;
while(i--){
var attr = this[i];
//console.log(attr.nodeName,key)
if(attr.nodeName == key){
return attr;
}
@ -373,7 +382,7 @@ Node.prototype = {
}
}
}
el = el.nodeType == 2?el.ownerDocument : el.parentNode;
el = el.nodeType == ATTRIBUTE_NODE?el.ownerDocument : el.parentNode;
}
return null;
},
@ -388,7 +397,7 @@ Node.prototype = {
return map[prefix] ;
}
}
el = el.nodeType == 2?el.ownerDocument : el.parentNode;
el = el.nodeType == ATTRIBUTE_NODE?el.ownerDocument : el.parentNode;
}
return null;
},
@ -573,7 +582,7 @@ Document.prototype = {
}
return newChild;
}
if(this.documentElement == null && newChild.nodeType == 1){
if(this.documentElement == null && newChild.nodeType == ELEMENT_NODE){
this.documentElement = newChild;
}
@ -593,7 +602,7 @@ Document.prototype = {
getElementById : function(id){
var rtv = null;
_visitNode(this.documentElement,function(node){
if(node.nodeType == 1){
if(node.nodeType == ELEMENT_NODE){
if(node.getAttribute('id') == id){
rtv = node;
return true;
@ -742,6 +751,7 @@ Element.prototype = {
return this.attributes.setNamedItemNS(newAttr);
},
removeAttributeNode : function(oldAttr){
//console.log(this == oldAttr.ownerElement)
return this.attributes.removeNamedItem(oldAttr.nodeName);
},
//get real attribute name,and remove it by removeAttributeNode
@ -786,6 +796,7 @@ Element.prototype = {
}
});
return ls;
});
}
};
@ -817,10 +828,7 @@ CharacterData.prototype = {
},
appendChild:function(newChild){
//if(!(newChild instanceof CharacterData)){
throw new Error(ExceptionMessage[3])
//}
return Node.prototype.appendChild.apply(this,arguments)
throw new Error(ExceptionMessage[HIERARCHY_REQUEST_ERR])
},
deleteData: function(offset, count) {
this.replaceData(offset,count,"");
@ -902,39 +910,132 @@ function ProcessingInstruction() {
ProcessingInstruction.prototype.nodeType = PROCESSING_INSTRUCTION_NODE;
_extends(ProcessingInstruction,Node);
function XMLSerializer(){}
XMLSerializer.prototype.serializeToString = function(node,attributeSorter){
return node.toString(attributeSorter);
XMLSerializer.prototype.serializeToString = function(node,isHtml,nodeFilter){
return nodeSerializeToString.call(node,isHtml,nodeFilter);
}
Node.prototype.toString =function(attributeSorter){
Node.prototype.toString = nodeSerializeToString;
function nodeSerializeToString(isHtml,nodeFilter){
var buf = [];
serializeToString(this,buf,attributeSorter);
var refNode = this.nodeType == 9 && this.documentElement || this;
var prefix = refNode.prefix;
var uri = refNode.namespaceURI;
if(uri && prefix == null){
//console.log(prefix)
var prefix = refNode.lookupPrefix(uri);
if(prefix == null){
//isHTML = true;
var visibleNamespaces=[
{namespace:uri,prefix:null}
//{namespace:uri,prefix:''}
]
}
}
serializeToString(this,buf,isHtml,nodeFilter,visibleNamespaces);
//console.log('###',this.nodeType,uri,prefix,buf.join(''))
return buf.join('');
}
function serializeToString(node,buf,attributeSorter,isHTML){
function needNamespaceDefine(node,isHTML, visibleNamespaces) {
var prefix = node.prefix||'';
var uri = node.namespaceURI;
if (!prefix && !uri){
return false;
}
if (prefix === "xml" && uri === "http://www.w3.org/XML/1998/namespace"
|| uri == 'http://www.w3.org/2000/xmlns/'){
return false;
}
var i = visibleNamespaces.length
//console.log('@@@@',node.tagName,prefix,uri,visibleNamespaces)
while (i--) {
var ns = visibleNamespaces[i];
// get namespace prefix
//console.log(node.nodeType,node.tagName,ns.prefix,prefix)
if (ns.prefix == prefix){
return ns.namespace != uri;
}
}
//console.log(isHTML,uri,prefix=='')
//if(isHTML && prefix ==null && uri == 'http://www.w3.org/1999/xhtml'){
// return false;
//}
//node.flag = '11111'
//console.error(3,true,node.flag,node.prefix,node.namespaceURI)
return true;
}
function serializeToString(node,buf,isHTML,nodeFilter,visibleNamespaces){
if(nodeFilter){
node = nodeFilter(node);
if(node){
if(typeof node == 'string'){
buf.push(node);
return;
}
}else{
return;
}
//buf.sort.apply(attrs, attributeSorter);
}
switch(node.nodeType){
case ELEMENT_NODE:
if (!visibleNamespaces) visibleNamespaces = [];
var startVisibleNamespaces = visibleNamespaces.length;
var attrs = node.attributes;
var len = attrs.length;
var child = node.firstChild;
var nodeName = node.tagName;
isHTML = (htmlns === node.namespaceURI) ||isHTML
buf.push('<',nodeName);
if(attributeSorter){
buf.sort.apply(attrs, attributeSorter);
for(var i=0;i<len;i++){
// add namespaces for attributes
var attr = attrs.item(i);
if (attr.prefix == 'xmlns') {
visibleNamespaces.push({ prefix: attr.localName, namespace: attr.value });
}else if(attr.nodeName == 'xmlns'){
visibleNamespaces.push({ prefix: '', namespace: attr.value });
}
}
for(var i=0;i<len;i++){
serializeToString(attrs.item(i),buf,attributeSorter,isHTML);
var attr = attrs.item(i);
if (needNamespaceDefine(attr,isHTML, visibleNamespaces)) {
var prefix = attr.prefix||'';
var uri = attr.namespaceURI;
var ns = prefix ? ' xmlns:' + prefix : " xmlns";
buf.push(ns, '="' , uri , '"');
visibleNamespaces.push({ prefix: prefix, namespace:uri });
}
serializeToString(attr,buf,isHTML,nodeFilter,visibleNamespaces);
}
if(child || isHTML && !/^(?:meta|link|img|br|hr|input|button)$/i.test(nodeName)){
// add namespace for current node
if (needNamespaceDefine(node,isHTML, visibleNamespaces)) {
var prefix = node.prefix||'';
var uri = node.namespaceURI;
var ns = prefix ? ' xmlns:' + prefix : " xmlns";
buf.push(ns, '="' , uri , '"');
visibleNamespaces.push({ prefix: prefix, namespace:uri });
}
if(child || isHTML && !/^(?:meta|link|img|br|hr|input)$/i.test(nodeName)){
buf.push('>');
//if is cdata child node
if(isHTML && /^script$/i.test(nodeName)){
if(child){
buf.push(child.data);
}
}else{
while(child){
serializeToString(child,buf,attributeSorter,isHTML);
if(child.data){
buf.push(child.data);
}else{
serializeToString(child,buf,isHTML,nodeFilter,visibleNamespaces);
}
child = child.nextSibling;
}
}else
{
while(child){
serializeToString(child,buf,isHTML,nodeFilter,visibleNamespaces);
child = child.nextSibling;
}
}
@ -942,12 +1043,14 @@ function serializeToString(node,buf,attributeSorter,isHTML){
}else{
buf.push('/>');
}
// remove added visible namespaces
//visibleNamespaces.length = startVisibleNamespaces;
return;
case DOCUMENT_NODE:
case DOCUMENT_FRAGMENT_NODE:
var child = node.firstChild;
while(child){
serializeToString(child,buf,attributeSorter,isHTML);
serializeToString(child,buf,isHTML,nodeFilter,visibleNamespaces);
child = child.nextSibling;
}
return;
@ -1092,8 +1195,8 @@ try{
},
set:function(data){
switch(this.nodeType){
case 1:
case 11:
case ELEMENT_NODE:
case DOCUMENT_FRAGMENT_NODE:
while(this.firstChild){
this.removeChild(this.firstChild);
}
@ -1104,7 +1207,7 @@ try{
default:
//TODO:
this.data = data;
this.value = value;
this.value = data;
this.nodeValue = data;
}
}
@ -1112,8 +1215,8 @@ try{
function getTextContent(node){
switch(node.nodeType){
case 1:
case 11:
case ELEMENT_NODE:
case DOCUMENT_FRAGMENT_NODE:
var buf = [];
node = node.firstChild;
while(node){
@ -1135,7 +1238,7 @@ try{
}catch(e){//ie8
}
if(typeof require == 'function'){
//if(typeof require == 'function'){
exports.DOMImplementation = DOMImplementation;
exports.XMLSerializer = XMLSerializer;
}
//}

Wyświetl plik

@ -0,0 +1,244 @@
exports.entityMap = {
lt: '<',
gt: '>',
amp: '&',
quot: '"',
apos: "'",
Agrave: "À",
Aacute: "Á",
Acirc: "Â",
Atilde: "Ã",
Auml: "Ä",
Aring: "Å",
AElig: "Æ",
Ccedil: "Ç",
Egrave: "È",
Eacute: "É",
Ecirc: "Ê",
Euml: "Ë",
Igrave: "Ì",
Iacute: "Í",
Icirc: "Î",
Iuml: "Ï",
ETH: "Ð",
Ntilde: "Ñ",
Ograve: "Ò",
Oacute: "Ó",
Ocirc: "Ô",
Otilde: "Õ",
Ouml: "Ö",
Oslash: "Ø",
Ugrave: "Ù",
Uacute: "Ú",
Ucirc: "Û",
Uuml: "Ü",
Yacute: "Ý",
THORN: "Þ",
szlig: "ß",
agrave: "à",
aacute: "á",
acirc: "â",
atilde: "ã",
auml: "ä",
aring: "å",
aelig: "æ",
ccedil: "ç",
egrave: "è",
eacute: "é",
ecirc: "ê",
euml: "ë",
igrave: "ì",
iacute: "í",
icirc: "î",
iuml: "ï",
eth: "ð",
ntilde: "ñ",
ograve: "ò",
oacute: "ó",
ocirc: "ô",
otilde: "õ",
ouml: "ö",
oslash: "ø",
ugrave: "ù",
uacute: "ú",
ucirc: "û",
uuml: "ü",
yacute: "ý",
thorn: "þ",
yuml: "ÿ",
nbsp: " ",
iexcl: "¡",
cent: "¢",
pound: "£",
curren: "¤",
yen: "¥",
brvbar: "¦",
sect: "§",
uml: "¨",
copy: "©",
ordf: "ª",
laquo: "«",
not: "¬",
shy: "­­",
reg: "®",
macr: "¯",
deg: "°",
plusmn: "±",
sup2: "²",
sup3: "³",
acute: "´",
micro: "µ",
para: "¶",
middot: "·",
cedil: "¸",
sup1: "¹",
ordm: "º",
raquo: "»",
frac14: "¼",
frac12: "½",
frac34: "¾",
iquest: "¿",
times: "×",
divide: "÷",
forall: "∀",
part: "∂",
exist: "∃",
empty: "∅",
nabla: "∇",
isin: "∈",
notin: "∉",
ni: "∋",
prod: "∏",
sum: "∑",
minus: "",
lowast: "",
radic: "√",
prop: "∝",
infin: "∞",
ang: "∠",
and: "∧",
or: "",
cap: "∩",
cup: "",
'int': "∫",
there4: "∴",
sim: "",
cong: "≅",
asymp: "≈",
ne: "≠",
equiv: "≡",
le: "≤",
ge: "≥",
sub: "⊂",
sup: "⊃",
nsub: "⊄",
sube: "⊆",
supe: "⊇",
oplus: "⊕",
otimes: "⊗",
perp: "⊥",
sdot: "⋅",
Alpha: "Α",
Beta: "Β",
Gamma: "Γ",
Delta: "Δ",
Epsilon: "Ε",
Zeta: "Ζ",
Eta: "Η",
Theta: "Θ",
Iota: "Ι",
Kappa: "Κ",
Lambda: "Λ",
Mu: "Μ",
Nu: "Ν",
Xi: "Ξ",
Omicron: "Ο",
Pi: "Π",
Rho: "Ρ",
Sigma: "Σ",
Tau: "Τ",
Upsilon: "Υ",
Phi: "Φ",
Chi: "Χ",
Psi: "Ψ",
Omega: "Ω",
alpha: "α",
beta: "β",
gamma: "γ",
delta: "δ",
epsilon: "ε",
zeta: "ζ",
eta: "η",
theta: "θ",
iota: "ι",
kappa: "κ",
lambda: "λ",
mu: "μ",
nu: "ν",
xi: "ξ",
omicron: "ο",
pi: "π",
rho: "ρ",
sigmaf: "ς",
sigma: "σ",
tau: "τ",
upsilon: "υ",
phi: "φ",
chi: "χ",
psi: "ψ",
omega: "ω",
thetasym: "ϑ",
upsih: "ϒ",
piv: "ϖ",
OElig: "Œ",
oelig: "œ",
Scaron: "Š",
scaron: "š",
Yuml: "Ÿ",
fnof: "ƒ",
circ: "ˆ",
tilde: "˜",
ensp: "",
emsp: "",
thinsp: "",
zwnj: "",
zwj: "",
lrm: "",
rlm: "",
ndash: "–",
mdash: "—",
lsquo: "",
rsquo: "",
sbquo: "",
ldquo: "“",
rdquo: "”",
bdquo: "„",
dagger: "†",
Dagger: "‡",
bull: "•",
hellip: "…",
permil: "‰",
prime: "",
Prime: "″",
lsaquo: "",
rsaquo: "",
oline: "‾",
euro: "€",
trade: "™",
larr: "←",
uarr: "↑",
rarr: "→",
darr: "↓",
harr: "↔",
crarr: "↵",
lceil: "⌈",
rceil: "⌉",
lfloor: "⌊",
rfloor: "⌋",
loz: "◊",
spades: "♠",
clubs: "♣",
hearts: "♥",
diams: "♦"
};
//for(var n in exports.entityMap){console.log(exports.entityMap[n].charCodeAt())}

Wyświetl plik

Wyświetl plik

@ -2,21 +2,21 @@
//[4a] NameChar ::= NameStartChar | "-" | "." | [0-9] | #xB7 | [#x0300-#x036F] | [#x203F-#x2040]
//[5] Name ::= NameStartChar (NameChar)*
var nameStartChar = /[A-Z_a-z\xC0-\xD6\xD8-\xF6\u00F8-\u02FF\u0370-\u037D\u037F-\u1FFF\u200C-\u200D\u2070-\u218F\u2C00-\u2FEF\u3001-\uD7FF\uF900-\uFDCF\uFDF0-\uFFFD]///\u10000-\uEFFFF
var nameChar = new RegExp("[\\-\\.0-9"+nameStartChar.source.slice(1,-1)+"\u00B7\u0300-\u036F\\u203F-\u2040]");
var nameChar = new RegExp("[\\-\\.0-9"+nameStartChar.source.slice(1,-1)+"\\u00B7\\u0300-\\u036F\\u203F-\\u2040]");
var tagNamePattern = new RegExp('^'+nameStartChar.source+nameChar.source+'*(?:\:'+nameStartChar.source+nameChar.source+'*)?$');
//var tagNamePattern = /^[a-zA-Z_][\w\-\.]*(?:\:[a-zA-Z_][\w\-\.]*)?$/
//var handlers = 'resolveEntity,getExternalSubset,characters,endDocument,endElement,endPrefixMapping,ignorableWhitespace,processingInstruction,setDocumentLocator,skippedEntity,startDocument,startElement,startPrefixMapping,notationDecl,unparsedEntityDecl,error,fatalError,warning,attributeDecl,elementDecl,externalEntityDecl,internalEntityDecl,comment,endCDATA,endDTD,endEntity,startCDATA,startDTD,startEntity'.split(',')
//S_TAG, S_ATTR, S_EQ, S_V
//S_ATTR_S, S_E, S_S, S_C
//S_TAG, S_ATTR, S_EQ, S_ATTR_NOQUOT_VALUE
//S_ATTR_SPACE, S_ATTR_END, S_TAG_SPACE, S_TAG_CLOSE
var S_TAG = 0;//tag name offerring
var S_ATTR = 1;//attr name offerring
var S_ATTR_S=2;//attr name end and space offer
var S_ATTR_SPACE=2;//attr name end and space offer
var S_EQ = 3;//=space?
var S_V = 4;//attr value(no quot value only)
var S_E = 5;//attr value end and no space(quot end)
var S_S = 6;//(attr value end || tag end ) && (space offer)
var S_C = 7;//closed el<el />
var S_ATTR_NOQUOT_VALUE = 4;//attr value(no quot value only)
var S_ATTR_END = 5;//attr value end and no space(quot end)
var S_TAG_SPACE = 6;//(attr value end || tag end ) && (space offer)
var S_TAG_CLOSE = 7;//closed el<el />
function XMLReader(){
@ -33,7 +33,7 @@ XMLReader.prototype = {
}
}
function parse(source,defaultNSMapCopy,entityMap,domBuilder,errorHandler){
function fixedFromCharCode(code) {
function fixedFromCharCode(code) {
// String.prototype.fromCharCode does not supports
// > 2 bytes unicode chars directly
if (code > 0xffff) {
@ -76,7 +76,7 @@ function parse(source,defaultNSMapCopy,entityMap,domBuilder,errorHandler){
}
var lineStart = 0;
var lineEnd = 0;
var linePattern = /.+(?:\r\n?|\n)|.*$/g
var linePattern = /.*(?:\r\n?|\n)|.*$/g
var locator = domBuilder.locator;
var parseStack = [{currentNSMap:defaultNSMapCopy}]
@ -87,7 +87,7 @@ function parse(source,defaultNSMapCopy,entityMap,domBuilder,errorHandler){
var tagStart = source.indexOf('<',start);
if(tagStart<0){
if(!source.substr(start).match(/^\s*$/)){
var doc = domBuilder.document;
var doc = domBuilder.doc;
var text = doc.createTextNode(source.substr(start));
doc.appendChild(text);
domBuilder.currentElement = text;
@ -102,16 +102,36 @@ function parse(source,defaultNSMapCopy,entityMap,domBuilder,errorHandler){
var end = source.indexOf('>',tagStart+3);
var tagName = source.substring(tagStart+2,end);
var config = parseStack.pop();
var localNSMap = config.localNSMap;
if(config.tagName != tagName){
errorHandler.fatalError("end tag name: "+tagName+' is not match the current start tagName:'+config.tagName );
}
domBuilder.endElement(config.uri,config.localName,tagName);
if(localNSMap){
for(var prefix in localNSMap){
domBuilder.endPrefixMapping(prefix) ;
}
if(end<0){
tagName = source.substring(tagStart+2).replace(/[\s<].*/,'');
//console.error('#@@@@@@'+tagName)
errorHandler.error("end tag name: "+tagName+' is not complete:'+config.tagName);
end = tagStart+1+tagName.length;
}else if(tagName.match(/\s</)){
tagName = tagName.replace(/[\s<].*/,'');
errorHandler.error("end tag name: "+tagName+' maybe not complete');
end = tagStart+1+tagName.length;
}
//console.error(parseStack.length,parseStack)
//console.error(config);
var localNSMap = config.localNSMap;
var endMatch = config.tagName == tagName;
var endIgnoreCaseMach = endMatch || config.tagName&&config.tagName.toLowerCase() == tagName.toLowerCase()
if(endIgnoreCaseMach){
domBuilder.endElement(config.uri,config.localName,tagName);
if(localNSMap){
for(var prefix in localNSMap){
domBuilder.endPrefixMapping(prefix) ;
}
}
if(!endMatch){
errorHandler.fatalError("end tag name: "+tagName+' is not match the current start tagName:'+config.tagName );
}
}else{
parseStack.push(config)
}
end++;
break;
// end elment
@ -124,33 +144,40 @@ function parse(source,defaultNSMapCopy,entityMap,domBuilder,errorHandler){
end = parseDCC(source,tagStart,domBuilder,errorHandler);
break;
default:
locator&&position(tagStart);
var el = new ElementAttributes();
var currentNSMap = parseStack[parseStack.length-1].currentNSMap;
//elStartEnd
var end = parseElementStartPart(source,tagStart,el,entityReplacer,errorHandler);
var end = parseElementStartPart(source,tagStart,el,currentNSMap,entityReplacer,errorHandler);
var len = el.length;
if(locator){
if(len){
//attribute position fixed
for(var i = 0;i<len;i++){
var a = el[i];
position(a.offset);
a.offset = copyLocator(locator,{});
}
}
position(end);
}
if(!el.closed && fixSelfClosed(source,end,el.tagName,closeMap)){
el.closed = true;
if(!entityMap.nbsp){
errorHandler.warning('unclosed xml attribute');
}
}
appendElement(el,domBuilder,parseStack);
if(locator && len){
var locator2 = copyLocator(locator,{});
//try{//attribute position fixed
for(var i = 0;i<len;i++){
var a = el[i];
position(a.offset);
a.locator = copyLocator(locator,{});
}
//}catch(e){console.error('@@@@@'+e)}
domBuilder.locator = locator2
if(appendElement(el,domBuilder,currentNSMap)){
parseStack.push(el)
}
domBuilder.locator = locator;
}else{
if(appendElement(el,domBuilder,currentNSMap)){
parseStack.push(el)
}
}
if(el.uri === 'http://www.w3.org/1999/xhtml' && !el.closed){
@ -160,8 +187,10 @@ function parse(source,defaultNSMapCopy,entityMap,domBuilder,errorHandler){
}
}
}catch(e){
errorHandler.error('element parse error: '+e);
errorHandler.error('element parse error: '+e)
//errorHandler.error('element parse error: '+e);
end = -1;
//throw e;
}
if(end>start){
start = end;
@ -181,7 +210,7 @@ function copyLocator(f,t){
* @see #appendElement(source,elStartEnd,el,selfClosed,entityReplacer,domBuilder,parseStack);
* @return end of the elementStartPart(end of elementEndPart for selfClosed el)
*/
function parseElementStartPart(source,start,el,entityReplacer,errorHandler){
function parseElementStartPart(source,start,el,currentNSMap,entityReplacer,errorHandler){
var attrName;
var value;
var p = ++start;
@ -193,7 +222,7 @@ function parseElementStartPart(source,start,el,entityReplacer,errorHandler){
if(s === S_ATTR){//attrName
attrName = source.slice(start,p);
s = S_EQ;
}else if(s === S_ATTR_S){
}else if(s === S_ATTR_SPACE){
s = S_EQ;
}else{
//fatalError: equal must after attrName or space after attrName
@ -202,25 +231,30 @@ function parseElementStartPart(source,start,el,entityReplacer,errorHandler){
break;
case '\'':
case '"':
if(s === S_EQ){//equal
if(s === S_EQ || s === S_ATTR //|| s == S_ATTR_SPACE
){//equal
if(s === S_ATTR){
errorHandler.warning('attribute value must after "="')
attrName = source.slice(start,p)
}
start = p+1;
p = source.indexOf(c,start)
if(p>0){
value = source.slice(start,p).replace(/&#?\w+;/g,entityReplacer);
el.add(attrName,value,start-1);
s = S_E;
s = S_ATTR_END;
}else{
//fatalError: no end quot match
throw new Error('attribute value no end \''+c+'\' match');
}
}else if(s == S_V){
}else if(s == S_ATTR_NOQUOT_VALUE){
value = source.slice(start,p).replace(/&#?\w+;/g,entityReplacer);
//console.log(attrName,value,start,p)
el.add(attrName,value,start);
//console.dir(el)
errorHandler.warning('attribute "'+attrName+'" missed start quot('+c+')!!');
start = p+1;
s = S_E
s = S_ATTR_END
}else{
//fatalError: no equal before
throw new Error('attribute value must after "="');
@ -230,14 +264,14 @@ function parseElementStartPart(source,start,el,entityReplacer,errorHandler){
switch(s){
case S_TAG:
el.setTagName(source.slice(start,p));
case S_E:
case S_S:
case S_C:
s = S_C;
case S_ATTR_END:
case S_TAG_SPACE:
case S_TAG_CLOSE:
s =S_TAG_CLOSE;
el.closed = true;
case S_V:
case S_ATTR_NOQUOT_VALUE:
case S_ATTR:
case S_ATTR_S:
case S_ATTR_SPACE:
break;
//case S_EQ:
default:
@ -247,30 +281,36 @@ function parseElementStartPart(source,start,el,entityReplacer,errorHandler){
case ''://end document
//throw new Error('unexpected end of input')
errorHandler.error('unexpected end of input');
if(s == S_TAG){
el.setTagName(source.slice(start,p));
}
return p;
case '>':
switch(s){
case S_TAG:
el.setTagName(source.slice(start,p));
case S_E:
case S_S:
case S_C:
case S_ATTR_END:
case S_TAG_SPACE:
case S_TAG_CLOSE:
break;//normal
case S_V://Compatible state
case S_ATTR_NOQUOT_VALUE://Compatible state
case S_ATTR:
value = source.slice(start,p);
if(value.slice(-1) === '/'){
el.closed = true;
value = value.slice(0,-1)
}
case S_ATTR_S:
if(s === S_ATTR_S){
case S_ATTR_SPACE:
if(s === S_ATTR_SPACE){
value = attrName;
}
if(s == S_V){
if(s == S_ATTR_NOQUOT_VALUE){
errorHandler.warning('attribute "'+value+'" missed quot(")!!');
el.add(attrName,value.replace(/&#?\w+;/g,entityReplacer),start)
}else{
errorHandler.warning('attribute "'+value+'" missed value!! "'+value+'" instead!!')
if(currentNSMap[''] !== 'http://www.w3.org/1999/xhtml' || !value.match(/^(?:disabled|checked|selected)$/i)){
errorHandler.warning('attribute "'+value+'" missed value!! "'+value+'" instead!!')
}
el.add(value,value,start)
}
break;
@ -287,64 +327,68 @@ function parseElementStartPart(source,start,el,entityReplacer,errorHandler){
switch(s){
case S_TAG:
el.setTagName(source.slice(start,p));//tagName
s = S_S;
s = S_TAG_SPACE;
break;
case S_ATTR:
attrName = source.slice(start,p)
s = S_ATTR_S;
s = S_ATTR_SPACE;
break;
case S_V:
case S_ATTR_NOQUOT_VALUE:
var value = source.slice(start,p).replace(/&#?\w+;/g,entityReplacer);
errorHandler.warning('attribute "'+value+'" missed quot(")!!');
el.add(attrName,value,start)
case S_E:
s = S_S;
case S_ATTR_END:
s = S_TAG_SPACE;
break;
//case S_S:
//case S_TAG_SPACE:
//case S_EQ:
//case S_ATTR_S:
//case S_ATTR_SPACE:
// void();break;
//case S_C:
//case S_TAG_CLOSE:
//ignore warning
}
}else{//not space
//S_TAG, S_ATTR, S_EQ, S_V
//S_ATTR_S, S_E, S_S, S_C
//S_TAG, S_ATTR, S_EQ, S_ATTR_NOQUOT_VALUE
//S_ATTR_SPACE, S_ATTR_END, S_TAG_SPACE, S_TAG_CLOSE
switch(s){
//case S_TAG:void();break;
//case S_ATTR:void();break;
//case S_V:void();break;
case S_ATTR_S:
errorHandler.warning('attribute "'+attrName+'" missed value!! "'+attrName+'" instead!!')
//case S_ATTR_NOQUOT_VALUE:void();break;
case S_ATTR_SPACE:
var tagName = el.tagName;
if(currentNSMap[''] !== 'http://www.w3.org/1999/xhtml' || !attrName.match(/^(?:disabled|checked|selected)$/i)){
errorHandler.warning('attribute "'+attrName+'" missed value!! "'+attrName+'" instead2!!')
}
el.add(attrName,attrName,start);
start = p;
s = S_ATTR;
break;
case S_E:
case S_ATTR_END:
errorHandler.warning('attribute space is required"'+attrName+'"!!')
case S_S:
case S_TAG_SPACE:
s = S_ATTR;
start = p;
break;
case S_EQ:
s = S_V;
s = S_ATTR_NOQUOT_VALUE;
start = p;
break;
case S_C:
case S_TAG_CLOSE:
throw new Error("elements closed character '/' and '>' must be connected to");
}
}
}
}//end outer switch
//console.log('p++',p)
p++;
}
}
/**
* @return end of the elementStartPart(end of elementEndPart for selfClosed el)
* @return true if has new namespace define
*/
function appendElement(el,domBuilder,parseStack){
function appendElement(el,domBuilder,currentNSMap){
var tagName = el.tagName;
var localNSMap = null;
var currentNSMap = parseStack[parseStack.length-1].currentNSMap;
//var currentNSMap = parseStack[parseStack.length-1].currentNSMap;
var i = el.length;
while(i--){
var a = el[i];
@ -383,7 +427,7 @@ function appendElement(el,domBuilder,parseStack){
if(prefix === 'xml'){
a.uri = 'http://www.w3.org/XML/1998/namespace';
}if(prefix !== 'xmlns'){
a.uri = currentNSMap[prefix]
a.uri = currentNSMap[prefix || '']
//{console.log('###'+a.qName,domBuilder.locator.systemId+'',currentNSMap,a.uri)}
}
@ -412,7 +456,8 @@ function appendElement(el,domBuilder,parseStack){
}else{
el.currentNSMap = currentNSMap;
el.localNSMap = localNSMap;
parseStack.push(el);
//parseStack.push(el);
return true;
}
}
function parseHtmlSpecialContent(source,elStartEnd,tagName,entityReplacer,domBuilder){
@ -442,7 +487,11 @@ function fixSelfClosed(source,elStartEnd,tagName,closeMap){
var pos = closeMap[tagName];
if(pos == null){
//console.log(tagName)
pos = closeMap[tagName] = source.lastIndexOf('</'+tagName+'>')
pos = source.lastIndexOf('</'+tagName+'>')
if(pos<elStartEnd){//忘记闭合
pos = source.lastIndexOf('</'+tagName)
}
closeMap[tagName] =pos
}
return pos<elStartEnd;
//}
@ -533,7 +582,7 @@ ElementAttributes.prototype = {
},
length:0,
getLocalName:function(i){return this[i].localName},
getOffset:function(i){return this[i].offset},
getLocator:function(i){return this[i].locator},
getQName:function(i){return this[i].qName},
getURI:function(i){return this[i].uri},
getValue:function(i){return this[i].value}
@ -580,7 +629,5 @@ function split(source,start){
}
}
if(typeof require == 'function'){
exports.XMLReader = XMLReader;
}
exports.XMLReader = XMLReader;

Wyświetl plik

@ -18,6 +18,15 @@
},
"prefix": "",
"suffix": ""
},{
"file": "entities.js",
"fields": {
"type": "application/javascript",
"title": "$:/plugins/tiddlywiki/xmldom/entities",
"module-type": "library"
},
"prefix": "",
"suffix": ""
},{
"file": "sax.js",
"fields": {