0
|
1 /* This file defines an XML parser, with a few kludges to make it
|
|
2 * useable for HTML. autoSelfClosers defines a set of tag names that
|
|
3 * are expected to not have a closing tag, and doNotIndent specifies
|
|
4 * the tags inside of which no indentation should happen (see Config
|
|
5 * object). These can be disabled by passing the editor an object like
|
|
6 * {useHTMLKludges: false} as parserConfig option.
|
|
7 */
|
|
8
|
|
9 var XMLParser = Editor.Parser = (function() {
|
|
10 var Kludges = {
|
|
11 autoSelfClosers: {"br": true, "img": true, "hr": true, "link": true, "input": true,
|
|
12 "meta": true, "col": true, "frame": true, "base": true, "area": true},
|
|
13 doNotIndent: {"pre": true, "!cdata": true}
|
|
14 };
|
|
15 var NoKludges = {autoSelfClosers: {}, doNotIndent: {"!cdata": true}};
|
|
16 var UseKludges = Kludges;
|
|
17 var alignCDATA = false;
|
|
18
|
|
19 // Simple stateful tokenizer for XML documents. Returns a
|
|
20 // MochiKit-style iterator, with a state property that contains a
|
|
21 // function encapsulating the current state. See tokenize.js.
|
|
22 var tokenizeXML = (function() {
|
|
23 function inText(source, setState) {
|
|
24 var ch = source.next();
|
|
25 if (ch == "<") {
|
|
26 if (source.equals("!")) {
|
|
27 source.next();
|
|
28 if (source.equals("[")) {
|
|
29 if (source.lookAhead("[CDATA[", true)) {
|
|
30 setState(inBlock("xml-cdata", "]]>"));
|
|
31 return null;
|
|
32 }
|
|
33 else {
|
|
34 return "xml-text";
|
|
35 }
|
|
36 }
|
|
37 else if (source.lookAhead("--", true)) {
|
|
38 setState(inBlock("xml-comment", "-->"));
|
|
39 return null;
|
|
40 }
|
|
41 else if (source.lookAhead("DOCTYPE", true)) {
|
|
42 source.nextWhileMatches(/[\w\._\-]/);
|
|
43 setState(inBlock("xml-doctype", ">"));
|
|
44 return "xml-doctype";
|
|
45 }
|
|
46 else {
|
|
47 return "xml-text";
|
|
48 }
|
|
49 }
|
|
50 else if (source.equals("?")) {
|
|
51 source.next();
|
|
52 source.nextWhileMatches(/[\w\._\-]/);
|
|
53 setState(inBlock("xml-processing", "?>"));
|
|
54 return "xml-processing";
|
|
55 }
|
|
56 else {
|
|
57 if (source.equals("/")) source.next();
|
|
58 setState(inTag);
|
|
59 return "xml-punctuation";
|
|
60 }
|
|
61 }
|
|
62 else if (ch == "&") {
|
|
63 while (!source.endOfLine()) {
|
|
64 if (source.next() == ";")
|
|
65 break;
|
|
66 }
|
|
67 return "xml-entity";
|
|
68 }
|
|
69 else {
|
|
70 source.nextWhileMatches(/[^&<\n]/);
|
|
71 return "xml-text";
|
|
72 }
|
|
73 }
|
|
74
|
|
75 function inTag(source, setState) {
|
|
76 var ch = source.next();
|
|
77 if (ch == ">") {
|
|
78 setState(inText);
|
|
79 return "xml-punctuation";
|
|
80 }
|
|
81 else if (/[?\/]/.test(ch) && source.equals(">")) {
|
|
82 source.next();
|
|
83 setState(inText);
|
|
84 return "xml-punctuation";
|
|
85 }
|
|
86 else if (ch == "=") {
|
|
87 return "xml-punctuation";
|
|
88 }
|
|
89 else if (/[\'\"]/.test(ch)) {
|
|
90 setState(inAttribute(ch));
|
|
91 return null;
|
|
92 }
|
|
93 else {
|
|
94 source.nextWhileMatches(/[^\s\u00a0=<>\"\'\/?]/);
|
|
95 return "xml-name";
|
|
96 }
|
|
97 }
|
|
98
|
|
99 function inAttribute(quote) {
|
|
100 return function(source, setState) {
|
|
101 while (!source.endOfLine()) {
|
|
102 if (source.next() == quote) {
|
|
103 setState(inTag);
|
|
104 break;
|
|
105 }
|
|
106 }
|
|
107 return "xml-attribute";
|
|
108 };
|
|
109 }
|
|
110
|
|
111 function inBlock(style, terminator) {
|
|
112 return function(source, setState) {
|
|
113 while (!source.endOfLine()) {
|
|
114 if (source.lookAhead(terminator, true)) {
|
|
115 setState(inText);
|
|
116 break;
|
|
117 }
|
|
118 source.next();
|
|
119 }
|
|
120 return style;
|
|
121 };
|
|
122 }
|
|
123
|
|
124 return function(source, startState) {
|
|
125 return tokenizer(source, startState || inText);
|
|
126 };
|
|
127 })();
|
|
128
|
|
129 // The parser. The structure of this function largely follows that of
|
|
130 // parseJavaScript in parsejavascript.js (there is actually a bit more
|
|
131 // shared code than I'd like), but it is quite a bit simpler.
|
|
132 function parseXML(source) {
|
|
133 var tokens = tokenizeXML(source), token;
|
|
134 var cc = [base];
|
|
135 var tokenNr = 0, indented = 0;
|
|
136 var currentTag = null, context = null;
|
|
137 var consume;
|
|
138
|
|
139 function push(fs) {
|
|
140 for (var i = fs.length - 1; i >= 0; i--)
|
|
141 cc.push(fs[i]);
|
|
142 }
|
|
143 function cont() {
|
|
144 push(arguments);
|
|
145 consume = true;
|
|
146 }
|
|
147 function pass() {
|
|
148 push(arguments);
|
|
149 consume = false;
|
|
150 }
|
|
151
|
|
152 function markErr() {
|
|
153 token.style += " xml-error";
|
|
154 }
|
|
155 function expect(text) {
|
|
156 return function(style, content) {
|
|
157 if (content == text) cont();
|
|
158 else {markErr(); cont(arguments.callee);}
|
|
159 };
|
|
160 }
|
|
161
|
|
162 function pushContext(tagname, startOfLine) {
|
|
163 var noIndent = UseKludges.doNotIndent.hasOwnProperty(tagname) || (context && context.noIndent);
|
|
164 context = {prev: context, name: tagname, indent: indented, startOfLine: startOfLine, noIndent: noIndent};
|
|
165 }
|
|
166 function popContext() {
|
|
167 context = context.prev;
|
|
168 }
|
|
169 function computeIndentation(baseContext) {
|
|
170 return function(nextChars, current) {
|
|
171 var context = baseContext;
|
|
172 if (context && context.noIndent)
|
|
173 return current;
|
|
174 if (alignCDATA && /<!\[CDATA\[/.test(nextChars))
|
|
175 return 0;
|
|
176 if (context && /^<\//.test(nextChars))
|
|
177 context = context.prev;
|
|
178 while (context && !context.startOfLine)
|
|
179 context = context.prev;
|
|
180 if (context)
|
|
181 return context.indent + indentUnit;
|
|
182 else
|
|
183 return 0;
|
|
184 };
|
|
185 }
|
|
186
|
|
187 function base() {
|
|
188 return pass(element, base);
|
|
189 }
|
|
190 var harmlessTokens = {"xml-text": true, "xml-entity": true, "xml-comment": true, "xml-processing": true, "xml-doctype": true};
|
|
191 function element(style, content) {
|
|
192 if (content == "<") cont(tagname, attributes, endtag(tokenNr == 1));
|
|
193 else if (content == "</") cont(closetagname, expect(">"));
|
|
194 else if (style == "xml-cdata") {
|
|
195 if (!context || context.name != "!cdata") pushContext("!cdata");
|
|
196 if (/\]\]>$/.test(content)) popContext();
|
|
197 cont();
|
|
198 }
|
|
199 else if (harmlessTokens.hasOwnProperty(style)) cont();
|
|
200 else {markErr(); cont();}
|
|
201 }
|
|
202 function tagname(style, content) {
|
|
203 if (style == "xml-name") {
|
|
204 currentTag = content.toLowerCase();
|
|
205 token.style = "xml-tagname";
|
|
206 cont();
|
|
207 }
|
|
208 else {
|
|
209 currentTag = null;
|
|
210 pass();
|
|
211 }
|
|
212 }
|
|
213 function closetagname(style, content) {
|
|
214 if (style == "xml-name") {
|
|
215 token.style = "xml-tagname";
|
|
216 if (context && content.toLowerCase() == context.name) popContext();
|
|
217 else markErr();
|
|
218 }
|
|
219 cont();
|
|
220 }
|
|
221 function endtag(startOfLine) {
|
|
222 return function(style, content) {
|
|
223 if (content == "/>" || (content == ">" && UseKludges.autoSelfClosers.hasOwnProperty(currentTag))) cont();
|
|
224 else if (content == ">") {pushContext(currentTag, startOfLine); cont();}
|
|
225 else {markErr(); cont(arguments.callee);}
|
|
226 };
|
|
227 }
|
|
228 function attributes(style) {
|
|
229 if (style == "xml-name") {token.style = "xml-attname"; cont(attribute, attributes);}
|
|
230 else pass();
|
|
231 }
|
|
232 function attribute(style, content) {
|
|
233 if (content == "=") cont(value);
|
|
234 else if (content == ">" || content == "/>") pass(endtag);
|
|
235 else pass();
|
|
236 }
|
|
237 function value(style) {
|
|
238 if (style == "xml-attribute") cont(value);
|
|
239 else pass();
|
|
240 }
|
|
241
|
|
242 return {
|
|
243 indentation: function() {return indented;},
|
|
244
|
|
245 next: function(){
|
|
246 token = tokens.next();
|
|
247 if (token.style == "whitespace" && tokenNr == 0)
|
|
248 indented = token.value.length;
|
|
249 else
|
|
250 tokenNr++;
|
|
251 if (token.content == "\n") {
|
|
252 indented = tokenNr = 0;
|
|
253 token.indentation = computeIndentation(context);
|
|
254 }
|
|
255
|
|
256 if (token.style == "whitespace" || token.type == "xml-comment")
|
|
257 return token;
|
|
258
|
|
259 while(true){
|
|
260 consume = false;
|
|
261 cc.pop()(token.style, token.content);
|
|
262 if (consume) return token;
|
|
263 }
|
|
264 },
|
|
265
|
|
266 copy: function(){
|
|
267 var _cc = cc.concat([]), _tokenState = tokens.state, _context = context;
|
|
268 var parser = this;
|
|
269
|
|
270 return function(input){
|
|
271 cc = _cc.concat([]);
|
|
272 tokenNr = indented = 0;
|
|
273 context = _context;
|
|
274 tokens = tokenizeXML(input, _tokenState);
|
|
275 return parser;
|
|
276 };
|
|
277 }
|
|
278 };
|
|
279 }
|
|
280
|
|
281 return {
|
|
282 make: parseXML,
|
|
283 electricChars: "/",
|
|
284 configure: function(config) {
|
|
285 if (config.useHTMLKludges != null)
|
|
286 UseKludges = config.useHTMLKludges ? Kludges : NoKludges;
|
|
287 if (config.alignCDATA)
|
|
288 alignCDATA = config.alignCDATA;
|
|
289 }
|
|
290 };
|
|
291 })();
|