Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(169)

Issues Search
    My Issues | Starred     Open | Closed | All

Side by Side Diff: bower_components/google-code-prettify/src/prettify.js

Issue 786953007: npm_modules: Fork bower_components into Polymer 0.4.0 and 0.5.0 versions (Closed) Base URL: https://chromium.googlesource.com/infra/third_party/npm_modules.git@master
Patch Set: Created 5 years, 11 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View unified diff | Download patch
« no previous file with comments | « bower_components/google-code-prettify/src/prettify.css ('k') | bower_components/google-code-prettify/src/run_prettify.js » ('j') | no next file with comments »
Toggle Intra-line Diffs ('i') | Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
OLDNEW
(Empty)
1 // Copyright (C) 2006 Google Inc.
2 //
3 // Licensed under the Apache License, Version 2.0 (the "License");
4 // you may not use this file except in compliance with the License.
5 // You may obtain a copy of the License at
6 //
7 // http://www.apache.org/licenses/LICENSE-2.0
8 //
9 // Unless required by applicable law or agreed to in writing, software
10 // distributed under the License is distributed on an "AS IS" BASIS,
11 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12 // See the License for the specific language governing permissions and
13 // limitations under the License.
14
15
16 /**
17 * @fileoverview
18 * some functions for browser-side pretty printing of code contained in html.
19 *
20 * <p>
21 * For a fairly comprehensive set of languages see the
22 * <a href="http://google-code-prettify.googlecode.com/svn/trunk/README.html#lan gs">README</a>
23 * file that came with this source. At a minimum, the lexer should work on a
24 * number of languages including C and friends, Java, Python, Bash, SQL, HTML,
25 * XML, CSS, Javascript, and Makefiles. It works passably on Ruby, PHP and Awk
26 * and a subset of Perl, but, because of commenting conventions, doesn't work on
27 * Smalltalk, Lisp-like, or CAML-like languages without an explicit lang class.
28 * <p>
29 * Usage: <ol>
30 * <li> include this source file in an html page via
31 * {@code <script type="text/javascript" src="/path/to/prettify.js"></script>}
32 * <li> define style rules. See the example page for examples.
33 * <li> mark the {@code <pre>} and {@code <code>} tags in your source with
34 * {@code class=prettyprint.}
35 * You can also use the (html deprecated) {@code <xmp>} tag, but the pretty
36 * printer needs to do more substantial DOM manipulations to support that, so
37 * some css styles may not be preserved.
38 * </ol>
39 * That's it. I wanted to keep the API as simple as possible, so there's no
40 * need to specify which language the code is in, but if you wish, you can add
41 * another class to the {@code <pre>} or {@code <code>} element to specify the
42 * language, as in {@code <pre class="prettyprint lang-java">}. Any class that
43 * starts with "lang-" followed by a file extension, specifies the file type.
44 * See the "lang-*.js" files in this directory for code that implements
45 * per-language file handlers.
46 * <p>
47 * Change log:<br>
48 * cbeust, 2006/08/22
49 * <blockquote>
50 * Java annotations (start with "@") are now captured as literals ("lit")
51 * </blockquote>
52 * @requires console
53 */
54
55 // JSLint declarations
56 /*global console, document, navigator, setTimeout, window, define */
57
58 /** @define {boolean} */
59 var IN_GLOBAL_SCOPE = true;
60
61 /**
62 * Split {@code prettyPrint} into multiple timeouts so as not to interfere with
63 * UI events.
64 * If set to {@code false}, {@code prettyPrint()} is synchronous.
65 */
66 window['PR_SHOULD_USE_CONTINUATION'] = true;
67
68 /**
69 * Pretty print a chunk of code.
70 * @param {string} sourceCodeHtml The HTML to pretty print.
71 * @param {string} opt_langExtension The language name to use.
72 * Typically, a filename extension like 'cpp' or 'java'.
73 * @param {number|boolean} opt_numberLines True to number lines,
74 * or the 1-indexed number of the first line in sourceCodeHtml.
75 * @return {string} code as html, but prettier
76 */
77 var prettyPrintOne;
78 /**
79 * Find all the {@code <pre>} and {@code <code>} tags in the DOM with
80 * {@code class=prettyprint} and prettify them.
81 *
82 * @param {Function} opt_whenDone called when prettifying is done.
83 * @param {HTMLElement|HTMLDocument} opt_root an element or document
84 * containing all the elements to pretty print.
85 * Defaults to {@code document.body}.
86 */
87 var prettyPrint;
88
89
90 (function () {
91 var win = window;
92 // Keyword lists for various languages.
93 // We use things that coerce to strings to make them compact when minified
94 // and to defeat aggressive optimizers that fold large string constants.
95 var FLOW_CONTROL_KEYWORDS = ["break,continue,do,else,for,if,return,while"];
96 var C_KEYWORDS = [FLOW_CONTROL_KEYWORDS,"auto,case,char,const,default," +
97 "double,enum,extern,float,goto,inline,int,long,register,short,signed," +
98 "sizeof,static,struct,switch,typedef,union,unsigned,void,volatile"];
99 var COMMON_KEYWORDS = [C_KEYWORDS,"catch,class,delete,false,import," +
100 "new,operator,private,protected,public,this,throw,true,try,typeof"];
101 var CPP_KEYWORDS = [COMMON_KEYWORDS,"alignof,align_union,asm,axiom,bool," +
102 "concept,concept_map,const_cast,constexpr,decltype,delegate," +
103 "dynamic_cast,explicit,export,friend,generic,late_check," +
104 "mutable,namespace,nullptr,property,reinterpret_cast,static_assert," +
105 "static_cast,template,typeid,typename,using,virtual,where"];
106 var JAVA_KEYWORDS = [COMMON_KEYWORDS,
107 "abstract,assert,boolean,byte,extends,final,finally,implements,import," +
108 "instanceof,interface,null,native,package,strictfp,super,synchronized," +
109 "throws,transient"];
110 var CSHARP_KEYWORDS = [JAVA_KEYWORDS,
111 "as,base,by,checked,decimal,delegate,descending,dynamic,event," +
112 "fixed,foreach,from,group,implicit,in,internal,into,is,let," +
113 "lock,object,out,override,orderby,params,partial,readonly,ref,sbyte," +
114 "sealed,stackalloc,string,select,uint,ulong,unchecked,unsafe,ushort," +
115 "var,virtual,where"];
116 var COFFEE_KEYWORDS = "all,and,by,catch,class,else,extends,false,finally," +
117 "for,if,in,is,isnt,loop,new,no,not,null,of,off,on,or,return,super,then," +
118 "throw,true,try,unless,until,when,while,yes";
119 var JSCRIPT_KEYWORDS = [COMMON_KEYWORDS,
120 "debugger,eval,export,function,get,null,set,undefined,var,with," +
121 "Infinity,NaN"];
122 var PERL_KEYWORDS = "caller,delete,die,do,dump,elsif,eval,exit,foreach,for," +
123 "goto,if,import,last,local,my,next,no,our,print,package,redo,require," +
124 "sub,undef,unless,until,use,wantarray,while,BEGIN,END";
125 var PYTHON_KEYWORDS = [FLOW_CONTROL_KEYWORDS, "and,as,assert,class,def,del," +
126 "elif,except,exec,finally,from,global,import,in,is,lambda," +
127 "nonlocal,not,or,pass,print,raise,try,with,yield," +
128 "False,True,None"];
129 var RUBY_KEYWORDS = [FLOW_CONTROL_KEYWORDS, "alias,and,begin,case,class," +
130 "def,defined,elsif,end,ensure,false,in,module,next,nil,not,or,redo," +
131 "rescue,retry,self,super,then,true,undef,unless,until,when,yield," +
132 "BEGIN,END"];
133 var RUST_KEYWORDS = [FLOW_CONTROL_KEYWORDS, "as,assert,const,copy,drop," +
134 "enum,extern,fail,false,fn,impl,let,log,loop,match,mod,move,mut,priv," +
135 "pub,pure,ref,self,static,struct,true,trait,type,unsafe,use"];
136 var SH_KEYWORDS = [FLOW_CONTROL_KEYWORDS, "case,done,elif,esac,eval,fi," +
137 "function,in,local,set,then,until"];
138 var ALL_KEYWORDS = [
139 CPP_KEYWORDS, CSHARP_KEYWORDS, JSCRIPT_KEYWORDS, PERL_KEYWORDS,
140 PYTHON_KEYWORDS, RUBY_KEYWORDS, SH_KEYWORDS];
141 var C_TYPES = /^(DIR|FILE|vector|(de|priority_)?queue|list|stack|(const_)?iter ator|(multi)?(set|map)|bitset|u?(int|float)\d*)\b/;
142
143 // token style names. correspond to css classes
144 /**
145 * token style for a string literal
146 * @const
147 */
148 var PR_STRING = 'str';
149 /**
150 * token style for a keyword
151 * @const
152 */
153 var PR_KEYWORD = 'kwd';
154 /**
155 * token style for a comment
156 * @const
157 */
158 var PR_COMMENT = 'com';
159 /**
160 * token style for a type
161 * @const
162 */
163 var PR_TYPE = 'typ';
164 /**
165 * token style for a literal value. e.g. 1, null, true.
166 * @const
167 */
168 var PR_LITERAL = 'lit';
169 /**
170 * token style for a punctuation string.
171 * @const
172 */
173 var PR_PUNCTUATION = 'pun';
174 /**
175 * token style for plain text.
176 * @const
177 */
178 var PR_PLAIN = 'pln';
179
180 /**
181 * token style for an sgml tag.
182 * @const
183 */
184 var PR_TAG = 'tag';
185 /**
186 * token style for a markup declaration such as a DOCTYPE.
187 * @const
188 */
189 var PR_DECLARATION = 'dec';
190 /**
191 * token style for embedded source.
192 * @const
193 */
194 var PR_SOURCE = 'src';
195 /**
196 * token style for an sgml attribute name.
197 * @const
198 */
199 var PR_ATTRIB_NAME = 'atn';
200 /**
201 * token style for an sgml attribute value.
202 * @const
203 */
204 var PR_ATTRIB_VALUE = 'atv';
205
206 /**
207 * A class that indicates a section of markup that is not code, e.g. to allow
208 * embedding of line numbers within code listings.
209 * @const
210 */
211 var PR_NOCODE = 'nocode';
212
213
214
215 /**
216 * A set of tokens that can precede a regular expression literal in
217 * javascript
218 * http://web.archive.org/web/20070717142515/http://www.mozilla.org/js/languag e/js20/rationale/syntax.html
219 * has the full list, but I've removed ones that might be problematic when
220 * seen in languages that don't support regular expression literals.
221 *
222 * <p>Specifically, I've removed any keywords that can't precede a regexp
223 * literal in a syntactically legal javascript program, and I've removed the
224 * "in" keyword since it's not a keyword in many languages, and might be used
225 * as a count of inches.
226 *
227 * <p>The link above does not accurately describe EcmaScript rules since
228 * it fails to distinguish between (a=++/b/i) and (a++/b/i) but it works
229 * very well in practice.
230 *
231 * @private
232 * @const
233 */
234 var REGEXP_PRECEDER_PATTERN = '(?:^^\\.?|[+-]|[!=]=?=?|\\#|%=?|&&?=?|\\(|\\*=? |[+\\-]=|->|\\/=?|::?|<<?=?|>>?>?=?|,|;|\\?|@|\\[|~|{|\\^\\^?=?|\\|\\|?=?|break| case|continue|delete|do|else|finally|instanceof|return|throw|try|typeof)\\s*';
235
236 // CAVEAT: this does not properly handle the case where a regular
237 // expression immediately follows another since a regular expression may
238 // have flags for case-sensitivity and the like. Having regexp tokens
239 // adjacent is not valid in any language I'm aware of, so I'm punting.
240 // TODO: maybe style special characters inside a regexp as punctuation.
241
242 /**
243 * Given a group of {@link RegExp}s, returns a {@code RegExp} that globally
244 * matches the union of the sets of strings matched by the input RegExp.
245 * Since it matches globally, if the input strings have a start-of-input
246 * anchor (/^.../), it is ignored for the purposes of unioning.
247 * @param {Array.<RegExp>} regexs non multiline, non-global regexs.
248 * @return {RegExp} a global regex.
249 */
250 function combinePrefixPatterns(regexs) {
251 var capturedGroupIndex = 0;
252
253 var needToFoldCase = false;
254 var ignoreCase = false;
255 for (var i = 0, n = regexs.length; i < n; ++i) {
256 var regex = regexs[i];
257 if (regex.ignoreCase) {
258 ignoreCase = true;
259 } else if (/[a-z]/i.test(regex.source.replace(
260 /\\u[0-9a-f]{4}|\\x[0-9a-f]{2}|\\[^ux]/gi, ''))) {
261 needToFoldCase = true;
262 ignoreCase = false;
263 break;
264 }
265 }
266
267 var escapeCharToCodeUnit = {
268 'b': 8,
269 't': 9,
270 'n': 0xa,
271 'v': 0xb,
272 'f': 0xc,
273 'r': 0xd
274 };
275
276 function decodeEscape(charsetPart) {
277 var cc0 = charsetPart.charCodeAt(0);
278 if (cc0 !== 92 /* \\ */) {
279 return cc0;
280 }
281 var c1 = charsetPart.charAt(1);
282 cc0 = escapeCharToCodeUnit[c1];
283 if (cc0) {
284 return cc0;
285 } else if ('0' <= c1 && c1 <= '7') {
286 return parseInt(charsetPart.substring(1), 8);
287 } else if (c1 === 'u' || c1 === 'x') {
288 return parseInt(charsetPart.substring(2), 16);
289 } else {
290 return charsetPart.charCodeAt(1);
291 }
292 }
293
294 function encodeEscape(charCode) {
295 if (charCode < 0x20) {
296 return (charCode < 0x10 ? '\\x0' : '\\x') + charCode.toString(16);
297 }
298 var ch = String.fromCharCode(charCode);
299 return (ch === '\\' || ch === '-' || ch === ']' || ch === '^')
300 ? "\\" + ch : ch;
301 }
302
303 function caseFoldCharset(charSet) {
304 var charsetParts = charSet.substring(1, charSet.length - 1).match(
305 new RegExp(
306 '\\\\u[0-9A-Fa-f]{4}'
307 + '|\\\\x[0-9A-Fa-f]{2}'
308 + '|\\\\[0-3][0-7]{0,2}'
309 + '|\\\\[0-7]{1,2}'
310 + '|\\\\[\\s\\S]'
311 + '|-'
312 + '|[^-\\\\]',
313 'g'));
314 var ranges = [];
315 var inverse = charsetParts[0] === '^';
316
317 var out = ['['];
318 if (inverse) { out.push('^'); }
319
320 for (var i = inverse ? 1 : 0, n = charsetParts.length; i < n; ++i) {
321 var p = charsetParts[i];
322 if (/\\[bdsw]/i.test(p)) { // Don't muck with named groups.
323 out.push(p);
324 } else {
325 var start = decodeEscape(p);
326 var end;
327 if (i + 2 < n && '-' === charsetParts[i + 1]) {
328 end = decodeEscape(charsetParts[i + 2]);
329 i += 2;
330 } else {
331 end = start;
332 }
333 ranges.push([start, end]);
334 // If the range might intersect letters, then expand it.
335 // This case handling is too simplistic.
336 // It does not deal with non-latin case folding.
337 // It works for latin source code identifiers though.
338 if (!(end < 65 || start > 122)) {
339 if (!(end < 65 || start > 90)) {
340 ranges.push([Math.max(65, start) | 32, Math.min(end, 90) | 32]);
341 }
342 if (!(end < 97 || start > 122)) {
343 ranges.push([Math.max(97, start) & ~32, Math.min(end, 122) & ~32]) ;
344 }
345 }
346 }
347 }
348
349 // [[1, 10], [3, 4], [8, 12], [14, 14], [16, 16], [17, 17]]
350 // -> [[1, 12], [14, 14], [16, 17]]
351 ranges.sort(function (a, b) { return (a[0] - b[0]) || (b[1] - a[1]); });
352 var consolidatedRanges = [];
353 var lastRange = [];
354 for (var i = 0; i < ranges.length; ++i) {
355 var range = ranges[i];
356 if (range[0] <= lastRange[1] + 1) {
357 lastRange[1] = Math.max(lastRange[1], range[1]);
358 } else {
359 consolidatedRanges.push(lastRange = range);
360 }
361 }
362
363 for (var i = 0; i < consolidatedRanges.length; ++i) {
364 var range = consolidatedRanges[i];
365 out.push(encodeEscape(range[0]));
366 if (range[1] > range[0]) {
367 if (range[1] + 1 > range[0]) { out.push('-'); }
368 out.push(encodeEscape(range[1]));
369 }
370 }
371 out.push(']');
372 return out.join('');
373 }
374
375 function allowAnywhereFoldCaseAndRenumberGroups(regex) {
376 // Split into character sets, escape sequences, punctuation strings
377 // like ('(', '(?:', ')', '^'), and runs of characters that do not
378 // include any of the above.
379 var parts = regex.source.match(
380 new RegExp(
381 '(?:'
382 + '\\[(?:[^\\x5C\\x5D]|\\\\[\\s\\S])*\\]' // a character set
383 + '|\\\\u[A-Fa-f0-9]{4}' // a unicode escape
384 + '|\\\\x[A-Fa-f0-9]{2}' // a hex escape
385 + '|\\\\[0-9]+' // a back-reference or octal escape
386 + '|\\\\[^ux0-9]' // other escape sequence
387 + '|\\(\\?[:!=]' // start of a non-capturing group
388 + '|[\\(\\)\\^]' // start/end of a group, or line start
389 + '|[^\\x5B\\x5C\\(\\)\\^]+' // run of other characters
390 + ')',
391 'g'));
392 var n = parts.length;
393
394 // Maps captured group numbers to the number they will occupy in
395 // the output or to -1 if that has not been determined, or to
396 // undefined if they need not be capturing in the output.
397 var capturedGroups = [];
398
399 // Walk over and identify back references to build the capturedGroups
400 // mapping.
401 for (var i = 0, groupIndex = 0; i < n; ++i) {
402 var p = parts[i];
403 if (p === '(') {
404 // groups are 1-indexed, so max group index is count of '('
405 ++groupIndex;
406 } else if ('\\' === p.charAt(0)) {
407 var decimalValue = +p.substring(1);
408 if (decimalValue) {
409 if (decimalValue <= groupIndex) {
410 capturedGroups[decimalValue] = -1;
411 } else {
412 // Replace with an unambiguous escape sequence so that
413 // an octal escape sequence does not turn into a backreference
414 // to a capturing group from an earlier regex.
415 parts[i] = encodeEscape(decimalValue);
416 }
417 }
418 }
419 }
420
421 // Renumber groups and reduce capturing groups to non-capturing groups
422 // where possible.
423 for (var i = 1; i < capturedGroups.length; ++i) {
424 if (-1 === capturedGroups[i]) {
425 capturedGroups[i] = ++capturedGroupIndex;
426 }
427 }
428 for (var i = 0, groupIndex = 0; i < n; ++i) {
429 var p = parts[i];
430 if (p === '(') {
431 ++groupIndex;
432 if (!capturedGroups[groupIndex]) {
433 parts[i] = '(?:';
434 }
435 } else if ('\\' === p.charAt(0)) {
436 var decimalValue = +p.substring(1);
437 if (decimalValue && decimalValue <= groupIndex) {
438 parts[i] = '\\' + capturedGroups[decimalValue];
439 }
440 }
441 }
442
443 // Remove any prefix anchors so that the output will match anywhere.
444 // ^^ really does mean an anchored match though.
445 for (var i = 0; i < n; ++i) {
446 if ('^' === parts[i] && '^' !== parts[i + 1]) { parts[i] = ''; }
447 }
448
449 // Expand letters to groups to handle mixing of case-sensitive and
450 // case-insensitive patterns if necessary.
451 if (regex.ignoreCase && needToFoldCase) {
452 for (var i = 0; i < n; ++i) {
453 var p = parts[i];
454 var ch0 = p.charAt(0);
455 if (p.length >= 2 && ch0 === '[') {
456 parts[i] = caseFoldCharset(p);
457 } else if (ch0 !== '\\') {
458 // TODO: handle letters in numeric escapes.
459 parts[i] = p.replace(
460 /[a-zA-Z]/g,
461 function (ch) {
462 var cc = ch.charCodeAt(0);
463 return '[' + String.fromCharCode(cc & ~32, cc | 32) + ']';
464 });
465 }
466 }
467 }
468
469 return parts.join('');
470 }
471
472 var rewritten = [];
473 for (var i = 0, n = regexs.length; i < n; ++i) {
474 var regex = regexs[i];
475 if (regex.global || regex.multiline) { throw new Error('' + regex); }
476 rewritten.push(
477 '(?:' + allowAnywhereFoldCaseAndRenumberGroups(regex) + ')');
478 }
479
480 return new RegExp(rewritten.join('|'), ignoreCase ? 'gi' : 'g');
481 }
482
483 /**
484 * Split markup into a string of source code and an array mapping ranges in
485 * that string to the text nodes in which they appear.
486 *
487 * <p>
488 * The HTML DOM structure:</p>
489 * <pre>
490 * (Element "p"
491 * (Element "b"
492 * (Text "print ")) ; #1
493 * (Text "'Hello '") ; #2
494 * (Element "br") ; #3
495 * (Text " + 'World';")) ; #4
496 * </pre>
497 * <p>
498 * corresponds to the HTML
499 * {@code <p><b>print </b>'Hello '<br> + 'World';</p>}.</p>
500 *
501 * <p>
502 * It will produce the output:</p>
503 * <pre>
504 * {
505 * sourceCode: "print 'Hello '\n + 'World';",
506 * // 1 2
507 * // 012345678901234 5678901234567
508 * spans: [0, #1, 6, #2, 14, #3, 15, #4]
509 * }
510 * </pre>
511 * <p>
512 * where #1 is a reference to the {@code "print "} text node above, and so
513 * on for the other text nodes.
514 * </p>
515 *
516 * <p>
517 * The {@code} spans array is an array of pairs. Even elements are the start
518 * indices of substrings, and odd elements are the text nodes (or BR elements)
519 * that contain the text for those substrings.
520 * Substrings continue until the next index or the end of the source.
521 * </p>
522 *
523 * @param {Node} node an HTML DOM subtree containing source-code.
524 * @param {boolean} isPreformatted true if white-space in text nodes should
525 * be considered significant.
526 * @return {Object} source code and the text nodes in which they occur.
527 */
528 function extractSourceSpans(node, isPreformatted) {
529 var nocode = /(?:^|\s)nocode(?:\s|$)/;
530
531 var chunks = [];
532 var length = 0;
533 var spans = [];
534 var k = 0;
535
536 function walk(node) {
537 var type = node.nodeType;
538 if (type == 1) { // Element
539 if (nocode.test(node.className)) { return; }
540 for (var child = node.firstChild; child; child = child.nextSibling) {
541 walk(child);
542 }
543 var nodeName = node.nodeName.toLowerCase();
544 if ('br' === nodeName || 'li' === nodeName) {
545 chunks[k] = '\n';
546 spans[k << 1] = length++;
547 spans[(k++ << 1) | 1] = node;
548 }
549 } else if (type == 3 || type == 4) { // Text
550 var text = node.nodeValue;
551 if (text.length) {
552 if (!isPreformatted) {
553 text = text.replace(/[ \t\r\n]+/g, ' ');
554 } else {
555 text = text.replace(/\r\n?/g, '\n'); // Normalize newlines.
556 }
557 // TODO: handle tabs here?
558 chunks[k] = text;
559 spans[k << 1] = length;
560 length += text.length;
561 spans[(k++ << 1) | 1] = node;
562 }
563 }
564 }
565
566 walk(node);
567
568 return {
569 sourceCode: chunks.join('').replace(/\n$/, ''),
570 spans: spans
571 };
572 }
573
574 /**
575 * Apply the given language handler to sourceCode and add the resulting
576 * decorations to out.
577 * @param {number} basePos the index of sourceCode within the chunk of source
578 * whose decorations are already present on out.
579 */
580 function appendDecorations(basePos, sourceCode, langHandler, out) {
581 if (!sourceCode) { return; }
582 var job = {
583 sourceCode: sourceCode,
584 basePos: basePos
585 };
586 langHandler(job);
587 out.push.apply(out, job.decorations);
588 }
589
590 var notWs = /\S/;
591
592 /**
593 * Given an element, if it contains only one child element and any text nodes
594 * it contains contain only space characters, return the sole child element.
595 * Otherwise returns undefined.
596 * <p>
597 * This is meant to return the CODE element in {@code <pre><code ...>} when
598 * there is a single child element that contains all the non-space textual
599 * content, but not to return anything where there are multiple child elements
600 * as in {@code <pre><code>...</code><code>...</code></pre>} or when there
601 * is textual content.
602 */
603 function childContentWrapper(element) {
604 var wrapper = undefined;
605 for (var c = element.firstChild; c; c = c.nextSibling) {
606 var type = c.nodeType;
607 wrapper = (type === 1) // Element Node
608 ? (wrapper ? element : c)
609 : (type === 3) // Text Node
610 ? (notWs.test(c.nodeValue) ? element : wrapper)
611 : wrapper;
612 }
613 return wrapper === element ? undefined : wrapper;
614 }
615
616 /** Given triples of [style, pattern, context] returns a lexing function,
617 * The lexing function interprets the patterns to find token boundaries and
618 * returns a decoration list of the form
619 * [index_0, style_0, index_1, style_1, ..., index_n, style_n]
620 * where index_n is an index into the sourceCode, and style_n is a style
621 * constant like PR_PLAIN. index_n-1 <= index_n, and style_n-1 applies to
622 * all characters in sourceCode[index_n-1:index_n].
623 *
624 * The stylePatterns is a list whose elements have the form
625 * [style : string, pattern : RegExp, DEPRECATED, shortcut : string].
626 *
627 * Style is a style constant like PR_PLAIN, or can be a string of the
628 * form 'lang-FOO', where FOO is a language extension describing the
629 * language of the portion of the token in $1 after pattern executes.
630 * E.g., if style is 'lang-lisp', and group 1 contains the text
631 * '(hello (world))', then that portion of the token will be passed to the
632 * registered lisp handler for formatting.
633 * The text before and after group 1 will be restyled using this decorator
634 * so decorators should take care that this doesn't result in infinite
635 * recursion. For example, the HTML lexer rule for SCRIPT elements looks
636 * something like ['lang-js', /<[s]cript>(.+?)<\/script>/]. This may match
637 * '<script>foo()<\/script>', which would cause the current decorator to
638 * be called with '<script>' which would not match the same rule since
639 * group 1 must not be empty, so it would be instead styled as PR_TAG by
640 * the generic tag rule. The handler registered for the 'js' extension would
641 * then be called with 'foo()', and finally, the current decorator would
642 * be called with '<\/script>' which would not match the original rule and
643 * so the generic tag rule would identify it as a tag.
644 *
645 * Pattern must only match prefixes, and if it matches a prefix, then that
646 * match is considered a token with the same style.
647 *
648 * Context is applied to the last non-whitespace, non-comment token
649 * recognized.
650 *
651 * Shortcut is an optional string of characters, any of which, if the first
652 * character, gurantee that this pattern and only this pattern matches.
653 *
654 * @param {Array} shortcutStylePatterns patterns that always start with
655 * a known character. Must have a shortcut string.
656 * @param {Array} fallthroughStylePatterns patterns that will be tried in
657 * order if the shortcut ones fail. May have shortcuts.
658 *
659 * @return {function (Object)} a
660 * function that takes source code and returns a list of decorations.
661 */
662 function createSimpleLexer(shortcutStylePatterns, fallthroughStylePatterns) {
663 var shortcuts = {};
664 var tokenizer;
665 (function () {
666 var allPatterns = shortcutStylePatterns.concat(fallthroughStylePatterns);
667 var allRegexs = [];
668 var regexKeys = {};
669 for (var i = 0, n = allPatterns.length; i < n; ++i) {
670 var patternParts = allPatterns[i];
671 var shortcutChars = patternParts[3];
672 if (shortcutChars) {
673 for (var c = shortcutChars.length; --c >= 0;) {
674 shortcuts[shortcutChars.charAt(c)] = patternParts;
675 }
676 }
677 var regex = patternParts[1];
678 var k = '' + regex;
679 if (!regexKeys.hasOwnProperty(k)) {
680 allRegexs.push(regex);
681 regexKeys[k] = null;
682 }
683 }
684 allRegexs.push(/[\0-\uffff]/);
685 tokenizer = combinePrefixPatterns(allRegexs);
686 })();
687
688 var nPatterns = fallthroughStylePatterns.length;
689
690 /**
691 * Lexes job.sourceCode and produces an output array job.decorations of
692 * style classes preceded by the position at which they start in
693 * job.sourceCode in order.
694 *
695 * @param {Object} job an object like <pre>{
696 * sourceCode: {string} sourceText plain text,
697 * basePos: {int} position of job.sourceCode in the larger chunk of
698 * sourceCode.
699 * }</pre>
700 */
701 var decorate = function (job) {
702 var sourceCode = job.sourceCode, basePos = job.basePos;
703 /** Even entries are positions in source in ascending order. Odd enties
704 * are style markers (e.g., PR_COMMENT) that run from that position until
705 * the end.
706 * @type {Array.<number|string>}
707 */
708 var decorations = [basePos, PR_PLAIN];
709 var pos = 0; // index into sourceCode
710 var tokens = sourceCode.match(tokenizer) || [];
711 var styleCache = {};
712
713 for (var ti = 0, nTokens = tokens.length; ti < nTokens; ++ti) {
714 var token = tokens[ti];
715 var style = styleCache[token];
716 var match = void 0;
717
718 var isEmbedded;
719 if (typeof style === 'string') {
720 isEmbedded = false;
721 } else {
722 var patternParts = shortcuts[token.charAt(0)];
723 if (patternParts) {
724 match = token.match(patternParts[1]);
725 style = patternParts[0];
726 } else {
727 for (var i = 0; i < nPatterns; ++i) {
728 patternParts = fallthroughStylePatterns[i];
729 match = token.match(patternParts[1]);
730 if (match) {
731 style = patternParts[0];
732 break;
733 }
734 }
735
736 if (!match) { // make sure that we make progress
737 style = PR_PLAIN;
738 }
739 }
740
741 isEmbedded = style.length >= 5 && 'lang-' === style.substring(0, 5);
742 if (isEmbedded && !(match && typeof match[1] === 'string')) {
743 isEmbedded = false;
744 style = PR_SOURCE;
745 }
746
747 if (!isEmbedded) { styleCache[token] = style; }
748 }
749
750 var tokenStart = pos;
751 pos += token.length;
752
753 if (!isEmbedded) {
754 decorations.push(basePos + tokenStart, style);
755 } else { // Treat group 1 as an embedded block of source code.
756 var embeddedSource = match[1];
757 var embeddedSourceStart = token.indexOf(embeddedSource);
758 var embeddedSourceEnd = embeddedSourceStart + embeddedSource.length;
759 if (match[2]) {
760 // If embeddedSource can be blank, then it would match at the
761 // beginning which would cause us to infinitely recurse on the
762 // entire token, so we catch the right context in match[2].
763 embeddedSourceEnd = token.length - match[2].length;
764 embeddedSourceStart = embeddedSourceEnd - embeddedSource.length;
765 }
766 var lang = style.substring(5);
767 // Decorate the left of the embedded source
768 appendDecorations(
769 basePos + tokenStart,
770 token.substring(0, embeddedSourceStart),
771 decorate, decorations);
772 // Decorate the embedded source
773 appendDecorations(
774 basePos + tokenStart + embeddedSourceStart,
775 embeddedSource,
776 langHandlerForExtension(lang, embeddedSource),
777 decorations);
778 // Decorate the right of the embedded section
779 appendDecorations(
780 basePos + tokenStart + embeddedSourceEnd,
781 token.substring(embeddedSourceEnd),
782 decorate, decorations);
783 }
784 }
785 job.decorations = decorations;
786 };
787 return decorate;
788 }
789
790 /** returns a function that produces a list of decorations from source text.
791 *
792 * This code treats ", ', and ` as string delimiters, and \ as a string
793 * escape. It does not recognize perl's qq() style strings.
794 * It has no special handling for double delimiter escapes as in basic, or
795 * the tripled delimiters used in python, but should work on those regardless
796 * although in those cases a single string literal may be broken up into
797 * multiple adjacent string literals.
798 *
799 * It recognizes C, C++, and shell style comments.
800 *
801 * @param {Object} options a set of optional parameters.
802 * @return {function (Object)} a function that examines the source code
803 * in the input job and builds the decoration list.
804 */
805 function sourceDecorator(options) {
806 var shortcutStylePatterns = [], fallthroughStylePatterns = [];
807 if (options['tripleQuotedStrings']) {
808 // '''multi-line-string''', 'single-line-string', and double-quoted
809 shortcutStylePatterns.push(
810 [PR_STRING, /^(?:\'\'\'(?:[^\'\\]|\\[\s\S]|\'{1,2}(?=[^\']))*(?:\'\'\ '|$)|\"\"\"(?:[^\"\\]|\\[\s\S]|\"{1,2}(?=[^\"]))*(?:\"\"\"|$)|\'(?:[^\\\']|\\[\s \S])*(?:\'|$)|\"(?:[^\\\"]|\\[\s\S])*(?:\"|$))/,
811 null, '\'"']);
812 } else if (options['multiLineStrings']) {
813 // 'multi-line-string', "multi-line-string"
814 shortcutStylePatterns.push(
815 [PR_STRING, /^(?:\'(?:[^\\\']|\\[\s\S])*(?:\'|$)|\"(?:[^\\\"]|\\[\s\S ])*(?:\"|$)|\`(?:[^\\\`]|\\[\s\S])*(?:\`|$))/,
816 null, '\'"`']);
817 } else {
818 // 'single-line-string', "single-line-string"
819 shortcutStylePatterns.push(
820 [PR_STRING,
821 /^(?:\'(?:[^\\\'\r\n]|\\.)*(?:\'|$)|\"(?:[^\\\"\r\n]|\\.)*(?:\"|$))/,
822 null, '"\'']);
823 }
824 if (options['verbatimStrings']) {
825 // verbatim-string-literal production from the C# grammar. See issue 93.
826 fallthroughStylePatterns.push(
827 [PR_STRING, /^@\"(?:[^\"]|\"\")*(?:\"|$)/, null]);
828 }
829 var hc = options['hashComments'];
830 if (hc) {
831 if (options['cStyleComments']) {
832 if (hc > 1) { // multiline hash comments
833 shortcutStylePatterns.push(
834 [PR_COMMENT, /^#(?:##(?:[^#]|#(?!##))*(?:###|$)|.*)/, null, '#']);
835 } else {
836 // Stop C preprocessor declarations at an unclosed open comment
837 shortcutStylePatterns.push(
838 [PR_COMMENT, /^#(?:(?:define|e(?:l|nd)if|else|error|ifn?def|includ e|line|pragma|undef|warning)\b|[^\r\n]*)/,
839 null, '#']);
840 }
841 // #include <stdio.h>
842 fallthroughStylePatterns.push(
843 [PR_STRING,
844 /^<(?:(?:(?:\.\.\/)*|\/?)(?:[\w-]+(?:\/[\w-]+)+)?[\w-]+\.h(?:h|pp|\ +\+)?|[a-z]\w*)>/,
845 null]);
846 } else {
847 shortcutStylePatterns.push([PR_COMMENT, /^#[^\r\n]*/, null, '#']);
848 }
849 }
850 if (options['cStyleComments']) {
851 fallthroughStylePatterns.push([PR_COMMENT, /^\/\/[^\r\n]*/, null]);
852 fallthroughStylePatterns.push(
853 [PR_COMMENT, /^\/\*[\s\S]*?(?:\*\/|$)/, null]);
854 }
855 var regexLiterals = options['regexLiterals'];
856 if (regexLiterals) {
857 /**
858 * @const
859 */
860 var regexExcls = regexLiterals > 1
861 ? '' // Multiline regex literals
862 : '\n\r';
863 /**
864 * @const
865 */
866 var regexAny = regexExcls ? '.' : '[\\S\\s]';
867 /**
868 * @const
869 */
870 var REGEX_LITERAL = (
871 // A regular expression literal starts with a slash that is
872 // not followed by * or / so that it is not confused with
873 // comments.
874 '/(?=[^/*' + regexExcls + '])'
875 // and then contains any number of raw characters,
876 + '(?:[^/\\x5B\\x5C' + regexExcls + ']'
877 // escape sequences (\x5C),
878 + '|\\x5C' + regexAny
879 // or non-nesting character sets (\x5B\x5D);
880 + '|\\x5B(?:[^\\x5C\\x5D' + regexExcls + ']'
881 + '|\\x5C' + regexAny + ')*(?:\\x5D|$))+'
882 // finally closed by a /.
883 + '/');
884 fallthroughStylePatterns.push(
885 ['lang-regex',
886 RegExp('^' + REGEXP_PRECEDER_PATTERN + '(' + REGEX_LITERAL + ')')
887 ]);
888 }
889
890 var types = options['types'];
891 if (types) {
892 fallthroughStylePatterns.push([PR_TYPE, types]);
893 }
894
895 var keywords = ("" + options['keywords']).replace(/^ | $/g, '');
896 if (keywords.length) {
897 fallthroughStylePatterns.push(
898 [PR_KEYWORD,
899 new RegExp('^(?:' + keywords.replace(/[\s,]+/g, '|') + ')\\b'),
900 null]);
901 }
902
903 shortcutStylePatterns.push([PR_PLAIN, /^\s+/, null, ' \r\n\t\xA0']);
904
905 var punctuation =
906 // The Bash man page says
907
908 // A word is a sequence of characters considered as a single
909 // unit by GRUB. Words are separated by metacharacters,
910 // which are the following plus space, tab, and newline: { }
911 // | & $ ; < >
912 // ...
913
914 // A word beginning with # causes that word and all remaining
915 // characters on that line to be ignored.
916
917 // which means that only a '#' after /(?:^|[{}|&$;<>\s])/ starts a
918 // comment but empirically
919 // $ echo {#}
920 // {#}
921 // $ echo \$#
922 // $#
923 // $ echo }#
924 // }#
925
926 // so /(?:^|[|&;<>\s])/ is more appropriate.
927
928 // http://gcc.gnu.org/onlinedocs/gcc-2.95.3/cpp_1.html#SEC3
929 // suggests that this definition is compatible with a
930 // default mode that tries to use a single token definition
931 // to recognize both bash/python style comments and C
932 // preprocessor directives.
933
934 // This definition of punctuation does not include # in the list of
935 // follow-on exclusions, so # will not be broken before if preceeded
936 // by a punctuation character. We could try to exclude # after
937 // [|&;<>] but that doesn't seem to cause many major problems.
938 // If that does turn out to be a problem, we should change the below
939 // when hc is truthy to include # in the run of punctuation characters
940 // only when not followint [|&;<>].
941 '^.[^\\s\\w.$@\'"`/\\\\]*';
942 if (options['regexLiterals']) {
943 punctuation += '(?!\s*\/)';
944 }
945
946 fallthroughStylePatterns.push(
947 // TODO(mikesamuel): recognize non-latin letters and numerals in idents
948 [PR_LITERAL, /^@[a-z_$][a-z_$@0-9]*/i, null],
949 [PR_TYPE, /^(?:[@_]?[A-Z]+[a-z][A-Za-z_$@0-9]*|\w+_t\b)/, null],
950 [PR_PLAIN, /^[a-z_$][a-z_$@0-9]*/i, null],
951 [PR_LITERAL,
952 new RegExp(
953 '^(?:'
954 // A hex number
955 + '0x[a-f0-9]+'
956 // or an octal or decimal number,
957 + '|(?:\\d(?:_\\d+)*\\d*(?:\\.\\d*)?|\\.\\d\\+)'
958 // possibly in scientific notation
959 + '(?:e[+\\-]?\\d+)?'
960 + ')'
961 // with an optional modifier like UL for unsigned long
962 + '[a-z]*', 'i'),
963 null, '0123456789'],
964 // Don't treat escaped quotes in bash as starting strings.
965 // See issue 144.
966 [PR_PLAIN, /^\\[\s\S]?/, null],
967 [PR_PUNCTUATION, new RegExp(punctuation), null]);
968
969 return createSimpleLexer(shortcutStylePatterns, fallthroughStylePatterns);
970 }
971
972 var decorateSource = sourceDecorator({
973 'keywords': ALL_KEYWORDS,
974 'hashComments': true,
975 'cStyleComments': true,
976 'multiLineStrings': true,
977 'regexLiterals': true
978 });
979
980 /**
981 * Given a DOM subtree, wraps it in a list, and puts each line into its own
982 * list item.
983 *
984 * @param {Node} node modified in place. Its content is pulled into an
985 * HTMLOListElement, and each line is moved into a separate list item.
986 * This requires cloning elements, so the input might not have unique
987 * IDs after numbering.
988 * @param {boolean} isPreformatted true iff white-space in text nodes should
989 * be treated as significant.
990 */
991 function numberLines(node, opt_startLineNum, isPreformatted) {
992 var nocode = /(?:^|\s)nocode(?:\s|$)/;
993 var lineBreak = /\r\n?|\n/;
994
995 var document = node.ownerDocument;
996
997 var li = document.createElement('li');
998 while (node.firstChild) {
999 li.appendChild(node.firstChild);
1000 }
1001 // An array of lines. We split below, so this is initialized to one
1002 // un-split line.
1003 var listItems = [li];
1004
1005 function walk(node) {
1006 var type = node.nodeType;
1007 if (type == 1 && !nocode.test(node.className)) { // Element
1008 if ('br' === node.nodeName) {
1009 breakAfter(node);
1010 // Discard the <BR> since it is now flush against a </LI>.
1011 if (node.parentNode) {
1012 node.parentNode.removeChild(node);
1013 }
1014 } else {
1015 for (var child = node.firstChild; child; child = child.nextSibling) {
1016 walk(child);
1017 }
1018 }
1019 } else if ((type == 3 || type == 4) && isPreformatted) { // Text
1020 var text = node.nodeValue;
1021 var match = text.match(lineBreak);
1022 if (match) {
1023 var firstLine = text.substring(0, match.index);
1024 node.nodeValue = firstLine;
1025 var tail = text.substring(match.index + match[0].length);
1026 if (tail) {
1027 var parent = node.parentNode;
1028 parent.insertBefore(
1029 document.createTextNode(tail), node.nextSibling);
1030 }
1031 breakAfter(node);
1032 if (!firstLine) {
1033 // Don't leave blank text nodes in the DOM.
1034 node.parentNode.removeChild(node);
1035 }
1036 }
1037 }
1038 }
1039
1040 // Split a line after the given node.
1041 function breakAfter(lineEndNode) {
1042 // If there's nothing to the right, then we can skip ending the line
1043 // here, and move root-wards since splitting just before an end-tag
1044 // would require us to create a bunch of empty copies.
1045 while (!lineEndNode.nextSibling) {
1046 lineEndNode = lineEndNode.parentNode;
1047 if (!lineEndNode) { return; }
1048 }
1049
1050 function breakLeftOf(limit, copy) {
1051 // Clone shallowly if this node needs to be on both sides of the break.
1052 var rightSide = copy ? limit.cloneNode(false) : limit;
1053 var parent = limit.parentNode;
1054 if (parent) {
1055 // We clone the parent chain.
1056 // This helps us resurrect important styling elements that cross lines .
1057 // E.g. in <i>Foo<br>Bar</i>
1058 // should be rewritten to <li><i>Foo</i></li><li><i>Bar</i></li>.
1059 var parentClone = breakLeftOf(parent, 1);
1060 // Move the clone and everything to the right of the original
1061 // onto the cloned parent.
1062 var next = limit.nextSibling;
1063 parentClone.appendChild(rightSide);
1064 for (var sibling = next; sibling; sibling = next) {
1065 next = sibling.nextSibling;
1066 parentClone.appendChild(sibling);
1067 }
1068 }
1069 return rightSide;
1070 }
1071
1072 var copiedListItem = breakLeftOf(lineEndNode.nextSibling, 0);
1073
1074 // Walk the parent chain until we reach an unattached LI.
1075 for (var parent;
1076 // Check nodeType since IE invents document fragments.
1077 (parent = copiedListItem.parentNode) && parent.nodeType === 1;) {
1078 copiedListItem = parent;
1079 }
1080 // Put it on the list of lines for later processing.
1081 listItems.push(copiedListItem);
1082 }
1083
1084 // Split lines while there are lines left to split.
1085 for (var i = 0; // Number of lines that have been split so far.
1086 i < listItems.length; // length updated by breakAfter calls.
1087 ++i) {
1088 walk(listItems[i]);
1089 }
1090
1091 // Make sure numeric indices show correctly.
1092 if (opt_startLineNum === (opt_startLineNum|0)) {
1093 listItems[0].setAttribute('value', opt_startLineNum);
1094 }
1095
1096 var ol = document.createElement('ol');
1097 ol.className = 'linenums';
1098 var offset = Math.max(0, ((opt_startLineNum - 1 /* zero index */)) | 0) || 0 ;
1099 for (var i = 0, n = listItems.length; i < n; ++i) {
1100 li = listItems[i];
1101 // Stick a class on the LIs so that stylesheets can
1102 // color odd/even rows, or any other row pattern that
1103 // is co-prime with 10.
1104 li.className = 'L' + ((i + offset) % 10);
1105 if (!li.firstChild) {
1106 li.appendChild(document.createTextNode('\xA0'));
1107 }
1108 ol.appendChild(li);
1109 }
1110
1111 node.appendChild(ol);
1112 }
1113 /**
1114 * Breaks {@code job.sourceCode} around style boundaries in
1115 * {@code job.decorations} and modifies {@code job.sourceNode} in place.
1116 * @param {Object} job like <pre>{
1117 * sourceCode: {string} source as plain text,
1118 * sourceNode: {HTMLElement} the element containing the source,
1119 * spans: {Array.<number|Node>} alternating span start indices into source
1120 * and the text node or element (e.g. {@code <BR>}) corresponding to tha t
1121 * span.
1122 * decorations: {Array.<number|string} an array of style classes preceded
1123 * by the position at which they start in job.sourceCode in order
1124 * }</pre>
1125 * @private
1126 */
1127 function recombineTagsAndDecorations(job) {
1128 var isIE8OrEarlier = /\bMSIE\s(\d+)/.exec(navigator.userAgent);
1129 isIE8OrEarlier = isIE8OrEarlier && +isIE8OrEarlier[1] <= 8;
1130 var newlineRe = /\n/g;
1131
1132 var source = job.sourceCode;
1133 var sourceLength = source.length;
1134 // Index into source after the last code-unit recombined.
1135 var sourceIndex = 0;
1136
1137 var spans = job.spans;
1138 var nSpans = spans.length;
1139 // Index into spans after the last span which ends at or before sourceIndex.
1140 var spanIndex = 0;
1141
1142 var decorations = job.decorations;
1143 var nDecorations = decorations.length;
1144 // Index into decorations after the last decoration which ends at or before
1145 // sourceIndex.
1146 var decorationIndex = 0;
1147
1148 // Remove all zero-length decorations.
1149 decorations[nDecorations] = sourceLength;
1150 var decPos, i;
1151 for (i = decPos = 0; i < nDecorations;) {
1152 if (decorations[i] !== decorations[i + 2]) {
1153 decorations[decPos++] = decorations[i++];
1154 decorations[decPos++] = decorations[i++];
1155 } else {
1156 i += 2;
1157 }
1158 }
1159 nDecorations = decPos;
1160
1161 // Simplify decorations.
1162 for (i = decPos = 0; i < nDecorations;) {
1163 var startPos = decorations[i];
1164 // Conflate all adjacent decorations that use the same style.
1165 var startDec = decorations[i + 1];
1166 var end = i + 2;
1167 while (end + 2 <= nDecorations && decorations[end + 1] === startDec) {
1168 end += 2;
1169 }
1170 decorations[decPos++] = startPos;
1171 decorations[decPos++] = startDec;
1172 i = end;
1173 }
1174
1175 nDecorations = decorations.length = decPos;
1176
1177 var sourceNode = job.sourceNode;
1178 var oldDisplay;
1179 if (sourceNode) {
1180 oldDisplay = sourceNode.style.display;
1181 sourceNode.style.display = 'none';
1182 }
1183 try {
1184 var decoration = null;
1185 while (spanIndex < nSpans) {
1186 var spanStart = spans[spanIndex];
1187 var spanEnd = spans[spanIndex + 2] || sourceLength;
1188
1189 var decEnd = decorations[decorationIndex + 2] || sourceLength;
1190
1191 var end = Math.min(spanEnd, decEnd);
1192
1193 var textNode = spans[spanIndex + 1];
1194 var styledText;
1195 if (textNode.nodeType !== 1 // Don't muck with <BR>s or <LI>s
1196 // Don't introduce spans around empty text nodes.
1197 && (styledText = source.substring(sourceIndex, end))) {
1198 // This may seem bizarre, and it is. Emitting LF on IE causes the
1199 // code to display with spaces instead of line breaks.
1200 // Emitting Windows standard issue linebreaks (CRLF) causes a blank
1201 // space to appear at the beginning of every line but the first.
1202 // Emitting an old Mac OS 9 line separator makes everything spiffy.
1203 if (isIE8OrEarlier) {
1204 styledText = styledText.replace(newlineRe, '\r');
1205 }
1206 textNode.nodeValue = styledText;
1207 var document = textNode.ownerDocument;
1208 var span = document.createElement('span');
1209 span.className = decorations[decorationIndex + 1];
1210 var parentNode = textNode.parentNode;
1211 parentNode.replaceChild(span, textNode);
1212 span.appendChild(textNode);
1213 if (sourceIndex < spanEnd) { // Split off a text node.
1214 spans[spanIndex + 1] = textNode
1215 // TODO: Possibly optimize by using '' if there's no flicker.
1216 = document.createTextNode(source.substring(end, spanEnd));
1217 parentNode.insertBefore(textNode, span.nextSibling);
1218 }
1219 }
1220
1221 sourceIndex = end;
1222
1223 if (sourceIndex >= spanEnd) {
1224 spanIndex += 2;
1225 }
1226 if (sourceIndex >= decEnd) {
1227 decorationIndex += 2;
1228 }
1229 }
1230 } finally {
1231 if (sourceNode) {
1232 sourceNode.style.display = oldDisplay;
1233 }
1234 }
1235 }
1236
1237 /** Maps language-specific file extensions to handlers. */
1238 var langHandlerRegistry = {};
1239 /** Register a language handler for the given file extensions.
1240 * @param {function (Object)} handler a function from source code to a list
1241 * of decorations. Takes a single argument job which describes the
1242 * state of the computation. The single parameter has the form
1243 * {@code {
1244 * sourceCode: {string} as plain text.
1245 * decorations: {Array.<number|string>} an array of style classes
1246 * preceded by the position at which they start in
1247 * job.sourceCode in order.
1248 * The language handler should assigned this field.
1249 * basePos: {int} the position of source in the larger source chunk.
1250 * All positions in the output decorations array are relative
1251 * to the larger source chunk.
1252 * } }
1253 * @param {Array.<string>} fileExtensions
1254 */
1255 function registerLangHandler(handler, fileExtensions) {
1256 for (var i = fileExtensions.length; --i >= 0;) {
1257 var ext = fileExtensions[i];
1258 if (!langHandlerRegistry.hasOwnProperty(ext)) {
1259 langHandlerRegistry[ext] = handler;
1260 } else if (win['console']) {
1261 console['warn']('cannot override language handler %s', ext);
1262 }
1263 }
1264 }
1265 function langHandlerForExtension(extension, source) {
1266 if (!(extension && langHandlerRegistry.hasOwnProperty(extension))) {
1267 // Treat it as markup if the first non whitespace character is a < and
1268 // the last non-whitespace character is a >.
1269 extension = /^\s*</.test(source)
1270 ? 'default-markup'
1271 : 'default-code';
1272 }
1273 return langHandlerRegistry[extension];
1274 }
1275 registerLangHandler(decorateSource, ['default-code']);
1276 registerLangHandler(
1277 createSimpleLexer(
1278 [],
1279 [
1280 [PR_PLAIN, /^[^<?]+/],
1281 [PR_DECLARATION, /^<!\w[^>]*(?:>|$)/],
1282 [PR_COMMENT, /^<\!--[\s\S]*?(?:-\->|$)/],
1283 // Unescaped content in an unknown language
1284 ['lang-', /^<\?([\s\S]+?)(?:\?>|$)/],
1285 ['lang-', /^<%([\s\S]+?)(?:%>|$)/],
1286 [PR_PUNCTUATION, /^(?:<[%?]|[%?]>)/],
1287 ['lang-', /^<xmp\b[^>]*>([\s\S]+?)<\/xmp\b[^>]*>/i],
1288 // Unescaped content in javascript. (Or possibly vbscript).
1289 ['lang-js', /^<script\b[^>]*>([\s\S]*?)(<\/script\b[^>]*>)/i],
1290 // Contains unescaped stylesheet content
1291 ['lang-css', /^<style\b[^>]*>([\s\S]*?)(<\/style\b[^>]*>)/i],
1292 ['lang-in.tag', /^(<\/?[a-z][^<>]*>)/i]
1293 ]),
1294 ['default-markup', 'htm', 'html', 'mxml', 'xhtml', 'xml', 'xsl']);
1295 registerLangHandler(
1296 createSimpleLexer(
1297 [
1298 [PR_PLAIN, /^[\s]+/, null, ' \t\r\n'],
1299 [PR_ATTRIB_VALUE, /^(?:\"[^\"]*\"?|\'[^\']*\'?)/, null, '\"\'']
1300 ],
1301 [
1302 [PR_TAG, /^^<\/?[a-z](?:[\w.:-]*\w)?|\/?>$/i],
1303 [PR_ATTRIB_NAME, /^(?!style[\s=]|on)[a-z](?:[\w:-]*\w)?/i],
1304 ['lang-uq.val', /^=\s*([^>\'\"\s]*(?:[^>\'\"\s\/]|\/(?=\s)))/],
1305 [PR_PUNCTUATION, /^[=<>\/]+/],
1306 ['lang-js', /^on\w+\s*=\s*\"([^\"]+)\"/i],
1307 ['lang-js', /^on\w+\s*=\s*\'([^\']+)\'/i],
1308 ['lang-js', /^on\w+\s*=\s*([^\"\'>\s]+)/i],
1309 ['lang-css', /^style\s*=\s*\"([^\"]+)\"/i],
1310 ['lang-css', /^style\s*=\s*\'([^\']+)\'/i],
1311 ['lang-css', /^style\s*=\s*([^\"\'>\s]+)/i]
1312 ]),
1313 ['in.tag']);
1314 registerLangHandler(
1315 createSimpleLexer([], [[PR_ATTRIB_VALUE, /^[\s\S]+/]]), ['uq.val']);
1316 registerLangHandler(sourceDecorator({
1317 'keywords': CPP_KEYWORDS,
1318 'hashComments': true,
1319 'cStyleComments': true,
1320 'types': C_TYPES
1321 }), ['c', 'cc', 'cpp', 'cxx', 'cyc', 'm']);
1322 registerLangHandler(sourceDecorator({
1323 'keywords': 'null,true,false'
1324 }), ['json']);
1325 registerLangHandler(sourceDecorator({
1326 'keywords': CSHARP_KEYWORDS,
1327 'hashComments': true,
1328 'cStyleComments': true,
1329 'verbatimStrings': true,
1330 'types': C_TYPES
1331 }), ['cs']);
1332 registerLangHandler(sourceDecorator({
1333 'keywords': JAVA_KEYWORDS,
1334 'cStyleComments': true
1335 }), ['java']);
1336 registerLangHandler(sourceDecorator({
1337 'keywords': SH_KEYWORDS,
1338 'hashComments': true,
1339 'multiLineStrings': true
1340 }), ['bash', 'bsh', 'csh', 'sh']);
1341 registerLangHandler(sourceDecorator({
1342 'keywords': PYTHON_KEYWORDS,
1343 'hashComments': true,
1344 'multiLineStrings': true,
1345 'tripleQuotedStrings': true
1346 }), ['cv', 'py', 'python']);
1347 registerLangHandler(sourceDecorator({
1348 'keywords': PERL_KEYWORDS,
1349 'hashComments': true,
1350 'multiLineStrings': true,
1351 'regexLiterals': 2 // multiline regex literals
1352 }), ['perl', 'pl', 'pm']);
1353 registerLangHandler(sourceDecorator({
1354 'keywords': RUBY_KEYWORDS,
1355 'hashComments': true,
1356 'multiLineStrings': true,
1357 'regexLiterals': true
1358 }), ['rb', 'ruby']);
1359 registerLangHandler(sourceDecorator({
1360 'keywords': JSCRIPT_KEYWORDS,
1361 'cStyleComments': true,
1362 'regexLiterals': true
1363 }), ['javascript', 'js']);
1364 registerLangHandler(sourceDecorator({
1365 'keywords': COFFEE_KEYWORDS,
1366 'hashComments': 3, // ### style block comments
1367 'cStyleComments': true,
1368 'multilineStrings': true,
1369 'tripleQuotedStrings': true,
1370 'regexLiterals': true
1371 }), ['coffee']);
1372 registerLangHandler(sourceDecorator({
1373 'keywords': RUST_KEYWORDS,
1374 'cStyleComments': true,
1375 'multilineStrings': true
1376 }), ['rc', 'rs', 'rust']);
1377 registerLangHandler(
1378 createSimpleLexer([], [[PR_STRING, /^[\s\S]+/]]), ['regex']);
1379
1380 function applyDecorator(job) {
1381 var opt_langExtension = job.langExtension;
1382
1383 try {
1384 // Extract tags, and convert the source code to plain text.
1385 var sourceAndSpans = extractSourceSpans(job.sourceNode, job.pre);
1386 /** Plain text. @type {string} */
1387 var source = sourceAndSpans.sourceCode;
1388 job.sourceCode = source;
1389 job.spans = sourceAndSpans.spans;
1390 job.basePos = 0;
1391
1392 // Apply the appropriate language handler
1393 langHandlerForExtension(opt_langExtension, source)(job);
1394
1395 // Integrate the decorations and tags back into the source code,
1396 // modifying the sourceNode in place.
1397 recombineTagsAndDecorations(job);
1398 } catch (e) {
1399 if (win['console']) {
1400 console['log'](e && e['stack'] || e);
1401 }
1402 }
1403 }
1404
1405 /**
1406 * Pretty print a chunk of code.
1407 * @param sourceCodeHtml {string} The HTML to pretty print.
1408 * @param opt_langExtension {string} The language name to use.
1409 * Typically, a filename extension like 'cpp' or 'java'.
1410 * @param opt_numberLines {number|boolean} True to number lines,
1411 * or the 1-indexed number of the first line in sourceCodeHtml.
1412 */
1413 function $prettyPrintOne(sourceCodeHtml, opt_langExtension, opt_numberLines) {
1414 var container = document.createElement('div');
1415 // This could cause images to load and onload listeners to fire.
1416 // E.g. <img onerror="alert(1337)" src="nosuchimage.png">.
1417 // We assume that the inner HTML is from a trusted source.
1418 // The pre-tag is required for IE8 which strips newlines from innerHTML
1419 // when it is injected into a <pre> tag.
1420 // http://stackoverflow.com/questions/451486/pre-tag-loses-line-breaks-when- setting-innerhtml-in-ie
1421 // http://stackoverflow.com/questions/195363/inserting-a-newline-into-a-pre- tag-ie-javascript
1422 container.innerHTML = '<pre>' + sourceCodeHtml + '</pre>';
1423 container = container.firstChild;
1424 if (opt_numberLines) {
1425 numberLines(container, opt_numberLines, true);
1426 }
1427
1428 var job = {
1429 langExtension: opt_langExtension,
1430 numberLines: opt_numberLines,
1431 sourceNode: container,
1432 pre: 1
1433 };
1434 applyDecorator(job);
1435 return container.innerHTML;
1436 }
1437
1438 /**
1439 * Find all the {@code <pre>} and {@code <code>} tags in the DOM with
1440 * {@code class=prettyprint} and prettify them.
1441 *
1442 * @param {Function} opt_whenDone called when prettifying is done.
1443 * @param {HTMLElement|HTMLDocument} opt_root an element or document
1444 * containing all the elements to pretty print.
1445 * Defaults to {@code document.body}.
1446 */
1447 function $prettyPrint(opt_whenDone, opt_root) {
1448 var root = opt_root || document.body;
1449 var doc = root.ownerDocument || document;
1450 function byTagName(tn) { return root.getElementsByTagName(tn); }
1451 // fetch a list of nodes to rewrite
1452 var codeSegments = [byTagName('pre'), byTagName('code'), byTagName('xmp')];
1453 var elements = [];
1454 for (var i = 0; i < codeSegments.length; ++i) {
1455 for (var j = 0, n = codeSegments[i].length; j < n; ++j) {
1456 elements.push(codeSegments[i][j]);
1457 }
1458 }
1459 codeSegments = null;
1460
1461 var clock = Date;
1462 if (!clock['now']) {
1463 clock = { 'now': function () { return +(new Date); } };
1464 }
1465
1466 // The loop is broken into a series of continuations to make sure that we
1467 // don't make the browser unresponsive when rewriting a large page.
1468 var k = 0;
1469 var prettyPrintingJob;
1470
1471 var langExtensionRe = /\blang(?:uage)?-([\w.]+)(?!\S)/;
1472 var prettyPrintRe = /\bprettyprint\b/;
1473 var prettyPrintedRe = /\bprettyprinted\b/;
1474 var preformattedTagNameRe = /pre|xmp/i;
1475 var codeRe = /^code$/i;
1476 var preCodeXmpRe = /^(?:pre|code|xmp)$/i;
1477 var EMPTY = {};
1478
1479 function doWork() {
1480 var endTime = (win['PR_SHOULD_USE_CONTINUATION'] ?
1481 clock['now']() + 250 /* ms */ :
1482 Infinity);
1483 for (; k < elements.length && clock['now']() < endTime; k++) {
1484 var cs = elements[k];
1485
1486 // Look for a preceding comment like
1487 // <?prettify lang="..." linenums="..."?>
1488 var attrs = EMPTY;
1489 {
1490 for (var preceder = cs; (preceder = preceder.previousSibling);) {
1491 var nt = preceder.nodeType;
1492 // <?foo?> is parsed by HTML 5 to a comment node (8)
1493 // like <!--?foo?-->, but in XML is a processing instruction
1494 var value = (nt === 7 || nt === 8) && preceder.nodeValue;
1495 if (value
1496 ? !/^\??prettify\b/.test(value)
1497 : (nt !== 3 || /\S/.test(preceder.nodeValue))) {
1498 // Skip over white-space text nodes but not others.
1499 break;
1500 }
1501 if (value) {
1502 attrs = {};
1503 value.replace(
1504 /\b(\w+)=([\w:.%+-]+)/g,
1505 function (_, name, value) { attrs[name] = value; });
1506 break;
1507 }
1508 }
1509 }
1510
1511 var className = cs.className;
1512 if ((attrs !== EMPTY || prettyPrintRe.test(className))
1513 // Don't redo this if we've already done it.
1514 // This allows recalling pretty print to just prettyprint elements
1515 // that have been added to the page since last call.
1516 && !prettyPrintedRe.test(className)) {
1517
1518 // make sure this is not nested in an already prettified element
1519 var nested = false;
1520 for (var p = cs.parentNode; p; p = p.parentNode) {
1521 var tn = p.tagName;
1522 if (preCodeXmpRe.test(tn)
1523 && p.className && prettyPrintRe.test(p.className)) {
1524 nested = true;
1525 break;
1526 }
1527 }
1528 if (!nested) {
1529 // Mark done. If we fail to prettyprint for whatever reason,
1530 // we shouldn't try again.
1531 cs.className += ' prettyprinted';
1532
1533 // If the classes includes a language extensions, use it.
1534 // Language extensions can be specified like
1535 // <pre class="prettyprint lang-cpp">
1536 // the language extension "cpp" is used to find a language handler
1537 // as passed to PR.registerLangHandler.
1538 // HTML5 recommends that a language be specified using "language-"
1539 // as the prefix instead. Google Code Prettify supports both.
1540 // http://dev.w3.org/html5/spec-author-view/the-code-element.html
1541 var langExtension = attrs['lang'];
1542 if (!langExtension) {
1543 langExtension = className.match(langExtensionRe);
1544 // Support <pre class="prettyprint"><code class="language-c">
1545 var wrapper;
1546 if (!langExtension && (wrapper = childContentWrapper(cs))
1547 && codeRe.test(wrapper.tagName)) {
1548 langExtension = wrapper.className.match(langExtensionRe);
1549 }
1550
1551 if (langExtension) { langExtension = langExtension[1]; }
1552 }
1553
1554 var preformatted;
1555 if (preformattedTagNameRe.test(cs.tagName)) {
1556 preformatted = 1;
1557 } else {
1558 var currentStyle = cs['currentStyle'];
1559 var defaultView = doc.defaultView;
1560 var whitespace = (
1561 currentStyle
1562 ? currentStyle['whiteSpace']
1563 : (defaultView
1564 && defaultView.getComputedStyle)
1565 ? defaultView.getComputedStyle(cs, null)
1566 .getPropertyValue('white-space')
1567 : 0);
1568 preformatted = whitespace
1569 && 'pre' === whitespace.substring(0, 3);
1570 }
1571
1572 // Look for a class like linenums or linenums:<n> where <n> is the
1573 // 1-indexed number of the first line.
1574 var lineNums = attrs['linenums'];
1575 if (!(lineNums = lineNums === 'true' || +lineNums)) {
1576 lineNums = className.match(/\blinenums\b(?::(\d+))?/);
1577 lineNums =
1578 lineNums
1579 ? lineNums[1] && lineNums[1].length
1580 ? +lineNums[1] : true
1581 : false;
1582 }
1583 if (lineNums) { numberLines(cs, lineNums, preformatted); }
1584
1585 // do the pretty printing
1586 prettyPrintingJob = {
1587 langExtension: langExtension,
1588 sourceNode: cs,
1589 numberLines: lineNums,
1590 pre: preformatted
1591 };
1592 applyDecorator(prettyPrintingJob);
1593 }
1594 }
1595 }
1596 if (k < elements.length) {
1597 // finish up in a continuation
1598 setTimeout(doWork, 250);
1599 } else if ('function' === typeof opt_whenDone) {
1600 opt_whenDone();
1601 }
1602 }
1603
1604 doWork();
1605 }
1606
1607 /**
1608 * Contains functions for creating and registering new language handlers.
1609 * @type {Object}
1610 */
1611 var PR = win['PR'] = {
1612 'createSimpleLexer': createSimpleLexer,
1613 'registerLangHandler': registerLangHandler,
1614 'sourceDecorator': sourceDecorator,
1615 'PR_ATTRIB_NAME': PR_ATTRIB_NAME,
1616 'PR_ATTRIB_VALUE': PR_ATTRIB_VALUE,
1617 'PR_COMMENT': PR_COMMENT,
1618 'PR_DECLARATION': PR_DECLARATION,
1619 'PR_KEYWORD': PR_KEYWORD,
1620 'PR_LITERAL': PR_LITERAL,
1621 'PR_NOCODE': PR_NOCODE,
1622 'PR_PLAIN': PR_PLAIN,
1623 'PR_PUNCTUATION': PR_PUNCTUATION,
1624 'PR_SOURCE': PR_SOURCE,
1625 'PR_STRING': PR_STRING,
1626 'PR_TAG': PR_TAG,
1627 'PR_TYPE': PR_TYPE,
1628 'prettyPrintOne':
1629 IN_GLOBAL_SCOPE
1630 ? (win['prettyPrintOne'] = $prettyPrintOne)
1631 : (prettyPrintOne = $prettyPrintOne),
1632 'prettyPrint': prettyPrint =
1633 IN_GLOBAL_SCOPE
1634 ? (win['prettyPrint'] = $prettyPrint)
1635 : (prettyPrint = $prettyPrint)
1636 };
1637
1638 // Make PR available via the Asynchronous Module Definition (AMD) API.
1639 // Per https://github.com/amdjs/amdjs-api/wiki/AMD:
1640 // The Asynchronous Module Definition (AMD) API specifies a
1641 // mechanism for defining modules such that the module and its
1642 // dependencies can be asynchronously loaded.
1643 // ...
1644 // To allow a clear indicator that a global define function (as
1645 // needed for script src browser loading) conforms to the AMD API,
1646 // any global define function SHOULD have a property called "amd"
1647 // whose value is an object. This helps avoid conflict with any
1648 // other existing JavaScript code that could have defined a define()
1649 // function that does not conform to the AMD API.
1650 if (typeof define === "function" && define['amd']) {
1651 define("google-code-prettify", [], function () {
1652 return PR;
1653 });
1654 }
1655 })();
OLDNEW
« no previous file with comments | « bower_components/google-code-prettify/src/prettify.css ('k') | bower_components/google-code-prettify/src/run_prettify.js » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698