+ * For a fairly comprehensive set of languages see the + * README + * file that came with this source. At a minimum, the lexer should work on a + * number of languages including C and friends, Java, Python, Bash, SQL, HTML, + * XML, CSS, Javascript, and Makefiles. It works passably on Ruby, PHP and Awk + * and a subset of Perl, but, because of commenting conventions, doesn't work on + * Smalltalk, Lisp-like, or CAML-like languages without an explicit lang class. + *
+ * Usage:
} and {@code } tags in your source with
+ * {@code class=prettyprint.}
+ * You can also use the (html deprecated) {@code } tag, but the pretty
+ * printer needs to do more substantial DOM manipulations to support that, so
+ * some css styles may not be preserved.
+ * </ol>
+ * That's it. I wanted to keep the API as simple as possible, so there's no
+ * need to specify which language the code is in, but if you wish, you can add
+ * another class to the {@code <pre>} or {@code <code>} element to specify the
+ * language, as in {@code <pre class="prettyprint lang-java">}. Any class that
+ * starts with "lang-" followed by a file extension, specifies the file type.
+ * See the "lang-*.js" files in this directory for code that implements
+ * per-language file handlers.
+ * <p>
+ * Change log:<br>
+ * cbeust, 2006/08/22
+ * <blockquote>
+ * Java annotations (start with "@") are now captured as literals ("lit")
+ * </blockquote>
+ * @requires console
+ */
+
+// JSLint declarations
+/*global console, document, navigator, setTimeout, window */
+
+/**
+ * Split {@code prettyPrint} into multiple timeouts so as not to interfere with
+ * UI events.
+ * If set to {@code false}, {@code prettyPrint()} is synchronous.
+ */
+window['PR_SHOULD_USE_CONTINUATION'] = true;
+
+(function () {
+ // Keyword lists for various languages.
+ // We use things that coerce to strings to make them compact when minified
+ // and to defeat aggressive optimizers that fold large string constants.
+ var FLOW_CONTROL_KEYWORDS = ["break,continue,do,else,for,if,return,while"];
+ var C_KEYWORDS = [FLOW_CONTROL_KEYWORDS,"auto,case,char,const,default," +
+ "double,enum,extern,float,goto,int,long,register,short,signed,sizeof," +
+ "static,struct,switch,typedef,union,unsigned,void,volatile"];
+ var COMMON_KEYWORDS = [C_KEYWORDS,"catch,class,delete,false,import," +
+ "new,operator,private,protected,public,this,throw,true,try,typeof"];
+ var CPP_KEYWORDS = [COMMON_KEYWORDS,"alignof,align_union,asm,axiom,bool," +
+ "concept,concept_map,const_cast,constexpr,decltype," +
+ "dynamic_cast,explicit,export,friend,inline,late_check," +
+ "mutable,namespace,nullptr,reinterpret_cast,static_assert,static_cast," +
+ "template,typeid,typename,using,virtual,where"];
+ var JAVA_KEYWORDS = [COMMON_KEYWORDS,
+ "abstract,boolean,byte,extends,final,finally,implements,import," +
+ "instanceof,null,native,package,strictfp,super,synchronized,throws," +
+ "transient"];
+ var CSHARP_KEYWORDS = [JAVA_KEYWORDS,
+ "as,base,by,checked,decimal,delegate,descending,dynamic,event," +
+ "fixed,foreach,from,group,implicit,in,interface,internal,into,is,lock," +
+ "object,out,override,orderby,params,partial,readonly,ref,sbyte,sealed," +
+ "stackalloc,string,select,uint,ulong,unchecked,unsafe,ushort,var"];
+ var COFFEE_KEYWORDS = "all,and,by,catch,class,else,extends,false,finally," +
+ "for,if,in,is,isnt,loop,new,no,not,null,of,off,on,or,return,super,then," +
+ "true,try,unless,until,when,while,yes";
+ var JSCRIPT_KEYWORDS = [COMMON_KEYWORDS,
+ "debugger,eval,export,function,get,null,set,undefined,var,with," +
+ "Infinity,NaN"];
+ var PERL_KEYWORDS = "caller,delete,die,do,dump,elsif,eval,exit,foreach,for," +
+ "goto,if,import,last,local,my,next,no,our,print,package,redo,require," +
+ "sub,undef,unless,until,use,wantarray,while,BEGIN,END";
+ var PYTHON_KEYWORDS = [FLOW_CONTROL_KEYWORDS, "and,as,assert,class,def,del," +
+ "elif,except,exec,finally,from,global,import,in,is,lambda," +
+ "nonlocal,not,or,pass,print,raise,try,with,yield," +
+ "False,True,None"];
+ var RUBY_KEYWORDS = [FLOW_CONTROL_KEYWORDS, "alias,and,begin,case,class," +
+ "def,defined,elsif,end,ensure,false,in,module,next,nil,not,or,redo," +
+ "rescue,retry,self,super,then,true,undef,unless,until,when,yield," +
+ "BEGIN,END"];
+ var SH_KEYWORDS = [FLOW_CONTROL_KEYWORDS, "case,done,elif,esac,eval,fi," +
+ "function,in,local,set,then,until"];
+ var ALL_KEYWORDS = [
+ CPP_KEYWORDS, CSHARP_KEYWORDS, JSCRIPT_KEYWORDS, PERL_KEYWORDS +
+ PYTHON_KEYWORDS, RUBY_KEYWORDS, SH_KEYWORDS];
+ var C_TYPES = /^(DIR|FILE|vector|(de|priority_)?queue|list|stack|(const_)?iterator|(multi)?(set|map)|bitset|u?(int|float)\d*)/;
+
+ // token style names. correspond to css classes
+ /**
+ * token style for a string literal
+ * @const
+ */
+ var PR_STRING = 'str';
+ /**
+ * token style for a keyword
+ * @const
+ */
+ var PR_KEYWORD = 'kwd';
+ /**
+ * token style for a comment
+ * @const
+ */
+ var PR_COMMENT = 'com';
+ /**
+ * token style for a type
+ * @const
+ */
+ var PR_TYPE = 'typ';
+ /**
+ * token style for a literal value. e.g. 1, null, true.
+ * @const
+ */
+ var PR_LITERAL = 'lit';
+ /**
+ * token style for a punctuation string.
+ * @const
+ */
+ var PR_PUNCTUATION = 'pun';
+ /**
+ * token style for a punctuation string.
+ * @const
+ */
+ var PR_PLAIN = 'pln';
+
+ /**
+ * token style for an sgml tag.
+ * @const
+ */
+ var PR_TAG = 'tag';
+ /**
+ * token style for a markup declaration such as a DOCTYPE.
+ * @const
+ */
+ var PR_DECLARATION = 'dec';
+ /**
+ * token style for embedded source.
+ * @const
+ */
+ var PR_SOURCE = 'src';
+ /**
+ * token style for an sgml attribute name.
+ * @const
+ */
+ var PR_ATTRIB_NAME = 'atn';
+ /**
+ * token style for an sgml attribute value.
+ * @const
+ */
+ var PR_ATTRIB_VALUE = 'atv';
+
+ /**
+ * A class that indicates a section of markup that is not code, e.g. to allow
+ * embedding of line numbers within code listings.
+ * @const
+ */
+ var PR_NOCODE = 'nocode';
+
+
+
+/**
+ * A set of tokens that can precede a regular expression literal in
+ * javascript
+ * http://web.archive.org/web/20070717142515/http://www.mozilla.org/js/language/js20/rationale/syntax.html
+ * has the full list, but I've removed ones that might be problematic when
+ * seen in languages that don't support regular expression literals.
+ *
+ * <p>Specifically, I've removed any keywords that can't precede a regexp
+ * literal in a syntactically legal javascript program, and I've removed the
+ * "in" keyword since it's not a keyword in many languages, and might be used
+ * as a count of inches.
+ *
+ * <p>The link a above does not accurately describe EcmaScript rules since
+ * it fails to distinguish between (a=++/b/i) and (a++/b/i) but it works
+ * very well in practice.
+ *
+ * @private
+ * @const
+ */
+var REGEXP_PRECEDER_PATTERN = '(?:^^\\.?|[+-]|\\!|\\!=|\\!==|\\#|\\%|\\%=|&|&&|&&=|&=|\\(|\\*|\\*=|\\+=|\\,|\\-=|\\->|\\/|\\/=|:|::|\\;|<|<<|<<=|<=|=|==|===|>|>=|>>|>>=|>>>|>>>=|\\?|\\@|\\[|\\^|\\^=|\\^\\^|\\^\\^=|\\{|\\||\\|=|\\|\\||\\|\\|=|\\~|break|case|continue|delete|do|else|finally|instanceof|return|throw|try|typeof)\\s*';
+
+// CAVEAT: this does not properly handle the case where a regular
+// expression immediately follows another since a regular expression may
+// have flags for case-sensitivity and the like. Having regexp tokens
+// adjacent is not valid in any language I'm aware of, so I'm punting.
+// TODO: maybe style special characters inside a regexp as punctuation.
+
+
+ /**
+ * Given a group of {@link RegExp}s, returns a {@code RegExp} that globally
+ * matches the union of the sets of strings matched by the input RegExp.
+ * Since it matches globally, if the input strings have a start-of-input
+ * anchor (/^.../), it is ignored for the purposes of unioning.
+ * @param {Array.<RegExp>} regexs non multiline, non-global regexs.
+ * @return {RegExp} a global regex.
+ */
+ function combinePrefixPatterns(regexs) {
+ var capturedGroupIndex = 0;
+
+ var needToFoldCase = false;
+ var ignoreCase = false;
+ for (var i = 0, n = regexs.length; i < n; ++i) {
+ var regex = regexs[i];
+ if (regex.ignoreCase) {
+ ignoreCase = true;
+ } else if (/[a-z]/i.test(regex.source.replace(
+ /\\u[0-9a-f]{4}|\\x[0-9a-f]{2}|\\[^ux]/gi, ''))) {
+ needToFoldCase = true;
+ ignoreCase = false;
+ break;
+ }
+ }
+
+ var escapeCharToCodeUnit = {
+ 'b': 8,
+ 't': 9,
+ 'n': 0xa,
+ 'v': 0xb,
+ 'f': 0xc,
+ 'r': 0xd
+ };
+
+ function decodeEscape(charsetPart) {
+ var cc0 = charsetPart.charCodeAt(0);
+ if (cc0 !== 92 /* \\ */) {
+ return cc0;
+ }
+ var c1 = charsetPart.charAt(1);
+ cc0 = escapeCharToCodeUnit[c1];
+ if (cc0) {
+ return cc0;
+ } else if ('0' <= c1 && c1 <= '7') {
+ return parseInt(charsetPart.substring(1), 8);
+ } else if (c1 === 'u' || c1 === 'x') {
+ return parseInt(charsetPart.substring(2), 16);
+ } else {
+ return charsetPart.charCodeAt(1);
+ }
+ }
+
+ function encodeEscape(charCode) {
+ if (charCode < 0x20) {
+ return (charCode < 0x10 ? '\\x0' : '\\x') + charCode.toString(16);
+ }
+ var ch = String.fromCharCode(charCode);
+ if (ch === '\\' || ch === '-' || ch === '[' || ch === ']') {
+ ch = '\\' + ch;
+ }
+ return ch;
+ }
+
+ function caseFoldCharset(charSet) {
+ var charsetParts = charSet.substring(1, charSet.length - 1).match(
+ new RegExp(
+ '\\\\u[0-9A-Fa-f]{4}'
+ + '|\\\\x[0-9A-Fa-f]{2}'
+ + '|\\\\[0-3][0-7]{0,2}'
+ + '|\\\\[0-7]{1,2}'
+ + '|\\\\[\\s\\S]'
+ + '|-'
+ + '|[^-\\\\]',
+ 'g'));
+ var groups = [];
+ var ranges = [];
+ var inverse = charsetParts[0] === '^';
+ for (var i = inverse ? 1 : 0, n = charsetParts.length; i < n; ++i) {
+ var p = charsetParts[i];
+ if (/\\[bdsw]/i.test(p)) { // Don't muck with named groups.
+ groups.push(p);
+ } else {
+ var start = decodeEscape(p);
+ var end;
+ if (i + 2 < n && '-' === charsetParts[i + 1]) {
+ end = decodeEscape(charsetParts[i + 2]);
+ i += 2;
+ } else {
+ end = start;
+ }
+ ranges.push([start, end]);
+ // If the range might intersect letters, then expand it.
+ // This case handling is too simplistic.
+ // It does not deal with non-latin case folding.
+ // It works for latin source code identifiers though.
+ if (!(end < 65 || start > 122)) {
+ if (!(end < 65 || start > 90)) {
+ ranges.push([Math.max(65, start) | 32, Math.min(end, 90) | 32]);
+ }
+ if (!(end < 97 || start > 122)) {
+ ranges.push([Math.max(97, start) & ~32, Math.min(end, 122) & ~32]);
+ }
+ }
+ }
+ }
+
+ // [[1, 10], [3, 4], [8, 12], [14, 14], [16, 16], [17, 17]]
+ // -> [[1, 12], [14, 14], [16, 17]]
+ ranges.sort(function (a, b) { return (a[0] - b[0]) || (b[1] - a[1]); });
+ var consolidatedRanges = [];
+ var lastRange = [NaN, NaN];
+ for (var i = 0; i < ranges.length; ++i) {
+ var range = ranges[i];
+ if (range[0] <= lastRange[1] + 1) {
+ lastRange[1] = Math.max(lastRange[1], range[1]);
+ } else {
+ consolidatedRanges.push(lastRange = range);
+ }
+ }
+
+ var out = ['['];
+ if (inverse) { out.push('^'); }
+ out.push.apply(out, groups);
+ for (var i = 0; i < consolidatedRanges.length; ++i) {
+ var range = consolidatedRanges[i];
+ out.push(encodeEscape(range[0]));
+ if (range[1] > range[0]) {
+ if (range[1] + 1 > range[0]) { out.push('-'); }
+ out.push(encodeEscape(range[1]));
+ }
+ }
+ out.push(']');
+ return out.join('');
+ }
+
+ function allowAnywhereFoldCaseAndRenumberGroups(regex) {
+ // Split into character sets, escape sequences, punctuation strings
+ // like ('(', '(?:', ')', '^'), and runs of characters that do not
+ // include any of the above.
+ var parts = regex.source.match(
+ new RegExp(
+ '(?:'
+ + '\\[(?:[^\\x5C\\x5D]|\\\\[\\s\\S])*\\]' // a character set
+ + '|\\\\u[A-Fa-f0-9]{4}' // a unicode escape
+ + '|\\\\x[A-Fa-f0-9]{2}' // a hex escape
+ + '|\\\\[0-9]+' // a back-reference or octal escape
+ + '|\\\\[^ux0-9]' // other escape sequence
+ + '|\\(\\?[:!=]' // start of a non-capturing group
+ + '|[\\(\\)\\^]' // start/emd of a group, or line start
+ + '|[^\\x5B\\x5C\\(\\)\\^]+' // run of other characters
+ + ')',
+ 'g'));
+ var n = parts.length;
+
+ // Maps captured group numbers to the number they will occupy in
+ // the output or to -1 if that has not been determined, or to
+ // undefined if they need not be capturing in the output.
+ var capturedGroups = [];
+
+ // Walk over and identify back references to build the capturedGroups
+ // mapping.
+ for (var i = 0, groupIndex = 0; i < n; ++i) {
+ var p = parts[i];
+ if (p === '(') {
+ // groups are 1-indexed, so max group index is count of '('
+ ++groupIndex;
+ } else if ('\\' === p.charAt(0)) {
+ var decimalValue = +p.substring(1);
+ if (decimalValue && decimalValue <= groupIndex) {
+ capturedGroups[decimalValue] = -1;
+ }
+ }
+ }
+
+ // Renumber groups and reduce capturing groups to non-capturing groups
+ // where possible.
+ for (var i = 1; i < capturedGroups.length; ++i) {
+ if (-1 === capturedGroups[i]) {
+ capturedGroups[i] = ++capturedGroupIndex;
+ }
+ }
+ for (var i = 0, groupIndex = 0; i < n; ++i) {
+ var p = parts[i];
+ if (p === '(') {
+ ++groupIndex;
+ if (capturedGroups[groupIndex] === undefined) {
+ parts[i] = '(?:';
+ }
+ } else if ('\\' === p.charAt(0)) {
+ var decimalValue = +p.substring(1);
+ if (decimalValue && decimalValue <= groupIndex) {
+ parts[i] = '\\' + capturedGroups[groupIndex];
+ }
+ }
+ }
+
+ // Remove any prefix anchors so that the output will match anywhere.
+ // ^^ really does mean an anchored match though.
+ for (var i = 0, groupIndex = 0; i < n; ++i) {
+ if ('^' === parts[i] && '^' !== parts[i + 1]) { parts[i] = ''; }
+ }
+
+ // Expand letters to groups to handle mixing of case-sensitive and
+ // case-insensitive patterns if necessary.
+ if (regex.ignoreCase && needToFoldCase) {
+ for (var i = 0; i < n; ++i) {
+ var p = parts[i];
+ var ch0 = p.charAt(0);
+ if (p.length >= 2 && ch0 === '[') {
+ parts[i] = caseFoldCharset(p);
+ } else if (ch0 !== '\\') {
+ // TODO: handle letters in numeric escapes.
+ parts[i] = p.replace(
+ /[a-zA-Z]/g,
+ function (ch) {
+ var cc = ch.charCodeAt(0);
+ return '[' + String.fromCharCode(cc & ~32, cc | 32) + ']';
+ });
+ }
+ }
+ }
+
+ return parts.join('');
+ }
+
+ var rewritten = [];
+ for (var i = 0, n = regexs.length; i < n; ++i) {
+ var regex = regexs[i];
+ if (regex.global || regex.multiline) { throw new Error('' + regex); }
+ rewritten.push(
+ '(?:' + allowAnywhereFoldCaseAndRenumberGroups(regex) + ')');
+ }
+
+ return new RegExp(rewritten.join('|'), ignoreCase ? 'gi' : 'g');
+ }
+
+
+ /**
+ * Split markup into a string of source code and an array mapping ranges in
+ * that string to the text nodes in which they appear.
+ *
+ * <p>
+ * The HTML DOM structure:</p>
+ * <pre>
+ * (Element "p"
+ * (Element "b"
+ * (Text "print ")) ; #1
+ * (Text "'Hello '") ; #2
+ * (Element "br") ; #3
+ * (Text " + 'World';")) ; #4
+ * </pre>
+ * <p>
+ * corresponds to the HTML
+ * {@code <p><b>print </b>'Hello '<br> + 'World';</p>}.</p>
+ *
+ * <p>
+ * It will produce the output:</p>
+ * <pre>
+ * {
+ * sourceCode: "print 'Hello '\n + 'World';",
+ * // 1 2
+ * // 012345678901234 5678901234567
+ * spans: [0, #1, 6, #2, 14, #3, 15, #4]
+ * }
+ * </pre>
+ * <p>
+ * where #1 is a reference to the {@code "print "} text node above, and so
+ * on for the other text nodes.
+ * </p>
+ *
+ * <p>
+ * The {@code} spans array is an array of pairs. Even elements are the start
+ * indices of substrings, and odd elements are the text nodes (or BR elements)
+ * that contain the text for those substrings.
+ * Substrings continue until the next index or the end of the source.
+ * </p>
+ *
+ * @param {Node} node an HTML DOM subtree containing source-code.
+ * @return {Object} source code and the text nodes in which they occur.
+ */
+ function extractSourceSpans(node) {
+ var nocode = /(?:^|\s)nocode(?:\s|$)/;
+
+ var chunks = [];
+ var length = 0;
+ var spans = [];
+ var k = 0;
+
+ var whitespace;
+ if (node.currentStyle) {
+ whitespace = node.currentStyle.whiteSpace;
+ } else if (window.getComputedStyle) {
+ whitespace = document.defaultView.getComputedStyle(node, null)
+ .getPropertyValue('white-space');
+ }
+ var isPreformatted = whitespace && 'pre' === whitespace.substring(0, 3);
+
+ function walk(node) {
+ switch (node.nodeType) {
+ case 1: // Element
+ if (nocode.test(node.className)) { return; }
+ for (var child = node.firstChild; child; child = child.nextSibling) {
+ walk(child);
+ }
+ var nodeName = node.nodeName;
+ if ('BR' === nodeName || 'LI' === nodeName) {
+ chunks[k] = '\n';
+ spans[k << 1] = length++;
+ spans[(k++ << 1) | 1] = node;
+ }
+ break;
+ case 3: case 4: // Text
+ var text = node.nodeValue;
+ if (text.length) {
+ if (!isPreformatted) {
+ text = text.replace(/[ \t\r\n]+/g, ' ');
+ } else {
+ text = text.replace(/\r\n?/g, '\n'); // Normalize newlines.
+ }
+ // TODO: handle tabs here?
+ chunks[k] = text;
+ spans[k << 1] = length;
+ length += text.length;
+ spans[(k++ << 1) | 1] = node;
+ }
+ break;
+ }
+ }
+
+ walk(node);
+
+ return {
+ sourceCode: chunks.join('').replace(/\n$/, ''),
+ spans: spans
+ };
+ }
+
+
+ /**
+ * Apply the given language handler to sourceCode and add the resulting
+ * decorations to out.
+ * @param {number} basePos the index of sourceCode within the chunk of source
+ * whose decorations are already present on out.
+ */
+ function appendDecorations(basePos, sourceCode, langHandler, out) {
+ if (!sourceCode) { return; }
+ var job = {
+ sourceCode: sourceCode,
+ basePos: basePos
+ };
+ langHandler(job);
+ out.push.apply(out, job.decorations);
+ }
+
+ var notWs = /\S/;
+
+ /**
+ * Given an element, if it contains only one child element and any text nodes
+ * it contains contain only space characters, return the sole child element.
+ * Otherwise returns undefined.
+ * <p>
+ * This is meant to return the CODE element in {@code <pre><code ...>} when
+ * there is a single child element that contains all the non-space textual
+ * content, but not to return anything where there are multiple child elements
+ * as in {@code <pre><code>...</code><code>...</code></pre>} or when there
+ * is textual content.
+ */
+ function childContentWrapper(element) {
+ var wrapper = undefined;
+ for (var c = element.firstChild; c; c = c.nextSibling) {
+ var type = c.nodeType;
+ wrapper = (type === 1) // Element Node
+ ? (wrapper ? element : c)
+ : (type === 3) // Text Node
+ ? (notWs.test(c.nodeValue) ? element : wrapper)
+ : wrapper;
+ }
+ return wrapper === element ? undefined : wrapper;
+ }
+
+ /** Given triples of [style, pattern, context] returns a lexing function,
+ * The lexing function interprets the patterns to find token boundaries and
+ * returns a decoration list of the form
+ * [index_0, style_0, index_1, style_1, ..., index_n, style_n]
+ * where index_n is an index into the sourceCode, and style_n is a style
+ * constant like PR_PLAIN. index_n-1 <= index_n, and style_n-1 applies to
+ * all characters in sourceCode[index_n-1:index_n].
+ *
+ * The stylePatterns is a list whose elements have the form
+ * [style : string, pattern : RegExp, DEPRECATED, shortcut : string].
+ *
+ * Style is a style constant like PR_PLAIN, or can be a string of the
+ * form 'lang-FOO', where FOO is a language extension describing the
+ * language of the portion of the token in $1 after pattern executes.
+ * E.g., if style is 'lang-lisp', and group 1 contains the text
+ * '(hello (world))', then that portion of the token will be passed to the
+ * registered lisp handler for formatting.
+ * The text before and after group 1 will be restyled using this decorator
+ * so decorators should take care that this doesn't result in infinite
+ * recursion. For example, the HTML lexer rule for SCRIPT elements looks
+ * something like ['lang-js', /<[s]cript>(.+?)<\/script>/]. This may match
+ * '<script>foo()<\/script>', which would cause the current decorator to
+ * be called with '<script>' which would not match the same rule since
+ * group 1 must not be empty, so it would be instead styled as PR_TAG by
+ * the generic tag rule. The handler registered for the 'js' extension would
+ * then be called with 'foo()', and finally, the current decorator would
+ * be called with '<\/script>' which would not match the original rule and
+ * so the generic tag rule would identify it as a tag.
+ *
+ * Pattern must only match prefixes, and if it matches a prefix, then that
+ * match is considered a token with the same style.
+ *
+ * Context is applied to the last non-whitespace, non-comment token
+ * recognized.
+ *
+ * Shortcut is an optional string of characters, any of which, if the first
+ * character, gurantee that this pattern and only this pattern matches.
+ *
+ * @param {Array} shortcutStylePatterns patterns that always start with
+ * a known character. Must have a shortcut string.
+ * @param {Array} fallthroughStylePatterns patterns that will be tried in
+ * order if the shortcut ones fail. May have shortcuts.
+ *
+ * @return {function (Object)} a
+ * function that takes source code and returns a list of decorations.
+ */
+ function createSimpleLexer(shortcutStylePatterns, fallthroughStylePatterns) {
+ var shortcuts = {};
+ var tokenizer;
+ (function () {
+ var allPatterns = shortcutStylePatterns.concat(fallthroughStylePatterns);
+ var allRegexs = [];
+ var regexKeys = {};
+ for (var i = 0, n = allPatterns.length; i < n; ++i) {
+ var patternParts = allPatterns[i];
+ var shortcutChars = patternParts[3];
+ if (shortcutChars) {
+ for (var c = shortcutChars.length; --c >= 0;) {
+ shortcuts[shortcutChars.charAt(c)] = patternParts;
+ }
+ }
+ var regex = patternParts[1];
+ var k = '' + regex;
+ if (!regexKeys.hasOwnProperty(k)) {
+ allRegexs.push(regex);
+ regexKeys[k] = null;
+ }
+ }
+ allRegexs.push(/[\0-\uffff]/);
+ tokenizer = combinePrefixPatterns(allRegexs);
+ })();
+
+ var nPatterns = fallthroughStylePatterns.length;
+
+ /**
+ * Lexes job.sourceCode and produces an output array job.decorations of
+ * style classes preceded by the position at which they start in
+ * job.sourceCode in order.
+ *
+ * @param {Object} job an object like <pre>{
+ * sourceCode: {string} sourceText plain text,
+ * basePos: {int} position of job.sourceCode in the larger chunk of
+ * sourceCode.
+ * }</pre>
+ */
+ var decorate = function (job) {
+ var sourceCode = job.sourceCode, basePos = job.basePos;
+ /** Even entries are positions in source in ascending order. Odd enties
+ * are style markers (e.g., PR_COMMENT) that run from that position until
+ * the end.
+ * @type {Array.<number|string>}
+ */
+ var decorations = [basePos, PR_PLAIN];
+ var pos = 0; // index into sourceCode
+ var tokens = sourceCode.match(tokenizer) || [];
+ var styleCache = {};
+
+ for (var ti = 0, nTokens = tokens.length; ti < nTokens; ++ti) {
+ var token = tokens[ti];
+ var style = styleCache[token];
+ var match = void 0;
+
+ var isEmbedded;
+ if (typeof style === 'string') {
+ isEmbedded = false;
+ } else {
+ var patternParts = shortcuts[token.charAt(0)];
+ if (patternParts) {
+ match = token.match(patternParts[1]);
+ style = patternParts[0];
+ } else {
+ for (var i = 0; i < nPatterns; ++i) {
+ patternParts = fallthroughStylePatterns[i];
+ match = token.match(patternParts[1]);
+ if (match) {
+ style = patternParts[0];
+ break;
+ }
+ }
+
+ if (!match) { // make sure that we make progress
+ style = PR_PLAIN;
+ }
+ }
+
+ isEmbedded = style.length >= 5 && 'lang-' === style.substring(0, 5);
+ if (isEmbedded && !(match && typeof match[1] === 'string')) {
+ isEmbedded = false;
+ style = PR_SOURCE;
+ }
+
+ if (!isEmbedded) { styleCache[token] = style; }
+ }
+
+ var tokenStart = pos;
+ pos += token.length;
+
+ if (!isEmbedded) {
+ decorations.push(basePos + tokenStart, style);
+ } else { // Treat group 1 as an embedded block of source code.
+ var embeddedSource = match[1];
+ var embeddedSourceStart = token.indexOf(embeddedSource);
+ var embeddedSourceEnd = embeddedSourceStart + embeddedSource.length;
+ if (match[2]) {
+ // If embeddedSource can be blank, then it would match at the
+ // beginning which would cause us to infinitely recurse on the
+ // entire token, so we catch the right context in match[2].
+ embeddedSourceEnd = token.length - match[2].length;
+ embeddedSourceStart = embeddedSourceEnd - embeddedSource.length;
+ }
+ var lang = style.substring(5);
+ // Decorate the left of the embedded source
+ appendDecorations(
+ basePos + tokenStart,
+ token.substring(0, embeddedSourceStart),
+ decorate, decorations);
+ // Decorate the embedded source
+ appendDecorations(
+ basePos + tokenStart + embeddedSourceStart,
+ embeddedSource,
+ langHandlerForExtension(lang, embeddedSource),
+ decorations);
+ // Decorate the right of the embedded section
+ appendDecorations(
+ basePos + tokenStart + embeddedSourceEnd,
+ token.substring(embeddedSourceEnd),
+ decorate, decorations);
+ }
+ }
+ job.decorations = decorations;
+ };
+ return decorate;
+ }
+
+ /** returns a function that produces a list of decorations from source text.
+ *
+ * This code treats ", ', and ` as string delimiters, and \ as a string
+ * escape. It does not recognize perl's qq() style strings.
+ * It has no special handling for double delimiter escapes as in basic, or
+ * the tripled delimiters used in python, but should work on those regardless
+ * although in those cases a single string literal may be broken up into
+ * multiple adjacent string literals.
+ *
+ * It recognizes C, C++, and shell style comments.
+ *
+ * @param {Object} options a set of optional parameters.
+ * @return {function (Object)} a function that examines the source code
+ * in the input job and builds the decoration list.
+ */
+ function sourceDecorator(options) {
+ var shortcutStylePatterns = [], fallthroughStylePatterns = [];
+ if (options['tripleQuotedStrings']) {
+ // '''multi-line-string''', 'single-line-string', and double-quoted
+ shortcutStylePatterns.push(
+ [PR_STRING, /^(?:\'\'\'(?:[^\'\\]|\\[\s\S]|\'{1,2}(?=[^\']))*(?:\'\'\'|$)|\"\"\"(?:[^\"\\]|\\[\s\S]|\"{1,2}(?=[^\"]))*(?:\"\"\"|$)|\'(?:[^\\\']|\\[\s\S])*(?:\'|$)|\"(?:[^\\\"]|\\[\s\S])*(?:\"|$))/,
+ null, '\'"']);
+ } else if (options['multiLineStrings']) {
+ // 'multi-line-string', "multi-line-string"
+ shortcutStylePatterns.push(
+ [PR_STRING, /^(?:\'(?:[^\\\']|\\[\s\S])*(?:\'|$)|\"(?:[^\\\"]|\\[\s\S])*(?:\"|$)|\`(?:[^\\\`]|\\[\s\S])*(?:\`|$))/,
+ null, '\'"`']);
+ } else {
+ // 'single-line-string', "single-line-string"
+ shortcutStylePatterns.push(
+ [PR_STRING,
+ /^(?:\'(?:[^\\\'\r\n]|\\.)*(?:\'|$)|\"(?:[^\\\"\r\n]|\\.)*(?:\"|$))/,
+ null, '"\'']);
+ }
+ if (options['verbatimStrings']) {
+ // verbatim-string-literal production from the C# grammar. See issue 93.
+ fallthroughStylePatterns.push(
+ [PR_STRING, /^@\"(?:[^\"]|\"\")*(?:\"|$)/, null]);
+ }
+ var hc = options['hashComments'];
+ if (hc) {
+ if (options['cStyleComments']) {
+ if (hc > 1) { // multiline hash comments
+ shortcutStylePatterns.push(
+ [PR_COMMENT, /^#(?:##(?:[^#]|#(?!##))*(?:###|$)|.*)/, null, '#']);
+ } else {
+ // Stop C preprocessor declarations at an unclosed open comment
+ shortcutStylePatterns.push(
+ [PR_COMMENT, /^#(?:(?:define|elif|else|endif|error|ifdef|include|ifndef|line|pragma|undef|warning)\b|[^\r\n]*)/,
+ null, '#']);
+ }
+ fallthroughStylePatterns.push(
+ [PR_STRING,
+ /^<(?:(?:(?:\.\.\/)*|\/?)(?:[\w-]+(?:\/[\w-]+)+)?[\w-]+\.h|[a-z]\w*)>/,
+ null]);
+ } else {
+ shortcutStylePatterns.push([PR_COMMENT, /^#[^\r\n]*/, null, '#']);
+ }
+ }
+ if (options['cStyleComments']) {
+ fallthroughStylePatterns.push([PR_COMMENT, /^\/\/[^\r\n]*/, null]);
+ fallthroughStylePatterns.push(
+ [PR_COMMENT, /^\/\*[\s\S]*?(?:\*\/|$)/, null]);
+ }
+ if (options['regexLiterals']) {
+ /**
+ * @const
+ */
+ var REGEX_LITERAL = (
+ // A regular expression literal starts with a slash that is
+ // not followed by * or / so that it is not confused with
+ // comments.
+ '/(?=[^/*])'
+ // and then contains any number of raw characters,
+ + '(?:[^/\\x5B\\x5C]'
+ // escape sequences (\x5C),
+ + '|\\x5C[\\s\\S]'
+ // or non-nesting character sets (\x5B\x5D);
+ + '|\\x5B(?:[^\\x5C\\x5D]|\\x5C[\\s\\S])*(?:\\x5D|$))+'
+ // finally closed by a /.
+ + '/');
+ fallthroughStylePatterns.push(
+ ['lang-regex',
+ new RegExp('^' + REGEXP_PRECEDER_PATTERN + '(' + REGEX_LITERAL + ')')
+ ]);
+ }
+
+ var types = options['types'];
+ if (types) {
+ fallthroughStylePatterns.push([PR_TYPE, types]);
+ }
+
+ var keywords = ("" + options['keywords']).replace(/^ | $/g, '');
+ if (keywords.length) {
+ fallthroughStylePatterns.push(
+ [PR_KEYWORD,
+ new RegExp('^(?:' + keywords.replace(/[\s,]+/g, '|') + ')\\b'),
+ null]);
+ }
+
+ shortcutStylePatterns.push([PR_PLAIN, /^\s+/, null, ' \r\n\t\xA0']);
+ fallthroughStylePatterns.push(
+ // TODO(mikesamuel): recognize non-latin letters and numerals in idents
+ [PR_LITERAL, /^@[a-z_$][a-z_$@0-9]*/i, null],
+ [PR_TYPE, /^(?:[@_]?[A-Z]+[a-z][A-Za-z_$@0-9]*|\w+_t\b)/, null],
+ [PR_PLAIN, /^[a-z_$][a-z_$@0-9]*/i, null],
+ [PR_LITERAL,
+ new RegExp(
+ '^(?:'
+ // A hex number
+ + '0x[a-f0-9]+'
+ // or an octal or decimal number,
+ + '|(?:\\d(?:_\\d+)*\\d*(?:\\.\\d*)?|\\.\\d\\+)'
+ // possibly in scientific notation
+ + '(?:e[+\\-]?\\d+)?'
+ + ')'
+ // with an optional modifier like UL for unsigned long
+ + '[a-z]*', 'i'),
+ null, '0123456789'],
+ // Don't treat escaped quotes in bash as starting strings. See issue 144.
+ [PR_PLAIN, /^\\[\s\S]?/, null],
+ [PR_PUNCTUATION, /^.[^\s\w\.$@\'\"\`\/\#\\]*/, null]);
+
+ return createSimpleLexer(shortcutStylePatterns, fallthroughStylePatterns);
+ }
+
+ var decorateSource = sourceDecorator({
+ 'keywords': ALL_KEYWORDS,
+ 'hashComments': true,
+ 'cStyleComments': true,
+ 'multiLineStrings': true,
+ 'regexLiterals': true
+ });
+
+ /**
+ * Given a DOM subtree, wraps it in a list, and puts each line into its own
+ * list item.
+ *
+ * @param {Node} node modified in place. Its content is pulled into an
+ * HTMLOListElement, and each line is moved into a separate list item.
+ * This requires cloning elements, so the input might not have unique
+ * IDs after numbering.
+ */
+ function numberLines(node, opt_startLineNum) {
+ var nocode = /(?:^|\s)nocode(?:\s|$)/;
+ var lineBreak = /\r\n?|\n/;
+
+ var document = node.ownerDocument;
+
+ var whitespace;
+ if (node.currentStyle) {
+ whitespace = node.currentStyle.whiteSpace;
+ } else if (window.getComputedStyle) {
+ whitespace = document.defaultView.getComputedStyle(node, null)
+ .getPropertyValue('white-space');
+ }
+ // If it's preformatted, then we need to split lines on line breaks
+ // in addition to <BR>s.
+ var isPreformatted = whitespace && 'pre' === whitespace.substring(0, 3);
+
+ var li = document.createElement('LI');
+ while (node.firstChild) {
+ li.appendChild(node.firstChild);
+ }
+ // An array of lines. We split below, so this is initialized to one
+ // un-split line.
+ var listItems = [li];
+
+ function walk(node) {
+ switch (node.nodeType) {
+ case 1: // Element
+ if (nocode.test(node.className)) { break; }
+ if ('BR' === node.nodeName) {
+ breakAfter(node);
+ // Discard the <BR> since it is now flush against a </LI>.
+ if (node.parentNode) {
+ node.parentNode.removeChild(node);
+ }
+ } else {
+ for (var child = node.firstChild; child; child = child.nextSibling) {
+ walk(child);
+ }
+ }
+ break;
+ case 3: case 4: // Text
+ if (isPreformatted) {
+ var text = node.nodeValue;
+ var match = text.match(lineBreak);
+ if (match) {
+ var firstLine = text.substring(0, match.index);
+ node.nodeValue = firstLine;
+ var tail = text.substring(match.index + match[0].length);
+ if (tail) {
+ var parent = node.parentNode;
+ parent.insertBefore(
+ document.createTextNode(tail), node.nextSibling);
+ }
+ breakAfter(node);
+ if (!firstLine) {
+ // Don't leave blank text nodes in the DOM.
+ node.parentNode.removeChild(node);
+ }
+ }
+ }
+ break;
+ }
+ }
+
+ // Split a line after the given node.
+ function breakAfter(lineEndNode) {
+ // If there's nothing to the right, then we can skip ending the line
+ // here, and move root-wards since splitting just before an end-tag
+ // would require us to create a bunch of empty copies.
+ while (!lineEndNode.nextSibling) {
+ lineEndNode = lineEndNode.parentNode;
+ if (!lineEndNode) { return; }
+ }
+
+ function breakLeftOf(limit, copy) {
+ // Clone shallowly if this node needs to be on both sides of the break.
+ var rightSide = copy ? limit.cloneNode(false) : limit;
+ var parent = limit.parentNode;
+ if (parent) {
+ // We clone the parent chain.
+ // This helps us resurrect important styling elements that cross lines.
+ // E.g. in <i>Foo<br>Bar</i>
+ // should be rewritten to <li><i>Foo</i></li><li><i>Bar</i></li>.
+ var parentClone = breakLeftOf(parent, 1);
+ // Move the clone and everything to the right of the original
+ // onto the cloned parent.
+ var next = limit.nextSibling;
+ parentClone.appendChild(rightSide);
+ for (var sibling = next; sibling; sibling = next) {
+ next = sibling.nextSibling;
+ parentClone.appendChild(sibling);
+ }
+ }
+ return rightSide;
+ }
+
+ var copiedListItem = breakLeftOf(lineEndNode.nextSibling, 0);
+
+ // Walk the parent chain until we reach an unattached LI.
+ for (var parent;
+ // Check nodeType since IE invents document fragments.
+ (parent = copiedListItem.parentNode) && parent.nodeType === 1;) {
+ copiedListItem = parent;
+ }
+ // Put it on the list of lines for later processing.
+ listItems.push(copiedListItem);
+ }
+
+ // Split lines while there are lines left to split.
+ for (var i = 0; // Number of lines that have been split so far.
+ i < listItems.length; // length updated by breakAfter calls.
+ ++i) {
+ walk(listItems[i]);
+ }
+
+ // Make sure numeric indices show correctly.
+ if (opt_startLineNum === (opt_startLineNum|0)) {
+ listItems[0].setAttribute('value', opt_startLineNum);
+ }
+
+ var ol = document.createElement('OL');
+ ol.className = 'linenums';
+ var offset = Math.max(0, ((opt_startLineNum - 1 /* zero index */)) | 0) || 0;
+ for (var i = 0, n = listItems.length; i < n; ++i) {
+ li = listItems[i];
+ // Stick a class on the LIs so that stylesheets can
+ // color odd/even rows, or any other row pattern that
+ // is co-prime with 10.
+ li.className = 'L' + ((i + offset) % 10);
+ if (!li.firstChild) {
+ li.appendChild(document.createTextNode('\xA0'));
+ }
+ ol.appendChild(li);
+ }
+
+ node.appendChild(ol);
+ }
+
+ /**
+ * Breaks {@code job.sourceCode} around style boundaries in
+ * {@code job.decorations} and modifies {@code job.sourceNode} in place.
+ * @param {Object} job like <pre>{
+ * sourceCode: {string} source as plain text,
+ * spans: {Array.<number|Node>} alternating span start indices into source
+ * and the text node or element (e.g. {@code <BR>}) corresponding to that
+ * span.
+ * decorations: {Array.<number|string} an array of style classes preceded
+ * by the position at which they start in job.sourceCode in order
+ * }</pre>
+ * @private
+ */
+ function recombineTagsAndDecorations(job) {
+ var isIE = /\bMSIE\b/.test(navigator.userAgent);
+ var newlineRe = /\n/g;
+
+ var source = job.sourceCode;
+ var sourceLength = source.length;
+ // Index into source after the last code-unit recombined.
+ var sourceIndex = 0;
+
+ var spans = job.spans;
+ var nSpans = spans.length;
+ // Index into spans after the last span which ends at or before sourceIndex.
+ var spanIndex = 0;
+
+ var decorations = job.decorations;
+ var nDecorations = decorations.length;
+ // Index into decorations after the last decoration which ends at or before
+ // sourceIndex.
+ var decorationIndex = 0;
+
+ // Remove all zero-length decorations.
+ decorations[nDecorations] = sourceLength;
+ var decPos, i;
+ for (i = decPos = 0; i < nDecorations;) {
+ if (decorations[i] !== decorations[i + 2]) {
+ decorations[decPos++] = decorations[i++];
+ decorations[decPos++] = decorations[i++];
+ } else {
+ i += 2;
+ }
+ }
+ nDecorations = decPos;
+
+ // Simplify decorations.
+ for (i = decPos = 0; i < nDecorations;) {
+ var startPos = decorations[i];
+ // Conflate all adjacent decorations that use the same style.
+ var startDec = decorations[i + 1];
+ var end = i + 2;
+ while (end + 2 <= nDecorations && decorations[end + 1] === startDec) {
+ end += 2;
+ }
+ decorations[decPos++] = startPos;
+ decorations[decPos++] = startDec;
+ i = end;
+ }
+
+ nDecorations = decorations.length = decPos;
+
+ var decoration = null;
+ while (spanIndex < nSpans) {
+ var spanStart = spans[spanIndex];
+ var spanEnd = spans[spanIndex + 2] || sourceLength;
+
+ var decStart = decorations[decorationIndex];
+ var decEnd = decorations[decorationIndex + 2] || sourceLength;
+
+ var end = Math.min(spanEnd, decEnd);
+
+ var textNode = spans[spanIndex + 1];
+ var styledText;
+ if (textNode.nodeType !== 1 // Don't muck with <BR>s or <LI>s
+ // Don't introduce spans around empty text nodes.
+ && (styledText = source.substring(sourceIndex, end))) {
+ // This may seem bizarre, and it is. Emitting LF on IE causes the
+ // code to display with spaces instead of line breaks.
+ // Emitting Windows standard issue linebreaks (CRLF) causes a blank
+ // space to appear at the beginning of every line but the first.
+ // Emitting an old Mac OS 9 line separator makes everything spiffy.
+ if (isIE) { styledText = styledText.replace(newlineRe, '\r'); }
+ textNode.nodeValue = styledText;
+ var document = textNode.ownerDocument;
+ var span = document.createElement('SPAN');
+ span.className = decorations[decorationIndex + 1];
+ var parentNode = textNode.parentNode;
+ parentNode.replaceChild(span, textNode);
+ span.appendChild(textNode);
+ if (sourceIndex < spanEnd) { // Split off a text node.
+ spans[spanIndex + 1] = textNode
+ // TODO: Possibly optimize by using '' if there's no flicker.
+ = document.createTextNode(source.substring(end, spanEnd));
+ parentNode.insertBefore(textNode, span.nextSibling);
+ }
+ }
+
+ sourceIndex = end;
+
+ if (sourceIndex >= spanEnd) {
+ spanIndex += 2;
+ }
+ if (sourceIndex >= decEnd) {
+ decorationIndex += 2;
+ }
+ }
+ }
+
+
+ /** Maps language-specific file extensions to handlers. */
+ var langHandlerRegistry = {};
+ /** Register a language handler for the given file extensions.
+ * @param {function (Object)} handler a function from source code to a list
+ * of decorations. Takes a single argument job which describes the
+ * state of the computation. The single parameter has the form
+ * {@code {
+ * sourceCode: {string} as plain text.
+ * decorations: {Array.<number|string>} an array of style classes
+ * preceded by the position at which they start in
+ * job.sourceCode in order.
+ * The language handler should assigned this field.
+ * basePos: {int} the position of source in the larger source chunk.
+ * All positions in the output decorations array are relative
+ * to the larger source chunk.
+ * } }
+ * @param {Array.<string>} fileExtensions
+ */
+ function registerLangHandler(handler, fileExtensions) {
+ for (var i = fileExtensions.length; --i >= 0;) {
+ var ext = fileExtensions[i];
+ if (!langHandlerRegistry.hasOwnProperty(ext)) {
+ langHandlerRegistry[ext] = handler;
+ } else if (window['console']) {
+ console['warn']('cannot override language handler %s', ext);
+ }
+ }
+ }
+ function langHandlerForExtension(extension, source) {
+ if (!(extension && langHandlerRegistry.hasOwnProperty(extension))) {
+ // Treat it as markup if the first non whitespace character is a < and
+ // the last non-whitespace character is a >.
+ extension = /^\s*</.test(source)
+ ? 'default-markup'
+ : 'default-code';
+ }
+ return langHandlerRegistry[extension];
+ }
+ registerLangHandler(decorateSource, ['default-code']);
+ registerLangHandler(
+ createSimpleLexer(
+ [],
+ [
+ [PR_PLAIN, /^[^<?]+/],
+ [PR_DECLARATION, /^<!\w[^>]*(?:>|$)/],
+ [PR_COMMENT, /^<\!--[\s\S]*?(?:-\->|$)/],
+ // Unescaped content in an unknown language
+ ['lang-', /^<\?([\s\S]+?)(?:\?>|$)/],
+ ['lang-', /^<%([\s\S]+?)(?:%>|$)/],
+ [PR_PUNCTUATION, /^(?:<[%?]|[%?]>)/],
+ ['lang-', /^<xmp\b[^>]*>([\s\S]+?)<\/xmp\b[^>]*>/i],
+ // Unescaped content in javascript. (Or possibly vbscript).
+ ['lang-js', /^<script\b[^>]*>([\s\S]*?)(<\/script\b[^>]*>)/i],
+ // Contains unescaped stylesheet content
+ ['lang-css', /^<style\b[^>]*>([\s\S]*?)(<\/style\b[^>]*>)/i],
+ ['lang-in.tag', /^(<\/?[a-z][^<>]*>)/i]
+ ]),
+ ['default-markup', 'htm', 'html', 'mxml', 'xhtml', 'xml', 'xsl']);
+ registerLangHandler(
+ createSimpleLexer(
+ [
+ [PR_PLAIN, /^[\s]+/, null, ' \t\r\n'],
+ [PR_ATTRIB_VALUE, /^(?:\"[^\"]*\"?|\'[^\']*\'?)/, null, '\"\'']
+ ],
+ [
+ [PR_TAG, /^^<\/?[a-z](?:[\w.:-]*\w)?|\/?>$/i],
+ [PR_ATTRIB_NAME, /^(?!style[\s=]|on)[a-z](?:[\w:-]*\w)?/i],
+ ['lang-uq.val', /^=\s*([^>\'\"\s]*(?:[^>\'\"\s\/]|\/(?=\s)))/],
+ [PR_PUNCTUATION, /^[=<>\/]+/],
+ ['lang-js', /^on\w+\s*=\s*\"([^\"]+)\"/i],
+ ['lang-js', /^on\w+\s*=\s*\'([^\']+)\'/i],
+ ['lang-js', /^on\w+\s*=\s*([^\"\'>\s]+)/i],
+ ['lang-css', /^style\s*=\s*\"([^\"]+)\"/i],
+ ['lang-css', /^style\s*=\s*\'([^\']+)\'/i],
+ ['lang-css', /^style\s*=\s*([^\"\'>\s]+)/i]
+ ]),
+ ['in.tag']);
+ registerLangHandler(
+ createSimpleLexer([], [[PR_ATTRIB_VALUE, /^[\s\S]+/]]), ['uq.val']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': CPP_KEYWORDS,
+ 'hashComments': true,
+ 'cStyleComments': true,
+ 'types': C_TYPES
+ }), ['c', 'cc', 'cpp', 'cxx', 'cyc', 'm']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': 'null,true,false'
+ }), ['json']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': CSHARP_KEYWORDS,
+ 'hashComments': true,
+ 'cStyleComments': true,
+ 'verbatimStrings': true,
+ 'types': C_TYPES
+ }), ['cs']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': JAVA_KEYWORDS,
+ 'cStyleComments': true
+ }), ['java']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': SH_KEYWORDS,
+ 'hashComments': true,
+ 'multiLineStrings': true
+ }), ['bsh', 'csh', 'sh']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': PYTHON_KEYWORDS,
+ 'hashComments': true,
+ 'multiLineStrings': true,
+ 'tripleQuotedStrings': true
+ }), ['cv', 'py']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': PERL_KEYWORDS,
+ 'hashComments': true,
+ 'multiLineStrings': true,
+ 'regexLiterals': true
+ }), ['perl', 'pl', 'pm']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': RUBY_KEYWORDS,
+ 'hashComments': true,
+ 'multiLineStrings': true,
+ 'regexLiterals': true
+ }), ['rb']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': JSCRIPT_KEYWORDS,
+ 'cStyleComments': true,
+ 'regexLiterals': true
+ }), ['js']);
+ registerLangHandler(sourceDecorator({
+ 'keywords': COFFEE_KEYWORDS,
+ 'hashComments': 3, // ### style block comments
+ 'cStyleComments': true,
+ 'multilineStrings': true,
+ 'tripleQuotedStrings': true,
+ 'regexLiterals': true
+ }), ['coffee']);
+ registerLangHandler(createSimpleLexer([], [[PR_STRING, /^[\s\S]+/]]), ['regex']);
+
+ function applyDecorator(job) {
+ var opt_langExtension = job.langExtension;
+
+ try {
+ // Extract tags, and convert the source code to plain text.
+ var sourceAndSpans = extractSourceSpans(job.sourceNode);
+ /** Plain text. @type {string} */
+ var source = sourceAndSpans.sourceCode;
+ job.sourceCode = source;
+ job.spans = sourceAndSpans.spans;
+ job.basePos = 0;
+
+ // Apply the appropriate language handler
+ langHandlerForExtension(opt_langExtension, source)(job);
+
+ // Integrate the decorations and tags back into the source code,
+ // modifying the sourceNode in place.
+ recombineTagsAndDecorations(job);
+ } catch (e) {
+ if ('console' in window) {
+ console['log'](e && e['stack'] ? e['stack'] : e);
+ }
+ }
+ }
+
+ /**
+ * @param sourceCodeHtml {string} The HTML to pretty print.
+ * @param opt_langExtension {string} The language name to use.
+ * Typically, a filename extension like 'cpp' or 'java'.
+ * @param opt_numberLines {number|boolean} True to number lines,
+ * or the 1-indexed number of the first line in sourceCodeHtml.
+ */
+ function prettyPrintOne(sourceCodeHtml, opt_langExtension, opt_numberLines) {
+ var container = document.createElement('PRE');
+ // This could cause images to load and onload listeners to fire.
+ // E.g. <img onerror="alert(1337)" src="nosuchimage.png">.
+ // We assume that the inner HTML is from a trusted source.
+ container.innerHTML = sourceCodeHtml;
+ if (opt_numberLines) {
+ numberLines(container, opt_numberLines);
+ }
+
+ var job = {
+ langExtension: opt_langExtension,
+ numberLines: opt_numberLines,
+ sourceNode: container
+ };
+ applyDecorator(job);
+ return container.innerHTML;
+ }
+
+ function prettyPrint(opt_whenDone) {
+ function byTagName(tn) { return document.getElementsByTagName(tn); }
+ // fetch a list of nodes to rewrite
+ var codeSegments = [byTagName('pre'), byTagName('code'), byTagName('xmp')];
+ var elements = [];
+ for (var i = 0; i < codeSegments.length; ++i) {
+ for (var j = 0, n = codeSegments[i].length; j < n; ++j) {
+ elements.push(codeSegments[i][j]);
+ }
+ }
+ codeSegments = null;
+
+ var clock = Date;
+ if (!clock['now']) {
+ clock = { 'now': function () { return +(new Date); } };
+ }
+
+ // The loop is broken into a series of continuations to make sure that we
+ // don't make the browser unresponsive when rewriting a large page.
+ var k = 0;
+ var prettyPrintingJob;
+
+ var langExtensionRe = /\blang(?:uage)?-([\w.]+)(?!\S)/;
+ var prettyPrintRe = /\bprettyprint\b/;
+
+ function doWork() {
+ var endTime = (window['PR_SHOULD_USE_CONTINUATION'] ?
+ clock['now']() + 250 /* ms */ :
+ Infinity);
+ for (; k < elements.length && clock['now']() < endTime; k++) {
+ var cs = elements[k];
+ var className = cs.className;
+ if (className.indexOf('prettyprint') >= 0) {
+ // If the classes includes a language extensions, use it.
+ // Language extensions can be specified like
+ // <pre class="prettyprint lang-cpp">
+ // the language extension "cpp" is used to find a language handler as
+ // passed to PR.registerLangHandler.
+ // HTML5 recommends that a language be specified using "language-"
+ // as the prefix instead. Google Code Prettify supports both.
+ // http://dev.w3.org/html5/spec-author-view/the-code-element.html
+ var langExtension = className.match(langExtensionRe);
+ // Support <pre class="prettyprint"><code class="language-c">
+ var wrapper;
+ if (!langExtension && (wrapper = childContentWrapper(cs))
+ && "CODE" === wrapper.tagName) {
+ langExtension = wrapper.className.match(langExtensionRe);
+ }
+
+ if (langExtension) {
+ langExtension = langExtension[1];
+ }
+
+ // make sure this is not nested in an already prettified element
+ var nested = false;
+ for (var p = cs.parentNode; p; p = p.parentNode) {
+ if ((p.tagName === 'pre' || p.tagName === 'code' ||
+ p.tagName === 'xmp') &&
+ p.className && p.className.indexOf('prettyprint') >= 0) {
+ nested = true;
+ break;
+ }
+ }
+ if (!nested) {
+ // Look for a class like linenums or linenums:<n> where <n> is the
+ // 1-indexed number of the first line.
+ var lineNums = cs.className.match(/\blinenums\b(?::(\d+))?/);
+ lineNums = lineNums
+ ? lineNums[1] && lineNums[1].length ? +lineNums[1] : true
+ : false;
+ if (lineNums) { numberLines(cs, lineNums); }
+
+ // do the pretty printing
+ prettyPrintingJob = {
+ langExtension: langExtension,
+ sourceNode: cs,
+ numberLines: lineNums
+ };
+ applyDecorator(prettyPrintingJob);
+ }
+ }
+ }
+ if (k < elements.length) {
+ // finish up in a continuation
+ setTimeout(doWork, 250);
+ } else if (opt_whenDone) {
+ opt_whenDone();
+ }
+ }
+
+ doWork();
+ }
+
+ /**
+ * Find all the {@code <pre>} and {@code <code>} tags in the DOM with
+ * {@code class=prettyprint} and prettify them.
+ *
+ * @param {Function?} opt_whenDone if specified, called when the last entry
+ * has been finished.
+ */
+ window['prettyPrintOne'] = prettyPrintOne;
+ /**
+ * Pretty print a chunk of code.
+ *
+ * @param {string} sourceCodeHtml code as html
+ * @return {string} code as html, but prettier
+ */
+ window['prettyPrint'] = prettyPrint;
+ /**
+ * Contains functions for creating and registering new language handlers.
+ * @type {Object}
+ */
+ window['PR'] = {
+ 'createSimpleLexer': createSimpleLexer,
+ 'registerLangHandler': registerLangHandler,
+ 'sourceDecorator': sourceDecorator,
+ 'PR_ATTRIB_NAME': PR_ATTRIB_NAME,
+ 'PR_ATTRIB_VALUE': PR_ATTRIB_VALUE,
+ 'PR_COMMENT': PR_COMMENT,
+ 'PR_DECLARATION': PR_DECLARATION,
+ 'PR_KEYWORD': PR_KEYWORD,
+ 'PR_LITERAL': PR_LITERAL,
+ 'PR_NOCODE': PR_NOCODE,
+ 'PR_PLAIN': PR_PLAIN,
+ 'PR_PUNCTUATION': PR_PUNCTUATION,
+ 'PR_SOURCE': PR_SOURCE,
+ 'PR_STRING': PR_STRING,
+ 'PR_TAG': PR_TAG,
+ 'PR_TYPE': PR_TYPE
+ };
+})();
diff --git a/js/zerobin.js b/js/zerobin.js
index 23329a2..5dc7551 100644
--- a/js/zerobin.js
+++ b/js/zerobin.js
@@ -109,12 +109,15 @@ function displayMessages(key, comments) {
var cleartext = zeroDecipher(key, comments[0].data);
} catch(err) {
$('div#cleartext').hide();
+ $('div#prettymessage').hide();
$('button#clonebutton').hide();
showError('Could not decrypt data (Wrong key ?)');
return;
}
setElementText($('div#cleartext'), cleartext);
+ setElementText($('pre#prettyprint'), cleartext);
urls2links($('div#cleartext')); // Convert URLs to clickable links.
+ prettyPrint();
// Display paste expiration.
if (comments[0].meta.expire_date) $('div#remainingtime').removeClass('foryoureyesonly').text('This document will expire in '+secondsToHuman(comments[0].meta.remaining_time)+'.').show();
@@ -257,8 +260,10 @@ function send_data() {
showStatus('');
$('div#pastelink').html('Your paste is <a href="' + url + '">' + url + '</a>').show();
setElementText($('div#cleartext'), $('textarea#message').val());
+ setElementText($('pre#prettyprint'), $('textarea#message').val());
urls2links($('div#cleartext'));
showStatus('');
+ prettyPrint();
}
else if (data.status==1) {
showError('Could not create paste: '+data.message);
@@ -287,6 +292,7 @@ function stateNewPaste() {
$('div#cleartext').hide();
$('div#message').focus();
$('div#discussion').hide();
+ $('div#prettymessage').hide();
}
/**
@@ -310,7 +316,8 @@ function stateExistingPaste() {
$('button#newbutton').show();
$('div#pastelink').hide();
$('textarea#message').hide();
- $('div#cleartext').show();
+ $('div#cleartext').hide();
+ $('div#prettymessage').show();
}
/**
diff --git a/tpl/page.html b/tpl/page.html
index 06873b0..442f0e0 100644
--- a/tpl/page.html
+++ b/tpl/page.html
@@ -2,11 +2,13 @@
<head>
<title>ZeroBin</title>
<link type="text/css" rel="stylesheet" href="css/zerobin.css?{$VERSION|rawurlencode}#" />
+<link type="text/css" rel="stylesheet" href="css/prettify.css?{$VERSION|rawurlencode}#" />
<script src="js/jquery.js#"></script>
<script src="js/sjcl.js#"></script>
<script src="js/base64.js#"></script>
<script src="js/rawdeflate.js#"></script>
<script src="js/rawinflate.js#"></script>
+<script src="js/prettify.js#"></script>
<script src="js/zerobin.js?{$VERSION|rawurlencode}#"></script>
<!--[if lt IE 10]>
@@ -70,6 +72,9 @@
</div>
</div>
<div id="pastelink" style="display:none;"></div>
+ <div id="prettymessage" style="display:none;">
+ <pre id="prettyprint" class="prettyprint linenums:1"></pre>
+ </div>
<div id="cleartext" style="display:none;"></div>
<textarea id="message" name="message" cols="80" rows="25" style="display:none;"></textarea>
<div id="discussion" style="display:none;">