* 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, define */
/** @define {boolean} */
var IN_GLOBAL_SCOPE = true;
/**
* 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;
/**
* Pretty print a chunk of code.
* @param {string} sourceCodeHtml The HTML to pretty print.
* @param {string} opt_langExtension The language name to use.
* Typically, a filename extension like 'cpp' or 'java'.
* @param {number|boolean} opt_numberLines True to number lines,
* or the 1-indexed number of the first line in sourceCodeHtml.
* @return {string} code as html, but prettier
*/
var prettyPrintOne;
/**
* Find all the {@code <pre>} and {@code <code>} tags in the DOM with
* {@code class=prettyprint} and prettify them.
*
* @param {Function} opt_whenDone called when prettifying is done.
* @param {HTMLElement|HTMLDocument} opt_root an element or document
* containing all the elements to pretty print.
* Defaults to {@code document.body}.
*/
var prettyPrint;
(function () {
var win = window;
// 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,inline,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,delegate," +
"dynamic_cast,explicit,export,friend,generic,late_check," +
"mutable,namespace,nullptr,property,reinterpret_cast,static_assert," +
"static_cast,template,typeid,typename,using,virtual,where"];
var JAVA_KEYWORDS = [COMMON_KEYWORDS,
"abstract,assert,boolean,byte,extends,final,finally,implements,import," +
"instanceof,interface,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,internal,into,is,let," +
"lock,object,out,override,orderby,params,partial,readonly,ref,sbyte," +
"sealed,stackalloc,string,select,uint,ulong,unchecked,unsafe,ushort," +
"var,virtual,where"];
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," +
"throw,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 RUST_KEYWORDS = [FLOW_CONTROL_KEYWORDS, "as,assert,const,copy,drop," +
"enum,extern,fail,false,fn,impl,let,log,loop,match,mod,move,mut,priv," +
"pub,pure,ref,self,static,struct,true,trait,type,unsafe,use"];
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*)\b/;
// 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 plain text.
* @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';
include("regexpPrecederPatterns.pl");
include("combinePrefixPatterns.js");
include("extractSourceSpans.js");
/**
* 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|e(?:l|nd)if|else|error|ifn?def|include|line|pragma|undef|warning)\b|[^\r\n]*)/,
null, '#']);
}
// #include <stdio.h>
fallthroughStylePatterns.push(
[PR_STRING,
/^<(?:(?:(?:\.\.\/)*|\/?)(?:[\w-]+(?:\/[\w-]+)+)?[\w-]+\.h(?:h|pp|\+\+)?|[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]*?(?:\*\/|$)/, null]);
}
var regexLiterals = options['regexLiterals'];
if (regexLiterals) {
/**
* @const
*/
var regexExcls = regexLiterals > 1
? '' // Multiline regex literals
: '\n\r';
/**
* @const
*/
var regexAny = regexExcls ? '.' : '[\\S\\s]';
/**
* @const
*/
var REGEX_LITERAL = (
// A regular expression literal starts with a slash that is
// not followed by * or / so that it is not confused with
// comments.
'/(?=[^/*' + regexExcls + '])'
// and then contains any number of raw characters,
+ '(?:[^/\\x5B\\x5C' + regexExcls + ']'
// escape sequences (\x5C),
+ '|\\x5C' + regexAny
// or non-nesting character sets (\x5B\x5D);
+ '|\\x5B(?:[^\\x5C\\x5D' + regexExcls + ']'
+ '|\\x5C' + regexAny + ')*(?:\\x5D|$))+'
// finally closed by a /.
+ '/');
fallthroughStylePatterns.push(
['lang-regex',
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']);
var punctuation =
// The Bash man page says
// A word is a sequence of characters considered as a single
// unit by GRUB. Words are separated by metacharacters,
// which are the following plus space, tab, and newline: { }
// | & $ ; < >
// ...
// A word beginning with # causes that word and all remaining
// characters on that line to be ignored.
// which means that only a '#' after /(?:^|[{}|&$;<>\s])/ starts a
// comment but empirically
// $ echo {#}
// {#}
// $ echo \$#
// $#
// $ echo }#
// }#
// so /(?:^|[|&;<>\s])/ is more appropriate.
// http://gcc.gnu.org/onlinedocs/gcc-2.95.3/cpp_1.html#SEC3
// suggests that this definition is compatible with a
// default mode that tries to use a single token definition
// to recognize both bash/python style comments and C
// preprocessor directives.
// This definition of punctuation does not include # in the list of
// follow-on exclusions, so # will not be broken before if preceeded
// by a punctuation character. We could try to exclude # after
// [|&;<>] but that doesn't seem to cause many major problems.
// If that does turn out to be a problem, we should change the below
// when hc is truthy to include # in the run of punctuation characters
// only when not followint [|&;<>].
'^.[^\\s\\w.$@\'"`/\\\\]*';
if (options['regexLiterals']) {
punctuation += '(?!\s*\/)';
}
fallthroughStylePatterns.push(
// TODO(mikesamuel): recognize non-latin letters and numerals in idents
[PR_LITERAL, /^@[a-z_$][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, new RegExp(punctuation), null]);
return createSimpleLexer(shortcutStylePatterns, fallthroughStylePatterns);
}
var decorateSource = sourceDecorator({
'keywords': ALL_KEYWORDS,
'hashComments': true,
'cStyleComments': true,
'multiLineStrings': true,
'regexLiterals': true
});
include("numberLines.js");
include("recombineTagsAndDecorations.js");
/** 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 (win['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
}), ['bash', 'bsh', 'csh', 'sh']);
registerLangHandler(sourceDecorator({
'keywords': PYTHON_KEYWORDS,
'hashComments': true,
'multiLineStrings': true,
'tripleQuotedStrings': true
}), ['cv', 'py', 'python']);
registerLangHandler(sourceDecorator({
'keywords': PERL_KEYWORDS,
'hashComments': true,
'multiLineStrings': true,
'regexLiterals': 2 // multiline regex literals
}), ['perl', 'pl', 'pm']);
registerLangHandler(sourceDecorator({
'keywords': RUBY_KEYWORDS,
'hashComments': true,
'multiLineStrings': true,
'regexLiterals': true
}), ['rb', 'ruby']);
registerLangHandler(sourceDecorator({
'keywords': JSCRIPT_KEYWORDS,
'cStyleComments': true,
'regexLiterals': true
}), ['javascript', 'js']);
registerLangHandler(sourceDecorator({
'keywords': COFFEE_KEYWORDS,
'hashComments': 3, // ### style block comments
'cStyleComments': true,
'multilineStrings': true,
'tripleQuotedStrings': true,
'regexLiterals': true
}), ['coffee']);
registerLangHandler(sourceDecorator({
'keywords': RUST_KEYWORDS,
'cStyleComments': true,
'multilineStrings': true
}), ['rc', 'rs', 'rust']);
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, job.pre);
/** 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 (win['console']) {
console['log'](e && e['stack'] || e);
}
}
}
/**
* Pretty print a chunk of code.
* @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('div');
// 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.
// The pre-tag is required for IE8 which strips newlines from innerHTML
// when it is injected into a <pre> tag.
// http://stackoverflow.com/questions/451486/pre-tag-loses-line-breaks-when-setting-innerhtml-in-ie
// http://stackoverflow.com/questions/195363/inserting-a-newline-into-a-pre-tag-ie-javascript
container.innerHTML = '<pre>' + sourceCodeHtml + '</pre>';
container = container.firstChild;
if (opt_numberLines) {
numberLines(container, opt_numberLines, true);
}
var job = {
langExtension: opt_langExtension,
numberLines: opt_numberLines,
sourceNode: container,
pre: 1
};
applyDecorator(job);
return container.innerHTML;
}
/**
* Find all the {@code <pre>} and {@code <code>} tags in the DOM with
* {@code class=prettyprint} and prettify them.
*
* @param {Function} opt_whenDone called when prettifying is done.
* @param {HTMLElement|HTMLDocument} opt_root an element or document
* containing all the elements to pretty print.
* Defaults to {@code document.body}.
*/
function $prettyPrint(opt_whenDone, opt_root) {
var root = opt_root || document.body;
var doc = root.ownerDocument || document;
function byTagName(tn) { return root.getElementsByTagName(tn); }
// fetch a list of nodes to rewrite
var codeSegments = [byTagName('pre'), byTagName('code'), byTagName('xmp')];
var elements = [];
for (var i = 0; i < codeSegments.length; ++i) {
for (var j = 0, n = codeSegments[i].length; j < n; ++j) {
elements.push(codeSegments[i][j]);
}
}
codeSegments = null;
var clock = Date;
if (!clock['now']) {
clock = { 'now': function () { return +(new Date); } };
}
// The loop is broken into a series of continuations to make sure that we
// don't make the browser unresponsive when rewriting a large page.
var k = 0;
var prettyPrintingJob;
var langExtensionRe = /\blang(?:uage)?-([\w.]+)(?!\S)/;
var prettyPrintRe = /\bprettyprint\b/;
var prettyPrintedRe = /\bprettyprinted\b/;
var preformattedTagNameRe = /pre|xmp/i;
var codeRe = /^code$/i;
var preCodeXmpRe = /^(?:pre|code|xmp)$/i;
var EMPTY = {};
function doWork() {
var endTime = (win['PR_SHOULD_USE_CONTINUATION'] ?
clock['now']() + 250 /* ms */ :
Infinity);
for (; k < elements.length && clock['now']() < endTime; k++) {
var cs = elements[k];
// Look for a preceding comment like
// <?prettify lang="..." linenums="..."?>
var attrs = EMPTY;
{
for (var preceder = cs; (preceder = preceder.previousSibling);) {
var nt = preceder.nodeType;
// <?foo?> is parsed by HTML 5 to a comment node (8)
// like <!--?foo?-->, but in XML is a processing instruction
var value = (nt === 7 || nt === 8) && preceder.nodeValue;
if (value
? !/^\??prettify\b/.test(value)
: (nt !== 3 || /\S/.test(preceder.nodeValue))) {
// Skip over white-space text nodes but not others.
break;
}
if (value) {
attrs = {};
value.replace(
/\b(\w+)=([\w:.%+-]+)/g,
function (_, name, value) { attrs[name] = value; });
break;
}
}
}
var className = cs.className;
if ((attrs !== EMPTY || prettyPrintRe.test(className))
// Don't redo this if we've already done it.
// This allows recalling pretty print to just prettyprint elements
// that have been added to the page since last call.
&& !prettyPrintedRe.test(className)) {
// make sure this is not nested in an already prettified element
var nested = false;
for (var p = cs.parentNode; p; p = p.parentNode) {
var tn = p.tagName;
if (preCodeXmpRe.test(tn)
&& p.className && prettyPrintRe.test(p.className)) {
nested = true;
break;
}
}
if (!nested) {
// Mark done. If we fail to prettyprint for whatever reason,
// we shouldn't try again.
cs.className += ' prettyprinted';
// 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 = attrs['lang'];
if (!langExtension) {
langExtension = className.match(langExtensionRe);
// Support <pre class="prettyprint"><code class="language-c">
var wrapper;
if (!langExtension && (wrapper = childContentWrapper(cs))
&& codeRe.test(wrapper.tagName)) {
langExtension = wrapper.className.match(langExtensionRe);
}
if (langExtension) { langExtension = langExtension[1]; }
}
var preformatted;
if (preformattedTagNameRe.test(cs.tagName)) {
preformatted = 1;
} else {
var currentStyle = cs['currentStyle'];
var defaultView = doc.defaultView;
var whitespace = (
currentStyle
? currentStyle['whiteSpace']
: (defaultView
&& defaultView.getComputedStyle)
? defaultView.getComputedStyle(cs, null)
.getPropertyValue('white-space')
: 0);
preformatted = whitespace
&& 'pre' === whitespace.substring(0, 3);
}
// Look for a class like linenums or linenums:<n> where <n> is the
// 1-indexed number of the first line.
var lineNums = attrs['linenums'];
if (!(lineNums = lineNums === 'true' || +lineNums)) {
lineNums = className.match(/\blinenums\b(?::(\d+))?/);
lineNums =
lineNums
? lineNums[1] && lineNums[1].length
? +lineNums[1] : true
: false;
}
if (lineNums) { numberLines(cs, lineNums, preformatted); }
// do the pretty printing
prettyPrintingJob = {
langExtension: langExtension,
sourceNode: cs,
numberLines: lineNums,
pre: preformatted
};
applyDecorator(prettyPrintingJob);
}
}
}
if (k < elements.length) {
// finish up in a continuation
setTimeout(doWork, 250);
} else if ('function' === typeof opt_whenDone) {
opt_whenDone();
}
}
doWork();
}
/**
* Contains functions for creating and registering new language handlers.
* @type {Object}
*/
var PR = win['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,
'prettyPrintOne':
IN_GLOBAL_SCOPE
? (win['prettyPrintOne'] = $prettyPrintOne)
: (prettyPrintOne = $prettyPrintOne),
'prettyPrint': prettyPrint =
IN_GLOBAL_SCOPE
? (win['prettyPrint'] = $prettyPrint)
: (prettyPrint = $prettyPrint)
};
// Make PR available via the Asynchronous Module Definition (AMD) API.
// Per https://github.com/amdjs/amdjs-api/wiki/AMD:
// The Asynchronous Module Definition (AMD) API specifies a
// mechanism for defining modules such that the module and its
// dependencies can be asynchronously loaded.
// ...
// To allow a clear indicator that a global define function (as
// needed for script src browser loading) conforms to the AMD API,
// any global define function SHOULD have a property called "amd"
// whose value is an object. This helps avoid conflict with any
// other existing JavaScript code that could have defined a define()
// function that does not conform to the AMD API.
if (typeof define === "function" && define['amd']) {
define("google-code-prettify", [], function () {
return PR;
});
}
})();