a

// Copyright (C) 2006 Google Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//      http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.


/**
 * @fileoverview
 * some functions for browser-side pretty printing of code contained in html.
 *
 * 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 decent subset of Perl, but,
 * because of commenting conventions, doesn't work on Smalltalk, Lisp-like, or
 * CAML-like languages.
 *
 * If there's a language not mentioned here, then I don't know it, and don't
 * know whether it works.  If it has a C-like, Bash-like, or XML-like syntax
 * then it should work passably.
 *
 * Usage:
 * 1) include this source file in an html page via
 * <script type="text/javascript" src="/path/to/prettify.js"></script>
 * 2) define style rules.  See the example page for examples.
 * 3) mark the <pre> and <code> tags in your source with class=prettyprint.
 *    You can also use the (html deprecated) <xmp> tag, but the pretty printer
 *    needs to do more substantial DOM manipulations to support that, so some
 *    css styles may not be preserved.
 * 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.
 *
 * Change log:
 * cbeust, 2006/08/22
 *   Java annotations (start with "@") are now captured as literals ("lit")
 */

var PR_keywords = {};
/** initialize the keyword list for our target languages. */
(function () {
  var CPP_KEYWORDS = "abstract bool break case catch char class const " +
    "const_cast continue default delete deprecated dllexport dllimport do " +
    "double dynamic_cast else enum explicit extern false float for friend " +
    "goto if inline int long mutable naked namespace new noinline noreturn " +
    "nothrow novtable operator private property protected public register " +
    "reinterpret_cast return selectany short signed sizeof static " +
    "static_cast struct switch template this thread throw true try typedef " +
    "typeid typename union unsigned using declaration, directive uuid " +
    "virtual void volatile while typeof";
  var CSHARP_KEYWORDS = "as base by byte checked decimal delegate descending " +
    "event finally fixed foreach from group implicit in interface internal " +
    "into is lock null object out override orderby params readonly ref sbyte " +
    "sealed stackalloc string select uint ulong unchecked unsafe ushort var";
  var JAVA_KEYWORDS = "package synchronized boolean implements import throws " +
    "instanceof transient extends final strictfp native super";
  var JSCRIPT_KEYWORDS = "debugger export function with NaN Infinity";
  var PERL_KEYWORDS = "require sub unless until use elsif BEGIN END";
  var PYTHON_KEYWORDS = "and assert def del elif except exec global lambda " +
    "not or pass print raise yield False True None";
  var RUBY_KEYWORDS = "then end begin rescue ensure module when undef next " +
    "redo retry alias defined has_many alias_method has_one belongs_to before_filter " +
    "after_filter"; 
  var SH_KEYWORDS = "done fi test tail head grep sed seq sendmail cat through " +
    "cd sudo time";

  var KEYWORDS = [CPP_KEYWORDS, CSHARP_KEYWORDS, JAVA_KEYWORDS,
                  JSCRIPT_KEYWORDS, PERL_KEYWORDS, PYTHON_KEYWORDS,
                  RUBY_KEYWORDS, SH_KEYWORDS];
  for (var k = 0; k < KEYWORDS.length; k++) {
    var kw = KEYWORDS[k].split(' ');
    for (var i = 0; i < kw.length; i++) {
      if (kw[i]) { PR_keywords[kw[i]] = true; }
    }
  }
}).call(this);

// token style names.  correspond to css classes
/** token style for a string literal */
var PR_STRING = 'str';
/** token style for a keyword */
var PR_KEYWORD = 'kwd';
/** token style for a comment */
var PR_COMMENT = 'com';
/** token style for a type */
var PR_TYPE = 'typ';
/** token style for a literal value.  e.g. 1, null, true. */
var PR_LITERAL = 'lit';
/** token style for a punctuation string. */
var PR_PUNCTUATION = 'pun';
/** token style for a punctuation string. */
var PR_PLAIN = 'pln';

/** token style for an sgml tag. */
var PR_TAG = 'tag';
/** token style for a markup declaration such as a DOCTYPE. */
var PR_DECLARATION = 'dec';
/** token style for embedded source. */
var PR_SOURCE = 'src';
/** token style for an sgml attribute name. */
var PR_ATTRIB_NAME = 'atn';
/** token style for an sgml attribute value. */
var PR_ATTRIB_VALUE = 'atv';

/** the number of characters between tab columns */
var PR_TAB_WIDTH = 8;

function PR_isWordChar(ch) {
  return (ch >= 'a' && ch <= 'z') || (ch >= 'A' && ch <= 'Z');
}

function PR_isIE6() {
  var isIE6 = navigator && navigator.userAgent
      && /\bMSIE 6\./.test(navigator.userAgent);
  PR_isIE6 = function () { return isIE6; }
  return isIE6;
}

/** Splice one array into another.
  * Like the python <code>
  * container[containerPosition:containerPosition + countReplaced] = inserted
  * </code>
  * @param {Array} inserted
  * @param {Array} container modified in place
  * @param {Number} containerPosition
  * @param {Number} countReplaced
  */
function PR_spliceArrayInto(
    inserted, container, containerPosition, countReplaced) {
  inserted.unshift(containerPosition, countReplaced || 0);
  try {
    container.splice.apply(container, inserted);
  } finally {
    inserted.splice(0, 2);
  }
}

/** a set of tokens that can precede a regular expression literal in javascript.
  * 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.
  * @private
  */
var REGEXP_PRECEDER_PATTERN = (function () {
    var preceders = [
        "!", "!=", "!==", "#", "%", "%=", "&", "&&", "&&=",
        "&=", "(", "*", "*=", /* "+", */ "+=", ",", /* "-", */ "-=",
        "->", /*".", "..", "...", handled below */ "/", "/=", ":", "::", ";",
        "<", "<<", "<<=", "<=", "=", "==", "===", ">",
        ">=", ">>", ">>=", ">>>", ">>>=", "?", "@", "[",
        "^", "^=", "^^", "^^=", "{", "|", "|=", "||",
        "||=", "~", "break", "case", "continue", "delete",
        "do", "else", "finally", "instanceof",
        "return", "throw", "try", "typeof"
        ];
    var pattern = '(?:' +
      '(?:(?:^|[^0-9\.])\\.{1,3})|' +  // a dot that's not part of a number
      '(?:(?:^|[^\\+])\\+)|' +  // allow + but not ++
      '(?:(?:^|[^\\-])-)'  // allow - but not --
      ;
    for (var i = 0; i < preceders.length; ++i) {
      var preceder = preceders[i];
      if (PR_isWordChar(preceder.charAt(0))) {
        pattern += '|\\b' + preceder;
      } else {
        pattern += '|' + preceder.replace(/([^=<>:&])/g, '\\$1');
      }
    }
    pattern += '|^)\\s*$';  // matches at end, and matches empty string
    return new RegExp(pattern);
    // 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.
  })();

// Define regexps here so that the interpreter doesn't have to create an object
// each time the function containing them is called.
// The language spec requires a new object created even if you don't access the
// $1 members.
var pr_amp = /&/g;
var pr_lt = /</g;
var pr_gt = />/g;
var pr_quot = /\"/g;
/** like textToHtml but escapes double quotes to be attribute safe. */
function PR_attribToHtml(str) {
  return str.replace(pr_amp, '&amp;')
    .replace(pr_lt, '&lt;')
    .replace(pr_gt, '&gt;')
    .replace(pr_quot, '&quot;');
}

/** escapest html special characters to html. */
function PR_textToHtml(str) {
  return str.replace(pr_amp, '&amp;')
    .replace(pr_lt, '&lt;')
    .replace(pr_gt, '&gt;');
}


var pr_ltEnt = /&lt;/g;
var pr_gtEnt = /&gt;/g;
var pr_aposEnt = /&apos;/g;
var pr_quotEnt = /&quot;/g;
var pr_ampEnt = /&amp;/g;
/** unescapes html to plain text. */
function PR_htmlToText(html) {
  var pos = html.indexOf('&');
  if (pos < 0) { return html; }
  // Handle numeric entities specially.  We can't use functional substitution
  // since that doesn't work in older versions of Safari.
  // These should be rare since most browsers convert them to normal chars.
  for (--pos; (pos = html.indexOf('&#', pos + 1)) >= 0;) {
    var end = html.indexOf(';', pos);
    if (end >= 0) {
      var num = html.substring(pos + 3, end);
      var radix = 10;
      if (num && num.charAt(0) == 'x') {
        num = num.substring(1);
        radix = 16;
      }
      var codePoint = parseInt(num, radix);
      if (!isNaN(codePoint)) {
        html = (html.substring(0, pos) + String.fromCharCode(codePoint) +
                html.substring(end + 1));
      }
    }
  }

  return html.replace(pr_ltEnt, '<')
    .replace(pr_gtEnt, '>')
    .replace(pr_aposEnt, "'")
    .replace(pr_quotEnt, '"')
    .replace(pr_ampEnt, '&');
}

/** is the given node's innerHTML normally unescaped? */
function PR_isRawContent(node) {
  return 'XMP' == node.tagName;
}

var PR_innerHtmlWorks = null;
function PR_getInnerHtml(node) {
  // inner html is hopelessly broken in Safari 2.0.4 when the content is
  // an html description of well formed XML and the containing tag is a PRE
  // tag, so we detect that case and emulate innerHTML.
  if (null === PR_innerHtmlWorks) {
    var testNode = document.createElement('PRE');
    testNode.appendChild(
        document.createTextNode('<!DOCTYPE foo PUBLIC "foo bar">\n<foo />'));
    PR_innerHtmlWorks = !/</.test(testNode.innerHTML);
  }

  if (PR_innerHtmlWorks) {
    var content = node.innerHTML;
    // XMP tags contain unescaped entities so require special handling.
    if (PR_isRawContent(node)) {
      content = PR_textToHtml(content);
    }
    return content;
  }

  var out = [];
  for (var child = node.firstChild; child; child = child.nextSibling) {
    PR_normalizedHtml(child, out);
  }
  return out.join('');
}

/** walks the DOM returning a properly escaped version of innerHTML.
  */
function PR_normalizedHtml(node, out) {
  switch (node.nodeType) {
    case 1:  // an element
      var name = node.tagName.toLowerCase();
      out.push('\074', name);
      for (var i = 0; i < node.attributes.length; ++i) {
        var attr = node.attributes[i];
        if (!attr.specified) { continue; }
        out.push(' ');
        PR_normalizedHtml(attr, out);
      }
      out.push('>');
      for (var child = node.firstChild; child; child = child.nextSibling) {
        PR_normalizedHtml(child, out);
      }
      if (node.firstChild || !/^(?:br|link|img)$/.test(name)) {
        out.push('<\/', name, '>');
      }
      break;
    case 2: // an attribute
      out.push(node.name.toLowerCase(), '="', PR_attribToHtml(node.value), '"');
      break;
    case 3: case 4: // text
      out.push(PR_textToHtml(node.nodeValue));
      break;
  }
}

/** returns a function that expand tabs to spaces.  This function can be fed
  * successive chunks of text, and will maintain its own internal state to
  * keep track of how tabs are expanded.
  * @return {function (plainText : String) : String } a function that takes
  *   plain text and return the text with tabs expanded.
  * @private
  */
function PR_tabExpander(tabWidth) {
  var SPACES = '                ';
  var charInLine = 0;

  return function (plainText) {
    // walk over each character looking for tabs and newlines.
    // On tabs, expand them.  On newlines, reset charInLine.
    // Otherwise increment charInLine
    var out = null;
    var pos = 0;
    for (var i = 0, n = plainText.length; i < n; ++i) {
      var ch = plainText.charAt(i);

      switch (ch) {
        case '\t':
          if (!out) { out = []; }
          out.push(plainText.substring(pos, i));
          // calculate how much space we need in front of this part
          // nSpaces is the amount of padding -- the number of spaces needed to
          // move us to the next column, where columns occur at factors of
          // tabWidth.
          var nSpaces = tabWidth - (charInLine % tabWidth);
          charInLine += nSpaces;
          for (; nSpaces >= 0; nSpaces -= SPACES.length) {
            out.push(SPACES.substring(0, nSpaces));
          }
          pos = i + 1;
          break;
        case '\n':
          charInLine = 0;
          break;
        default:
          ++charInLine;
      }
    }
    if (!out) { return plainText; }
    out.push(plainText.substring(pos));
    return out.join('');
  };
}

// The below pattern matches one of the following
// (1) /[^<]+/ : A run of characters other than '<'
// (2) /<!--.*?-->/: an HTML comment
// (3) /<!\[CDATA\[.*?\]\]>/: a cdata section
// (3) /<\/?[a-zA-Z][^>]*>/ : A probably tag that should not be highlighted
// (4) /</ : A '<' that does not begin a larger chunk.  Treated as 1