From aaba58ceb0eb3316361592a441ea6c866adcd4f3 Mon Sep 17 00:00:00 2001
From: Tim Schaub <tim.schaub@gmail.com>
Date: Fri, 10 Apr 2015 18:10:20 -0600
Subject: [PATCH 1/3] Generate index.js when building examples

---
 examples_src/index.html |  2 +-
 tasks/build-examples.js | 81 ++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 80 insertions(+), 3 deletions(-)

diff --git a/examples_src/index.html b/examples_src/index.html
index 696db3930d..43af471197 100644
--- a/examples_src/index.html
+++ b/examples_src/index.html
@@ -48,7 +48,7 @@
       }
     </style>
     <script type="text/javascript" src="Jugl.js"></script>
-    <script type="text/javascript" src="example-list.js"></script>
+    <script type="text/javascript" src="index.js"></script>
     <script type="text/javascript">
       var template, target;
 
diff --git a/tasks/build-examples.js b/tasks/build-examples.js
index 789fa3f6fb..75da2c12de 100644
--- a/tasks/build-examples.js
+++ b/tasks/build-examples.js
@@ -81,14 +81,88 @@ function augmentExamples(files, metalsmith, done) {
   }
 }
 
+/**
+ * Create an inverted index of keywords from examples.  Property names are
+ * lowercased words.  Property values are objects mapping example index to word
+ * count.
+ * @param {Array.<Object>} exampleInfos Array of example info objects.
+ * @return {Object} Word index.
+ */
+function createWordIndex(exampleInfos) {
+  var index = {};
+  var keys = ['shortdesc', 'title', 'tags'];
+  exampleInfos.forEach(function(info, i) {
+    keys.forEach(function(key) {
+      var text = info[key];
+      var words = text ? text.split(/\W+/) : [];
+      words.forEach(function(word) {
+        if (word) {
+          word = word.toLowerCase();
+          var counts = index[word];
+          if (counts) {
+            if (index in counts) {
+              counts[i] += 1;
+            } else {
+              counts[i] = 1;
+            }
+          } else {
+            counts = {};
+            counts[i] = 1;
+            index[word] = counts;
+          }
+        }
+      });
+    });
+  });
+  return index;
+}
+
+/**
+ * A plugin that generates the example index.js file.  This file includes a
+ * list of example metadata objects and a word index used when searching for
+ * examples.
+ * @param {Object} files The file lookup provided by Metalsmith.  Property names
+ *     are file paths relative to the source directory.  The file objects
+ *     include any existing metadata (e.g. from YAML front-matter), the file
+ *     contents, and stats.
+ * @param {Object} metalsmith The metalsmith instance the plugin is being used
+ *     with.
+ * @param {function(Error)} done Called when done (with any error).
+ */
+function createIndex(files, metalsmith, done) {
+  setImmediate(done); // all remaining code is synchronous
+  var exampleInfos = [];
+  for (var filename in files) {
+    var example = files[filename];
+    if (markupRegEx.test(filename)) {
+      exampleInfos.push({
+        link: filename,
+        example: filename,
+        title: example.title,
+        shortdesc: example.shortdesc,
+        tags: example.tags
+      });
+    }
+  }
+  var info = {
+    examples: exampleInfos,
+    index: createWordIndex(exampleInfos)
+  };
+  files['index.js'] = {
+    contents: new Buffer('var info = ' + JSON.stringify(info)),
+    mode: '0644'
+  };
+}
+
 function main(callback) {
-  new Metalsmith('.')
+  var smith = new Metalsmith('.')
       .source(srcDir)
       .destination(destDir)
       .metadata({
         olVersion: pkg.version
       })
       .use(augmentExamples)
+      .use(createIndex)
       .use(templates({
         engine: 'handlebars',
         directory: templatesDir,
@@ -101,12 +175,15 @@ function main(callback) {
       .build(function(err) {
         callback(err);
       });
+  return smith;
 }
 
 if (require.main === module) {
   main(function(err) {
     if (err) {
-      process.stderr.write(err.message + '\n');
+      process.stderr.write(
+          'Building examples failed.  See the full trace below.\n\n' +
+          err.stack + '\n');
       process.exit(1);
     } else {
       process.exit(0);

From 8ddec980758fb60780c0640d6401f5ea15112686 Mon Sep 17 00:00:00 2001
From: Tim Schaub <tim.schaub@gmail.com>
Date: Sun, 12 Apr 2015 14:42:32 -0600
Subject: [PATCH 2/3] Rebuild examples when the sources change

---
 package.json            |   6 +-
 tasks/install.js        |  11 +--
 tasks/parse-examples.js | 163 ----------------------------------------
 tasks/readme.md         |   7 +-
 tasks/serve.js          |  60 +++++++++++----
 5 files changed, 55 insertions(+), 192 deletions(-)
 delete mode 100644 tasks/parse-examples.js

diff --git a/package.json b/package.json
index ff89ab49da..f1889acfa4 100644
--- a/package.json
+++ b/package.json
@@ -46,7 +46,9 @@
   "devDependencies": {
     "clean-css": "2.2.16",
     "coveralls": "2.11.2",
+    "debounce": "^1.0.0",
     "expect.js": "0.3.1",
+    "gaze": "^0.5.1",
     "istanbul": "0.3.13",
     "jquery": "2.1.1",
     "jshint": "2.5.6",
@@ -54,9 +56,9 @@
     "mocha-phantomjs": "3.5.1",
     "phantomjs": "1.9.10",
     "proj4": "2.3.3",
+    "resemblejs": "1.2.0",
     "sinon": "1.10.3",
-    "slimerjs-edge": "0.10.0-pre-2",
-    "resemblejs": "1.2.0"
+    "slimerjs-edge": "0.10.0-pre-2"
   },
   "ext": [
     "rbush"
diff --git a/tasks/install.js b/tasks/install.js
index 52cce80f8f..d3850dc968 100644
--- a/tasks/install.js
+++ b/tasks/install.js
@@ -1,17 +1,12 @@
 var async = require('async');
 
 var buildExt = require('./build-ext');
-var parseExamples = require('./parse-examples');
 
-/**
- * Parse examples and build external modules.
- */
-async.waterfall([
-  parseExamples,
-  buildExt
-], function(err) {
+buildExt(function(err) {
   if (err) {
     process.stderr.write(err + '\n');
     process.exit(1);
+  } else {
+    process.exit(0);
   }
 });
diff --git a/tasks/parse-examples.js b/tasks/parse-examples.js
deleted file mode 100644
index a99eea4d83..0000000000
--- a/tasks/parse-examples.js
+++ /dev/null
@@ -1,163 +0,0 @@
-var fs = require('fs');
-var path = require('path');
-
-var async = require('async');
-var Parser = require('htmlparser2').Parser;
-
-var buildExamples = require('./build-examples');
-
-var exampleDir = path.join(__dirname, '..', 'examples');
-
-
-/**
- * List all .html files in the example directory (excluding index.html).
- * @param {function(Error, Array.<string>)} callback Called with any error or
- *     the list of paths to examples.
- */
-function listExamples(callback) {
-  fs.readdir(exampleDir, function(err, items) {
-    if (err) {
-      return callback(err);
-    }
-    var examplePaths = items.filter(function(item) {
-      return /\.html$/i.test(item) && item !== 'index.html';
-    }).map(function(item) {
-      return path.join(exampleDir, item);
-    });
-    callback(null, examplePaths);
-  });
-}
-
-
-/**
- * Parse info from examples.
- * @param {Array.<string>} examplePaths Paths to examples.
- * @param {function(Error, Array.<Object>)} callback Called with any error or
- *     the list of example info objects.
- */
-function parseExamples(examplePaths, callback) {
-  async.map(examplePaths, function(examplePath, next) {
-    fs.readFile(examplePath, function(err, data) {
-      if (err) {
-        return next(err);
-      }
-      var name = path.basename(examplePath);
-      var info = {
-        link: name,
-        example: name,
-        title: '',
-        shortdesc: '',
-        tags: ''
-      };
-      var key;
-      var openTag;
-      var parser = new Parser({
-        onopentag: function(tag, attrs) {
-          if (attrs.id in info) {
-            key = attrs.id;
-            openTag = tag;
-          }
-        },
-        ontext: function(text) {
-          if (key) {
-            info[key] += text.replace(/\n/g, '').trim() + ' ';
-          }
-        },
-        onclosetag: function(tag) {
-          if (tag === openTag) {
-            info[key] = info[key].trim();
-            key = undefined;
-            openTag = undefined;
-          }
-        },
-        onerror: function(err2) {
-          var message = 'Trouble parsing ' + examplePath + '\n' + err2.message;
-          next(new Error(message));
-        }
-      });
-      parser.write(data.toString('utf8'));
-      parser.end();
-      next(null, info);
-    });
-  }, callback);
-}
-
-
-/**
- * Create an inverted index of keywords from examples.  Property names are
- * lowercased words.  Property values are objects mapping example index to word
- * count.
- * @param {Array.<Object>} exampleInfos Array of example info objects.
- * @return {Object} Word index.
- */
-function createWordIndex(exampleInfos) {
-  var index = {};
-  var keys = ['shortdesc', 'title', 'tags'];
-  exampleInfos.forEach(function(info, i) {
-    keys.forEach(function(key) {
-      var text = info[key];
-      var words = text ? text.split(/\W+/) : [];
-      words.forEach(function(word) {
-        if (word) {
-          word = word.toLowerCase();
-          var counts = index[word];
-          if (counts) {
-            if (index in counts) {
-              counts[i] += 1;
-            } else {
-              counts[i] = 1;
-            }
-          } else {
-            counts = {};
-            counts[i] = 1;
-            index[word] = counts;
-          }
-        }
-      });
-    });
-  });
-  return index;
-}
-
-
-/**
- * Write the example-list.js file with example info and word index.
- * @param {Array.<Object>} exampleInfos Array of example info objects.
- * @param {function(Error)} callback Called with any error.
- */
-function writeExampleList(exampleInfos, callback) {
-  var info = {
-    examples: exampleInfos,
-    index: createWordIndex(exampleInfos)
-  };
-  var indexPath = path.join(exampleDir, 'example-list.js');
-  var str = 'var info = ' + JSON.stringify(info);
-  fs.writeFile(indexPath, str, callback);
-}
-
-
-/**
- * List examples, parse them, and write example list.
- * @param {function(Error)} callback Called with any error.
- */
-function main(callback) {
-  async.waterfall([
-    buildExamples,
-    listExamples,
-    parseExamples,
-    writeExampleList
-  ], callback);
-}
-
-if (require.main === module) {
-  main(function(err) {
-    if (err) {
-      process.stderr.write(err.message + '\n');
-      process.exit(1);
-    } else {
-      process.exit(0);
-    }
-  });
-}
-
-module.exports = main;
diff --git a/tasks/readme.md b/tasks/readme.md
index 6bfc5507d8..aefbaedd63 100644
--- a/tasks/readme.md
+++ b/tasks/readme.md
@@ -107,13 +107,10 @@ This is useful for projects that use the Closure Compiler to build, but want to
 Called internally to parse the library for annotations and write out a `build/info.json` file.
 
 
-## `parse-examples.js`
-
-Called after install to generate an example index.  After new examples are added, run `node tasks/parse-examples.js` to regenerate the example index.
-
 ## `build-examples.js`
 
-Called internally by `parse-examples.js` to build the examples from templates.
+Builds examples and the example index.
+
 
 ## `serve.js`
 
diff --git a/tasks/serve.js b/tasks/serve.js
index dfed1e5e8d..c301c9a1b3 100644
--- a/tasks/serve.js
+++ b/tasks/serve.js
@@ -7,9 +7,12 @@
 var path = require('path');
 var url = require('url');
 
+var Gaze = require('gaze').Gaze;
 var closure = require('closure-util');
+var debounce = require('debounce');
 var nomnom = require('nomnom');
 
+var buildExamples = require('./build-examples');
 var log = closure.log;
 
 
@@ -68,6 +71,26 @@ var createServer = exports.createServer = function(callback) {
   });
 };
 
+/**
+ * Build the examples and exit on any error.
+ * @param {Function=} opt_callback Called when done building examples.
+ */
+function buildExamplesOrFatal(opt_callback) {
+  log.info('serve', 'Building examples.');
+  buildExamples(function(err) {
+    if (err) {
+      log.error('serve', 'Building examples failed.');
+      log.error('serve', err.message);
+      log.error('serve', 'Use "verbose" logging to see the full stack trace.');
+      log.verbose('serve', err.stack);
+      process.exit(1);
+    }
+    log.verbose('serve', 'Done building examples.');
+    if (opt_callback) {
+      opt_callback();
+    }
+  });
+}
 
 /**
  * If running this module directly start the server.
@@ -92,21 +115,30 @@ if (require.main === module) {
   /** @type {string} */
   log.level = options.loglevel;
 
-  log.info('serve', 'Parsing dependencies ...');
-  createServer(function(err, server) {
-    if (err) {
-      log.error('serve', 'Parsing failed');
-      log.error('serve', err.message);
-      process.exit(1);
-    }
-    server.listen(options.port, function() {
-      log.info('serve', 'Listening on http://localhost:' +
-          options.port + '/ (Ctrl+C to stop)');
-    });
-    server.on('error', function(err) {
-      log.error('serve', 'Server failed to start: ' + err.message);
-      process.exit(1);
+  buildExamplesOrFatal(function() {
+    log.info('serve', 'Parsing dependencies.');
+    createServer(function(err, server) {
+      if (err) {
+        log.error('serve', 'Parsing failed');
+        log.error('serve', err.message);
+        process.exit(1);
+      }
+      server.listen(options.port, function() {
+        log.info('serve', 'Listening on http://localhost:' +
+            options.port + '/ (Ctrl+C to stop)');
+      });
+      server.on('error', function(err) {
+        log.error('serve', 'Server failed to start: ' + err.message);
+        process.exit(1);
+      });
     });
 
+    var gaze = new Gaze('examples_src/**/*');
+    var debouncedBuild = debounce(buildExamplesOrFatal, 250);
+    gaze.on('all', function(event, filepath) {
+      log.verbose('serve', 'Watch event: ' + event + ' ' + filepath);
+      debouncedBuild();
+    });
   });
+
 }

From a85de86ec28deb139ccf94f99c201d17b0a88087 Mon Sep 17 00:00:00 2001
From: Tim Schaub <tim.schaub@gmail.com>
Date: Sun, 12 Apr 2015 15:48:46 -0600
Subject: [PATCH 3/3] Make build.py call tasks/build-examples.js

---
 .npmignore           |    3 -
 bin/BeautifulSoup.py | 1767 ------------------------------------------
 bin/exampleparser.py |  269 -------
 build.py             |   72 +-
 4 files changed, 32 insertions(+), 2079 deletions(-)
 delete mode 100644 bin/BeautifulSoup.py
 delete mode 100755 bin/exampleparser.py

diff --git a/.npmignore b/.npmignore
index cdb35d22d6..864fa4bc73 100644
--- a/.npmignore
+++ b/.npmignore
@@ -1,5 +1,2 @@
 *.pyc
 /build/
-/examples/*.html.png
-/examples/example-list.js
-/examples/example-list.xml
diff --git a/bin/BeautifulSoup.py b/bin/BeautifulSoup.py
deleted file mode 100644
index 5c7d31dddd..0000000000
--- a/bin/BeautifulSoup.py
+++ /dev/null
@@ -1,1767 +0,0 @@
-"""Beautiful Soup
-Elixir and Tonic
-"The Screen-Scraper's Friend"
-http://www.crummy.com/software/BeautifulSoup/
-
-Beautiful Soup parses a (possibly invalid) XML or HTML document into a
-tree representation. It provides methods and Pythonic idioms that make
-it easy to navigate, search, and modify the tree.
-
-A well-formed XML/HTML document yields a well-formed data
-structure. An ill-formed XML/HTML document yields a correspondingly
-ill-formed data structure. If your document is only locally
-well-formed, you can use this library to find and process the
-well-formed part of it. The BeautifulSoup class 
-
-Beautiful Soup works with Python 2.2 and up. It has no external
-dependencies, but you'll have more success at converting data to UTF-8
-if you also install these three packages:
-
-* chardet, for auto-detecting character encodings
-  http://chardet.feedparser.org/
-* cjkcodecs and iconv_codec, which add more encodings to the ones supported
-  by stock Python.
-  http://cjkpython.i18n.org/
-
-Beautiful Soup defines classes for two main parsing strategies:
-    
- * BeautifulStoneSoup, for parsing XML, SGML, or your domain-specific
-   language that kind of looks like XML.
-
- * BeautifulSoup, for parsing run-of-the-mill HTML code, be it valid
-   or invalid. This class has web browser-like heuristics for
-   obtaining a sensible parse tree in the face of common HTML errors.
-
-Beautiful Soup also defines a class (UnicodeDammit) for autodetecting
-the encoding of an HTML or XML document, and converting it to
-Unicode. Much of this code is taken from Mark Pilgrim's Universal Feed Parser.
-
-For more than you ever wanted to know about Beautiful Soup, see the
-documentation:
-http://www.crummy.com/software/BeautifulSoup/documentation.html
-
-"""
-from __future__ import generators
-
-__author__ = "Leonard Richardson (leonardr@segfault.org)"
-__version__ = "3.0.4"
-__copyright__ = "Copyright (c) 2004-2007 Leonard Richardson"
-__license__ = "PSF"
-
-from sgmllib import SGMLParser, SGMLParseError
-import codecs
-import types
-import re
-import sgmllib
-try:
-  from htmlentitydefs import name2codepoint
-except ImportError:
-  name2codepoint = {}
-
-#This hack makes Beautiful Soup able to parse XML with namespaces
-sgmllib.tagfind = re.compile('[a-zA-Z][-_.:a-zA-Z0-9]*')
-
-DEFAULT_OUTPUT_ENCODING = "utf-8"
-
-# First, the classes that represent markup elements.
-
-class PageElement:
-    """Contains the navigational information for some part of the page
-    (either a tag or a piece of text)"""
-
-    def setup(self, parent=None, previous=None):
-        """Sets up the initial relations between this element and
-        other elements."""        
-        self.parent = parent
-        self.previous = previous
-        self.next = None
-        self.previousSibling = None
-        self.nextSibling = None
-        if self.parent and self.parent.contents:
-            self.previousSibling = self.parent.contents[-1]
-            self.previousSibling.nextSibling = self
-
-    def replaceWith(self, replaceWith):        
-        oldParent = self.parent
-        myIndex = self.parent.contents.index(self)
-        if hasattr(replaceWith, 'parent') and replaceWith.parent == self.parent:
-            # We're replacing this element with one of its siblings.
-            index = self.parent.contents.index(replaceWith)
-            if index and index < myIndex:
-                # Furthermore, it comes before this element. That
-                # means that when we extract it, the index of this
-                # element will change.
-                myIndex = myIndex - 1
-        self.extract()        
-        oldParent.insert(myIndex, replaceWith)
-        
-    def extract(self):
-        """Destructively rips this element out of the tree."""        
-        if self.parent:
-            try:
-                self.parent.contents.remove(self)
-            except ValueError:
-                pass
-
-        #Find the two elements that would be next to each other if
-        #this element (and any children) hadn't been parsed. Connect
-        #the two.        
-        lastChild = self._lastRecursiveChild()
-        nextElement = lastChild.next
-
-        if self.previous:
-            self.previous.next = nextElement
-        if nextElement:
-            nextElement.previous = self.previous
-        self.previous = None
-        lastChild.next = None
-
-        self.parent = None        
-        if self.previousSibling:
-            self.previousSibling.nextSibling = self.nextSibling
-        if self.nextSibling:
-            self.nextSibling.previousSibling = self.previousSibling
-        self.previousSibling = self.nextSibling = None       
-
-    def _lastRecursiveChild(self):
-        "Finds the last element beneath this object to be parsed."
-        lastChild = self
-        while hasattr(lastChild, 'contents') and lastChild.contents:
-            lastChild = lastChild.contents[-1]
-        return lastChild
-
-    def insert(self, position, newChild):
-        if (isinstance(newChild, basestring)
-            or isinstance(newChild, unicode)) \
-            and not isinstance(newChild, NavigableString):
-            newChild = NavigableString(newChild)        
-
-        position =  min(position, len(self.contents))
-        if hasattr(newChild, 'parent') and newChild.parent != None:
-            # We're 'inserting' an element that's already one
-            # of this object's children. 
-            if newChild.parent == self:
-                index = self.find(newChild)
-                if index and index < position:
-                    # Furthermore we're moving it further down the
-                    # list of this object's children. That means that
-                    # when we extract this element, our target index
-                    # will jump down one.
-                    position = position - 1
-            newChild.extract()
-            
-        newChild.parent = self
-        previousChild = None
-        if position == 0:
-            newChild.previousSibling = None
-            newChild.previous = self
-        else:
-            previousChild = self.contents[position-1]
-            newChild.previousSibling = previousChild
-            newChild.previousSibling.nextSibling = newChild
-            newChild.previous = previousChild._lastRecursiveChild()
-        if newChild.previous:
-            newChild.previous.next = newChild        
-
-        newChildsLastElement = newChild._lastRecursiveChild()
-
-        if position >= len(self.contents):
-            newChild.nextSibling = None
-            
-            parent = self
-            parentsNextSibling = None
-            while not parentsNextSibling:
-                parentsNextSibling = parent.nextSibling
-                parent = parent.parent
-                if not parent: # This is the last element in the document.
-                    break
-            if parentsNextSibling:
-                newChildsLastElement.next = parentsNextSibling
-            else:
-                newChildsLastElement.next = None
-        else:
-            nextChild = self.contents[position]            
-            newChild.nextSibling = nextChild            
-            if newChild.nextSibling:
-                newChild.nextSibling.previousSibling = newChild
-            newChildsLastElement.next = nextChild
-
-        if newChildsLastElement.next:
-            newChildsLastElement.next.previous = newChildsLastElement
-        self.contents.insert(position, newChild)
-
-    def findNext(self, name=None, attrs={}, text=None, **kwargs):
-        """Returns the first item that matches the given criteria and
-        appears after this Tag in the document."""
-        return self._findOne(self.findAllNext, name, attrs, text, **kwargs)
-
-    def findAllNext(self, name=None, attrs={}, text=None, limit=None,
-                    **kwargs):
-        """Returns all items that match the given criteria and appear
-        before after Tag in the document."""
-        return self._findAll(name, attrs, text, limit, self.nextGenerator)
-
-    def findNextSibling(self, name=None, attrs={}, text=None, **kwargs):
-        """Returns the closest sibling to this Tag that matches the
-        given criteria and appears after this Tag in the document."""
-        return self._findOne(self.findNextSiblings, name, attrs, text,
-                             **kwargs)
-
-    def findNextSiblings(self, name=None, attrs={}, text=None, limit=None,
-                         **kwargs):
-        """Returns the siblings of this Tag that match the given
-        criteria and appear after this Tag in the document."""
-        return self._findAll(name, attrs, text, limit,
-                             self.nextSiblingGenerator, **kwargs)
-    fetchNextSiblings = findNextSiblings # Compatibility with pre-3.x
-
-    def findPrevious(self, name=None, attrs={}, text=None, **kwargs):
-        """Returns the first item that matches the given criteria and
-        appears before this Tag in the document."""
-        return self._findOne(self.findAllPrevious, name, attrs, text, **kwargs)
-
-    def findAllPrevious(self, name=None, attrs={}, text=None, limit=None,
-                        **kwargs):
-        """Returns all items that match the given criteria and appear
-        before this Tag in the document."""
-        return self._findAll(name, attrs, text, limit, self.previousGenerator,
-                           **kwargs)
-    fetchPrevious = findAllPrevious # Compatibility with pre-3.x
-
-    def findPreviousSibling(self, name=None, attrs={}, text=None, **kwargs):
-        """Returns the closest sibling to this Tag that matches the
-        given criteria and appears before this Tag in the document."""
-        return self._findOne(self.findPreviousSiblings, name, attrs, text,
-                             **kwargs)
-
-    def findPreviousSiblings(self, name=None, attrs={}, text=None,
-                             limit=None, **kwargs):
-        """Returns the siblings of this Tag that match the given
-        criteria and appear before this Tag in the document."""
-        return self._findAll(name, attrs, text, limit,
-                             self.previousSiblingGenerator, **kwargs)
-    fetchPreviousSiblings = findPreviousSiblings # Compatibility with pre-3.x
-
-    def findParent(self, name=None, attrs={}, **kwargs):
-        """Returns the closest parent of this Tag that matches the given
-        criteria."""
-        # NOTE: We can't use _findOne because findParents takes a different
-        # set of arguments.
-        r = None
-        l = self.findParents(name, attrs, 1)
-        if l:
-            r = l[0]
-        return r
-
-    def findParents(self, name=None, attrs={}, limit=None, **kwargs):
-        """Returns the parents of this Tag that match the given
-        criteria."""
-
-        return self._findAll(name, attrs, None, limit, self.parentGenerator,
-                             **kwargs)
-    fetchParents = findParents # Compatibility with pre-3.x
-
-    #These methods do the real heavy lifting.
-
-    def _findOne(self, method, name, attrs, text, **kwargs):
-        r = None
-        l = method(name, attrs, text, 1, **kwargs)
-        if l:
-            r = l[0]
-        return r
-    
-    def _findAll(self, name, attrs, text, limit, generator, **kwargs):
-        "Iterates over a generator looking for things that match."
-
-        if isinstance(name, SoupStrainer):
-            strainer = name
-        else:
-            # Build a SoupStrainer
-            strainer = SoupStrainer(name, attrs, text, **kwargs)
-        results = ResultSet(strainer)
-        g = generator()
-        while True:
-            try:
-                i = g.next()
-            except StopIteration:
-                break
-            if i:
-                found = strainer.search(i)
-                if found:
-                    results.append(found)
-                    if limit and len(results) >= limit:
-                        break
-        return results
-
-    #These Generators can be used to navigate starting from both
-    #NavigableStrings and Tags.                
-    def nextGenerator(self):
-        i = self
-        while i:
-            i = i.next
-            yield i
-
-    def nextSiblingGenerator(self):
-        i = self
-        while i:
-            i = i.nextSibling
-            yield i
-
-    def previousGenerator(self):
-        i = self
-        while i:
-            i = i.previous
-            yield i
-
-    def previousSiblingGenerator(self):
-        i = self
-        while i:
-            i = i.previousSibling
-            yield i
-
-    def parentGenerator(self):
-        i = self
-        while i:
-            i = i.parent
-            yield i
-
-    # Utility methods
-    def substituteEncoding(self, str, encoding=None):
-        encoding = encoding or "utf-8"
-        return str.replace("%SOUP-ENCODING%", encoding)    
-
-    def toEncoding(self, s, encoding=None):
-        """Encodes an object to a string in some encoding, or to Unicode.
-        ."""
-        if isinstance(s, unicode):
-            if encoding:
-                s = s.encode(encoding)
-        elif isinstance(s, str):
-            if encoding:
-                s = s.encode(encoding)
-            else:
-                s = unicode(s)
-        else:
-            if encoding:
-                s  = self.toEncoding(str(s), encoding)
-            else:
-                s = unicode(s)
-        return s
-
-class NavigableString(unicode, PageElement):
-
-    def __getattr__(self, attr):
-        """text.string gives you text. This is for backwards
-        compatibility for Navigable*String, but for CData* it lets you
-        get the string without the CData wrapper."""
-        if attr == 'string':
-            return self
-        else:
-            raise AttributeError, "'%s' object has no attribute '%s'" % (self.__class__.__name__, attr)
-
-    def __unicode__(self):
-        return self.__str__(None)
-
-    def __str__(self, encoding=DEFAULT_OUTPUT_ENCODING):
-        if encoding:
-            return self.encode(encoding)
-        else:
-            return self
-        
-class CData(NavigableString):
-
-    def __str__(self, encoding=DEFAULT_OUTPUT_ENCODING):
-        return "<![CDATA[%s]]>" % NavigableString.__str__(self, encoding)
-
-class ProcessingInstruction(NavigableString):
-    def __str__(self, encoding=DEFAULT_OUTPUT_ENCODING):
-        output = self
-        if "%SOUP-ENCODING%" in output:
-            output = self.substituteEncoding(output, encoding)
-        return "<?%s?>" % self.toEncoding(output, encoding)
-
-class Comment(NavigableString):
-    def __str__(self, encoding=DEFAULT_OUTPUT_ENCODING):
-        return "<!--%s-->" % NavigableString.__str__(self, encoding)    
-
-class Declaration(NavigableString):
-    def __str__(self, encoding=DEFAULT_OUTPUT_ENCODING):
-        return "<!%s>" % NavigableString.__str__(self, encoding)        
-
-class Tag(PageElement):
-
-    """Represents a found HTML tag with its attributes and contents."""
-
-    XML_SPECIAL_CHARS_TO_ENTITIES = { "'" : "squot",
-                                      '"' : "quote",
-                                      "&" : "amp",
-                                      "<" : "lt",
-                                      ">" : "gt" }
-
-    def __init__(self, parser, name, attrs=None, parent=None,
-                 previous=None):
-        "Basic constructor."
-
-        # We don't actually store the parser object: that lets extracted
-        # chunks be garbage-collected
-        self.parserClass = parser.__class__
-        self.isSelfClosing = parser.isSelfClosingTag(name)
-        self.name = name
-        if attrs == None:
-            attrs = []
-        self.attrs = attrs
-        self.contents = []
-        self.setup(parent, previous)
-        self.hidden = False
-        self.containsSubstitutions = False
-
-    def get(self, key, default=None):
-        """Returns the value of the 'key' attribute for the tag, or
-        the value given for 'default' if it doesn't have that
-        attribute."""
-        return self._getAttrMap().get(key, default)    
-
-    def has_key(self, key):
-        return self._getAttrMap().has_key(key)
-
-    def __getitem__(self, key):
-        """tag[key] returns the value of the 'key' attribute for the tag,
-        and throws an exception if it's not there."""
-        return self._getAttrMap()[key]
-
-    def __iter__(self):
-        "Iterating over a tag iterates over its contents."
-        return iter(self.contents)
-
-    def __len__(self):
-        "The length of a tag is the length of its list of contents."
-        return len(self.contents)
-
-    def __contains__(self, x):
-        return x in self.contents
-
-    def __nonzero__(self):
-        "A tag is non-None even if it has no contents."
-        return True
-
-    def __setitem__(self, key, value):        
-        """Setting tag[key] sets the value of the 'key' attribute for the
-        tag."""
-        self._getAttrMap()
-        self.attrMap[key] = value
-        found = False
-        for i in range(0, len(self.attrs)):
-            if self.attrs[i][0] == key:
-                self.attrs[i] = (key, value)
-                found = True
-        if not found:
-            self.attrs.append((key, value))
-        self._getAttrMap()[key] = value
-
-    def __delitem__(self, key):
-        "Deleting tag[key] deletes all 'key' attributes for the tag."
-        for item in self.attrs:
-            if item[0] == key:
-                self.attrs.remove(item)
-                #We don't break because bad HTML can define the same
-                #attribute multiple times.
-            self._getAttrMap()
-            if self.attrMap.has_key(key):
-                del self.attrMap[key]
-
-    def __call__(self, *args, **kwargs):
-        """Calling a tag like a function is the same as calling its
-        findAll() method. Eg. tag('a') returns a list of all the A tags
-        found within this tag."""
-        return apply(self.findAll, args, kwargs)
-
-    def __getattr__(self, tag):
-        #print "Getattr %s.%s" % (self.__class__, tag)
-        if len(tag) > 3 and tag.rfind('Tag') == len(tag)-3:
-            return self.find(tag[:-3])
-        elif tag.find('__') != 0:
-            return self.find(tag)
-
-    def __eq__(self, other):
-        """Returns true iff this tag has the same name, the same attributes,
-        and the same contents (recursively) as the given tag.
-
-        NOTE: right now this will return false if two tags have the
-        same attributes in a different order. Should this be fixed?"""
-        if not hasattr(other, 'name') or not hasattr(other, 'attrs') or not hasattr(other, 'contents') or self.name != other.name or self.attrs != other.attrs or len(self) != len(other):
-            return False
-        for i in range(0, len(self.contents)):
-            if self.contents[i] != other.contents[i]:
-                return False
-        return True
-
-    def __ne__(self, other):
-        """Returns true iff this tag is not identical to the other tag,
-        as defined in __eq__."""
-        return not self == other
-
-    def __repr__(self, encoding=DEFAULT_OUTPUT_ENCODING):
-        """Renders this tag as a string."""
-        return self.__str__(encoding)
-
-    def __unicode__(self):
-        return self.__str__(None)
-
-    def __str__(self, encoding=DEFAULT_OUTPUT_ENCODING,
-                prettyPrint=False, indentLevel=0):
-        """Returns a string or Unicode representation of this tag and
-        its contents. To get Unicode, pass None for encoding.
-
-        NOTE: since Python's HTML parser consumes whitespace, this
-        method is not certain to reproduce the whitespace present in
-        the original string."""
-
-        encodedName = self.toEncoding(self.name, encoding)
-
-        attrs = []
-        if self.attrs:
-            for key, val in self.attrs:
-                fmt = '%s="%s"'
-                if isString(val):                    
-                    if self.containsSubstitutions and '%SOUP-ENCODING%' in val:
-                        val = self.substituteEncoding(val, encoding)
-
-                    # The attribute value either:
-                    #
-                    # * Contains no embedded double quotes or single quotes.
-                    #   No problem: we enclose it in double quotes.
-                    # * Contains embedded single quotes. No problem:
-                    #   double quotes work here too.
-                    # * Contains embedded double quotes. No problem:
-                    #   we enclose it in single quotes.
-                    # * Embeds both single _and_ double quotes. This
-                    #   can't happen naturally, but it can happen if
-                    #   you modify an attribute value after parsing
-                    #   the document. Now we have a bit of a
-                    #   problem. We solve it by enclosing the
-                    #   attribute in single quotes, and escaping any
-                    #   embedded single quotes to XML entities.
-                    if '"' in val:
-                        fmt = "%s='%s'"
-                        # This can't happen naturally, but it can happen
-                        # if you modify an attribute value after parsing.
-                        if "'" in val:
-                            val = val.replace("'", "&squot;")
-
-                    # Now we're okay w/r/t quotes. But the attribute
-                    # value might also contain angle brackets, or
-                    # ampersands that aren't part of entities. We need
-                    # to escape those to XML entities too.
-                    val = re.sub("([<>]|&(?![^\s]+;))",
-                                 lambda x: "&" + self.XML_SPECIAL_CHARS_TO_ENTITIES[x.group(0)[0]] + ";",
-                                 val)
-                                      
-                attrs.append(fmt % (self.toEncoding(key, encoding),
-                                    self.toEncoding(val, encoding)))
-        close = ''
-        closeTag = ''
-        if self.isSelfClosing:
-            close = ' /'
-        else:
-            closeTag = '</%s>' % encodedName
-
-        indentTag, indentContents = 0, 0
-        if prettyPrint:
-            indentTag = indentLevel
-            space = (' ' * (indentTag-1))
-            indentContents = indentTag + 1
-        contents = self.renderContents(encoding, prettyPrint, indentContents)
-        if self.hidden:
-            s = contents
-        else:
-            s = []
-            attributeString = ''
-            if attrs:
-                attributeString = ' ' + ' '.join(attrs)            
-            if prettyPrint:
-                s.append(space)
-            s.append('<%s%s%s>' % (encodedName, attributeString, close))
-            if prettyPrint:
-                s.append("\n")
-            s.append(contents)
-            if prettyPrint and contents and contents[-1] != "\n":
-                s.append("\n")
-            if prettyPrint and closeTag:
-                s.append(space)
-            s.append(closeTag)
-            if prettyPrint and closeTag and self.nextSibling:
-                s.append("\n")
-            s = ''.join(s)
-        return s
-
-    def prettify(self, encoding=DEFAULT_OUTPUT_ENCODING):
-        return self.__str__(encoding, True)
-
-    def renderContents(self, encoding=DEFAULT_OUTPUT_ENCODING,
-                       prettyPrint=False, indentLevel=0):
-        """Renders the contents of this tag as a string in the given
-        encoding. If encoding is None, returns a Unicode string.."""
-        s=[]
-        for c in self:
-            text = None
-            if isinstance(c, NavigableString):
-                text = c.__str__(encoding)
-            elif isinstance(c, Tag):
-                s.append(c.__str__(encoding, prettyPrint, indentLevel))
-            if text and prettyPrint:
-                text = text.strip()              
-            if text:
-                if prettyPrint:
-                    s.append(" " * (indentLevel-1))
-                s.append(text)
-                if prettyPrint:
-                    s.append("\n")
-        return ''.join(s)    
-
-    #Soup methods
-
-    def find(self, name=None, attrs={}, recursive=True, text=None,
-             **kwargs):
-        """Return only the first child of this Tag matching the given
-        criteria."""
-        r = None
-        l = self.findAll(name, attrs, recursive, text, 1, **kwargs)
-        if l:
-            r = l[0]
-        return r
-    findChild = find
-
-    def findAll(self, name=None, attrs={}, recursive=True, text=None,
-                limit=None, **kwargs):
-        """Extracts a list of Tag objects that match the given
-        criteria.  You can specify the name of the Tag and any
-        attributes you want the Tag to have.
-
-        The value of a key-value pair in the 'attrs' map can be a
-        string, a list of strings, a regular expression object, or a
-        callable that takes a string and returns whether or not the
-        string matches for some custom definition of 'matches'. The
-        same is true of the tag name."""
-        generator = self.recursiveChildGenerator
-        if not recursive:
-            generator = self.childGenerator
-        return self._findAll(name, attrs, text, limit, generator, **kwargs)
-    findChildren = findAll
-
-    # Pre-3.x compatibility methods
-    first = find
-    fetch = findAll
-    
-    def fetchText(self, text=None, recursive=True, limit=None):
-        return self.findAll(text=text, recursive=recursive, limit=limit)
-
-    def firstText(self, text=None, recursive=True):
-        return self.find(text=text, recursive=recursive)
-    
-    #Utility methods
-
-    def append(self, tag):
-        """Appends the given tag to the contents of this tag."""
-        self.contents.append(tag)
-
-    #Private methods
-
-    def _getAttrMap(self):
-        """Initializes a map representation of this tag's attributes,
-        if not already initialized."""
-        if not getattr(self, 'attrMap'):
-            self.attrMap = {}
-            for (key, value) in self.attrs:
-                self.attrMap[key] = value 
-        return self.attrMap
-
-    #Generator methods
-    def childGenerator(self):
-        for i in range(0, len(self.contents)):
-            yield self.contents[i]
-        raise StopIteration
-    
-    def recursiveChildGenerator(self):
-        stack = [(self, 0)]
-        while stack:
-            tag, start = stack.pop()
-            if isinstance(tag, Tag):            
-                for i in range(start, len(tag.contents)):
-                    a = tag.contents[i]
-                    yield a
-                    if isinstance(a, Tag) and tag.contents:
-                        if i < len(tag.contents) - 1:
-                            stack.append((tag, i+1))
-                        stack.append((a, 0))
-                        break
-        raise StopIteration
-
-# Next, a couple classes to represent queries and their results.
-class SoupStrainer:
-    """Encapsulates a number of ways of matching a markup element (tag or
-    text)."""
-
-    def __init__(self, name=None, attrs={}, text=None, **kwargs):
-        self.name = name
-        if isString(attrs):
-            kwargs['class'] = attrs
-            attrs = None
-        if kwargs:
-            if attrs:
-                attrs = attrs.copy()
-                attrs.update(kwargs)
-            else:
-                attrs = kwargs
-        self.attrs = attrs
-        self.text = text
-
-    def __str__(self):
-        if self.text:
-            return self.text
-        else:
-            return "%s|%s" % (self.name, self.attrs)
-    
-    def searchTag(self, markupName=None, markupAttrs={}):
-        found = None
-        markup = None
-        if isinstance(markupName, Tag):
-            markup = markupName
-            markupAttrs = markup
-        callFunctionWithTagData = callable(self.name) \
-                                and not isinstance(markupName, Tag)
-
-        if (not self.name) \
-               or callFunctionWithTagData \
-               or (markup and self._matches(markup, self.name)) \
-               or (not markup and self._matches(markupName, self.name)):
-            if callFunctionWithTagData:
-                match = self.name(markupName, markupAttrs)
-            else:
-                match = True            
-                markupAttrMap = None
-                for attr, matchAgainst in self.attrs.items():
-                    if not markupAttrMap:
-                         if hasattr(markupAttrs, 'get'):
-                            markupAttrMap = markupAttrs
-                         else:
-                            markupAttrMap = {}
-                            for k,v in markupAttrs:
-                                markupAttrMap[k] = v
-                    attrValue = markupAttrMap.get(attr)
-                    if not self._matches(attrValue, matchAgainst):
-                        match = False
-                        break
-            if match:
-                if markup:
-                    found = markup
-                else:
-                    found = markupName
-        return found
-
-    def search(self, markup):
-        #print 'looking for %s in %s' % (self, markup)
-        found = None
-        # If given a list of items, scan it for a text element that
-        # matches.        
-        if isList(markup) and not isinstance(markup, Tag):
-            for element in markup:
-                if isinstance(element, NavigableString) \
-                       and self.search(element):
-                    found = element
-                    break
-        # If it's a Tag, make sure its name or attributes match.
-        # Don't bother with Tags if we're searching for text.
-        elif isinstance(markup, Tag):
-            if not self.text:
-                found = self.searchTag(markup)
-        # If it's text, make sure the text matches.
-        elif isinstance(markup, NavigableString) or \
-                 isString(markup):
-            if self._matches(markup, self.text):
-                found = markup
-        else:
-            raise Exception, "I don't know how to match against a %s" \
-                  % markup.__class__
-        return found
-        
-    def _matches(self, markup, matchAgainst):    
-        #print "Matching %s against %s" % (markup, matchAgainst)
-        result = False
-        if matchAgainst == True and type(matchAgainst) == types.BooleanType:
-            result = markup != None
-        elif callable(matchAgainst):
-            result = matchAgainst(markup)
-        else:
-            #Custom match methods take the tag as an argument, but all
-            #other ways of matching match the tag name as a string.
-            if isinstance(markup, Tag):
-                markup = markup.name
-            if markup and not isString(markup):
-                markup = unicode(markup)
-            #Now we know that chunk is either a string, or None.
-            if hasattr(matchAgainst, 'match'):
-                # It's a regexp object.
-                result = markup and matchAgainst.search(markup)
-            elif isList(matchAgainst):
-                result = markup in matchAgainst
-            elif hasattr(matchAgainst, 'items'):
-                result = markup.has_key(matchAgainst)
-            elif matchAgainst and isString(markup):
-                if isinstance(markup, unicode):
-                    matchAgainst = unicode(matchAgainst)
-                else:
-                    matchAgainst = str(matchAgainst)
-
-            if not result:
-                result = matchAgainst == markup
-        return result
-
-class ResultSet(list):
-    """A ResultSet is just a list that keeps track of the SoupStrainer
-    that created it."""
-    def __init__(self, source):
-        list.__init__([])
-        self.source = source
-
-# Now, some helper functions.
-
-def isList(l):
-    """Convenience method that works with all 2.x versions of Python
-    to determine whether or not something is listlike."""
-    return hasattr(l, '__iter__') \
-           or (type(l) in (types.ListType, types.TupleType))
-
-def isString(s):
-    """Convenience method that works with all 2.x versions of Python
-    to determine whether or not something is stringlike."""
-    try:
-        return isinstance(s, unicode) or isintance(s, basestring) 
-    except NameError:
-        return isinstance(s, str)
-
-def buildTagMap(default, *args):
-    """Turns a list of maps, lists, or scalars into a single map.
-    Used to build the SELF_CLOSING_TAGS, NESTABLE_TAGS, and
-    NESTING_RESET_TAGS maps out of lists and partial maps."""
-    built = {}
-    for portion in args:
-        if hasattr(portion, 'items'):
-            #It's a map. Merge it.
-            for k,v in portion.items():
-                built[k] = v
-        elif isList(portion):
-            #It's a list. Map each item to the default.
-            for k in portion:
-                built[k] = default
-        else:
-            #It's a scalar. Map it to the default.
-            built[portion] = default
-    return built
-
-# Now, the parser classes.
-
-class BeautifulStoneSoup(Tag, SGMLParser):
-
-    """This class contains the basic parser and search code. It defines
-    a parser that knows nothing about tag behavior except for the
-    following:
-   
-      You can't close a tag without closing all the tags it encloses.
-      That is, "<foo><bar></foo>" actually means
-      "<foo><bar></bar></foo>".
-
-    [Another possible explanation is "<foo><bar /></foo>", but since
-    this class defines no SELF_CLOSING_TAGS, it will never use that
-    explanation.]
-
-    This class is useful for parsing XML or made-up markup languages,
-    or when BeautifulSoup makes an assumption counter to what you were
-    expecting."""
-
-    XML_ENTITY_LIST = {}
-    for i in Tag.XML_SPECIAL_CHARS_TO_ENTITIES.values():
-        XML_ENTITY_LIST[i] = True 
-
-    SELF_CLOSING_TAGS = {}
-    NESTABLE_TAGS = {}
-    RESET_NESTING_TAGS = {}
-    QUOTE_TAGS = {}
-
-    MARKUP_MASSAGE = [(re.compile('(<[^<>]*)/>'),
-                       lambda x: x.group(1) + ' />'),
-                      (re.compile('<!\s+([^<>]*)>'),
-                       lambda x: '<!' + x.group(1) + '>')
-                      ]
-
-    ROOT_TAG_NAME = u'[document]'
-
-    HTML_ENTITIES = "html"
-    XML_ENTITIES = "xml"
-
-    def __init__(self, markup="", parseOnlyThese=None, fromEncoding=None,
-                 markupMassage=True, smartQuotesTo=XML_ENTITIES,
-                 convertEntities=None, selfClosingTags=None):
-        """The Soup object is initialized as the 'root tag', and the
-        provided markup (which can be a string or a file-like object)
-        is fed into the underlying parser. 
-
-        sgmllib will process most bad HTML, and the BeautifulSoup
-        class has some tricks for dealing with some HTML that kills
-        sgmllib, but Beautiful Soup can nonetheless choke or lose data
-        if your data uses self-closing tags or declarations
-        incorrectly.
-
-        By default, Beautiful Soup uses regexes to sanitize input,
-        avoiding the vast majority of these problems. If the problems
-        don't apply to you, pass in False for markupMassage, and
-        you'll get better performance.
-
-        The default parser massage techniques fix the two most common
-        instances of invalid HTML that choke sgmllib:
-
-         <br/> (No space between name of closing tag and tag close)
-         <! --Comment--> (Extraneous whitespace in declaration)
-
-        You can pass in a custom list of (RE object, replace method)
-        tuples to get Beautiful Soup to scrub your input the way you
-        want."""
-
-        self.parseOnlyThese = parseOnlyThese
-        self.fromEncoding = fromEncoding
-        self.smartQuotesTo = smartQuotesTo
-        self.convertEntities = convertEntities
-        if self.convertEntities:
-            # It doesn't make sense to convert encoded characters to
-            # entities even while you're converting entities to Unicode.
-            # Just convert it all to Unicode.
-            self.smartQuotesTo = None
-        self.instanceSelfClosingTags = buildTagMap(None, selfClosingTags)
-        SGMLParser.__init__(self)
-            
-        if hasattr(markup, 'read'):        # It's a file-type object.
-            markup = markup.read()
-        self.markup = markup
-        self.markupMassage = markupMassage
-        try:
-            self._feed()
-        except StopParsing:
-            pass
-        self.markup = None                 # The markup can now be GCed
-        
-    def _feed(self, inDocumentEncoding=None):
-        # Convert the document to Unicode.
-        markup = self.markup
-        if isinstance(markup, unicode):
-            if not hasattr(self, 'originalEncoding'):
-                self.originalEncoding = None
-        else:
-            dammit = UnicodeDammit\
-                     (markup, [self.fromEncoding, inDocumentEncoding],
-                      smartQuotesTo=self.smartQuotesTo)
-            markup = dammit.unicode
-            self.originalEncoding = dammit.originalEncoding
-        if markup:
-            if self.markupMassage:
-                if not isList(self.markupMassage):
-                    self.markupMassage = self.MARKUP_MASSAGE            
-                for fix, m in self.markupMassage:
-                    markup = fix.sub(m, markup)
-        self.reset()
-
-        SGMLParser.feed(self, markup)
-        # Close out any unfinished strings and close all the open tags.
-        self.endData()
-        while self.currentTag.name != self.ROOT_TAG_NAME:
-            self.popTag()
-
-    def __getattr__(self, methodName):
-        """This method routes method call requests to either the SGMLParser
-        superclass or the Tag superclass, depending on the method name."""
-        #print "__getattr__ called on %s.%s" % (self.__class__, methodName)
-
-        if methodName.find('start_') == 0 or methodName.find('end_') == 0 \
-               or methodName.find('do_') == 0:
-            return SGMLParser.__getattr__(self, methodName)
-        elif methodName.find('__') != 0:
-            return Tag.__getattr__(self, methodName)
-        else:
-            raise AttributeError
-
-    def isSelfClosingTag(self, name):
-        """Returns true iff the given string is the name of a
-        self-closing tag according to this parser."""
-        return self.SELF_CLOSING_TAGS.has_key(name) \
-               or self.instanceSelfClosingTags.has_key(name)
-            
-    def reset(self):
-        Tag.__init__(self, self, self.ROOT_TAG_NAME)
-        self.hidden = 1
-        SGMLParser.reset(self)
-        self.currentData = []
-        self.currentTag = None
-        self.tagStack = []
-        self.quoteStack = []
-        self.pushTag(self)
-    
-    def popTag(self):
-        tag = self.tagStack.pop()
-        # Tags with just one string-owning child get the child as a
-        # 'string' property, so that soup.tag.string is shorthand for
-        # soup.tag.contents[0]
-        if len(self.currentTag.contents) == 1 and \
-           isinstance(self.currentTag.contents[0], NavigableString):
-            self.currentTag.string = self.currentTag.contents[0]
-
-        #print "Pop", tag.name
-        if self.tagStack:
-            self.currentTag = self.tagStack[-1]
-        return self.currentTag
-
-    def pushTag(self, tag):
-        #print "Push", tag.name
-        if self.currentTag:
-            self.currentTag.append(tag)
-        self.tagStack.append(tag)
-        self.currentTag = self.tagStack[-1]
-
-    def endData(self, containerClass=NavigableString):
-        if self.currentData:
-            currentData = ''.join(self.currentData)
-            if not currentData.strip():
-                if '\n' in currentData:
-                    currentData = '\n'
-                else:
-                    currentData = ' '
-            self.currentData = []
-            if self.parseOnlyThese and len(self.tagStack) <= 1 and \
-                   (not self.parseOnlyThese.text or \
-                    not self.parseOnlyThese.search(currentData)):
-                return
-            o = containerClass(currentData)
-            o.setup(self.currentTag, self.previous)
-            if self.previous:
-                self.previous.next = o
-            self.previous = o
-            self.currentTag.contents.append(o)
-
-
-    def _popToTag(self, name, inclusivePop=True):
-        """Pops the tag stack up to and including the most recent
-        instance of the given tag. If inclusivePop is false, pops the tag
-        stack up to but *not* including the most recent instqance of
-        the given tag."""
-        #print "Popping to %s" % name
-        if name == self.ROOT_TAG_NAME:
-            return            
-
-        numPops = 0
-        mostRecentTag = None
-        for i in range(len(self.tagStack)-1, 0, -1):
-            if name == self.tagStack[i].name:
-                numPops = len(self.tagStack)-i
-                break
-        if not inclusivePop:
-            numPops = numPops - 1
-
-        for i in range(0, numPops):
-            mostRecentTag = self.popTag()
-        return mostRecentTag    
-
-    def _smartPop(self, name):
-
-        """We need to pop up to the previous tag of this type, unless
-        one of this tag's nesting reset triggers comes between this
-        tag and the previous tag of this type, OR unless this tag is a
-        generic nesting trigger and another generic nesting trigger
-        comes between this tag and the previous tag of this type.
-
-        Examples:
-         <p>Foo<b>Bar<p> should pop to 'p', not 'b'.
-         <p>Foo<table>Bar<p> should pop to 'table', not 'p'.
-         <p>Foo<table><tr>Bar<p> should pop to 'tr', not 'p'.
-         <p>Foo<b>Bar<p> should pop to 'p', not 'b'.
-
-         <li><ul><li> *<li>* should pop to 'ul', not the first 'li'.
-         <tr><table><tr> *<tr>* should pop to 'table', not the first 'tr'
-         <td><tr><td> *<td>* should pop to 'tr', not the first 'td'
-        """
-
-        nestingResetTriggers = self.NESTABLE_TAGS.get(name)
-        isNestable = nestingResetTriggers != None
-        isResetNesting = self.RESET_NESTING_TAGS.has_key(name)
-        popTo = None
-        inclusive = True
-        for i in range(len(self.tagStack)-1, 0, -1):
-            p = self.tagStack[i]
-            if (not p or p.name == name) and not isNestable:
-                #Non-nestable tags get popped to the top or to their
-                #last occurrence.
-                popTo = name
-                break
-            if (nestingResetTriggers != None
-                and p.name in nestingResetTriggers) \
-                or (nestingResetTriggers == None and isResetNesting
-                    and self.RESET_NESTING_TAGS.has_key(p.name)):
-                
-                #If we encounter one of the nesting reset triggers
-                #peculiar to this tag, or we encounter another tag
-                #that causes nesting to reset, pop up to but not
-                #including that tag.
-                popTo = p.name
-                inclusive = False
-                break
-            p = p.parent
-        if popTo:
-            self._popToTag(popTo, inclusive)
-
-    def unknown_starttag(self, name, attrs, selfClosing=0):
-        #print "Start tag %s: %s" % (name, attrs)
-        if self.quoteStack:
-            #This is not a real tag.
-            #print "<%s> is not real!" % name
-            attrs = ''.join(map(lambda(x, y): ' %s="%s"' % (x, y), attrs))
-            self.handle_data('<%s%s>' % (name, attrs))
-            return        
-        self.endData()
-
-        if not self.isSelfClosingTag(name) and not selfClosing:
-            self._smartPop(name)
-
-        if self.parseOnlyThese and len(self.tagStack) <= 1 \
-               and (self.parseOnlyThese.text or not self.parseOnlyThese.searchTag(name, attrs)):
-            return
-
-        tag = Tag(self, name, attrs, self.currentTag, self.previous)
-        if self.previous:
-            self.previous.next = tag
-        self.previous = tag
-        self.pushTag(tag)
-        if selfClosing or self.isSelfClosingTag(name):
-            self.popTag()                
-        if name in self.QUOTE_TAGS:
-            #print "Beginning quote (%s)" % name
-            self.quoteStack.append(name)
-            self.literal = 1
-        return tag
-
-    def unknown_endtag(self, name):
-        #print "End tag %s" % name
-        if self.quoteStack and self.quoteStack[-1] != name:
-            #This is not a real end tag.
-            #print "</%s> is not real!" % name
-            self.handle_data('</%s>' % name)
-            return
-        self.endData()
-        self._popToTag(name)
-        if self.quoteStack and self.quoteStack[-1] == name:
-            self.quoteStack.pop()
-            self.literal = (len(self.quoteStack) > 0)
-
-    def handle_data(self, data):
-        self.currentData.append(data)
-
-    def _toStringSubclass(self, text, subclass):
-        """Adds a certain piece of text to the tree as a NavigableString
-        subclass."""
-        self.endData()
-        self.handle_data(text)
-        self.endData(subclass)
-
-    def handle_pi(self, text):
-        """Handle a processing instruction as a ProcessingInstruction
-        object, possibly one with a %SOUP-ENCODING% slot into which an
-        encoding will be plugged later."""
-        if text[:3] == "xml":
-            text = "xml version='1.0' encoding='%SOUP-ENCODING%'"
-        self._toStringSubclass(text, ProcessingInstruction)
-
-    def handle_comment(self, text):
-        "Handle comments as Comment objects."
-        self._toStringSubclass(text, Comment)
-
-    def handle_charref(self, ref):
-        "Handle character references as data."
-        if self.convertEntities in [self.HTML_ENTITIES,
-                                    self.XML_ENTITIES]:
-            data = unichr(int(ref))
-        else:
-            data = '&#%s;' % ref
-        self.handle_data(data)
-
-    def handle_entityref(self, ref):
-        """Handle entity references as data, possibly converting known
-        HTML entity references to the corresponding Unicode
-        characters."""
-        data = None
-        if self.convertEntities == self.HTML_ENTITIES or \
-               (self.convertEntities == self.XML_ENTITIES and \
-                self.XML_ENTITY_LIST.get(ref)):
-            try:
-                data = unichr(name2codepoint[ref])
-            except KeyError:
-                pass
-        if not data:
-            data = '&%s;' % ref
-        self.handle_data(data)
-        
-    def handle_decl(self, data):
-        "Handle DOCTYPEs and the like as Declaration objects."
-        self._toStringSubclass(data, Declaration)
-
-    def parse_declaration(self, i):
-        """Treat a bogus SGML declaration as raw data. Treat a CDATA
-        declaration as a CData object."""
-        j = None
-        if self.rawdata[i:i+9] == '<![CDATA[':
-             k = self.rawdata.find(']]>', i)
-             if k == -1:
-                 k = len(self.rawdata)
-             data = self.rawdata[i+9:k]
-             j = k+3
-             self._toStringSubclass(data, CData)
-        else:
-            try:
-                j = SGMLParser.parse_declaration(self, i)
-            except SGMLParseError:
-                toHandle = self.rawdata[i:]
-                self.handle_data(toHandle)
-                j = i + len(toHandle)
-        return j
-
-class BeautifulSoup(BeautifulStoneSoup):
-
-    """This parser knows the following facts about HTML:
-
-    * Some tags have no closing tag and should be interpreted as being
-      closed as soon as they are encountered.
-
-    * The text inside some tags (ie. 'script') may contain tags which
-      are not really part of the document and which should be parsed
-      as text, not tags. If you want to parse the text as tags, you can
-      always fetch it and parse it explicitly.
-
-    * Tag nesting rules:
-
-      Most tags can't be nested at all. For instance, the occurrence of
-      a <p> tag should implicitly close the previous <p> tag.
-
-       <p>Para1<p>Para2
-        should be transformed into:
-       <p>Para1</p><p>Para2
-
-      Some tags can be nested arbitrarily. For instance, the occurrence
-      of a <blockquote> tag should _not_ implicitly close the previous
-      <blockquote> tag.
-
-       Alice said: <blockquote>Bob said: <blockquote>Blah
-        should NOT be transformed into:
-       Alice said: <blockquote>Bob said: </blockquote><blockquote>Blah
-
-      Some tags can be nested, but the nesting is reset by the
-      interposition of other tags. For instance, a <tr> tag should
-      implicitly close the previous <tr> tag within the same <table>,
-      but not close a <tr> tag in another table.
-
-       <table><tr>Blah<tr>Blah
-        should be transformed into:
-       <table><tr>Blah</tr><tr>Blah
-        but,
-       <tr>Blah<table><tr>Blah
-        should NOT be transformed into
-       <tr>Blah<table></tr><tr>Blah
-
-    Differing assumptions about tag nesting rules are a major source
-    of problems with the BeautifulSoup class. If BeautifulSoup is not
-    treating as nestable a tag your page author treats as nestable,
-    try ICantBelieveItsBeautifulSoup, MinimalSoup, or
-    BeautifulStoneSoup before writing your own subclass."""
-
-    def __init__(self, *args, **kwargs):
-        if not kwargs.has_key('smartQuotesTo'):
-            kwargs['smartQuotesTo'] = self.HTML_ENTITIES
-        BeautifulStoneSoup.__init__(self, *args, **kwargs)
-
-    SELF_CLOSING_TAGS = buildTagMap(None,
-                                    ['br' , 'hr', 'input', 'img', 'meta',
-                                    'spacer', 'link', 'frame', 'base'])
-
-    QUOTE_TAGS = {'script': None}
-    
-    #According to the HTML standard, each of these inline tags can
-    #contain another tag of the same type. Furthermore, it's common
-    #to actually use these tags this way.
-    NESTABLE_INLINE_TAGS = ['span', 'font', 'q', 'object', 'bdo', 'sub', 'sup',
-                            'center']
-
-    #According to the HTML standard, these block tags can contain
-    #another tag of the same type. Furthermore, it's common
-    #to actually use these tags this way.
-    NESTABLE_BLOCK_TAGS = ['blockquote', 'div', 'fieldset', 'ins', 'del']
-
-    #Lists can contain other lists, but there are restrictions.    
-    NESTABLE_LIST_TAGS = { 'ol' : [],
-                           'ul' : [],
-                           'li' : ['ul', 'ol'],
-                           'dl' : [],
-                           'dd' : ['dl'],
-                           'dt' : ['dl'] }
-
-    #Tables can contain other tables, but there are restrictions.    
-    NESTABLE_TABLE_TAGS = {'table' : [], 
-                           'tr' : ['table', 'tbody', 'tfoot', 'thead'],
-                           'td' : ['tr'],
-                           'th' : ['tr'],
-                           'thead' : ['table'],
-                           'tbody' : ['table'],
-                           'tfoot' : ['table'],
-                           }
-
-    NON_NESTABLE_BLOCK_TAGS = ['address', 'form', 'p', 'pre']
-
-    #If one of these tags is encountered, all tags up to the next tag of
-    #this type are popped.
-    RESET_NESTING_TAGS = buildTagMap(None, NESTABLE_BLOCK_TAGS, 'noscript',
-                                     NON_NESTABLE_BLOCK_TAGS,
-                                     NESTABLE_LIST_TAGS,
-                                     NESTABLE_TABLE_TAGS)
-
-    NESTABLE_TAGS = buildTagMap([], NESTABLE_INLINE_TAGS, NESTABLE_BLOCK_TAGS,
-                                NESTABLE_LIST_TAGS, NESTABLE_TABLE_TAGS)
-
-    # Used to detect the charset in a META tag; see start_meta
-    CHARSET_RE = re.compile("((^|;)\s*charset=)([^;]*)")
-
-    def start_meta(self, attrs):
-        """Beautiful Soup can detect a charset included in a META tag,
-        try to convert the document to that charset, and re-parse the
-        document from the beginning."""
-        httpEquiv = None
-        contentType = None
-        contentTypeIndex = None
-        tagNeedsEncodingSubstitution = False
-
-        for i in range(0, len(attrs)):
-            key, value = attrs[i]
-            key = key.lower()
-            if key == 'http-equiv':
-                httpEquiv = value
-            elif key == 'content':
-                contentType = value
-                contentTypeIndex = i
-
-        if httpEquiv and contentType: # It's an interesting meta tag.
-            match = self.CHARSET_RE.search(contentType)
-            if match:
-                if getattr(self, 'declaredHTMLEncoding') or \
-                       (self.originalEncoding == self.fromEncoding):
-                    # This is our second pass through the document, or
-                    # else an encoding was specified explicitly and it
-                    # worked. Rewrite the meta tag.
-                    newAttr = self.CHARSET_RE.sub\
-                              (lambda(match):match.group(1) +
-                               "%SOUP-ENCODING%", value)
-                    attrs[contentTypeIndex] = (attrs[contentTypeIndex][0],
-                                               newAttr)
-                    tagNeedsEncodingSubstitution = True
-                else:
-                    # This is our first pass through the document.
-                    # Go through it again with the new information.
-                    newCharset = match.group(3)
-                    if newCharset and newCharset != self.originalEncoding:
-                        self.declaredHTMLEncoding = newCharset
-                        self._feed(self.declaredHTMLEncoding)
-                        raise StopParsing
-        tag = self.unknown_starttag("meta", attrs)
-        if tag and tagNeedsEncodingSubstitution:
-            tag.containsSubstitutions = True
-
-class StopParsing(Exception):
-    pass
-   
-class ICantBelieveItsBeautifulSoup(BeautifulSoup):
-
-    """The BeautifulSoup class is oriented towards skipping over
-    common HTML errors like unclosed tags. However, sometimes it makes
-    errors of its own. For instance, consider this fragment:
-
-     <b>Foo<b>Bar</b></b>
-
-    This is perfectly valid (if bizarre) HTML. However, the
-    BeautifulSoup class will implicitly close the first b tag when it
-    encounters the second 'b'. It will think the author wrote
-    "<b>Foo<b>Bar", and didn't close the first 'b' tag, because
-    there's no real-world reason to bold something that's already
-    bold. When it encounters '</b></b>' it will close two more 'b'
-    tags, for a grand total of three tags closed instead of two. This
-    can throw off the rest of your document structure. The same is
-    true of a number of other tags, listed below.
-
-    It's much more common for someone to forget to close a 'b' tag
-    than to actually use nested 'b' tags, and the BeautifulSoup class
-    handles the common case. This class handles the not-co-common
-    case: where you can't believe someone wrote what they did, but
-    it's valid HTML and BeautifulSoup screwed up by assuming it
-    wouldn't be."""
-
-    I_CANT_BELIEVE_THEYRE_NESTABLE_INLINE_TAGS = \
-     ['em', 'big', 'i', 'small', 'tt', 'abbr', 'acronym', 'strong',
-      'cite', 'code', 'dfn', 'kbd', 'samp', 'strong', 'var', 'b',
-      'big']
-
-    I_CANT_BELIEVE_THEYRE_NESTABLE_BLOCK_TAGS = ['noscript']
-
-    NESTABLE_TAGS = buildTagMap([], BeautifulSoup.NESTABLE_TAGS,
-                                I_CANT_BELIEVE_THEYRE_NESTABLE_BLOCK_TAGS,
-                                I_CANT_BELIEVE_THEYRE_NESTABLE_INLINE_TAGS)
-
-class MinimalSoup(BeautifulSoup):
-    """The MinimalSoup class is for parsing HTML that contains
-    pathologically bad markup. It makes no assumptions about tag
-    nesting, but it does know which tags are self-closing, that
-    <script> tags contain Javascript and should not be parsed, that
-    META tags may contain encoding information, and so on.
-
-    This also makes it better for subclassing than BeautifulStoneSoup
-    or BeautifulSoup."""
-    
-    RESET_NESTING_TAGS = buildTagMap('noscript')
-    NESTABLE_TAGS = {}
-
-class BeautifulSOAP(BeautifulStoneSoup):
-    """This class will push a tag with only a single string child into
-    the tag's parent as an attribute. The attribute's name is the tag
-    name, and the value is the string child. An example should give
-    the flavor of the change:
-
-    <foo><bar>baz</bar></foo>
-     =>
-    <foo bar="baz"><bar>baz</bar></foo>
-
-    You can then access fooTag['bar'] instead of fooTag.barTag.string.
-
-    This is, of course, useful for scraping structures that tend to
-    use subelements instead of attributes, such as SOAP messages. Note
-    that it modifies its input, so don't print the modified version
-    out.
-
-    I'm not sure how many people really want to use this class; let me
-    know if you do. Mainly I like the name."""
-
-    def popTag(self):
-        if len(self.tagStack) > 1:
-            tag = self.tagStack[-1]
-            parent = self.tagStack[-2]
-            parent._getAttrMap()
-            if (isinstance(tag, Tag) and len(tag.contents) == 1 and
-                isinstance(tag.contents[0], NavigableString) and 
-                not parent.attrMap.has_key(tag.name)):
-                parent[tag.name] = tag.contents[0]
-        BeautifulStoneSoup.popTag(self)
-
-#Enterprise class names! It has come to our attention that some people
-#think the names of the Beautiful Soup parser classes are too silly
-#and "unprofessional" for use in enterprise screen-scraping. We feel
-#your pain! For such-minded folk, the Beautiful Soup Consortium And
-#All-Night Kosher Bakery recommends renaming this file to
-#"RobustParser.py" (or, in cases of extreme enterprisness,
-#"RobustParserBeanInterface.class") and using the following
-#enterprise-friendly class aliases:
-class RobustXMLParser(BeautifulStoneSoup):
-    pass
-class RobustHTMLParser(BeautifulSoup):
-    pass
-class RobustWackAssHTMLParser(ICantBelieveItsBeautifulSoup):
-    pass
-class RobustInsanelyWackAssHTMLParser(MinimalSoup):
-    pass
-class SimplifyingSOAPParser(BeautifulSOAP):
-    pass
-
-######################################################
-#
-# Bonus library: Unicode, Dammit
-#
-# This class forces XML data into a standard format (usually to UTF-8
-# or Unicode).  It is heavily based on code from Mark Pilgrim's
-# Universal Feed Parser. It does not rewrite the XML or HTML to
-# reflect a new encoding: that happens in BeautifulStoneSoup.handle_pi
-# (XML) and BeautifulSoup.start_meta (HTML).
-
-# Autodetects character encodings.
-# Download from http://chardet.feedparser.org/
-try:
-    import chardet
-#    import chardet.constants
-#    chardet.constants._debug = 1
-except:
-    chardet = None
-chardet = None
-
-# cjkcodecs and iconv_codec make Python know about more character encodings.
-# Both are available from http://cjkpython.i18n.org/
-# They're built in if you use Python 2.4.
-try:
-    import cjkcodecs.aliases
-except:
-    pass
-try:
-    import iconv_codec
-except:
-    pass
-
-class UnicodeDammit:
-    """A class for detecting the encoding of a *ML document and
-    converting it to a Unicode string. If the source encoding is
-    windows-1252, can replace MS smart quotes with their HTML or XML
-    equivalents."""
-
-    # This dictionary maps commonly seen values for "charset" in HTML
-    # meta tags to the corresponding Python codec names. It only covers
-    # values that aren't in Python's aliases and can't be determined
-    # by the heuristics in find_codec.
-    CHARSET_ALIASES = { "macintosh" : "mac-roman",
-                        "x-sjis" : "shift-jis" }
-
-    def __init__(self, markup, overrideEncodings=[],
-                 smartQuotesTo='xml'):
-        self.markup, documentEncoding, sniffedEncoding = \
-                     self._detectEncoding(markup)
-        self.smartQuotesTo = smartQuotesTo
-        self.triedEncodings = []
-        if markup == '' or isinstance(markup, unicode):
-            self.originalEncoding = None
-            self.unicode = unicode(markup)            
-            return
-        
-        u = None
-        for proposedEncoding in overrideEncodings:
-            u = self._convertFrom(proposedEncoding)
-            if u: break
-        if not u:
-            for proposedEncoding in (documentEncoding, sniffedEncoding):
-                u = self._convertFrom(proposedEncoding)
-                if u: break
-                
-        # If no luck and we have auto-detection library, try that:
-        if not u and chardet and not isinstance(self.markup, unicode):
-            u = self._convertFrom(chardet.detect(self.markup)['encoding'])
-
-        # As a last resort, try utf-8 and windows-1252:
-        if not u:
-            for proposed_encoding in ("utf-8", "windows-1252"):
-                u = self._convertFrom(proposed_encoding)
-                if u: break
-        self.unicode = u
-        if not u: self.originalEncoding = None
-
-    def _subMSChar(self, orig):
-        """Changes a MS smart quote character to an XML or HTML
-        entity."""
-        sub = self.MS_CHARS.get(orig)
-        if type(sub) == types.TupleType:
-            if self.smartQuotesTo == 'xml':
-                sub = '&#x%s;' % sub[1]
-            else:
-                sub = '&%s;' % sub[0]
-        return sub            
-
-    def _convertFrom(self, proposed):        
-        proposed = self.find_codec(proposed)
-        if not proposed or proposed in self.triedEncodings:
-            return None
-        self.triedEncodings.append(proposed)
-        markup = self.markup
-
-        # Convert smart quotes to HTML if coming from an encoding
-        # that might have them.
-        if self.smartQuotesTo and proposed.lower() in("windows-1252",
-                                                      "iso-8859-1",
-                                                      "iso-8859-2"):
-            markup = re.compile("([\x80-\x9f])").sub \
-                     (lambda(x): self._subMSChar(x.group(1)),
-                      markup)
-
-        try:
-            # print "Trying to convert document to %s" % proposed
-            u = self._toUnicode(markup, proposed)
-            self.markup = u       
-            self.originalEncoding = proposed
-        except Exception, e:
-            # print "That didn't work!"
-            # print e
-            return None        
-        #print "Correct encoding: %s" % proposed
-        return self.markup
-
-    def _toUnicode(self, data, encoding):
-        '''Given a string and its encoding, decodes the string into Unicode.
-        %encoding is a string recognized by encodings.aliases'''
-
-        # strip Byte Order Mark (if present)
-        if (len(data) >= 4) and (data[:2] == '\xfe\xff') \
-               and (data[2:4] != '\x00\x00'):
-            encoding = 'utf-16be'
-            data = data[2:]
-        elif (len(data) >= 4) and (data[:2] == '\xff\xfe') \
-                 and (data[2:4] != '\x00\x00'):
-            encoding = 'utf-16le'
-            data = data[2:]
-        elif data[:3] == '\xef\xbb\xbf':
-            encoding = 'utf-8'
-            data = data[3:]
-        elif data[:4] == '\x00\x00\xfe\xff':
-            encoding = 'utf-32be'
-            data = data[4:]
-        elif data[:4] == '\xff\xfe\x00\x00':
-            encoding = 'utf-32le'
-            data = data[4:]
-        newdata = unicode(data, encoding)
-        return newdata
-    
-    def _detectEncoding(self, xml_data):
-        """Given a document, tries to detect its XML encoding."""
-        xml_encoding = sniffed_xml_encoding = None
-        try:
-            if xml_data[:4] == '\x4c\x6f\xa7\x94':
-                # EBCDIC
-                xml_data = self._ebcdic_to_ascii(xml_data)
-            elif xml_data[:4] == '\x00\x3c\x00\x3f':
-                # UTF-16BE
-                sniffed_xml_encoding = 'utf-16be'
-                xml_data = unicode(xml_data, 'utf-16be').encode('utf-8')
-            elif (len(xml_data) >= 4) and (xml_data[:2] == '\xfe\xff') \
-                     and (xml_data[2:4] != '\x00\x00'):
-                # UTF-16BE with BOM
-                sniffed_xml_encoding = 'utf-16be'
-                xml_data = unicode(xml_data[2:], 'utf-16be').encode('utf-8')
-            elif xml_data[:4] == '\x3c\x00\x3f\x00':
-                # UTF-16LE
-                sniffed_xml_encoding = 'utf-16le'
-                xml_data = unicode(xml_data, 'utf-16le').encode('utf-8')
-            elif (len(xml_data) >= 4) and (xml_data[:2] == '\xff\xfe') and \
-                     (xml_data[2:4] != '\x00\x00'):
-                # UTF-16LE with BOM
-                sniffed_xml_encoding = 'utf-16le'
-                xml_data = unicode(xml_data[2:], 'utf-16le').encode('utf-8')
-            elif xml_data[:4] == '\x00\x00\x00\x3c':
-                # UTF-32BE
-                sniffed_xml_encoding = 'utf-32be'
-                xml_data = unicode(xml_data, 'utf-32be').encode('utf-8')
-            elif xml_data[:4] == '\x3c\x00\x00\x00':
-                # UTF-32LE
-                sniffed_xml_encoding = 'utf-32le'
-                xml_data = unicode(xml_data, 'utf-32le').encode('utf-8')
-            elif xml_data[:4] == '\x00\x00\xfe\xff':
-                # UTF-32BE with BOM
-                sniffed_xml_encoding = 'utf-32be'
-                xml_data = unicode(xml_data[4:], 'utf-32be').encode('utf-8')
-            elif xml_data[:4] == '\xff\xfe\x00\x00':
-                # UTF-32LE with BOM
-                sniffed_xml_encoding = 'utf-32le'
-                xml_data = unicode(xml_data[4:], 'utf-32le').encode('utf-8')
-            elif xml_data[:3] == '\xef\xbb\xbf':
-                # UTF-8 with BOM
-                sniffed_xml_encoding = 'utf-8'
-                xml_data = unicode(xml_data[3:], 'utf-8').encode('utf-8')
-            else:
-                sniffed_xml_encoding = 'ascii'
-                pass
-            xml_encoding_match = re.compile \
-                                 ('^<\?.*encoding=[\'"](.*?)[\'"].*\?>')\
-                                 .match(xml_data)
-        except:
-            xml_encoding_match = None
-        if xml_encoding_match:
-            xml_encoding = xml_encoding_match.groups()[0].lower()
-            if sniffed_xml_encoding and \
-               (xml_encoding in ('iso-10646-ucs-2', 'ucs-2', 'csunicode',
-                                 'iso-10646-ucs-4', 'ucs-4', 'csucs4',
-                                 'utf-16', 'utf-32', 'utf_16', 'utf_32',
-                                 'utf16', 'u16')):
-                xml_encoding = sniffed_xml_encoding
-        return xml_data, xml_encoding, sniffed_xml_encoding
-
-
-    def find_codec(self, charset):
-        return self._codec(self.CHARSET_ALIASES.get(charset, charset)) \
-               or (charset and self._codec(charset.replace("-", ""))) \
-               or (charset and self._codec(charset.replace("-", "_"))) \
-               or charset
-
-    def _codec(self, charset):
-        if not charset: return charset 
-        codec = None
-        try:
-            codecs.lookup(charset)
-            codec = charset
-        except LookupError:
-            pass
-        return codec
-
-    EBCDIC_TO_ASCII_MAP = None
-    def _ebcdic_to_ascii(self, s):
-        c = self.__class__
-        if not c.EBCDIC_TO_ASCII_MAP:
-            emap = (0,1,2,3,156,9,134,127,151,141,142,11,12,13,14,15,
-                    16,17,18,19,157,133,8,135,24,25,146,143,28,29,30,31,
-                    128,129,130,131,132,10,23,27,136,137,138,139,140,5,6,7,
-                    144,145,22,147,148,149,150,4,152,153,154,155,20,21,158,26,
-                    32,160,161,162,163,164,165,166,167,168,91,46,60,40,43,33,
-                    38,169,170,171,172,173,174,175,176,177,93,36,42,41,59,94,
-                    45,47,178,179,180,181,182,183,184,185,124,44,37,95,62,63,
-                    186,187,188,189,190,191,192,193,194,96,58,35,64,39,61,34,
-                    195,97,98,99,100,101,102,103,104,105,196,197,198,199,200,
-                    201,202,106,107,108,109,110,111,112,113,114,203,204,205,
-                    206,207,208,209,126,115,116,117,118,119,120,121,122,210,
-                    211,212,213,214,215,216,217,218,219,220,221,222,223,224,
-                    225,226,227,228,229,230,231,123,65,66,67,68,69,70,71,72,
-                    73,232,233,234,235,236,237,125,74,75,76,77,78,79,80,81,
-                    82,238,239,240,241,242,243,92,159,83,84,85,86,87,88,89,
-                    90,244,245,246,247,248,249,48,49,50,51,52,53,54,55,56,57,
-                    250,251,252,253,254,255)
-            import string
-            c.EBCDIC_TO_ASCII_MAP = string.maketrans( \
-            ''.join(map(chr, range(256))), ''.join(map(chr, emap)))
-        return s.translate(c.EBCDIC_TO_ASCII_MAP)
-
-    MS_CHARS = { '\x80' : ('euro', '20AC'),
-                 '\x81' : ' ',
-                 '\x82' : ('sbquo', '201A'),
-                 '\x83' : ('fnof', '192'),
-                 '\x84' : ('bdquo', '201E'),
-                 '\x85' : ('hellip', '2026'),
-                 '\x86' : ('dagger', '2020'),
-                 '\x87' : ('Dagger', '2021'),
-                 '\x88' : ('circ', '2C6'),
-                 '\x89' : ('permil', '2030'),
-                 '\x8A' : ('Scaron', '160'),
-                 '\x8B' : ('lsaquo', '2039'),
-                 '\x8C' : ('OElig', '152'),
-                 '\x8D' : '?',
-                 '\x8E' : ('#x17D', '17D'),
-                 '\x8F' : '?',
-                 '\x90' : '?',
-                 '\x91' : ('lsquo', '2018'),
-                 '\x92' : ('rsquo', '2019'),
-                 '\x93' : ('ldquo', '201C'),
-                 '\x94' : ('rdquo', '201D'),
-                 '\x95' : ('bull', '2022'),
-                 '\x96' : ('ndash', '2013'),
-                 '\x97' : ('mdash', '2014'),
-                 '\x98' : ('tilde', '2DC'),
-                 '\x99' : ('trade', '2122'),
-                 '\x9a' : ('scaron', '161'),
-                 '\x9b' : ('rsaquo', '203A'),
-                 '\x9c' : ('oelig', '153'),
-                 '\x9d' : '?',
-                 '\x9e' : ('#x17E', '17E'),
-                 '\x9f' : ('Yuml', ''),}
-
-#######################################################################
-
-
-#By default, act as an HTML pretty-printer.
-if __name__ == '__main__':
-    import sys
-    soup = BeautifulSoup(sys.stdin.read())
-    print soup.prettify()
diff --git a/bin/exampleparser.py b/bin/exampleparser.py
deleted file mode 100755
index 7a7352ec27..0000000000
--- a/bin/exampleparser.py
+++ /dev/null
@@ -1,269 +0,0 @@
-#!/usr/bin/env python
-
-import sys
-import os
-import re
-import time
-from xml.dom.minidom import Document
-
-try:
-    import xml.etree.ElementTree as ElementTree
-except ImportError:
-    try:
-        import cElementTree as ElementTree  # NOQA
-    except ImportError:
-        try:
-            import elementtree.ElementTree as ElementTree # NOQA
-        except ImportError:
-            import lxml.etree as ElementTree  # NOQA
-
-missing_deps = False
-try:
-    import json
-except ImportError:
-    try:
-        import simplejson as json  # NOQA
-    except ImportError, E:
-        missing_deps = E
-
-try:
-    from BeautifulSoup import BeautifulSoup
-except ImportError, E:
-    missing_deps = E
-
-feedName = "example-list.xml"
-feedPath = "http://openlayers.github.io/ol3/master/examples/"
-
-
-def getListOfExamples(relPath):
-    """
-    returns list of .html filenames within a given path - excludes
-    index.html
-    """
-    examples = os.listdir(relPath)
-    examples = [example for example in examples if
-                example.endswith('.html') and example != "index.html"]
-    return examples
-
-
-def getExampleHtml(path):
-    """
-    returns html of a specific example
-    """
-    print '.',
-    f = open(path)
-    html = f.read()
-    f.close()
-    return html
-
-
-def extractById(soup, tagId, value=None):
-    """
-    returns full contents of a particular tag id
-    """
-    beautifulTag = soup.find(id=tagId)
-    if beautifulTag:
-        if beautifulTag.contents:
-            value = str(beautifulTag.renderContents()).strip()
-            value = value.replace('\t', '')
-            value = value.replace('\n', '')
-    return value
-
-
-def getRelatedClasses(html):
-    """
-    parses the html, and returns a list of all OpenLayers Classes
-    used within (ie what parts of OL the javascript uses).
-    """
-    rawstr = r'''(?P<class>ol\..*?)\('''
-    return re.findall(rawstr, html)
-
-
-def parseHtml(html, ids):
-    """
-    returns dictionary of items of interest
-    """
-    soup = BeautifulSoup(html)
-    d = {}
-    for tagId in ids:
-        d[tagId] = extractById(soup, tagId)
-    #classes should eventually be parsed from docs - not automatically created.
-    classes = getRelatedClasses(html)
-    d['classes'] = classes
-    return d
-
-
-def getGitInfo(exampleDir, exampleName):
-    orig = os.getcwd()
-    os.chdir(exampleDir)
-    h = os.popen("git log -n 1 --pretty=format:'%an|%ai' " + exampleName)
-    os.chdir(orig)
-    log = h.read()
-    h.close()
-    d = {}
-    if log:
-        parts = log.split("|")
-        d["author"] = parts[0]
-        # compensate for spaces in git log time
-        td = parts[1].split(" ")
-        td.insert(1, "T")
-        d["date"] = "".join(td)
-    else:
-        d["author"] = ""
-        d["date"] = ""
-    return d
-
-
-def createFeed(examples):
-    doc = Document()
-    atomuri = "http://www.w3.org/2005/Atom"
-    feed = doc.createElementNS(atomuri, "feed")
-    feed.setAttribute("xmlns", atomuri)
-    title = doc.createElementNS(atomuri, "title")
-    title.appendChild(doc.createTextNode("OpenLayers Examples"))
-    feed.appendChild(title)
-    link = doc.createElementNS(atomuri, "link")
-    link.setAttribute("rel", "self")
-    link.setAttribute("href", feedPath + feedName)
-
-    modtime = time.strftime("%Y-%m-%dT%I:%M:%SZ", time.gmtime())
-    id = doc.createElementNS(atomuri, "id")
-    id.appendChild(doc.createTextNode(
-        "%s%s#%s" % (feedPath, feedName, modtime)))
-    feed.appendChild(id)
-
-    updated = doc.createElementNS(atomuri, "updated")
-    updated.appendChild(doc.createTextNode(modtime))
-    feed.appendChild(updated)
-
-    examples.sort(key=lambda x: x["modified"])
-    for example in sorted(examples, key=lambda x: x["modified"], reverse=True):
-        entry = doc.createElementNS(atomuri, "entry")
-
-        title = doc.createElementNS(atomuri, "title")
-        title.appendChild(doc.createTextNode(example["title"] or
-                                             example["example"]))
-        entry.appendChild(title)
-
-        tags = doc.createElementNS(atomuri, "tags")
-        tags.appendChild(doc.createTextNode(example["tags"] or
-                                            example["example"]))
-        entry.appendChild(tags)
-
-        link = doc.createElementNS(atomuri, "link")
-        link.setAttribute("href", "%s%s" % (feedPath, example["example"]))
-        entry.appendChild(link)
-
-        summary = doc.createElementNS(atomuri, "summary")
-        summary.appendChild(doc.createTextNode(example["shortdesc"] or
-                                               example["example"]))
-        entry.appendChild(summary)
-
-        updated = doc.createElementNS(atomuri, "updated")
-        updated.appendChild(doc.createTextNode(example["modified"]))
-        entry.appendChild(updated)
-
-        author = doc.createElementNS(atomuri, "author")
-        name = doc.createElementNS(atomuri, "name")
-        name.appendChild(doc.createTextNode(example["author"]))
-        author.appendChild(name)
-        entry.appendChild(author)
-
-        id = doc.createElementNS(atomuri, "id")
-        id.appendChild(doc.createTextNode("%s%s#%s" % (feedPath,
-                                                       example["example"],
-                                                       example["modified"])))
-        entry.appendChild(id)
-
-        feed.appendChild(entry)
-
-    doc.appendChild(feed)
-    return doc
-
-
-def wordIndex(examples):
-    """
-    Create an inverted index based on words in title and shortdesc.  Keys are
-    lower cased words.  Values are dictionaries with example index keys and
-    count values.
-    """
-    index = {}
-    unword = re.compile("\\W+")
-    keys = ["shortdesc", "title", "tags"]
-    for i in range(len(examples)):
-        for key in keys:
-            text = examples[i][key]
-            if text:
-                words = unword.split(text)
-                for word in words:
-                    if word:
-                        word = word.lower()
-                        if word in index:
-                            if i in index[word]:
-                                index[word][i] += 1
-                            else:
-                                index[word][i] = 1
-                        else:
-                            index[word] = {i: 1}
-    return index
-
-if __name__ == "__main__":
-
-    if missing_deps:
-        print """This script requires json or simplejson and BeautifulSoup.
-You don't have them. \n(%s)""" % E
-        sys.exit()
-
-    if len(sys.argv) == 3:
-        inExampleDir = sys.argv[1]
-        outExampleDir = sys.argv[2]
-    else:
-        inExampleDir = "../examples"
-        outExampleDir = "../examples"
-
-    outFile = open(os.path.join(outExampleDir, "example-list.js"), "w")
-
-    print 'Reading examples from %s and writing out to %s' % (inExampleDir,
-                                                              outFile.name)
-
-    exampleList = []
-    docIds = ['title', 'shortdesc', 'tags']
-
-    examples = getListOfExamples(inExampleDir)
-
-    modtime = time.strftime("%Y-%m-%dT%I:%M:%SZ", time.gmtime())
-
-    for example in examples:
-        path = os.path.join(inExampleDir, example)
-        html = getExampleHtml(path)
-        tagvalues = parseHtml(html, docIds)
-        tagvalues['example'] = example
-        # add in author/date info
-        d = getGitInfo(inExampleDir, example)
-        tagvalues["author"] = d["author"] or "anonymous"
-        tagvalues["modified"] = d["date"] or modtime
-        tagvalues['link'] = example
-
-        exampleList.append(tagvalues)
-
-    print
-
-    exampleList.sort(key=lambda x: x['example'].lower())
-
-    index = wordIndex(exampleList)
-
-    json = json.dumps({"examples": exampleList, "index": index})
-    # Give the json a global variable we can use in our js.
-    # This should be replaced or made optional.
-    json = 'var info=' + json + ';'
-    outFile.write(json)
-    outFile.close()
-
-    outFeedPath = os.path.join(outExampleDir, feedName)
-    print "writing feed to %s " % outFeedPath
-    atom = open(outFeedPath, 'w')
-    doc = createFeed(exampleList)
-    atom.write(doc.toxml())
-    atom.close()
-
-    print 'complete'
diff --git a/build.py b/build.py
index 058d61bc76..d2337cc2c7 100755
--- a/build.py
+++ b/build.py
@@ -112,23 +112,24 @@ EXECUTABLES = [variables.CLEANCSS, variables.GIT, variables.GJSLINT,
                variables.JSDOC, variables.JSHINT, variables.PYTHON,
                variables.PHANTOMJS]
 
-EXAMPLES = [path
-            for path in ifind('examples')
-            if path.endswith('.html')
-            if path != 'examples/index.html']
+EXAMPLES_SRC_ALL = [path for path in ifind('examples_src')]
 
-EXAMPLES_SRC = [path
-                for path in ifind('examples')
-                if path.endswith('.js')
-                if not path.endswith('.combined.js')
-                if path != 'examples/Jugl.js'
-                if path != 'examples/example-list.js']
+EXAMPLES_SRC_HTML = [path
+            for path in EXAMPLES_SRC_ALL
+            if path.endswith('.html')
+            if path != 'examples_src/index.html']
+
+EXAMPLES_SRC_JS = [example.replace('.html', '.js')
+            for example in EXAMPLES_SRC_HTML]
+
+EXAMPLES_DEST_ALL = [path.replace('examples_src', 'examples')
+            for path in EXAMPLES_SRC_ALL]
 
 EXAMPLES_JSON = ['build/' + example.replace('.html', '.json')
-                 for example in EXAMPLES]
+                 for example in EXAMPLES_SRC_HTML]
 
 EXAMPLES_COMBINED = ['build/' + example.replace('.html', '.combined.js')
-                     for example in EXAMPLES]
+                     for example in EXAMPLES_SRC_HTML]
 
 GLSL_SRC = [path
             for path in ifind('src')
@@ -261,17 +262,15 @@ virtual('build-examples', 'examples', 'build/examples/all.combined.js',
         EXAMPLES_COMBINED)
 
 
-virtual('examples', 'examples/example-list.xml', EXAMPLES_JSON)
+virtual('examples', EXAMPLES_DEST_ALL)
 
 
-@target('examples/example-list.xml', 'examples/example-list.js')
-def examples_examples_list_xml(t):
-    t.touch()  # already generated by bin/exampleparser.py
-
-
-@target('examples/example-list.js', 'bin/exampleparser.py', EXAMPLES)
-def examples_examples_list_js(t):
-    t.run('%(PYTHON)s', 'bin/exampleparser.py', 'examples', 'examples')
+@rule(r'\Aexamples/(?P<filepath>.*)\Z')
+def examples_dest(name, match):
+    def action(t):
+        t.run('node', 'tasks/build-examples.js')
+    dependencies = ['examples_src/%(filepath)s' % match.groupdict()]
+    return Target(name, action=action, dependencies=dependencies)
 
 
 @target('build/examples/all.combined.js', 'build/examples/all.js',
@@ -282,7 +281,7 @@ def build_examples_all_combined_js(t):
     report_sizes(t)
 
 
-@target('build/examples/all.js', EXAMPLES_SRC)
+@target('build/examples/all.js', EXAMPLES_SRC_JS)
 def build_examples_all_js(t):
     t.output('%(PYTHON)s', 'bin/combine-examples.py', t.dependencies)
 
@@ -405,7 +404,7 @@ def examples_star_combined_js(name, match):
     return Target(name, action=action, dependencies=dependencies)
 
 
-@target('serve', 'examples', 'build/test_requires.js', 'build/test_rendering_requires.js',
+@target('serve', 'build/test_requires.js', 'build/test_rendering_requires.js',
         NPM_INSTALL)
 def serve(t):
     t.run('node', 'tasks/serve.js')
@@ -415,7 +414,7 @@ virtual('lint', 'build/lint-timestamp', 'build/check-requires-timestamp',
     'build/check-whitespace-timestamp', 'jshint')
 
 
-@target('build/lint-timestamp', SRC, EXAMPLES_SRC, SPEC, SPEC_RENDERING,
+@target('build/lint-timestamp', SRC, EXAMPLES_SRC_JS, SPEC, SPEC_RENDERING,
         precious=True)
 def build_lint_src_timestamp(t):
     t.run('%(GJSLINT)s',
@@ -427,7 +426,7 @@ def build_lint_src_timestamp(t):
 
 virtual('jshint', 'build/jshint-timestamp')
 
-@target('build/jshint-timestamp', SRC, EXAMPLES_SRC, SPEC, SPEC_RENDERING,
+@target('build/jshint-timestamp', SRC, EXAMPLES_SRC_JS, SPEC, SPEC_RENDERING,
         TASKS, NPM_INSTALL, precious=True)
 def build_jshint_timestamp(t):
     t.run(variables.JSHINT, '--verbose', t.newer(t.dependencies))
@@ -457,7 +456,7 @@ def _strip_comments(lines):
                 yield lineno, line
 
 
-@target('build/check-requires-timestamp', SRC, EXAMPLES_SRC, SHADER_SRC,
+@target('build/check-requires-timestamp', SRC, EXAMPLES_SRC_JS, SHADER_SRC,
         SPEC, SPEC_RENDERING)
 def build_check_requires_timestamp(t):
     unused_count = 0
@@ -598,7 +597,7 @@ def build_check_requires_timestamp(t):
     t.touch()
 
 
-@target('build/check-whitespace-timestamp', SRC, EXAMPLES_SRC,
+@target('build/check-whitespace-timestamp', SRC, EXAMPLES_SRC_JS,
         SPEC, SPEC_RENDERING, JSDOC_SRC, precious=True)
 def build_check_whitespace_timestamp(t):
     CR_RE = re.compile(r'\r')
@@ -691,23 +690,16 @@ def host_examples(t):
     closure_lib_path = output('node', '-e',
         'process.stdout.write(require("closure-util").getLibraryPath())')
     t.rm_rf(examples_dir)
-    t.makedirs(examples_dir)
+    t.cp_r('examples', examples_dir)
+    for example in EXAMPLES_SRC_JS:
+        split_example_file(example, examples_dir % vars(variables))
+    t.cp('bin/loader_hosted_examples.js', examples_dir + '/loader.js')
     t.rm_rf(build_dir)
     t.makedirs(build_dir)
     t.rm_rf(css_dir)
     t.makedirs(css_dir)
-    t.cp(EXAMPLES, examples_dir)
-    for example in [path.replace('.html', '.js') for path in EXAMPLES]:
-        split_example_file(example, examples_dir % vars(variables))
-    for example in [path.replace('.html', '.css') for path in EXAMPLES]:
-        if os.path.isfile(example):
-            t.cp(example, examples_dir)
-    t.cp_r('examples/data', examples_dir + '/data')
-    t.cp('bin/loader_hosted_examples.js', examples_dir + '/loader.js')
     t.cp('build/ol.js', 'build/ol-debug.js', build_dir)
     t.cp('build/ol.css', css_dir)
-    t.cp('examples/index.html', 'examples/example-list.js',
-         'examples/example-list.xml', 'examples/Jugl.js', examples_dir)
     t.rm_rf('build/hosted/%(BRANCH)s/closure-library')
     t.cp_r(closure_lib_path, 'build/hosted/%(BRANCH)s/closure-library')
     t.rm_rf('build/hosted/%(BRANCH)s/ol')
@@ -726,8 +718,8 @@ def host_examples(t):
 
 @target('check-examples', 'host-examples', phony=True)
 def check_examples(t):
-    examples = ['build/hosted/%(BRANCH)s/' + e
-                for e in EXAMPLES
+    examples = ['build/hosted/%(BRANCH)s/' + e.replace('examples_src', 'examples')
+                for e in EXAMPLES_SRC_HTML
                 if not open(e.replace('.html', '.js'), 'rU').readline().startswith('// NOCOMPILE')]
     all_examples = [e + '?mode=advanced' for e in examples]
     # Run the examples checks in a pool of threads