Initial commit.

lib/htmlpurifier-3.0.0/maintenance/.htaccess

@@ -0,0 +1 @@
+Deny from all

lib/htmlpurifier-3.0.0/maintenance/PH5P.patch

@@ -0,0 +1,64 @@
+--- C:\Users\Edward\Webs\htmlpurifier\maintenance\PH5P.php	2007-11-04 23:41:49.074543700 -0500
++++ C:\Users\Edward\Webs\htmlpurifier\maintenance/PH5P.new.php	2007-11-05 00:23:52.839543700 -0500
+@@ -211,7 +211,10 @@
+         // If nothing is returned, emit a U+0026 AMPERSAND character token.
+         // Otherwise, emit the character token that was returned.
+         $char = (!$entity) ? '&' : $entity;
+-        $this->emitToken($char);
++        $this->emitToken(array(
++            'type' => self::CHARACTR,
++            'data' => $char
++        ));
+ 
+         // Finally, switch to the data state.
+         $this->state = 'data';
+@@ -708,7 +711,7 @@
+         } elseif($char === '&') {
+             /* U+0026 AMPERSAND (&)
+             Switch to the entity in attribute value state. */
+-            $this->entityInAttributeValueState('non');
++            $this->entityInAttributeValueState();
+ 
+         } elseif($char === '>') {
+             /* U+003E GREATER-THAN SIGN (>)
+@@ -738,7 +741,8 @@
+             ? '&'
+             : $entity;
+ 
+-        $this->emitToken($char);
++        $last = count($this->token['attr']) - 1;
++        $this->token['attr'][$last]['value'] .= $char;
+     }
+ 
+     private function bogusCommentState() {
+@@ -1066,6 +1070,11 @@
+                     $this->char++;
+ 
+                     if(in_array($id, $this->entities)) {
++                        if ($e_name[$c-1] !== ';') {
++                            if ($c < $len && $e_name[$c] == ';') {
++                                $this->char++; // consume extra semicolon
++                            }
++                        }
+                         $entity = $id;
+                         break;
+                     }
+@@ -3659,7 +3668,7 @@
+         }
+     }
+ 
+-    private function generateImpliedEndTags(array $exclude = array()) {
++    private function generateImpliedEndTags($exclude = array()) {
+         /* When the steps below require the UA to generate implied end tags,
+         then, if the current node is a dd element, a dt element, an li element,
+         a p element, a td element, a th  element, or a tr element, the UA must
+@@ -3673,7 +3682,8 @@
+         }
+     }
+ 
+-    private function getElementCategory($name) {
++    private function getElementCategory($node) {
++        $name = $node->tagName;
+         if(in_array($name, $this->special))
+             return self::SPECIAL;
+ 

lib/htmlpurifier-3.0.0/maintenance/common.php

@@ -0,0 +1,141 @@
+<?php
+
+function assertCli() {
+    if (php_sapi_name() != 'cli' && !getenv('PHP_IS_CLI')) {
+        echo 'Script cannot be called from web-browser (if you are calling via cli,
+set environment variable PHP_IS_CLI to work around this).';
+        exit;
+    }
+}
+
+/**
+ * Filesystem tools not provided by default; can recursively create, copy
+ * and delete folders. Some template methods are provided for extensibility.
+ * @note This class must be instantiated to be used, although it does
+ *       not maintain state.
+ */
+class FSTools
+{
+    
+    /**
+     * Recursively creates a directory
+     * @param string $folder Name of folder to create
+     * @note Adapted from the PHP manual comment 76612
+     */
+    function mkdir($folder) {
+        $folders = preg_split("#[\\\\/]#", $folder);
+        $base = '';
+        for($i = 0, $c = count($folders); $i < $c; $i++) {
+            if(empty($folders[$i])) {
+                if (!$i) {
+                    // special case for root level
+                    $base .= DIRECTORY_SEPARATOR;
+                }
+                continue;
+            }
+            $base .= $folders[$i];
+            if(!is_dir($base)){
+                mkdir($base);
+            }
+            $base .= DIRECTORY_SEPARATOR;
+        }
+    }
+    
+    /**
+     * Copy a file, or recursively copy a folder and its contents; modified
+     * so that copied files, if PHP, have includes removed
+     *
+     * @author      Aidan Lister <aidan@php.net>
+     * @version     1.0.1-modified
+     * @link        http://aidanlister.com/repos/v/function.copyr.php
+     * @param       string   $source    Source path
+     * @param       string   $dest      Destination path
+     * @return      bool     Returns TRUE on success, FALSE on failure
+     */
+    function copyr($source, $dest) {
+        // Simple copy for a file
+        if (is_file($source)) {
+            return $this->copy($source, $dest);
+        }
+        // Make destination directory
+        if (!is_dir($dest)) {
+            mkdir($dest);
+        }
+        // Loop through the folder
+        $dir = dir($source);
+        while (false !== $entry = $dir->read()) {
+            // Skip pointers
+            if ($entry == '.' || $entry == '..') {
+                continue;
+            }
+            if (!$this->copyable($entry)) {
+                continue;
+            }
+            // Deep copy directories
+            if ($dest !== "$source/$entry") {
+                $this->copyr("$source/$entry", "$dest/$entry");
+            }
+        }
+        // Clean up
+        $dir->close();
+        return true;
+    }
+    
+    /**
+     * Stub for PHP's built-in copy function, can be used to overload
+     * functionality
+     */
+    function copy($source, $dest) {
+        return copy($source, $dest);
+    }
+    
+    /**
+     * Overloadable function that tests a filename for copyability. By
+     * default, everything should be copied; you can restrict things to
+     * ignore hidden files, unreadable files, etc.
+     */
+    function copyable($file) {
+        return true;
+    }
+    
+    /**
+     * Delete a file, or a folder and its contents
+     *
+     * @author      Aidan Lister <aidan@php.net>
+     * @version     1.0.3
+     * @link        http://aidanlister.com/repos/v/function.rmdirr.php
+     * @param       string   $dirname    Directory to delete
+     * @return      bool     Returns TRUE on success, FALSE on failure
+     */
+    function rmdirr($dirname)
+    {
+        // Sanity check
+        if (!file_exists($dirname)) {
+            return false;
+        }
+     
+        // Simple delete for a file
+        if (is_file($dirname) || is_link($dirname)) {
+            return unlink($dirname);
+        }
+     
+        // Loop through the folder
+        $dir = dir($dirname);
+        while (false !== $entry = $dir->read()) {
+            // Skip pointers
+            if ($entry == '.' || $entry == '..') {
+                continue;
+            }
+            // Recurse
+            $this->rmdirr($dirname . DIRECTORY_SEPARATOR . $entry);
+        }
+     
+        // Clean up
+        $dir->close();
+        return rmdir($dirname);
+    }
+    
+    
+}
+
+

lib/htmlpurifier-3.0.0/maintenance/flush-definition-cache.php

@@ -0,0 +1,37 @@
+#!/usr/bin/php
+<?php
+
+chdir(dirname(__FILE__));
+require_once 'common.php';
+assertCli();
+
+/**
+ * Flushes the default HTMLDefinition serial cache
+ * @param Accepts one argument, cache type to flush; otherwise flushes all
+ *      the caches.
+ */
+
+echo "Flushing cache... \n";
+
+require_once(dirname(__FILE__) . '/../library/HTMLPurifier.auto.php');
+
+$config = HTMLPurifier_Config::createDefault();
+
+$names = array('HTML', 'CSS', 'URI', 'Test');
+if (isset($argv[1])) {
+    if (in_array($argv[1], $names)) {
+        $names = array($argv[1]);
+    } else {
+        echo "Did not recognized cache parameter {$argv[1]} as valid cache, aborting.\n";
+        exit;
+    }
+}
+
+foreach ($names as $name) {
+    echo " - Flushing $name\n";
+    $cache = new HTMLPurifier_DefinitionCache_Serializer($name);
+    $cache->flush($config);
+}
+
+echo "Cache flushed successfully.\n";
+

lib/htmlpurifier-3.0.0/maintenance/generate-entity-file.php

@@ -0,0 +1,72 @@
+#!/usr/bin/php
+<?php
+
+chdir(dirname(__FILE__));
+require_once 'common.php';
+assertCli();
+
+/**
+ * Parses *.ent files into an entity lookup table, and then serializes and
+ * writes the whole kaboodle to a file. The resulting file should be versioned.
+ */
+
+chdir( dirname(__FILE__) );
+
+// here's where the entity files are located, assuming working directory
+// is the same as the location of this PHP file. Needs trailing slash.
+$entity_dir = '../docs/entities/';
+
+// defines the output file for the serialized content.
+$output_file = '../library/HTMLPurifier/EntityLookup/entities.ser';
+
+// courtesy of a PHP manual comment
+function unichr($dec) {
+    if ($dec < 128) {
+        $utf  = chr($dec);
+    } else if ($dec < 2048) {
+        $utf  = chr(192 + (($dec - ($dec % 64)) / 64));
+        $utf .= chr(128 + ($dec % 64));
+    } else {
+        $utf  = chr(224 + (($dec - ($dec % 4096)) / 4096));
+        $utf .= chr(128 + ((($dec % 4096) - ($dec % 64)) / 64));
+        $utf .= chr(128 + ($dec % 64));
+    }
+    return $utf;
+}
+
+if ( !is_dir($entity_dir) ) exit("Fatal Error: Can't find entity directory.\n");
+if ( file_exists($output_file) ) exit("Fatal Error: entity-lookup.txt already exists.\n");
+
+$dh = @opendir($entity_dir);
+if ( !$dh ) exit("Fatal Error: Cannot read entity directory.\n");
+
+$entity_files = array();
+while (($file = readdir($dh)) !== false) {
+    if (@$file[0] === '.') continue;
+    if (substr(strrchr($file, "."), 1) !== 'ent') continue;
+    $entity_files[] = $file;
+}
+closedir($dh);
+
+if ( !$entity_files ) exit("Fatal Error: No entity files to parse.\n");
+
+$entity_table = array();
+$regexp = '/<!ENTITY\s+([A-Za-z]+)\s+"&#(?:38;#)?([0-9]+);">/';
+
+foreach ( $entity_files as $file ) {
+    $contents = file_get_contents($entity_dir . $file);
+    $matches = array();
+    preg_match_all($regexp, $contents, $matches, PREG_SET_ORDER);
+    foreach ($matches as $match) {
+        $entity_table[$match[1]] = unichr($match[2]);
+    }
+}
+
+$output = serialize($entity_table);
+
+$fh = fopen($output_file, 'w');
+fwrite($fh, $output);
+fclose($fh);
+
+echo "Completed successfully.";
+

lib/htmlpurifier-3.0.0/maintenance/generate-ph5p-patch.php

@@ -0,0 +1,13 @@
+<?php
+
+$orig = realpath(dirname(__FILE__) . '/PH5P.php');
+$new  = realpath(dirname(__FILE__) . '/../library/HTMLPurifier/Lexer/PH5P.php');
+$newt = dirname(__FILE__) . '/PH5P.new.php'; // temporary file
+
+// minor text-processing of new file to get into same format as original
+$new_src = file_get_contents($new);
+$new_src = '<?php' . PHP_EOL . substr($new_src, strpos($new_src, 'class HTML5 {'));
+
+file_put_contents($newt, $new_src);
+shell_exec("diff -u \"$orig\" \"$newt\" > PH5P.patch");
+unlink($newt);

lib/htmlpurifier-3.0.0/maintenance/merge-library.php

@@ -0,0 +1,163 @@
+#!/usr/bin/php
+<?php
+
+chdir(dirname(__FILE__));
+require_once 'common.php';
+assertCli();
+
+/**
+ * Compiles all of HTML Purifier's library files into one big file
+ * named HTMLPurifier.standalone.php.
+ */
+
+/**
+ * Global hash that tracks already loaded includes
+ */
+$GLOBALS['loaded'] = array('HTMLPurifier.php' => true);
+
+/**
+ * Custom FSTools for this script that overloads some behavior
+ * @warning The overloading of copy() is not necessarily global for
+ *          this script. Watch out!
+ */
+class MergeLibraryFSTools extends FSTools
+{
+    function copyable($entry) {
+        // Skip hidden files
+        if ($entry[0] == '.') {
+            return false;
+        }
+        return true;
+    }
+    function copy($source, $dest) {
+        copy_and_remove_includes($source, $dest);
+    }
+}
+$FS = new MergeLibraryFSTools();
+
+/**
+ * Replaces the includes inside PHP source code with the corresponding
+ * source.
+ * @param string $text PHP source code to replace includes from
+ */
+function replace_includes($text) {
+    return preg_replace_callback(
+        "/require_once ['\"]([^'\"]+)['\"];/",
+        'replace_includes_callback',
+        $text
+    );
+}
+
+/**
+ * Removes leading PHP tags from included files. Assumes that there is
+ * no trailing tag.
+ * @note This is safe for files that have internal <?php
+ * @param string $text Text to have leading PHP tag from
+ */
+function remove_php_tags($text) {
+    return substr($text, 5);
+}
+
+/**
+ * Creates an appropriate blank file, recursively generating directories
+ * if necessary
+ * @param string $file Filename to create blank for
+ */
+function create_blank($file) {
+    global $FS;
+    $dir = dirname($file);
+    $base = realpath('../tests/blanks/') . DIRECTORY_SEPARATOR ;
+    if ($dir != '.') {
+        $FS->mkdir($base . $dir);
+    }
+    file_put_contents($base . $file, '');
+}
+
+/**
+ * Copies the contents of a directory to the standalone directory
+ * @param string $dir Directory to copy
+ */
+function make_dir_standalone($dir) {
+    global $FS;
+    return $FS->copyr($dir, 'standalone/' . $dir);
+}
+
+/**
+ * Copies the contents of a file to the standalone directory
+ * @param string $file File to copy
+ */
+function make_file_standalone($file) {
+    global $FS;
+    $FS->mkdir('standalone/' . dirname($file));
+    copy_and_remove_includes($file, 'standalone/' . $file);
+    return true;
+}
+
+/**
+ * Copies a file to another location recursively, if it is a PHP file
+ * remove includes
+ * @param string $file Original file
+ * @param string $sfile New location of file
+ */
+function copy_and_remove_includes($file, $sfile) {
+    $contents = file_get_contents($file);
+    if (strrchr($file, '.') === '.php') $contents = replace_includes($contents);
+    return file_put_contents($sfile, $contents);
+}
+
+/**
+ * @param $matches preg_replace_callback matches array, where index 1
+ *        is the filename to include
+ */
+function replace_includes_callback($matches) {
+    $file = $matches[1];
+    $preserve = array(
+      // PHP 5 only
+      'HTMLPurifier/Lexer/DOMLex.php' => 1,
+      'HTMLPurifier/Printer.php' => 1,
+      // PEAR (external)
+      'XML/HTMLSax3.php' => 1
+    );
+    if (isset($preserve[$file])) {
+        return $matches[0];
+    }
+    if (isset($GLOBALS['loaded'][$file])) return '';
+    $GLOBALS['loaded'][$file] = true;
+    create_blank($file);
+    return replace_includes(remove_php_tags(file_get_contents($file)));
+}
+
+chdir(dirname(__FILE__) . '/../library/');
+create_blank('HTMLPurifier.php');
+
+echo 'Creating full file...';
+$contents = replace_includes(file_get_contents('HTMLPurifier.php'));
+$contents = str_replace(
+    "define('HTMLPURIFIER_PREFIX', dirname(__FILE__));",
+    "define('HTMLPURIFIER_PREFIX', dirname(__FILE__) . '/standalone');
+set_include_path(HTMLPURIFIER_PREFIX . PATH_SEPARATOR . get_include_path());",
+    $contents
+);
+file_put_contents('HTMLPurifier.standalone.php', $contents);
+echo ' done!' . PHP_EOL;
+
+echo 'Creating standalone directory...';
+$FS->rmdirr('standalone'); // ensure a clean copy
+
+// data files
+$FS->mkdir('standalone/HTMLPurifier/DefinitionCache/Serializer');
+make_dir_standalone('HTMLPurifier/EntityLookup');
+
+// non-standard inclusion setup
+make_dir_standalone('HTMLPurifier/Language');
+
+// optional components
+make_file_standalone('HTMLPurifier/Printer.php'); 
+make_dir_standalone('HTMLPurifier/Printer');
+make_dir_standalone('HTMLPurifier/Filter');
+make_file_standalone('HTMLPurifier/Lexer/PEARSax3.php');
+
+// PHP 5 only files
+make_file_standalone('HTMLPurifier/Lexer/DOMLex.php');
+make_file_standalone('HTMLPurifier/Lexer/PH5P.php');
+echo ' done!' . PHP_EOL;