From 349c4de75bc11c2c186ba1133d5022d4d0731a79 Mon Sep 17 00:00:00 2001
From: "Edward Z. Yang" <edwardzyang@thewritingpot.com>
Date: Mon, 30 Jul 2007 16:56:50 +0000
Subject: [PATCH] [2.1.0] Standalone file now can be generated using
 maintenance/merge-library.php. Also: - HTMLPURIFIER_PREFIX constant added,
 and relevant files transitioned over - Custom ChildDef added to default
 include list - Tester accepts ?standalone parameter

git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1316 48356398-32a2-884e-a903-53898d9a118a
---
 NEWS                                          |   7 +
 library/HTMLPurifier.php                      |   3 +
 library/HTMLPurifier/ContentSets.php          |   1 +
 .../DefinitionCache/Serializer.php            |   2 +-
 library/HTMLPurifier/EntityLookup.php         |   2 +-
 library/HTMLPurifier/LanguageFactory.php      |   2 +-
 library/HTMLPurifier/Printer/ConfigForm.php   |   4 +-
 library/HTMLPurifier/URISchemeRegistry.php    |   6 +-
 maintenance/merge-library.php                 | 190 ++++++++++++++++++
 .../DefinitionCache/SerializerTest.php        |   3 +-
 tests/index.php                               |   8 +-
 11 files changed, 218 insertions(+), 10 deletions(-)
 create mode 100644 maintenance/merge-library.php

diff --git a/NEWS b/NEWS
index 1ae3baa5..7e48c753 100644
--- a/NEWS
+++ b/NEWS
@@ -14,6 +14,9 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 ! With %Core.AggressivelyFixLt, <3 and similar emoticons no longer
   trigger HTML removal in PHP5 (DOMLex). This directive is not necessary
   for PHP4 (DirectLex).
+! Standalone file now available, which greatly reduces the amount of
+  includes (although there are still a few files that reside in the
+  standalone folder)
 - AutoFormatters emit friendly error messages if tags or attributes they
   need are not allowed
 - ConfigForm's compactification of directive names is now configurable
@@ -34,6 +37,10 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 . Hidden element content removal genericized: %Core.HiddenElements can
   be used to customize this behavior, by default <script> and <style> are
   hidden
+. Added HTMLPURIFIER_PREFIX constant, should be used instead of dirname(__FILE__)
+. Custom ChildDef added to default include list
+. URIScheme reflection improved: will not attempt to include file if class
+  already exists. May clobber autoload, so I need to keep an eye on it
 
 2.0.1, released 2007-06-27
 ! Tag auto-closing now based on a ChildDef heuristic rather than a
diff --git a/library/HTMLPurifier.php b/library/HTMLPurifier.php
index ebbde869..0d1d9fc2 100644
--- a/library/HTMLPurifier.php
+++ b/library/HTMLPurifier.php
@@ -40,6 +40,9 @@
     Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
  */
 
+// constants are slow, but we'll make one exception
+define('HTMLPURIFIER_PREFIX', dirname(__FILE__));
+
 // almost every class has an undocumented dependency to these, so make sure
 // they get included
 require_once 'HTMLPurifier/ConfigSchema.php'; // important
diff --git a/library/HTMLPurifier/ContentSets.php b/library/HTMLPurifier/ContentSets.php
index 001f4814..7baf7a31 100644
--- a/library/HTMLPurifier/ContentSets.php
+++ b/library/HTMLPurifier/ContentSets.php
@@ -5,6 +5,7 @@ require_once 'HTMLPurifier/ChildDef.php';
 require_once 'HTMLPurifier/ChildDef/Empty.php';
 require_once 'HTMLPurifier/ChildDef/Required.php';
 require_once 'HTMLPurifier/ChildDef/Optional.php';
+require_once 'HTMLPurifier/ChildDef/Custom.php';
 
 // NOT UNIT TESTED!!!
 
diff --git a/library/HTMLPurifier/DefinitionCache/Serializer.php b/library/HTMLPurifier/DefinitionCache/Serializer.php
index 2b07da13..1830e37e 100644
--- a/library/HTMLPurifier/DefinitionCache/Serializer.php
+++ b/library/HTMLPurifier/DefinitionCache/Serializer.php
@@ -99,7 +99,7 @@ class HTMLPurifier_DefinitionCache_Serializer extends
      */
     function generateBaseDirectoryPath($config) {
         $base = $config->get('Cache', 'SerializerPath');
-        $base = is_null($base) ? dirname(__FILE__) . '/Serializer' : $base;
+        $base = is_null($base) ? HTMLPURIFIER_PREFIX . '/HTMLPurifier/DefinitionCache/Serializer' : $base;
         return $base;
     }
     
diff --git a/library/HTMLPurifier/EntityLookup.php b/library/HTMLPurifier/EntityLookup.php
index e6b24c5c..8204867b 100644
--- a/library/HTMLPurifier/EntityLookup.php
+++ b/library/HTMLPurifier/EntityLookup.php
@@ -19,7 +19,7 @@ class HTMLPurifier_EntityLookup {
      */
     function setup($file = false) {
         if (!$file) {
-            $file = dirname(__FILE__) . '/EntityLookup/entities.ser';
+            $file = HTMLPURIFIER_PREFIX . '/HTMLPurifier/EntityLookup/entities.ser';
         }
         $this->table = unserialize(file_get_contents($file));
     }
diff --git a/library/HTMLPurifier/LanguageFactory.php b/library/HTMLPurifier/LanguageFactory.php
index 0721e386..9d26cd70 100644
--- a/library/HTMLPurifier/LanguageFactory.php
+++ b/library/HTMLPurifier/LanguageFactory.php
@@ -82,7 +82,7 @@ class HTMLPurifier_LanguageFactory
      */
     function setup() {
         $this->validator = new HTMLPurifier_AttrDef_Lang();
-        $this->dir = dirname(__FILE__);
+        $this->dir = HTMLPURIFIER_PREFIX . '/HTMLPurifier';
     }
     
     /**
diff --git a/library/HTMLPurifier/Printer/ConfigForm.php b/library/HTMLPurifier/Printer/ConfigForm.php
index 4305ea86..31da35f8 100644
--- a/library/HTMLPurifier/Printer/ConfigForm.php
+++ b/library/HTMLPurifier/Printer/ConfigForm.php
@@ -59,14 +59,14 @@ class HTMLPurifier_Printer_ConfigForm extends HTMLPurifier_Printer
      * available
      */
     function getCSS() {
-        return file_get_contents(dirname(__FILE__) . '/ConfigForm.css');
+        return file_get_contents(HTMLPURIFIER_PREFIX . '/HTMLPurifier/Printer/ConfigForm.css');
     }
     
     /**
      * Retrieves JavaScript, in case directory is not public
      */
     function getJavaScript() {
-        return file_get_contents(dirname(__FILE__) . '/ConfigForm.js');
+        return file_get_contents(HTMLPURIFIER_PREFIX . '/HTMLPurifier/Printer/ConfigForm.js');
     }
     
     /**
diff --git a/library/HTMLPurifier/URISchemeRegistry.php b/library/HTMLPurifier/URISchemeRegistry.php
index 7227b558..8cddb7f9 100644
--- a/library/HTMLPurifier/URISchemeRegistry.php
+++ b/library/HTMLPurifier/URISchemeRegistry.php
@@ -79,12 +79,14 @@ class HTMLPurifier_URISchemeRegistry
         }
         
         if (isset($this->schemes[$scheme])) return $this->schemes[$scheme];
-        if (empty($this->_dir)) $this->_dir = dirname(__FILE__) . '/URIScheme/';
+        if (empty($this->_dir)) $this->_dir = HTMLPURIFIER_PREFIX . '/HTMLPurifier/URIScheme/';
         
         if (!isset($allowed_schemes[$scheme])) return $null;
         
-        @include_once $this->_dir . $scheme . '.php';
+        // this bit of reflection is not very efficient, and a bit
+        // hacky too
         $class = 'HTMLPurifier_URIScheme_' . $scheme;
+        if (!class_exists($class)) include_once $this->_dir . $scheme . '.php';
         if (!class_exists($class)) return $null;
         $this->schemes[$scheme] = new $class();
         return $this->schemes[$scheme];
diff --git a/maintenance/merge-library.php b/maintenance/merge-library.php
new file mode 100644
index 00000000..903da869
--- /dev/null
+++ b/maintenance/merge-library.php
@@ -0,0 +1,190 @@
+#!/usr/bin/php
+<?php
+
+/**
+ * Compiles all of HTML Purifier's library files into one big file
+ * named HTMLPurifier.standalone.php. Operates recursively, and will
+ * barf if there are conditional includes.
+ * 
+ * Details: also creates blank "include" files in the test/blank directory
+ * in order to simulate require_once's inside the test files.
+ */
+
+/**
+ * Global array that tracks already loaded includes
+ */
+$GLOBALS['loaded'] = array('HTMLPurifier.php' => true);
+
+/**
+ * @param $text Text to replace includes from
+ */
+function replace_includes($text) {
+    return preg_replace_callback(
+        "/require_once ['\"]([^'\"]+)['\"];/",
+        'replace_includes_callback',
+        $text
+    );
+}
+
+/**
+ * Removes leading PHP tags from included files. Assumes that there is
+ * no trailing tag.
+ */
+function remove_php_tags($text) {
+    return substr($text, 5);
+}
+
+/**
+ * Creates an appropriate blank file, recursively generating directories
+ * if necessary
+ */
+function create_blank($file) {
+    $dir = dirname($file);
+    $base = realpath('../tests/blanks/') . DIRECTORY_SEPARATOR ;
+    if ($dir != '.') mkdir_deep($base . $dir);
+    file_put_contents($base . $file, '');
+}
+
+/**
+ * Recursively creates a directory
+ * @note Adapted from the PHP manual comment 76612
+ */
+function mkdir_deep($folder) {
+    $folders = preg_split("#[\\\\/]#", $folder);
+    $base = '';
+    for($i = 0, $c = count($folders); $i < $c; $i++) {
+        if(empty($folders[$i])) continue;
+        $base .= $folders[$i];
+        if(!is_dir($base)){
+            mkdir($base);
+        }
+        $base .= DIRECTORY_SEPARATOR;
+    }
+}
+
+/**
+ * Copy a file, or recursively copy a folder and its contents
+ *
+ * @author      Aidan Lister <aidan@php.net>
+ * @version     1.0.1
+ * @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 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 == '.' || $entry == '..') {
+            continue;
+        }
+        // Skip hidden files
+        if ($entry[0] == '.') {
+            continue;
+        }
+        // Deep copy directories
+        if ($dest !== "$source/$entry") {
+            copyr("$source/$entry", "$dest/$entry");
+        }
+    }
+    // Clean up
+    $dir->close();
+    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 == '.' || $entry == '..') {
+            continue;
+        }
+ 
+        // Recurse
+        rmdirr($dirname . DIRECTORY_SEPARATOR . $entry);
+    }
+ 
+    // Clean up
+    $dir->close();
+    return rmdir($dirname);
+}
+
+/**
+ * Copies the contents of a directory to the standalone directory
+ */
+function make_dir_standalone($dir) {
+    return copyr($dir, 'standalone/' . $dir);
+}
+
+function make_file_standalone($file) {
+    mkdir_deep('standalone/' . dirname($file));
+    return copy($file, 'standalone/' . $file);
+}
+
+/**
+ * @param $matches preg_replace_callback matches array, where index 1
+ *        is the filename to include
+ */
+function replace_includes_callback($matches) {
+    $file = $matches[1];
+    if (isset($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'));
+file_put_contents('HTMLPurifier.standalone.php', $contents);
+echo ' done!' . PHP_EOL;
+
+echo 'Creating standalone directory...';
+// rearrange things that can't be put in our standalone file
+$contents = str_replace(
+    "define('HTMLPURIFIER_PREFIX', dirname(__FILE__));",
+    "define('HTMLPURIFIER_PREFIX', dirname(__FILE__) . '/standalone');",
+    $contents
+);
+rmdirr('standalone'); // ensure a clean copy
+mkdir_deep('standalone/HTMLPurifier/DefinitionCache/Serializer');
+make_dir_standalone('HTMLPurifier/EntityLookup');
+make_dir_standalone('HTMLPurifier/Language');
+make_file_standalone('HTMLPurifier/Printer/ConfigForm.js');
+make_file_standalone('HTMLPurifier/Printer/ConfigForm.css');
+make_dir_standalone('HTMLPurifier/URIScheme');
+echo ' done!' . PHP_EOL;
+
diff --git a/tests/HTMLPurifier/DefinitionCache/SerializerTest.php b/tests/HTMLPurifier/DefinitionCache/SerializerTest.php
index 4925ff8e..cf4249af 100644
--- a/tests/HTMLPurifier/DefinitionCache/SerializerTest.php
+++ b/tests/HTMLPurifier/DefinitionCache/SerializerTest.php
@@ -17,8 +17,7 @@ class HTMLPurifier_DefinitionCache_SerializerTest extends HTMLPurifier_Definitio
         $config_md5   = '1.0.0-serial-2';
         
         $file = realpath(
-            $rel_file = dirname(__FILE__) .
-            '/../../../library/HTMLPurifier/DefinitionCache/Serializer/Test/' .
+            $rel_file = HTMLPURIFIER_PREFIX . '/HTMLPurifier/DefinitionCache/Serializer/Test/' .
             $config_md5 . '.ser'
         );
         if($file && file_exists($file)) unlink($file); // prevent previous failures from causing problems
diff --git a/tests/index.php b/tests/index.php
index 134b3b17..d5841c50 100644
--- a/tests/index.php
+++ b/tests/index.php
@@ -38,7 +38,13 @@ if ( is_string($GLOBALS['HTMLPurifierTest']['PEAR']) ) {
 }
 
 // initialize and load HTML Purifier
-require_once '../library/HTMLPurifier.auto.php';
+// use ?standalone to load the alterative standalone stub
+if (isset($_GET['standalone']) || (isset($argv[1]) && $argv[1] == 'standalone')) {
+    set_include_path(realpath('blanks') . PATH_SEPARATOR . get_include_path());
+    require_once '../library/HTMLPurifier.standalone.php';
+} else {
+    require_once '../library/HTMLPurifier.auto.php';
+}
 
 // setup special DefinitionCacheFactory decorator
 $factory =& HTMLPurifier_DefinitionCacheFactory::instance();