[2.1.2] Refactory merge-library.php script

git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1396 48356398-32a2-884e-a903-53898d9a118a

NEWS

@@ -32,6 +32,8 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
   the $config and $context member variables
 . HTML wrapper in DOMLex now takes DTD identifiers into account; doesn't
   really make a difference, but is good for completeness sake
+. merge-library.php script refactored for greater code reusability and
+  PHP4 compatibility
 
 2.1.1, released 2007-08-04
 - Fix show-stopper bug in %URI.MakeAbsolute functionality

maintenance/common.php

@@ -1,5 +1,7 @@
 <?php
 
+require_once 'compat-function-file-put-contents.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,
@@ -7,3 +9,135 @@ 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);
+    }
+    
+    
+}
+
+

maintenance/compat-function-file-put-contents.php

@@ -0,0 +1,107 @@
+<?php
+// $Id: file_put_contents.php,v 1.27 2007/04/17 10:09:56 arpad Exp $
+
+
+if (!defined('FILE_USE_INCLUDE_PATH')) {
+    define('FILE_USE_INCLUDE_PATH', 1);
+}
+
+if (!defined('LOCK_EX')) {
+    define('LOCK_EX', 2);
+}
+
+if (!defined('FILE_APPEND')) {
+    define('FILE_APPEND', 8);
+}
+
+
+/**
+ * Replace file_put_contents()
+ *
+ * @category    PHP
+ * @package     PHP_Compat
+ * @license     LGPL - http://www.gnu.org/licenses/lgpl.html
+ * @copyright   2004-2007 Aidan Lister <aidan@php.net>, Arpad Ray <arpad@php.net>
+ * @link        http://php.net/function.file_put_contents
+ * @author      Aidan Lister <aidan@php.net>
+ * @version     $Revision: 1.27 $
+ * @internal    resource_context is not supported
+ * @since       PHP 5
+ * @require     PHP 4.0.0 (user_error)
+ */
+function php_compat_file_put_contents($filename, $content, $flags = null, $resource_context = null)
+{
+    // If $content is an array, convert it to a string
+    if (is_array($content)) {
+        $content = implode('', $content);
+    }
+
+    // If we don't have a string, throw an error
+    if (!is_scalar($content)) {
+        user_error('file_put_contents() The 2nd parameter should be either a string or an array',
+            E_USER_WARNING);
+        return false;
+    }
+
+    // Get the length of data to write
+    $length = strlen($content);
+
+    // Check what mode we are using
+    $mode = ($flags & FILE_APPEND) ?
+                'a' :
+                'wb';
+
+    // Check if we're using the include path
+    $use_inc_path = ($flags & FILE_USE_INCLUDE_PATH) ?
+                true :
+                false;
+
+    // Open the file for writing
+    if (($fh = @fopen($filename, $mode, $use_inc_path)) === false) {
+        user_error('file_put_contents() failed to open stream: Permission denied',
+            E_USER_WARNING);
+        return false;
+    }
+
+    // Attempt to get an exclusive lock
+    $use_lock = ($flags & LOCK_EX) ? true : false ;
+    if ($use_lock === true) {
+        if (!flock($fh, LOCK_EX)) {
+            return false;
+        }
+    }
+
+    // Write to the file
+    $bytes = 0;
+    if (($bytes = @fwrite($fh, $content)) === false) {
+        $errormsg = sprintf('file_put_contents() Failed to write %d bytes to %s',
+                        $length,
+                        $filename);
+        user_error($errormsg, E_USER_WARNING);
+        return false;
+    }
+
+    // Close the handle
+    @fclose($fh);
+
+    // Check all the data was written
+    if ($bytes != $length) {
+        $errormsg = sprintf('file_put_contents() Only %d of %d bytes written, possibly out of free disk space.',
+                        $bytes,
+                        $length);
+        user_error($errormsg, E_USER_WARNING);
+        return false;
+    }
+
+    // Return length
+    return $bytes;
+}
+
+
+// Define
+if (!function_exists('file_put_contents')) {
+    function file_put_contents($filename, $content, $flags = null, $resource_context = null)
+    {
+        return php_compat_file_put_contents($filename, $content, $flags, $resource_context);
+    }
+}

maintenance/merge-library.php

@@ -19,7 +19,29 @@ assertCli();
 $GLOBALS['loaded'] = array('HTMLPurifier.php' => true);
 
 /**
- * @param $text Text to replace includes from
+ * Custom FSTools for this script that overloads some behavior
+ * @warning The overloading of copy() is not necessarily global for
+ *          this script. Watch out!
+ */
+class MergeLibraryFSTools extends FSTools
+{
+    function copyable($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(
@@ -32,6 +54,8 @@ function replace_includes($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($text, 5);
@@ -40,128 +64,44 @@ function remove_php_tags($text) {
 /**
  * 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 != '.') mkdir_deep($base . $dir);
+    if ($dir != '.') {
+        $FS->mkdir($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[$i])) {
-            if (!$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
- *
- * @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_and_remove_includes($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
+ * @param string $dir Directory to copy
  */
 function make_dir_standalone($dir) {
-    return copyr($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) {
-    mkdir_deep('standalone/' . dirname($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);
@@ -174,8 +114,13 @@ function copy_and_remove_includes($file, $sfile) {
  */
 function replace_includes_callback($matches) {
     $file = $matches[1];
-    // PHP 5 only file / PEAR files
+    $preserve = array(
-    $preserve = array('HTMLPurifier/Lexer/DOMLex.php'=>1, 'HTMLPurifier/Printer.php'=>1, 'XML/HTMLSax3.php'=>1);
+      // PHP 5 only
+      'HTMLPurifier/Lexer/DOMLex.php' => 1,
+      'HTMLPurifier/Printer.php' => 1,
+      // PEAR (external)
+      'XML/HTMLSax3.php' => 1
+    );
     if (isset($preserve[$file])) {
         return $matches[0];
     }
@@ -200,16 +145,22 @@ file_put_contents('HTMLPurifier.standalone.php', $contents);
 echo ' done!' . PHP_EOL;
 
 echo 'Creating standalone directory...';
-rmdirr('standalone'); // ensure a clean copy
+$FS->rmdirr('standalone'); // ensure a clean copy
-mkdir_deep('standalone/HTMLPurifier/DefinitionCache/Serializer');
+
+// data files
+$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'); // not incl by default
+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;
-