maintenance/config-scanner.php

   1 #!/usr/bin/php
   2 <?php
   3
   4 chdir(dirname(__FILE__));
   5 require_once 'common.php';
   6 require_once '../library/HTMLPurifier.auto.php';
   7 assertCli();
   8
   9 if (version_compare(PHP_VERSION, '5.2.2', '<')) {
  10     echo "This script requires PHP 5.2.2 or later, for tokenizer line numbers.";
  11     exit(1);
  12 }
  13
  14 /**
  15  * @file
  16  * Scans HTML Purifier source code for $config tokens and records the
  17  * directive being used; configdoc can use this info later.
  18  *
  19  * Currently, this just dumps all the info onto the console. Eventually, it
  20  * will create an XML file that our XSLT transform can use.
  21  */
  22
  23 $FS = new FSTools();
  24 chdir(dirname(__FILE__) . '/../library/');
  25 $raw_files = $FS->globr('.', '*.php');
  26 $files = array();
  27 foreach ($raw_files as $file) {
  28     $file = substr($file, 2); // rm leading './'
  29     if (strncmp('standalone/', $file, 11) === 0) continue; // rm generated files
  30     if (substr_count($file, '.') > 1) continue; // rm meta files
  31     $files[] = $file;
  32 }
  33
  34 /**
  35  * Moves the $i cursor to the next non-whitespace token
  36  */
  37 function consumeWhitespace($tokens, &$i)
  38 {
  39     do {$i++;} while (is_array($tokens[$i]) && $tokens[$i][0] === T_WHITESPACE);
  40 }
  41
  42 /**
  43  * Tests whether or not a token is a particular type. There are three run-cases:
  44  *      - ($token, $expect_token): tests if the token is $expect_token type;
  45  *      - ($token, $expect_value): tests if the token is the string $expect_value;
  46  *      - ($token, $expect_token, $expect_value): tests if token is $expect_token type, and
  47  *        its string representation is $expect_value
  48  */
  49 function testToken($token, $value_or_token, $value = null)
  50 {
  51     if (is_null($value)) {
  52         if (is_int($value_or_token)) return is_array($token) && $token[0] === $value_or_token;
  53         else return $token === $value_or_token;
  54     } else {
  55         return is_array($token) && $token[0] === $value_or_token && $token[1] === $value;
  56     }
  57 }
  58
  59 $counter = 0;
  60 $full_counter = 0;
  61 $tracker = array();
  62
  63 foreach ($files as $file) {
  64     $tokens = token_get_all(file_get_contents($file));
  65     $file = str_replace('\\', '/', $file);
  66     for ($i = 0, $c = count($tokens); $i < $c; $i++) {
  67         $ok = false;
  68         // Match $config
  69         if (!$ok && testToken($tokens[$i], T_VARIABLE, '$config')) $ok = true;
  70         // Match $this->config
  71         while (!$ok && testToken($tokens[$i], T_VARIABLE, '$this')) {
  72             consumeWhitespace($tokens, $i);
  73             if (!testToken($tokens[$i], T_OBJECT_OPERATOR)) break;
  74             consumeWhitespace($tokens, $i);
  75             if (testToken($tokens[$i], T_STRING, 'config')) $ok = true;
  76             break;
  77         }
  78         if (!$ok) continue;
  79
  80         $ok = false;
  81         for($i++; $i < $c; $i++) {
  82             if ($tokens[$i] === ',' || $tokens[$i] === ')' || $tokens[$i] === ';') {
  83                 break;
  84             }
  85             if (is_string($tokens[$i])) continue;
  86             if ($tokens[$i][0] === T_OBJECT_OPERATOR) {
  87                 $ok = true;
  88                 break;
  89             }
  90         }
  91         if (!$ok) continue;
  92
  93         $line = $tokens[$i][2];
  94
  95         consumeWhitespace($tokens, $i);
  96         if (!testToken($tokens[$i], T_STRING, 'get')) continue;
  97
  98         consumeWhitespace($tokens, $i);
  99         if (!testToken($tokens[$i], '(')) continue;
 100
 101         $full_counter++;
 102
 103         $matched = false;
 104         do {
 105
 106             // What we currently don't match are batch retrievals, and
 107             // wildcard retrievals. This data might be useful in the future,
 108             // which is why we have a do {} while loop that doesn't actually
 109             // do anything.
 110
 111             consumeWhitespace($tokens, $i);
 112             if (!testToken($tokens[$i], T_CONSTANT_ENCAPSED_STRING)) continue;
 113             $id = substr($tokens[$i][1], 1, -1);
 114
 115             $counter++;
 116             $matched = true;
 117
 118             if (!isset($tracker[$id])) $tracker[$id] = array();
 119             if (!isset($tracker[$id][$file])) $tracker[$id][$file] = array();
 120             $tracker[$id][$file][] = $line;
 121
 122         } while (0);
 123
 124         //echo "$file:$line uses $namespace.$directive\n";
 125     }
 126 }
 127
 128 echo "\n$counter/$full_counter instances of \$config or \$this->config found in source code.\n";
 129
 130 echo "Generating XML... ";
 131
 132 $xw = new XMLWriter();
 133 $xw->openURI('../configdoc/usage.xml');
 134 $xw->setIndent(true);
 135 $xw->startDocument('1.0', 'UTF-8');
 136 $xw->startElement('usage');
 137 foreach ($tracker as $id => $files) {
 138     $xw->startElement('directive');
 139     $xw->writeAttribute('id', $id);
 140     foreach ($files as $file => $lines) {
 141         $xw->startElement('file');
 142         $xw->writeAttribute('name', $file);
 143         foreach ($lines as $line) {
 144             $xw->writeElement('line', $line);
 145         }
 146         $xw->endElement();
 147     }
 148     $xw->endElement();
 149 }
 150 $xw->endElement();
 151 $xw->flush();
 152
 153 echo "done!\n";
 154
 155 // vim: et sw=4 sts=4