[Summary view]

Differences Between: [Versions 310 and 402] [Versions 311 and 402] [Versions 39 and 402] [Versions 400 and 402] [Versions 401 and 402]

   1  <?php
   2  
   3  /**
   4   * Abstract minifier class.
   5   *
   6   * Please report bugs on https://github.com/matthiasmullie/minify/issues
   7   *
   8   * @author Matthias Mullie <minify@mullie.eu>
   9   * @copyright Copyright (c) 2012, Matthias Mullie. All rights reserved
  10   * @license MIT License
  11   */
  12  
  13  namespace MatthiasMullie\Minify;
  14  
  15  use MatthiasMullie\Minify\Exceptions\IOException;
  16  use Psr\Cache\CacheItemInterface;
  17  
  18  /**
  19   * Abstract minifier class.
  20   *
  21   * Please report bugs on https://github.com/matthiasmullie/minify/issues
  22   *
  23   * @author Matthias Mullie <minify@mullie.eu>
  24   * @copyright Copyright (c) 2012, Matthias Mullie. All rights reserved
  25   * @license MIT License
  26   */
  27  abstract class Minify
  28  {
  29      /**
  30       * The data to be minified.
  31       *
  32       * @var string[]
  33       */
  34      protected $data = array();
  35  
  36      /**
  37       * Array of patterns to match.
  38       *
  39       * @var string[]
  40       */
  41      protected $patterns = array();
  42  
  43      /**
  44       * This array will hold content of strings and regular expressions that have
  45       * been extracted from the JS source code, so we can reliably match "code",
  46       * without having to worry about potential "code-like" characters inside.
  47       *
  48       * @internal
  49       *
  50       * @var string[]
  51       */
  52      public $extracted = array();
  53  
  54      /**
  55       * Init the minify class - optionally, code may be passed along already.
  56       */
  57      public function __construct(/* $data = null, ... */)
  58      {
  59          // it's possible to add the source through the constructor as well ;)
  60          if (func_num_args()) {
  61              call_user_func_array(array($this, 'add'), func_get_args());
  62          }
  63      }
  64  
  65      /**
  66       * Add a file or straight-up code to be minified.
  67       *
  68       * @param string|string[] $data
  69       *
  70       * @return static
  71       */
  72      public function add($data /* $data = null, ... */)
  73      {
  74          // bogus "usage" of parameter $data: scrutinizer warns this variable is
  75          // not used (we're using func_get_args instead to support overloading),
  76          // but it still needs to be defined because it makes no sense to have
  77          // this function without argument :)
  78          $args = array($data) + func_get_args();
  79  
  80          // this method can be overloaded
  81          foreach ($args as $data) {
  82              if (is_array($data)) {
  83                  call_user_func_array(array($this, 'add'), $data);
  84                  continue;
  85              }
  86  
  87              // redefine var
  88              $data = (string) $data;
  89  
  90              // load data
  91              $value = $this->load($data);
  92              $key = ($data != $value) ? $data : count($this->data);
  93  
  94              // replace CR linefeeds etc.
  95              // @see https://github.com/matthiasmullie/minify/pull/139
  96              $value = str_replace(array("\r\n", "\r"), "\n", $value);
  97  
  98              // store data
  99              $this->data[$key] = $value;
 100          }
 101  
 102          return $this;
 103      }
 104  
 105      /**
 106       * Add a file to be minified.
 107       *
 108       * @param string|string[] $data
 109       *
 110       * @return static
 111       *
 112       * @throws IOException
 113       */
 114      public function addFile($data /* $data = null, ... */)
 115      {
 116          // bogus "usage" of parameter $data: scrutinizer warns this variable is
 117          // not used (we're using func_get_args instead to support overloading),
 118          // but it still needs to be defined because it makes no sense to have
 119          // this function without argument :)
 120          $args = array($data) + func_get_args();
 121  
 122          // this method can be overloaded
 123          foreach ($args as $path) {
 124              if (is_array($path)) {
 125                  call_user_func_array(array($this, 'addFile'), $path);
 126                  continue;
 127              }
 128  
 129              // redefine var
 130              $path = (string) $path;
 131  
 132              // check if we can read the file
 133              if (!$this->canImportFile($path)) {
 134                  throw new IOException('The file "' . $path . '" could not be opened for reading. Check if PHP has enough permissions.');
 135              }
 136  
 137              $this->add($path);
 138          }
 139  
 140          return $this;
 141      }
 142  
 143      /**
 144       * Minify the data & (optionally) saves it to a file.
 145       *
 146       * @param string[optional] $path Path to write the data to
 147       *
 148       * @return string The minified data
 149       */
 150      public function minify($path = null)
 151      {
 152          $content = $this->execute($path);
 153  
 154          // save to path
 155          if ($path !== null) {
 156              $this->save($content, $path);
 157          }
 158  
 159          return $content;
 160      }
 161  
 162      /**
 163       * Minify & gzip the data & (optionally) saves it to a file.
 164       *
 165       * @param string[optional] $path  Path to write the data to
 166       * @param int[optional]    $level Compression level, from 0 to 9
 167       *
 168       * @return string The minified & gzipped data
 169       */
 170      public function gzip($path = null, $level = 9)
 171      {
 172          $content = $this->execute($path);
 173          $content = gzencode($content, $level, FORCE_GZIP);
 174  
 175          // save to path
 176          if ($path !== null) {
 177              $this->save($content, $path);
 178          }
 179  
 180          return $content;
 181      }
 182  
 183      /**
 184       * Minify the data & write it to a CacheItemInterface object.
 185       *
 186       * @param CacheItemInterface $item Cache item to write the data to
 187       *
 188       * @return CacheItemInterface Cache item with the minifier data
 189       */
 190      public function cache(CacheItemInterface $item)
 191      {
 192          $content = $this->execute();
 193          $item->set($content);
 194  
 195          return $item;
 196      }
 197  
 198      /**
 199       * Minify the data.
 200       *
 201       * @param string[optional] $path Path to write the data to
 202       *
 203       * @return string The minified data
 204       */
 205      abstract public function execute($path = null);
 206  
 207      /**
 208       * Load data.
 209       *
 210       * @param string $data Either a path to a file or the content itself
 211       *
 212       * @return string
 213       */
 214      protected function load($data)
 215      {
 216          // check if the data is a file
 217          if ($this->canImportFile($data)) {
 218              $data = file_get_contents($data);
 219  
 220              // strip BOM, if any
 221              if (substr($data, 0, 3) == "\xef\xbb\xbf") {
 222                  $data = substr($data, 3);
 223              }
 224          }
 225  
 226          return $data;
 227      }
 228  
 229      /**
 230       * Save to file.
 231       *
 232       * @param string $content The minified data
 233       * @param string $path    The path to save the minified data to
 234       *
 235       * @throws IOException
 236       */
 237      protected function save($content, $path)
 238      {
 239          $handler = $this->openFileForWriting($path);
 240  
 241          $this->writeToFile($handler, $content);
 242  
 243          @fclose($handler);
 244      }
 245  
 246      /**
 247       * Register a pattern to execute against the source content.
 248       *
 249       * If $replacement is a string, it must be plain text. Placeholders like $1 or \2 don't work.
 250       * If you need that functionality, use a callback instead.
 251       *
 252       * @param string          $pattern     PCRE pattern
 253       * @param string|callable $replacement Replacement value for matched pattern
 254       */
 255      protected function registerPattern($pattern, $replacement = '')
 256      {
 257          // study the pattern, we'll execute it more than once
 258          $pattern .= 'S';
 259  
 260          $this->patterns[] = array($pattern, $replacement);
 261      }
 262  
 263      /**
 264       * We can't "just" run some regular expressions against JavaScript: it's a
 265       * complex language. E.g. having an occurrence of // xyz would be a comment,
 266       * unless it's used within a string. Of you could have something that looks
 267       * like a 'string', but inside a comment.
 268       * The only way to accurately replace these pieces is to traverse the JS one
 269       * character at a time and try to find whatever starts first.
 270       *
 271       * @param string $content The content to replace patterns in
 272       *
 273       * @return string The (manipulated) content
 274       */
 275      protected function replace($content)
 276      {
 277          $contentLength = strlen($content);
 278          $output = '';
 279          $processedOffset = 0;
 280          $positions = array_fill(0, count($this->patterns), -1);
 281          $matches = array();
 282  
 283          while ($processedOffset < $contentLength) {
 284              // find first match for all patterns
 285              foreach ($this->patterns as $i => $pattern) {
 286                  list($pattern, $replacement) = $pattern;
 287  
 288                  // we can safely ignore patterns for positions we've unset earlier,
 289                  // because we know these won't show up anymore
 290                  if (array_key_exists($i, $positions) == false) {
 291                      continue;
 292                  }
 293  
 294                  // no need to re-run matches that are still in the part of the
 295                  // content that hasn't been processed
 296                  if ($positions[$i] >= $processedOffset) {
 297                      continue;
 298                  }
 299  
 300                  $match = null;
 301                  if (preg_match($pattern, $content, $match, PREG_OFFSET_CAPTURE, $processedOffset)) {
 302                      $matches[$i] = $match;
 303  
 304                      // we'll store the match position as well; that way, we
 305                      // don't have to redo all preg_matches after changing only
 306                      // the first (we'll still know where those others are)
 307                      $positions[$i] = $match[0][1];
 308                  } else {
 309                      // if the pattern couldn't be matched, there's no point in
 310                      // executing it again in later runs on this same content;
 311                      // ignore this one until we reach end of content
 312                      unset($matches[$i], $positions[$i]);
 313                  }
 314              }
 315  
 316              // no more matches to find: everything's been processed, break out
 317              if (!$matches) {
 318                  // output the remaining content
 319                  $output .= substr($content, $processedOffset);
 320                  break;
 321              }
 322  
 323              // see which of the patterns actually found the first thing (we'll
 324              // only want to execute that one, since we're unsure if what the
 325              // other found was not inside what the first found)
 326              $matchOffset = min($positions);
 327              $firstPattern = array_search($matchOffset, $positions);
 328              $match = $matches[$firstPattern];
 329  
 330              // execute the pattern that matches earliest in the content string
 331              list(, $replacement) = $this->patterns[$firstPattern];
 332  
 333              // add the part of the input between $processedOffset and the first match;
 334              // that content wasn't matched by anything
 335              $output .= substr($content, $processedOffset, $matchOffset - $processedOffset);
 336              // add the replacement for the match
 337              $output .= $this->executeReplacement($replacement, $match);
 338              // advance $processedOffset past the match
 339              $processedOffset = $matchOffset + strlen($match[0][0]);
 340          }
 341  
 342          return $output;
 343      }
 344  
 345      /**
 346       * If $replacement is a callback, execute it, passing in the match data.
 347       * If it's a string, just pass it through.
 348       *
 349       * @param string|callable $replacement Replacement value
 350       * @param array           $match       Match data, in PREG_OFFSET_CAPTURE form
 351       *
 352       * @return string
 353       */
 354      protected function executeReplacement($replacement, $match)
 355      {
 356          if (!is_callable($replacement)) {
 357              return $replacement;
 358          }
 359          // convert $match from the PREG_OFFSET_CAPTURE form to the form the callback expects
 360          foreach ($match as &$matchItem) {
 361              $matchItem = $matchItem[0];
 362          }
 363  
 364          return $replacement($match);
 365      }
 366  
 367      /**
 368       * Strings are a pattern we need to match, in order to ignore potential
 369       * code-like content inside them, but we just want all of the string
 370       * content to remain untouched.
 371       *
 372       * This method will replace all string content with simple STRING#
 373       * placeholder text, so we've rid all strings from characters that may be
 374       * misinterpreted. Original string content will be saved in $this->extracted
 375       * and after doing all other minifying, we can restore the original content
 376       * via restoreStrings().
 377       *
 378       * @param string[optional] $chars
 379       * @param string[optional] $placeholderPrefix
 380       */
 381      protected function extractStrings($chars = '\'"', $placeholderPrefix = '')
 382      {
 383          // PHP only supports $this inside anonymous functions since 5.4
 384          $minifier = $this;
 385          $callback = function ($match) use ($minifier, $placeholderPrefix) {
 386              // check the second index here, because the first always contains a quote
 387              if ($match[2] === '') {
 388                  /*
 389                   * Empty strings need no placeholder; they can't be confused for
 390                   * anything else anyway.
 391                   * But we still needed to match them, for the extraction routine
 392                   * to skip over this particular string.
 393                   */
 394                  return $match[0];
 395              }
 396  
 397              $count = count($minifier->extracted);
 398              $placeholder = $match[1] . $placeholderPrefix . $count . $match[1];
 399              $minifier->extracted[$placeholder] = $match[1] . $match[2] . $match[1];
 400  
 401              return $placeholder;
 402          };
 403  
 404          /*
 405           * The \\ messiness explained:
 406           * * Don't count ' or " as end-of-string if it's escaped (has backslash
 407           * in front of it)
 408           * * Unless... that backslash itself is escaped (another leading slash),
 409           * in which case it's no longer escaping the ' or "
 410           * * So there can be either no backslash, or an even number
 411           * * multiply all of that times 4, to account for the escaping that has
 412           * to be done to pass the backslash into the PHP string without it being
 413           * considered as escape-char (times 2) and to get it in the regex,
 414           * escaped (times 2)
 415           */
 416          $this->registerPattern('/([' . $chars . '])(.*?(?<!\\\\)(\\\\\\\\)*+)\\1/s', $callback);
 417      }
 418  
 419      /**
 420       * This method will restore all extracted data (strings, regexes) that were
 421       * replaced with placeholder text in extract*(). The original content was
 422       * saved in $this->extracted.
 423       *
 424       * @param string $content
 425       *
 426       * @return string
 427       */
 428      protected function restoreExtractedData($content)
 429      {
 430          if (!$this->extracted) {
 431              // nothing was extracted, nothing to restore
 432              return $content;
 433          }
 434  
 435          $content = strtr($content, $this->extracted);
 436  
 437          $this->extracted = array();
 438  
 439          return $content;
 440      }
 441  
 442      /**
 443       * Check if the path is a regular file and can be read.
 444       *
 445       * @param string $path
 446       *
 447       * @return bool
 448       */
 449      protected function canImportFile($path)
 450      {
 451          $parsed = parse_url($path);
 452          if (
 453              // file is elsewhere
 454              isset($parsed['host']) ||
 455              // file responds to queries (may change, or need to bypass cache)
 456              isset($parsed['query'])
 457          ) {
 458              return false;
 459          }
 460  
 461          return strlen($path) < PHP_MAXPATHLEN && @is_file($path) && is_readable($path);
 462      }
 463  
 464      /**
 465       * Attempts to open file specified by $path for writing.
 466       *
 467       * @param string $path The path to the file
 468       *
 469       * @return resource Specifier for the target file
 470       *
 471       * @throws IOException
 472       */
 473      protected function openFileForWriting($path)
 474      {
 475          if ($path === '' || ($handler = @fopen($path, 'w')) === false) {
 476              throw new IOException('The file "' . $path . '" could not be opened for writing. Check if PHP has enough permissions.');
 477          }
 478  
 479          return $handler;
 480      }
 481  
 482      /**
 483       * Attempts to write $content to the file specified by $handler. $path is used for printing exceptions.
 484       *
 485       * @param resource $handler The resource to write to
 486       * @param string   $content The content to write
 487       * @param string   $path    The path to the file (for exception printing only)
 488       *
 489       * @throws IOException
 490       */
 491      protected function writeToFile($handler, $content, $path = '')
 492      {
 493          if (
 494              !is_resource($handler) ||
 495              ($result = @fwrite($handler, $content)) === false ||
 496              ($result < strlen($content))
 497          ) {
 498              throw new IOException('The file "' . $path . '" could not be written to. Check your disk space and file permissions.');
 499          }
 500      }
 501  
 502      protected static function str_replace_first($search, $replace, $subject)
 503      {
 504          $pos = strpos($subject, $search);
 505          if ($pos !== false) {
 506              return substr_replace($subject, $replace, $pos, strlen($search));
 507          }
 508  
 509          return $subject;
 510      }
 511  }