HtmlMin - Code Metrics - Inspection of "[+]: fixed html-encoding" - voku/HtmlMin - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( c927d1...313909 )

by Lars

created 2016-10-17 15:28 UTC

HtmlMin D

↳ Parent: Project

Complexity

Total Complexity

Size/Duplication

Total Lines	556
Duplicated Lines	0 %

Coupling/Cohesion

Components	1
Dependencies	4

Test Coverage

Coverage

82.63%

Importance

Changes

Metric	Value
wmc	82
lcom	1
cbo	4
dl	0
loc	556
ccs	176
cts	213
cp	0.8263
rs	4.8717
c	0
b	0
f	0

11 Methods

Rating	Name	Size	Complexity
A	__construct()	8	1
B	minify()	103	5
C	optimizeAttributes()	45	13
C	optimizeAttributesFilters()	54	32
A	optimizeAttributesInDom()	8	2
B	protectTagsInDom()	31	6
A	removeCommentsInDom()	14	3
B	removeWhitespaceInDom()	26	5
A	restoreProtectedHtml()	13	3
B	sortCssClasses()	23	6
B	trimTagsInDom()	30	6

How to fix Complexity

<?php

namespace voku\helper;

/**
 * Class HtmlMin
 *
 * Inspired by:
 * - JS: https://github.com/kangax/html-minifier/blob/gh-pages/src/htmlminifier.js
 * - PHP: https://github.com/searchturbine/phpwee-php-minifier
 * - PHP: https://github.com/WyriHaximus/HtmlCompress
 * - PHP: https://github.com/zaininnari/html-minifier
 * - Java: https://code.google.com/archive/p/htmlcompressor/
 *
 * @package voku\helper
 */
class HtmlMin
{
  /**
   * // https://mathiasbynens.be/demo/javascript-mime-type
   * // https://developer.mozilla.org/en/docs/Web/HTML/Element/script#attr-type
   *
   * @var array
   */
  private static $executableScriptsMimeTypes = array(
      'text/javascript'          => '',
      'text/ecmascript'          => '',
      'text/jscript'             => '',
      'application/javascript'   => '',
      'application/x-javascript' => '',
      'application/ecmascript'   => '',
  );

  private static $selfClosingTags = array(
      'area',
      'base',
      'br',
      'col',
      'command',
      'embed',
      'hr',
      'img',
      'input',
      'keygen',
      'link',
      'meta',
      'param',
      'source',
      'track',
      'wbr',
  );


  /**
   * @var array
   */
  private static $booleanAttributes = array(
      'allowfullscreen' => '',
      'async'           => '',
      'autofocus'       => '',
      'autoplay'        => '',
      'checked'         => '',
      'compact'         => '',
      'controls'        => '',
      'declare'         => '',
      'default'         => '',
      'defaultchecked'  => '',
      'defaultmuted'    => '',
      'defaultselected' => '',
      'defer'           => '',
      'disabled'        => '',
      'enabled'         => '',
      'formnovalidate'  => '',
      'hidden'          => '',
      'indeterminate'   => '',
      'inert'           => '',
      'ismap'           => '',
      'itemscope'       => '',
      'loop'            => '',
      'multiple'        => '',
      'muted'           => '',
      'nohref'          => '',
      'noresize'        => '',
      'noshade'         => '',
      'novalidate'      => '',
      'nowrap'          => '',
      'open'            => '',
      'pauseonexit'     => '',
      'readonly'        => '',
      'required'        => '',
      'reversed'        => '',
      'scoped'          => '',
      'seamless'        => '',
      'selected'        => '',
      'sortable'        => '',
      'truespeed'       => '',
      'typemustmatch'   => '',
      'visible'         => '',
  );
  /**
   * @var array
   */
  private static $skipTagsForRemoveWhitespace = array(
      'style',
      'pre',
      'code',
      'script',
      'textarea',
  );

  /**
   * An random md5-hash, generated via "random_bytes()".
   *
   * @var string
   */
  private $randomHash;

  /**
   * @var array
   */
  private $protectedChildNodes;

  /**
   * @var string
   */
  private $protectedChildNodesHelper;

  /**
   * @var string
   */
  private $booleanAttributesHelper;

  /**
   * HtmlMin constructor.
   */
  public function __construct()
  {
    $this->protectedChildNodes = array();
    $this->randomHash = md5(Bootup::get_random_bytes(16));

    $this->protectedChildNodesHelper = 'html-min--saved-content-' . $this->randomHash;
    $this->booleanAttributesHelper = 'html-min--delete-this-' . $this->randomHash;
  }

  /**
   * @param string $html
   *
   * @return string
   */
  public function minify($html)
  {
    $html = (string)$html;
    if (!isset($html[0])) {
      return '';
    }

    $html = trim($html);
    if (!$html) {
      return '';
    }

    // init
    $this->protectedChildNodes = array();
    $origHtml = $html;
    $origHtmlLength = UTF8::strlen($html);

    $dom = new HtmlDomParser();
    $dom->getDocument()->preserveWhiteSpace = false; // remove redundant white space
    $dom->getDocument()->formatOutput = false; // do not formats output with indentation

    $dom->loadHtml($html);

    $dom = $this->protectTagsInDom($dom);
    $dom = $this->optimizeAttributesInDom($dom);
    $dom = $this->removeCommentsInDom($dom);
    $dom = $this->removeWhitespaceInDom($dom);
    $dom = $this->trimTagsInDom($dom);

    $html = $dom->html();

    // -------------------------------------------------------------------------
    // Trim whitespace from html-string. [protected html is still protected]
    // -------------------------------------------------------------------------

    // Remove spaces that are followed by either > or <
    $html = preg_replace('/ (>)/', '$1', $html);
    // Remove spaces that are preceded by either > or <
    $html = preg_replace('/(<) /', '$1', $html);
    // Remove spaces that are between > and <
    $html = preg_replace('/(>) (<)/', '>$2', $html);

    // -------------------------------------------------------------------------
    // Restore protected HTML-code.
    // -------------------------------------------------------------------------

    $html = preg_replace_callback(
        '/<(?<element>' . $this->protectedChildNodesHelper . ')(?<attributes> [^>]*)?>(?<value>.*?)<\/' . $this->protectedChildNodesHelper . '>/',
        array($this, 'restoreProtectedHtml'),
        $html
    );
    $html = $dom::putReplacedBackToPreserveHtmlEntities($html);

    // ------------------------------------
    // final clean-up
    // ------------------------------------

    $html = UTF8::cleanup($html);

    $html = str_replace(
        array(
            'html>' . "\n",
            "\n" . '<html',
            '<!doctype',
            '="' . $this->booleanAttributesHelper . '"',
            '</' . $this->protectedChildNodesHelper . '>',
        ),
        array(
            'html>',
            '<html',
            '<!DOCTYPE',
            '',
            '',
        ),
        $html
    );

    $html = preg_replace(
        array(
            '/<(?:' . $this->protectedChildNodesHelper . ')(:? [^>]*)?>/',
        ),
        array(
            '',
        ),
        $html
    );

    static $cacheSelfClosingTags = null;
    if ($cacheSelfClosingTags === null) {
      $cacheSelfClosingTags = implode('|', self::$selfClosingTags);
    }
    $html = preg_replace('#<\b(' . $cacheSelfClosingTags . ')([^>]+)><\/\b\1>#', '<\\1\\2/>', $html);

    // ------------------------------------
    // check if compression worked
    // ------------------------------------

    if ($origHtmlLength < UTF8::strlen($html)) {
      $html = $origHtml;
    }

    return $html;
  }

  /**
   * Sort HTML-Attributes, so that gzip can do better work
   *  and remove some default attributes.
   *
   * @param SimpleHtmlDom $element
   *
   * @return bool
   */
  private function optimizeAttributes(SimpleHtmlDom $element)
  {
    $attributs = $element->getAllAttributes();
    if ($attributs === null) {
      return false;
    }

    $attrs = array();
    foreach ((array)$attributs as $attrName => $attrValue) {

      if (isset(self::$booleanAttributes[$attrName])) {
        $attrs[$attrName] = $this->booleanAttributesHelper;
        $element->{$attrName} = null;
        continue;
      }

      if (
          ($attrName === 'href' || $attrName === 'src' || $attrName === 'action')
          &&
          !(isset($attributs['rel']) && $attributs['rel'] === 'external')
          &&
          !(isset($attributs['target']) && $attributs['target'] === '_blank')
      ) {
        $attrValue = str_replace('http://', '//', $attrValue);
      }

      if ($this->optimizeAttributesFilters($element->tag, $attrName, $attrValue, $attributs)) {
function acceptsInteger($int) { }

$x = '123'; // string "123"

// Instead of
acceptsInteger($x);

// we recommend to use
acceptsInteger((integer) $x);
        $element->{$attrName} = null;
        continue;
      }

      $attrValue = $this->sortCssClasses($attrName, $attrValue);

      $attrs[$attrName] = $attrValue;
      $element->{$attrName} = null;
    }

    ksort($attrs);
    foreach ($attrs as $attrName => $attrValue) {
      $attrValue = HtmlDomParser::replaceToPreserveHtmlEntities($attrValue);

      $element->setAttribute($attrName, $attrValue, true);
    }

    return true;
  }

  /**
   * Check if the attribute (key / value) is default and can be skipped.
   *
   * @param string $tag
   * @param string $attrName
   * @param string $attrValue
   * @param string $allAttr
   *
   * @return bool
   */
  private function optimizeAttributesFilters($tag, $attrName, $attrValue, $allAttr)
  {
    // remove default
    if ($tag === 'script' && $attrName === 'language' && $attrValue === 'javascript') {
      return true;
    }

    // remove default
    if ($tag === 'form' && $attrName === 'method' && $attrValue === 'get') {
      return true;
    }

    // remove default
    if ($tag === 'input' && $attrName === 'type' && $attrValue === 'text') {
      return true;
    }

    // remove default
    if ($tag === 'area' && $attrName === 'shape' && $attrValue === 'rect') {
      return true;
    }

    // remove deprecated charset-attribute (the Browser will use the charset from the HTTP-Header, anyway)
    if ($tag === 'script' && $attrName === 'charset' && !isset($allAttr['src'])) {
      return true;
    }

    // remove deprecated anchor-jump
    if ($tag === 'a' && $attrName === 'name' && isset($allAttr['id'])) {
      return true;
    }

    // remove "type=text/css" for css links
    if ($tag === 'link' && $attrName === 'type' && $attrValue === 'text/css' && isset($allAttr['rel']) && $allAttr['rel'] === 'stylesheet') {
      return true;
    }

    // remove deprecated script-mime-types
    if ($tag === 'script' && $attrName === 'type' && isset($allAttr['src'], self::$executableScriptsMimeTypes[$attrValue])) {
      return true;
    }

    // remove empty value from <input>
    if ($tag === 'input' && $attrName === 'value' && $attrValue === '') {
      return true;
    }

    // remove some empty attribute
    if ($attrValue === '' && preg_match('/^(?:class|id|style|title|lang|dir|on(?:focus|blur|change|click|dblclick|mouse(?:down|up|over|move|out)|key(?:press|down|up)))$/', $attrName)) {
      return true;
    }

    return false;
  }

  /**
   * Optimize HTML-tag attributes in the dom.
   *
   * @param HtmlDomParser $dom
   *
   * @return HtmlDomParser
   */
  private function optimizeAttributesInDom(HtmlDomParser $dom)
  {
    foreach ($dom->find('*') as $element) {
      $this->optimizeAttributes($element);
    }

    return $dom;
  }

  /**
   * Prevent changes of inline "styles" and "scripts".
   *
   * @param HtmlDomParser $dom
   *
   * @return HtmlDomParser
   */
  private function protectTagsInDom(HtmlDomParser $dom)
  {
    // init
    $i = 0;

    foreach ($dom->find('script, style') as $element) {

      // skip external links
      if ($element->tag === 'script' || $element->tag === 'style') {
        $attributs = $element->getAllAttributes();
        if (isset($attributs['src'])) {
          continue;
        }
      }

      $node = $element->getNode();
      while ($node->childNodes->length > 0) {
        $this->protectedChildNodes[$i][] = $node->firstChild->nodeValue;
        $node->removeChild($node->firstChild);
      }

      $child = new \DOMElement($this->protectedChildNodesHelper);
      $node = $element->getNode()->appendChild($child);
      /* @var $node \DOMElement */
      $node->setAttribute('data-html-min--saved-content', $i);

      ++$i;
    }

    return $dom;
  }

  /**
   * Remove comments in the dom.
   *
   * @param HtmlDomParser $dom
   *
   * @return HtmlDomParser
   */
  private function removeCommentsInDom(HtmlDomParser $dom)
  {
    foreach ($dom->find('//comment()') as $commentWrapper) {
      $comment = $commentWrapper->getNode();
      $val = $comment->nodeValue;
      if (strpos($val, '[') !== 0) {
        $comment->parentNode->removeChild($comment);
      }
    }

    $dom->getDocument()->normalizeDocument();

    return $dom;
  }

  /**
   * Remove whitespace from dom-nodes.
   *
   * @param HtmlDomParser $dom
   *
   * @return HtmlDomParser
   */
  private function removeWhitespaceInDom(HtmlDomParser $dom)
  {
    $textnodes = $dom->find('//text()');
    foreach ($textnodes as $textnodeWrapper) {
      $textnode = $textnodeWrapper->getNode();
      $xp = $textnode->getNodePath();

      $doSkip = false;
      foreach (self::$skipTagsForRemoveWhitespace as $pattern) {
        if (strpos($xp, "/$pattern") !== false) {
          $doSkip = true;
          break;
        }
      }

      if ($doSkip) {
        continue;
      }

      $textnode->nodeValue = preg_replace("/\s{2,}/", ' ', $textnode->nodeValue);
    }

    $dom->getDocument()->normalizeDocument();

    return $dom;
  }

  /**
   * Callback function for preg_replace_callback use.
   *
   * @param  array $matches PREG matches
   *
   * @return string
   */
  private function restoreProtectedHtml($matches)
  {
    preg_match('/.*"(?<id>\d*)"/', $matches['attributes'], $matchesInner);

    $htmlChild = '';
    if (isset($this->protectedChildNodes[$matchesInner['id']])) {
      foreach ($this->protectedChildNodes[$matchesInner['id']] as $childNode) {
        $htmlChild .= $childNode;
      }
    }

    return $htmlChild;
  }

  /**
   * @param $attrName
   * @param $attrValue
   *
   * @return string
   */
  private function sortCssClasses($attrName, $attrValue)
  {
    if ($attrName !== 'class' || !$attrValue) {
      return $attrValue;
    }

    $classes = explode(' ', $attrValue);
    if (!$classes) {

      return '';
    }

    sort($classes);
    $attrValue = '';
    foreach ($classes as $class) {
      if (!$class) {
        continue;
      }
      $attrValue .= trim($class) . ' ';
    }
    $attrValue = trim($attrValue);

    return $attrValue;
  }

  /**
   * Trim tags in the dom.
   *
   * @param HtmlDomParser $dom
   *
   * @return HtmlDomParser
   */
  private function trimTagsInDom(HtmlDomParser $dom)
  {
    $divnodes = $dom->find('//div|//p|//nav|//footer|//article|//script|//hr|//br');
    foreach ($divnodes as $divnodeWrapper) {
      $divnode = $divnodeWrapper->getNode();

      $candidates = array();
      /** @noinspection PhpParamsInspection */
      if (count($divnode->childNodes) > 0) {
        $candidates[] = $divnode->firstChild;
        $candidates[] = $divnode->lastChild;
        $candidates[] = $divnode->previousSibling;
        $candidates[] = $divnode->nextSibling;
      }

      foreach ($candidates as $candidate) {
        if ($candidate === null) {
          continue;
        }

        if ($candidate->nodeType === 3) {
          $candidate->nodeValue = trim($candidate->nodeValue);
        }
      }
    }

    $dom->getDocument()->normalizeDocument();

    return $dom;
  }
}


1		<?php
2
3		namespace voku\helper;
4
5		/**
6		* Class HtmlMin
7		*
8		* Inspired by:
9		* - JS: https://github.com/kangax/html-minifier/blob/gh-pages/src/htmlminifier.js
10		* - PHP: https://github.com/searchturbine/phpwee-php-minifier
11		* - PHP: https://github.com/WyriHaximus/HtmlCompress
12		* - PHP: https://github.com/zaininnari/html-minifier
13		* - Java: https://code.google.com/archive/p/htmlcompressor/
14		*
15		* @package voku\helper
16		*/
17		class HtmlMin
18		{
19		/**
20		* // https://mathiasbynens.be/demo/javascript-mime-type
21		* // https://developer.mozilla.org/en/docs/Web/HTML/Element/script#attr-type
22		*
23		* @var array
24		*/
25		private static $executableScriptsMimeTypes = array(
26		'text/javascript' => '',
27		'text/ecmascript' => '',
28		'text/jscript' => '',
29		'application/javascript' => '',
30		'application/x-javascript' => '',
31		'application/ecmascript' => '',
32		);
33
34		private static $selfClosingTags = array(
35		'area',
36		'base',
37		'br',
38		'col',
39		'command',
40		'embed',
41		'hr',
42		'img',
43		'input',
44		'keygen',
45		'link',
46		'meta',
47		'param',
48		'source',
49		'track',
50		'wbr',
51		);
52
53
54		/**
55		* @var array
56		*/
57		private static $booleanAttributes = array(
58		'allowfullscreen' => '',
59		'async' => '',
60		'autofocus' => '',
61		'autoplay' => '',
62		'checked' => '',
63		'compact' => '',
64		'controls' => '',
65		'declare' => '',
66		'default' => '',
67		'defaultchecked' => '',
68		'defaultmuted' => '',
69		'defaultselected' => '',
70		'defer' => '',
71		'disabled' => '',
72		'enabled' => '',
73		'formnovalidate' => '',
74		'hidden' => '',
75		'indeterminate' => '',
76		'inert' => '',
77		'ismap' => '',
78		'itemscope' => '',
79		'loop' => '',
80		'multiple' => '',
81		'muted' => '',
82		'nohref' => '',
83		'noresize' => '',
84		'noshade' => '',
85		'novalidate' => '',
86		'nowrap' => '',
87		'open' => '',
88		'pauseonexit' => '',
89		'readonly' => '',
90		'required' => '',
91		'reversed' => '',
92		'scoped' => '',
93		'seamless' => '',
94		'selected' => '',
95		'sortable' => '',
96		'truespeed' => '',
97		'typemustmatch' => '',
98		'visible' => '',
99		);
100		/**
101		* @var array
102		*/
103		private static $skipTagsForRemoveWhitespace = array(
104		'style',
105		'pre',
106		'code',
107		'script',
108		'textarea',
109		);
110
111		/**
112		* An random md5-hash, generated via "random_bytes()".
113		*
114		* @var string
115		*/
116		private $randomHash;
117
118		/**
119		* @var array
120		*/
121		private $protectedChildNodes;
122
123		/**
124		* @var string
125		*/
126		private $protectedChildNodesHelper;
127
128		/**
129		* @var string
130		*/
131		private $booleanAttributesHelper;
132
133		/**
134		* HtmlMin constructor.
135		*/
136	23	public function __construct()
137		{
138	23	$this->protectedChildNodes = array();
139	23	$this->randomHash = md5(Bootup::get_random_bytes(16));
140
141	23	$this->protectedChildNodesHelper = 'html-min--saved-content-' . $this->randomHash;
142	23	$this->booleanAttributesHelper = 'html-min--delete-this-' . $this->randomHash;
143	23	}
144
145		/**
146		* @param string $html
147		*
148		* @return string
149		*/
150	23	public function minify($html)
151		{
152	23	$html = (string)$html;
153	23	if (!isset($html[0])) {
154	1	return '';
155		}
156
157	23	$html = trim($html);
158	23	if (!$html) {
159	3	return '';
160		}
161
162		// init
163	20	$this->protectedChildNodes = array();
164	20	$origHtml = $html;
165	20	$origHtmlLength = UTF8::strlen($html);
166
167	20	$dom = new HtmlDomParser();
168	20	$dom->getDocument()->preserveWhiteSpace = false; // remove redundant white space
169	20	$dom->getDocument()->formatOutput = false; // do not formats output with indentation
170
171	20	$dom->loadHtml($html);
172
173	20	$dom = $this->protectTagsInDom($dom);
174	20	$dom = $this->optimizeAttributesInDom($dom);
175	11	$dom = $this->removeCommentsInDom($dom);
176	11	$dom = $this->removeWhitespaceInDom($dom);
177	11	$dom = $this->trimTagsInDom($dom);
178
179	11	$html = $dom->html();
180
181		// -------------------------------------------------------------------------
182		// Trim whitespace from html-string. [protected html is still protected]
183		// -------------------------------------------------------------------------
184
185		// Remove spaces that are followed by either > or <
186	11	$html = preg_replace('/ (>)/', '$1', $html);
187		// Remove spaces that are preceded by either > or <
188	11	$html = preg_replace('/(<) /', '$1', $html);
189		// Remove spaces that are between > and <
190	11	$html = preg_replace('/(>) (<)/', '>$2', $html);
191
192		// -------------------------------------------------------------------------
193		// Restore protected HTML-code.
194		// -------------------------------------------------------------------------
195
196	11	$html = preg_replace_callback(
197	11	'/<(?<element>' . $this->protectedChildNodesHelper . ')(?<attributes> [^>])?>(?<value>.?)<\/' . $this->protectedChildNodesHelper . '>/',
198	11	array($this, 'restoreProtectedHtml'),
199		$html
200	11	);
201	11	$html = $dom::putReplacedBackToPreserveHtmlEntities($html);
202
203		// ------------------------------------
204		// final clean-up
205		// ------------------------------------
206
207	11	$html = UTF8::cleanup($html);
208
209	11	$html = str_replace(
210		array(
211	11	'html>' . "\n",
212	11	"\n" . '<html',
213	11	'<!doctype',
214	11	'="' . $this->booleanAttributesHelper . '"',
215	11	'</' . $this->protectedChildNodesHelper . '>',
216	11	),
217		array(
218	11	'html>',
219	11	'<html',
220	11	'<!DOCTYPE',
221	11	'',
222	11	'',
223	11	),
224		$html
225	11	);
226
227	11	$html = preg_replace(
228		array(
229	11	'/<(?:' . $this->protectedChildNodesHelper . ')(:? [^>]*)?>/',
230	11	),
231		array(
232	11	'',
233	11	),
234		$html
235	11	);
236
237	11	static $cacheSelfClosingTags = null;
238	11	if ($cacheSelfClosingTags === null) {
239	1	$cacheSelfClosingTags = implode('\|', self::$selfClosingTags);
240	1	}
241	11	$html = preg_replace('#<\b(' . $cacheSelfClosingTags . ')([^>]+)><\/\b\1>#', '<\\1\\2/>', $html);
242
243		// ------------------------------------
244		// check if compression worked
245		// ------------------------------------
246
247	11	if ($origHtmlLength < UTF8::strlen($html)) {
248	3	$html = $origHtml;
249	3	}
250
251	11	return $html;
252		}
253
254		/**
255		* Sort HTML-Attributes, so that gzip can do better work
256		* and remove some default attributes.
257		*
258		* @param SimpleHtmlDom $element
259		*
260		* @return bool
261		*/
262	20	private function optimizeAttributes(SimpleHtmlDom $element)
263		{
264	20	$attributs = $element->getAllAttributes();
265	20	if ($attributs === null) {
266	17	return false;
267		}
268
269	9	$attrs = array();
270	9	foreach ((array)$attributs as $attrName => $attrValue) {
271
272	9	if (isset(self::$booleanAttributes[$attrName])) {
273	5	$attrs[$attrName] = $this->booleanAttributesHelper;
274	5	$element->{$attrName} = null;
275	5	continue;
276		}
277
278		if (
279	9	($attrName === 'href' \|\| $attrName === 'src' \|\| $attrName === 'action')
280	9	&&
281		!(isset($attributs['rel']) && $attributs['rel'] === 'external')
282	9	&&
283		!(isset($attributs['target']) && $attributs['target'] === '_blank')
284	9	) {
285		$attrValue = str_replace('http://', '//', $attrValue);
286		}
287
288	9	if ($this->optimizeAttributesFilters($element->tag, $attrName, $attrValue, $attributs)) {
		0 ignored issues – show Documentation introduced 2016-10-16 03:44 UTC by Report Bug Copy Issue Report `$attributs` is of type `array`, but the function expects a `string`. It seems like the type of the argument is not accepted by the function/method which you are calling. In some cases, in particular if PHP’s automatic type-juggling kicks in this might be fine. In other cases, however this might be a bug. We suggest to add an explicit type cast like in the following example: function acceptsInteger($int) { } $x = '123'; // string "123" // Instead of acceptsInteger($x); // we recommend to use acceptsInteger((integer) $x); Loading history...
289		$element->{$attrName} = null;
290		continue;
291		}
292
293	9	$attrValue = $this->sortCssClasses($attrName, $attrValue);
294
295	9	$attrs[$attrName] = $attrValue;
296	9	$element->{$attrName} = null;
297	9	}
298
299	9	ksort($attrs);
300	9	foreach ($attrs as $attrName => $attrValue) {
301	9	$attrValue = HtmlDomParser::replaceToPreserveHtmlEntities($attrValue);
		0 ignored issues – show Bug introduced 2016-10-17 15:32 UTC by Report Bug Copy Issue Report The method `replaceToPreserveHtmlEntities()` cannot be called from this context as it is declared protected in class `voku\helper\HtmlDomParser`. This check looks for access to methods that are not accessible from the current context. If you need to make a method accessible to another context you can raise its visibility level in the defining class. Loading history...
302		$element->setAttribute($attrName, $attrValue, true);
303		}
304
305		return true;
306		}
307
308		/**
309		* Check if the attribute (key / value) is default and can be skipped.
310		*
311		* @param string $tag
312		* @param string $attrName
313		* @param string $attrValue
314		* @param string $allAttr
315		*
316		* @return bool
317		*/
318	9	private function optimizeAttributesFilters($tag, $attrName, $attrValue, $allAttr)
319		{
320		// remove default
321	9	if ($tag === 'script' && $attrName === 'language' && $attrValue === 'javascript') {
322		return true;
323		}
324
325		// remove default
326	9	if ($tag === 'form' && $attrName === 'method' && $attrValue === 'get') {
327		return true;
328		}
329
330		// remove default
331	9	if ($tag === 'input' && $attrName === 'type' && $attrValue === 'text') {
332		return true;
333		}
334
335		// remove default
336	9	if ($tag === 'area' && $attrName === 'shape' && $attrValue === 'rect') {
337		return true;
338		}
339
340		// remove deprecated charset-attribute (the Browser will use the charset from the HTTP-Header, anyway)
341	9	if ($tag === 'script' && $attrName === 'charset' && !isset($allAttr['src'])) {
342		return true;
343		}
344
345		// remove deprecated anchor-jump
346	9	if ($tag === 'a' && $attrName === 'name' && isset($allAttr['id'])) {
347		return true;
348		}
349
350		// remove "type=text/css" for css links
351	9	if ($tag === 'link' && $attrName === 'type' && $attrValue === 'text/css' && isset($allAttr['rel']) && $allAttr['rel'] === 'stylesheet') {
352		return true;
353		}
354
355		// remove deprecated script-mime-types
356	9	if ($tag === 'script' && $attrName === 'type' && isset($allAttr['src'], self::$executableScriptsMimeTypes[$attrValue])) {
357		return true;
358		}
359
360		// remove empty value from <input>
361	9	if ($tag === 'input' && $attrName === 'value' && $attrValue === '') {
362		return true;
363		}
364
365		// remove some empty attribute
366	9	if ($attrValue === '' && preg_match('/^(?:class\|id\|style\|title\|lang\|dir\|on(?:focus\|blur\|change\|click\|dblclick\|mouse(?:down\|up\|over\|move\|out)\|key(?:press\|down\|up)))$/', $attrName)) {
367		return true;
368		}
369
370	9	return false;
371		}
372
373		/**
374		* Optimize HTML-tag attributes in the dom.
375		*
376		* @param HtmlDomParser $dom
377		*
378		* @return HtmlDomParser
379		*/
380	20	private function optimizeAttributesInDom(HtmlDomParser $dom)
381		{
382	20	foreach ($dom->find('*') as $element) {
383	20	$this->optimizeAttributes($element);
384	17	}
385
386	11	return $dom;
387		}
388
389		/**
390		* Prevent changes of inline "styles" and "scripts".
391		*
392		* @param HtmlDomParser $dom
393		*
394		* @return HtmlDomParser
395		*/
396	20	private function protectTagsInDom(HtmlDomParser $dom)
397		{
398		// init
399	20	$i = 0;
400
401	20	foreach ($dom->find('script, style') as $element) {
402
403		// skip external links
404	4	if ($element->tag === 'script' \|\| $element->tag === 'style') {
405	4	$attributs = $element->getAllAttributes();
406	4	if (isset($attributs['src'])) {
407	2	continue;
408		}
409	3	}
410
411	3	$node = $element->getNode();
412	3	while ($node->childNodes->length > 0) {
413	3	$this->protectedChildNodes[$i][] = $node->firstChild->nodeValue;
414	3	$node->removeChild($node->firstChild);
415	3	}
416
417	3	$child = new \DOMElement($this->protectedChildNodesHelper);
418	3	$node = $element->getNode()->appendChild($child);
419		/* @var $node \DOMElement */
420	3	$node->setAttribute('data-html-min--saved-content', $i);
421
422	3	++$i;
423	20	}
424
425	20	return $dom;
426		}
427
428		/**
429		* Remove comments in the dom.
430		*
431		* @param HtmlDomParser $dom
432		*
433		* @return HtmlDomParser
434		*/
435	11	private function removeCommentsInDom(HtmlDomParser $dom)
436		{
437	11	foreach ($dom->find('//comment()') as $commentWrapper) {
438		$comment = $commentWrapper->getNode();
439		$val = $comment->nodeValue;
440		if (strpos($val, '[') !== 0) {
441		$comment->parentNode->removeChild($comment);
442		}
443	11	}
444
445	11	$dom->getDocument()->normalizeDocument();
446
447	11	return $dom;
448		}
449
450		/**
451		* Remove whitespace from dom-nodes.
452		*
453		* @param HtmlDomParser $dom
454		*
455		* @return HtmlDomParser
456		*/
457	12	private function removeWhitespaceInDom(HtmlDomParser $dom)
458		{
459	11	$textnodes = $dom->find('//text()');
460	11	foreach ($textnodes as $textnodeWrapper) {
461	11	$textnode = $textnodeWrapper->getNode();
462	11	$xp = $textnode->getNodePath();
463
464	11	$doSkip = false;
465	11	foreach (self::$skipTagsForRemoveWhitespace as $pattern) {
466	11	if (strpos($xp, "/$pattern") !== false) {
467		$doSkip = true;
468		break;
469		}
470	11	}
471
472	11	if ($doSkip) {
473	1	continue;
474		}
475
476	11	$textnode->nodeValue = preg_replace("/\s{2,}/", ' ', $textnode->nodeValue);
477	12	}
478
479	11	$dom->getDocument()->normalizeDocument();
480
481	11	return $dom;
482		}
483
484		/**
485		* Callback function for preg_replace_callback use.
486		*
487		* @param array $matches PREG matches
488		*
489		* @return string
490		*/
491		private function restoreProtectedHtml($matches)
492		{
493		preg_match('/."(?<id>\d)"/', $matches['attributes'], $matchesInner);
494
495		$htmlChild = '';
496		if (isset($this->protectedChildNodes[$matchesInner['id']])) {
497		foreach ($this->protectedChildNodes[$matchesInner['id']] as $childNode) {
498		$htmlChild .= $childNode;
499		}
500		}
501
502		return $htmlChild;
503		}
504
505		/**
506		* @param $attrName
507		* @param $attrValue
508		*
509		* @return string
510		*/
511	9	private function sortCssClasses($attrName, $attrValue)
512		{
513	9	if ($attrName !== 'class' \|\| !$attrValue) {
514	8	return $attrValue;
515		}
516
517	3	$classes = explode(' ', $attrValue);
518	3	if (!$classes) {
		0 ignored issues – show Bug Best Practice introduced 2016-10-16 03:44 UTC by Report Bug Copy Issue Report The expression `$classes` of type `array` is implicitly converted to a boolean; are you sure this is intended? If so, consider using `empty($expr)` instead to make it clear that you intend to check for an array without elements. This check marks implicit conversions of arrays to boolean values in a comparison. While in PHP an empty array is considered to be equal (but not identical) to false, this is not always apparent. Consider making the comparison explicit by using `empty(..)` or `! empty(...)` instead. Loading history...
519		return '';
520		}
521
522	3	sort($classes);
523	3	$attrValue = '';
524	3	foreach ($classes as $class) {
525	3	if (!$class) {
526		continue;
527		}
528	3	$attrValue .= trim($class) . ' ';
529	3	}
530	3	$attrValue = trim($attrValue);
531
532	3	return $attrValue;
533		}
534
535		/**
536		* Trim tags in the dom.
537		*
538		* @param HtmlDomParser $dom
539		*
540		* @return HtmlDomParser
541		*/
542	11	private function trimTagsInDom(HtmlDomParser $dom)
543		{
544	11	$divnodes = $dom->find('//div\|//p\|//nav\|//footer\|//article\|//script\|//hr\|//br');
545	11	foreach ($divnodes as $divnodeWrapper) {
546	3	$divnode = $divnodeWrapper->getNode();
547
548	3	$candidates = array();
549		/** @noinspection PhpParamsInspection */
550	3	if (count($divnode->childNodes) > 0) {
551	3	$candidates[] = $divnode->firstChild;
552	3	$candidates[] = $divnode->lastChild;
553	3	$candidates[] = $divnode->previousSibling;
554	3	$candidates[] = $divnode->nextSibling;
555	3	}
556
557	3	foreach ($candidates as $candidate) {
558	3	if ($candidate === null) {
559	3	continue;
560		}
561
562	3	if ($candidate->nodeType === 3) {
563	3	$candidate->nodeValue = trim($candidate->nodeValue);
564	3	}
565	3	}
566	11	}
567
568	11	$dom->getDocument()->normalizeDocument();
569
570	11	return $dom;
571		}
572		}
573

voku / HtmlMin

Push — master ( c927d1...313909 )

HtmlMin D

Complexity

Size/Duplication

Coupling/Cohesion

Test Coverage

Importance

11 Methods

How to fix Complexity

Complex Class

Duplication Side-by-Side

Filter issues like