2014-10-08 21:32:46 +00:00
|
|
|
<?php
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* This class provides functions for converting CSS styles into inline style attributes in your HTML code.
|
2014-10-08 21:32:46 +00:00
|
|
|
*
|
|
|
|
|
* For more information, please see the README.md file.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @version 1.2.0
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
2014-10-08 21:32:46 +00:00
|
|
|
* @author Cameron Brooks
|
|
|
|
|
* @author Jaime Prado
|
2016-07-25 13:51:13 +00:00
|
|
|
* @author Oliver Klee <typo3-coding@oliverklee.de>
|
2014-10-08 21:32:46 +00:00
|
|
|
* @author Roman Ožana <ozana@omdesign.cz>
|
2017-03-31 20:19:35 +00:00
|
|
|
* @author Sander Kruger <s.kruger@invessel.com>
|
2014-10-08 21:32:46 +00:00
|
|
|
*/
|
2016-09-02 01:33:57 +00:00
|
|
|
// @codingStandardsIgnoreFile
|
2016-07-25 13:51:13 +00:00
|
|
|
class Emogrifier
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var int
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
|
|
|
|
const CACHE_KEY_CSS = 0;
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var int
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
|
|
|
|
const CACHE_KEY_SELECTOR = 1;
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var int
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
|
|
|
|
const CACHE_KEY_XPATH = 2;
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var int
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2016-07-25 13:51:13 +00:00
|
|
|
const CACHE_KEY_CSS_DECLARATIONS_BLOCK = 3;
|
2015-02-12 11:42:15 +00:00
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var int
|
|
|
|
|
*/
|
|
|
|
|
const CACHE_KEY_COMBINED_STYLES = 4;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* for calculating nth-of-type and nth-child selectors
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var int
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
|
|
|
|
const INDEX = 0;
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* for calculating nth-of-type and nth-child selectors
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var int
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
|
|
|
|
const MULTIPLIER = 1;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @var string
|
|
|
|
|
*/
|
|
|
|
|
const ID_ATTRIBUTE_MATCHER = '/(\\w+)?\\#([\\w\\-]+)/';
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @var string
|
|
|
|
|
*/
|
|
|
|
|
const CLASS_ATTRIBUTE_MATCHER = '/(\\w+|[\\*\\]])?((\\.[\\w\\-]+)+)/';
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
/**
|
|
|
|
|
* @var string
|
|
|
|
|
*/
|
|
|
|
|
const CONTENT_TYPE_META_TAG = '<meta http-equiv="Content-Type" content="text/html; charset=utf-8">';
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @var string
|
|
|
|
|
*/
|
|
|
|
|
const DEFAULT_DOCUMENT_TYPE = '<!DOCTYPE html>';
|
|
|
|
|
|
2015-02-12 11:42:15 +00:00
|
|
|
/**
|
|
|
|
|
* @var string
|
|
|
|
|
*/
|
|
|
|
|
private $html = '';
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @var string
|
|
|
|
|
*/
|
|
|
|
|
private $css = '';
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var bool[]
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2016-07-26 10:43:54 +00:00
|
|
|
private $excludedSelectors = array();
|
2015-02-12 11:42:15 +00:00
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var string[]
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2016-07-26 10:43:54 +00:00
|
|
|
private $unprocessableHtmlTags = array( 'wbr' );
|
2015-02-12 11:42:15 +00:00
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var bool[]
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2016-07-26 10:43:54 +00:00
|
|
|
private $allowedMediaTypes = array( 'all' => true, 'screen' => true, 'print' => true );
|
2015-02-12 11:42:15 +00:00
|
|
|
|
|
|
|
|
/**
|
2017-03-31 20:19:35 +00:00
|
|
|
* @var mixed[]
|
2016-07-25 13:51:13 +00:00
|
|
|
*/
|
2016-07-26 10:43:54 +00:00
|
|
|
private $caches = array(
|
|
|
|
|
self::CACHE_KEY_CSS => array(),
|
|
|
|
|
self::CACHE_KEY_SELECTOR => array(),
|
|
|
|
|
self::CACHE_KEY_XPATH => array(),
|
|
|
|
|
self::CACHE_KEY_CSS_DECLARATIONS_BLOCK => array(),
|
|
|
|
|
self::CACHE_KEY_COMBINED_STYLES => array(),
|
|
|
|
|
);
|
2016-07-25 13:51:13 +00:00
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* the visited nodes with the XPath paths as array keys
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @var \DOMElement[]
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2016-07-26 10:43:54 +00:00
|
|
|
private $visitedNodes = array();
|
2015-02-12 11:42:15 +00:00
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* the styles to apply to the nodes with the XPath paths as array keys for the outer array
|
|
|
|
|
* and the attribute names/values as key/value pairs for the inner array
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @var string[][]
|
2016-07-25 13:51:13 +00:00
|
|
|
*/
|
2016-07-26 10:43:54 +00:00
|
|
|
private $styleAttributesForNodes = array();
|
2016-07-25 13:51:13 +00:00
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Determines whether the "style" attributes of tags in the the HTML passed to this class should be preserved.
|
|
|
|
|
* If set to false, the value of the style attributes will be discarded.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var bool
|
|
|
|
|
*/
|
|
|
|
|
private $isInlineStyleAttributesParsingEnabled = true;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Determines whether the <style> blocks in the HTML passed to this class should be parsed.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* If set to true, the <style> blocks will be removed from the HTML and their contents will be applied to the HTML
|
|
|
|
|
* via inline styles.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* If set to false, the <style> blocks will be left as they are in the HTML.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @var bool
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2016-07-25 13:51:13 +00:00
|
|
|
private $isStyleBlocksParsingEnabled = true;
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
/**
|
|
|
|
|
* Determines whether elements with the `display: none` property are
|
|
|
|
|
* removed from the DOM.
|
|
|
|
|
*
|
|
|
|
|
* @var bool
|
|
|
|
|
*/
|
|
|
|
|
private $shouldKeepInvisibleNodes = true;
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
/**
|
|
|
|
|
* @var string[]
|
|
|
|
|
*/
|
|
|
|
|
private $xPathRules = array(
|
|
|
|
|
// child
|
|
|
|
|
'/\\s*>\\s*/' => '/',
|
|
|
|
|
// adjacent sibling
|
|
|
|
|
'/\\s+\\+\\s+/' => '/following-sibling::*[1]/self::',
|
|
|
|
|
// descendant
|
|
|
|
|
'/\\s+(?=.*[^\\]]{1}$)/' => '//',
|
|
|
|
|
// :first-child
|
|
|
|
|
'/([^\\/]+):first-child/i' => '*[1]/self::\\1',
|
|
|
|
|
// :last-child
|
|
|
|
|
'/([^\\/]+):last-child/i' => '*[last()]/self::\\1',
|
|
|
|
|
// attribute only
|
|
|
|
|
'/^\\[(\\w+|\\w+\\=[\'"]?\\w+[\'"]?)\\]/' => '*[@\\1]',
|
|
|
|
|
// attribute
|
|
|
|
|
'/(\\w)\\[(\\w+)\\]/' => '\\1[@\\2]',
|
|
|
|
|
// exact attribute
|
|
|
|
|
'/(\\w)\\[(\\w+)\\=[\'"]?([\\w\\s]+)[\'"]?\\]/' => '\\1[@\\2="\\3"]',
|
|
|
|
|
// element attribute~=
|
|
|
|
|
'/([\\w\\*]+)\\[(\\w+)[\\s]*\\~\\=[\\s]*[\'"]?([\\w-_\\/]+)[\'"]?\\]/'
|
|
|
|
|
=> '\\1[contains(concat(" ", @\\2, " "), concat(" ", "\\3", " "))]',
|
|
|
|
|
// element attribute^=
|
|
|
|
|
'/([\\w\\*]+)\\[(\\w+)[\\s]*\\^\\=[\\s]*[\'"]?([\\w-_\\/]+)[\'"]?\\]/' => '\\1[starts-with(@\\2, "\\3")]',
|
|
|
|
|
// element attribute*=
|
|
|
|
|
'/([\\w\\*]+)\\[(\\w+)[\\s]*\\*\\=[\\s]*[\'"]?([\\w-_\\s\\/:;]+)[\'"]?\\]/' => '\\1[contains(@\\2, "\\3")]',
|
|
|
|
|
// element attribute$=
|
|
|
|
|
'/([\\w\\*]+)\\[(\\w+)[\\s]*\\$\\=[\\s]*[\'"]?([\\w-_\\s\\/]+)[\'"]?\\]/'
|
|
|
|
|
=> '\\1[substring(@\\2, string-length(@\\2) - string-length("\\3") + 1) = "\\3"]',
|
|
|
|
|
// element attribute|=
|
|
|
|
|
'/([\\w\\*]+)\\[(\\w+)[\\s]*\\|\\=[\\s]*[\'"]?([\\w-_\\s\\/]+)[\'"]?\\]/'
|
|
|
|
|
=> '\\1[@\\2="\\3" or starts-with(@\\2, concat("\\3", "-"))]',
|
|
|
|
|
);
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Determines whether CSS styles that have an equivalent HTML attribute
|
|
|
|
|
* should be mapped and attached to those elements.
|
|
|
|
|
*
|
|
|
|
|
* @var bool
|
|
|
|
|
*/
|
|
|
|
|
private $shouldMapCssToHtml = false;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* This multi-level array contains simple mappings of CSS properties to
|
|
|
|
|
* HTML attributes. If a mapping only applies to certain HTML nodes or
|
|
|
|
|
* only for certain values, the mapping is an object with a whitelist
|
|
|
|
|
* of nodes and values.
|
|
|
|
|
*
|
|
|
|
|
* @var mixed[][]
|
|
|
|
|
*/
|
|
|
|
|
private $cssToHtmlMap = array(
|
|
|
|
|
'background-color' => array(
|
|
|
|
|
'attribute' => 'bgcolor',
|
|
|
|
|
),
|
|
|
|
|
'text-align' => array(
|
|
|
|
|
'attribute' => 'align',
|
|
|
|
|
'nodes' => array('p', 'div', 'td'),
|
|
|
|
|
'values' => array('left', 'right', 'center', 'justify'),
|
|
|
|
|
),
|
|
|
|
|
'float' => array(
|
|
|
|
|
'attribute' => 'align',
|
|
|
|
|
'nodes' => array('table', 'img'),
|
|
|
|
|
'values' => array('left', 'right'),
|
|
|
|
|
),
|
|
|
|
|
'border-spacing' => array(
|
|
|
|
|
'attribute' => 'cellspacing',
|
|
|
|
|
'nodes' => array('table'),
|
|
|
|
|
),
|
|
|
|
|
);
|
|
|
|
|
|
2016-07-26 16:03:16 +00:00
|
|
|
public static $_media = '';
|
|
|
|
|
|
2015-02-12 11:42:15 +00:00
|
|
|
/**
|
|
|
|
|
* The constructor.
|
|
|
|
|
*
|
|
|
|
|
* @param string $html the HTML to emogrify, must be UTF-8-encoded
|
|
|
|
|
* @param string $css the CSS to merge, must be UTF-8-encoded
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function __construct($html = '', $css = '')
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
$this->setHtml($html);
|
|
|
|
|
$this->setCss($css);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* The destructor.
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function __destruct()
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
$this->purgeVisitedNodes();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Sets the HTML to emogrify.
|
|
|
|
|
*
|
|
|
|
|
* @param string $html the HTML to emogrify, must be UTF-8-encoded
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
|
|
|
|
* @return void
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function setHtml($html)
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
$this->html = $html;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Sets the CSS to merge with the HTML.
|
|
|
|
|
*
|
|
|
|
|
* @param string $css the CSS to merge, must be UTF-8-encoded
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
|
|
|
|
* @return void
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function setCss($css)
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
$this->css = $css;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* Applies $this->css to $this->html and returns the HTML with the CSS
|
|
|
|
|
* applied.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* This method places the CSS inline.
|
|
|
|
|
*
|
|
|
|
|
* @return string
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @throws \BadMethodCallException
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function emogrify()
|
|
|
|
|
{
|
|
|
|
|
if ($this->html === '') {
|
2016-07-26 15:34:17 +00:00
|
|
|
throw new BadMethodCallException('Please set some HTML first before calling emogrify.', 1390393096);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
self::$_media = ''; // reset.
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
$xmlDocument = $this->createXmlDocument();
|
|
|
|
|
$this->process($xmlDocument);
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
return $xmlDocument->saveHTML();
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* Applies $this->css to $this->html and returns only the HTML content
|
|
|
|
|
* within the <body> tag.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* This method places the CSS inline.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @return string
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @throws \BadMethodCallException
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function emogrifyBodyContent()
|
|
|
|
|
{
|
|
|
|
|
if ($this->html === '') {
|
2016-07-26 15:34:17 +00:00
|
|
|
throw new BadMethodCallException('Please set some HTML first before calling emogrify.', 1390393096);
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
$xmlDocument = $this->createXmlDocument();
|
|
|
|
|
$this->process($xmlDocument);
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
$innerDocument = new DOMDocument();
|
|
|
|
|
foreach ($xmlDocument->documentElement->getElementsByTagName('body')->item(0)->childNodes as $childNode) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$innerDocument->appendChild($innerDocument->importNode($childNode, true));
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
return html_entity_decode($innerDocument->saveHTML());
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* Applies $this->css to $xmlDocument.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* This method places the CSS inline.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @param \DOMDocument $xmlDocument
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @return void
|
2017-03-31 20:19:35 +00:00
|
|
|
*
|
|
|
|
|
* @throws \InvalidArgumentException
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
protected function process(DOMDocument $xmlDocument)
|
|
|
|
|
{
|
|
|
|
|
$xPath = new DOMXPath($xmlDocument);
|
2015-02-12 11:42:15 +00:00
|
|
|
$this->clearAllCaches();
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
// Before be begin processing the CSS file, parse the document and normalize all existing CSS attributes.
|
|
|
|
|
// This changes 'DISPLAY: none' to 'display: none'.
|
|
|
|
|
// We wouldn't have to do this if DOMXPath supported XPath 2.0.
|
|
|
|
|
// Also store a reference of nodes with existing inline styles so we don't overwrite them.
|
2015-02-12 11:42:15 +00:00
|
|
|
$this->purgeVisitedNodes();
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
set_error_handler(array($this, 'handleXpathError'), E_WARNING);
|
|
|
|
|
|
|
|
|
|
$nodesWithStyleAttributes = $xPath->query('//*[@style]');
|
|
|
|
|
if ($nodesWithStyleAttributes !== false) {
|
|
|
|
|
/** @var \DOMElement $node */
|
|
|
|
|
foreach ($nodesWithStyleAttributes as $node) {
|
|
|
|
|
if ($this->isInlineStyleAttributesParsingEnabled) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$this->normalizeStyleAttributes($node);
|
|
|
|
|
} else {
|
|
|
|
|
$node->removeAttribute('style');
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// grab any existing style blocks from the html and append them to the existing CSS
|
|
|
|
|
// (these blocks should be appended so as to have precedence over conflicting styles in the existing CSS)
|
|
|
|
|
$allCss = $this->css;
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($this->isStyleBlocksParsingEnabled) {
|
|
|
|
|
$allCss .= $this->getCssFromAllStyleNodes($xPath);
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
|
|
|
|
$cssParts = $this->splitCssAndMediaQuery($allCss);
|
2017-03-31 20:19:35 +00:00
|
|
|
$excludedNodes = $this->getNodesToExclude($xPath);
|
2016-07-25 13:51:13 +00:00
|
|
|
$cssRules = $this->parseCssRules($cssParts['css']);
|
2017-03-31 20:19:35 +00:00
|
|
|
foreach ($cssRules as $cssRule) {
|
2016-07-25 13:51:13 +00:00
|
|
|
// query the body for the xpath selector
|
2017-03-31 20:19:35 +00:00
|
|
|
$nodesMatchingCssSelectors = $xPath->query($this->translateCssToXpath($cssRule['selector']));
|
2016-07-25 13:51:13 +00:00
|
|
|
// ignore invalid selectors
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($nodesMatchingCssSelectors === false) {
|
2016-07-25 13:51:13 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
/** @var \DOMElement $node */
|
|
|
|
|
foreach ($nodesMatchingCssSelectors as $node) {
|
|
|
|
|
if (in_array($node, $excludedNodes, true)) {
|
2016-07-25 13:51:13 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// if it has a style attribute, get it, process it, and append (overwrite) new stuff
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($node->hasAttribute('style')) {
|
2016-07-25 13:51:13 +00:00
|
|
|
// break it up into an associative array
|
|
|
|
|
$oldStyleDeclarations = $this->parseCssDeclarationsBlock($node->getAttribute('style'));
|
|
|
|
|
} else {
|
2016-07-26 10:43:54 +00:00
|
|
|
$oldStyleDeclarations = array();
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
$newStyleDeclarations = $this->parseCssDeclarationsBlock($cssRule['declarationsBlock']);
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($this->shouldMapCssToHtml) {
|
|
|
|
|
$this->mapCssToHtmlAttributes($newStyleDeclarations, $node);
|
|
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
$node->setAttribute(
|
|
|
|
|
'style',
|
|
|
|
|
$this->generateStyleStringFromDeclarationsArrays($oldStyleDeclarations, $newStyleDeclarations)
|
|
|
|
|
);
|
|
|
|
|
}
|
|
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
restore_error_handler();
|
|
|
|
|
|
|
|
|
|
if ($this->isInlineStyleAttributesParsingEnabled) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$this->fillStyleAttributesWithMergedStyles();
|
|
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($this->shouldKeepInvisibleNodes) {
|
|
|
|
|
$this->removeInvisibleNodes($xPath);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
$this->copyCssWithMediaToStyleNode($xmlDocument, $xPath, $cssParts['media']);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Applies $styles to $node.
|
|
|
|
|
*
|
|
|
|
|
* This method maps CSS styles to HTML attributes and adds those to the
|
|
|
|
|
* node.
|
|
|
|
|
*
|
|
|
|
|
* @param string[] $styles the new CSS styles taken from the global styles to be applied to this node
|
|
|
|
|
* @param \DOMNode $node node to apply styles to
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
|
|
|
|
private function mapCssToHtmlAttributes(array $styles, DOMNode $node)
|
|
|
|
|
{
|
|
|
|
|
foreach ($styles as $property => $value) {
|
|
|
|
|
// Strip !important indicator
|
|
|
|
|
$value = trim(str_replace('!important', '', $value));
|
|
|
|
|
$this->mapCssToHtmlAttribute($property, $value, $node);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Tries to apply the CSS style to $node as an attribute.
|
|
|
|
|
*
|
|
|
|
|
* This method maps a CSS rule to HTML attributes and adds those to the node.
|
|
|
|
|
*
|
|
|
|
|
* @param string $property the name of the CSS property to map
|
|
|
|
|
* @param string $value the value of the style rule to map
|
|
|
|
|
* @param \DOMNode $node node to apply styles to
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
|
|
|
|
private function mapCssToHtmlAttribute($property, $value, DOMNode $node)
|
|
|
|
|
{
|
|
|
|
|
if (!$this->mapSimpleCssProperty($property, $value, $node)) {
|
|
|
|
|
$this->mapComplexCssProperty($property, $value, $node);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Looks up the CSS property in the mapping table and maps it if it matches the conditions.
|
|
|
|
|
*
|
|
|
|
|
* @param string $property the name of the CSS property to map
|
|
|
|
|
* @param string $value the value of the style rule to map
|
|
|
|
|
* @param \DOMNode $node node to apply styles to
|
|
|
|
|
*
|
|
|
|
|
* @return bool true if the property cab be mapped using the simple mapping table
|
|
|
|
|
*/
|
|
|
|
|
private function mapSimpleCssProperty($property, $value, DOMNode $node)
|
|
|
|
|
{
|
|
|
|
|
if (!isset($this->cssToHtmlMap[$property])) {
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
$mapping = $this->cssToHtmlMap[$property];
|
|
|
|
|
$nodesMatch = !isset($mapping['nodes']) || in_array($node->nodeName, $mapping['nodes'], true);
|
|
|
|
|
$valuesMatch = !isset($mapping['values']) || in_array($value, $mapping['values'], true);
|
|
|
|
|
if (!$nodesMatch || !$valuesMatch) {
|
|
|
|
|
return false;
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
$node->setAttribute($mapping['attribute'], $value);
|
|
|
|
|
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Maps CSS properties that need special transformation to an HTML attribute.
|
|
|
|
|
*
|
|
|
|
|
* @param string $property the name of the CSS property to map
|
|
|
|
|
* @param string $value the value of the style rule to map
|
|
|
|
|
* @param \DOMNode $node node to apply styles to
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
|
|
|
|
private function mapComplexCssProperty($property, $value, DOMNode $node)
|
|
|
|
|
{
|
|
|
|
|
$nodeName = $node->nodeName;
|
|
|
|
|
$isTable = $nodeName === 'table';
|
|
|
|
|
$isImage = $nodeName === 'img';
|
|
|
|
|
$isTableOrImage = $isTable || $isImage;
|
|
|
|
|
|
|
|
|
|
switch ($property) {
|
|
|
|
|
case 'background':
|
|
|
|
|
// Parse out the color, if any
|
|
|
|
|
$styles = explode(' ', $value);
|
|
|
|
|
$first = $styles[0];
|
|
|
|
|
if (!is_numeric(substr($first, 0, 1)) && substr($first, 0, 3) !== 'url') {
|
|
|
|
|
// This is not a position or image, assume it's a color
|
|
|
|
|
$node->setAttribute('bgcolor', $first);
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case 'width':
|
|
|
|
|
// intentional fall-through
|
|
|
|
|
case 'height':
|
|
|
|
|
// Only parse values in px and %, but not values like "auto".
|
|
|
|
|
if (preg_match('/^\d+(px|%)$/', $value)) {
|
|
|
|
|
// Remove 'px'. This regex only conserves numbers and %
|
|
|
|
|
$number = preg_replace('/[^0-9.%]/', '', $value);
|
|
|
|
|
$node->setAttribute($property, $number);
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case 'margin':
|
|
|
|
|
if ($isTableOrImage) {
|
|
|
|
|
$margins = $this->parseCssShorthandValue($value);
|
|
|
|
|
if ($margins['left'] === 'auto' && $margins['right'] === 'auto') {
|
|
|
|
|
$node->setAttribute('align', 'center');
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case 'border':
|
|
|
|
|
if ($isTableOrImage) {
|
|
|
|
|
if ($value === 'none' || $value === '0') {
|
|
|
|
|
$node->setAttribute('border', '0');
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Parses a shorthand CSS value and splits it into individual values
|
|
|
|
|
*
|
|
|
|
|
* @param string $value a string of CSS value with 1, 2, 3 or 4 sizes
|
|
|
|
|
* For example: padding: 0 auto;
|
|
|
|
|
* '0 auto' is split into top: 0, left: auto, bottom: 0,
|
|
|
|
|
* right: auto.
|
|
|
|
|
*
|
|
|
|
|
* @return string[] an array of values for top, right, bottom and left (using these as associative array keys)
|
|
|
|
|
*/
|
|
|
|
|
private function parseCssShorthandValue($value)
|
|
|
|
|
{
|
|
|
|
|
$values = preg_split('/\\s+/', $value);
|
|
|
|
|
|
|
|
|
|
$css = array();
|
|
|
|
|
$css['top'] = $values[0];
|
|
|
|
|
$css['right'] = (count($values) > 1) ? $values[1] : $css['top'];
|
|
|
|
|
$css['bottom'] = (count($values) > 2) ? $values[2] : $css['top'];
|
|
|
|
|
$css['left'] = (count($values) > 3) ? $values[3] : $css['right'];
|
|
|
|
|
|
|
|
|
|
return $css;
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Extracts and parses the individual rules from a CSS string.
|
|
|
|
|
*
|
|
|
|
|
* @param string $css a string of raw CSS code
|
|
|
|
|
*
|
|
|
|
|
* @return string[][] an array of string sub-arrays with the keys
|
|
|
|
|
* "selector" (the CSS selector(s), e.g., "*" or "h1"),
|
|
|
|
|
* "declarationsBLock" (the semicolon-separated CSS declarations for that selector(s),
|
|
|
|
|
* e.g., "color: red; height: 4px;"),
|
|
|
|
|
* and "line" (the line number e.g. 42)
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function parseCssRules($css)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$cssKey = md5($css);
|
2017-03-31 20:19:35 +00:00
|
|
|
if (!isset($this->caches[self::CACHE_KEY_CSS][$cssKey])) {
|
2015-02-12 11:42:15 +00:00
|
|
|
// process the CSS file for selectors and definitions
|
2016-07-25 13:51:13 +00:00
|
|
|
preg_match_all('/(?:^|[\\s^{}]*)([^{]+){([^}]*)}/mis', $css, $matches, PREG_SET_ORDER);
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2016-07-26 10:43:54 +00:00
|
|
|
$cssRules = array();
|
2016-07-25 13:51:13 +00:00
|
|
|
/** @var string[] $cssRule */
|
2017-03-31 20:19:35 +00:00
|
|
|
foreach ($matches as $key => $cssRule) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$cssDeclaration = trim($cssRule[2]);
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($cssDeclaration === '') {
|
2015-02-12 11:42:15 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
$selectors = explode(',', $cssRule[1]);
|
2017-03-31 20:19:35 +00:00
|
|
|
foreach ($selectors as $selector) {
|
2016-07-25 13:51:13 +00:00
|
|
|
// don't process pseudo-elements and behavioral (dynamic) pseudo-classes;
|
|
|
|
|
// only allow structural pseudo-classes
|
2017-03-31 20:19:35 +00:00
|
|
|
$hasPseudoElement = strpos($selector, '::') !== false;
|
|
|
|
|
$hasAnyPseudoClass = (bool) preg_match('/:[a-zA-Z]/', $selector);
|
|
|
|
|
$hasSupportedPseudoClass = (bool) preg_match('/:\\S+\\-(child|type\\()/i', $selector);
|
|
|
|
|
if ($hasPseudoElement || ($hasAnyPseudoClass && !$hasSupportedPseudoClass)) {
|
2015-02-12 11:42:15 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
2016-07-26 10:43:54 +00:00
|
|
|
$cssRules[] = array(
|
2016-07-25 13:51:13 +00:00
|
|
|
'selector' => trim($selector),
|
|
|
|
|
'declarationsBlock' => $cssDeclaration,
|
|
|
|
|
// keep track of where it appears in the file, since order is important
|
|
|
|
|
'line' => $key,
|
2016-07-26 10:43:54 +00:00
|
|
|
);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
usort($cssRules, array($this, 'sortBySelectorPrecedence'));
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
$this->caches[self::CACHE_KEY_CSS][$cssKey] = $cssRules;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
return $this->caches[self::CACHE_KEY_CSS][$cssKey];
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
/**
|
|
|
|
|
* Disables the parsing of inline styles.
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function disableInlineStyleAttributesParsing()
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$this->isInlineStyleAttributesParsingEnabled = false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Disables the parsing of <style> blocks.
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function disableStyleBlocksParsing()
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$this->isStyleBlocksParsingEnabled = false;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Disables the removal of elements with `display: none` properties.
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function disableInvisibleNodeRemoval()
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$this->shouldKeepInvisibleNodes = false;
|
|
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
/**
|
|
|
|
|
* Enables the attachment/override of HTML attributes for which a
|
|
|
|
|
* corresponding CSS property has been set.
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
|
|
|
|
public function enableCssToHtmlMapping()
|
|
|
|
|
{
|
|
|
|
|
$this->shouldMapCssToHtml = true;
|
|
|
|
|
}
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
/**
|
|
|
|
|
* Clears all caches.
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function clearAllCaches()
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$this->clearCache(self::CACHE_KEY_CSS);
|
|
|
|
|
$this->clearCache(self::CACHE_KEY_SELECTOR);
|
|
|
|
|
$this->clearCache(self::CACHE_KEY_XPATH);
|
|
|
|
|
$this->clearCache(self::CACHE_KEY_CSS_DECLARATIONS_BLOCK);
|
|
|
|
|
$this->clearCache(self::CACHE_KEY_COMBINED_STYLES);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Clears a single cache by key.
|
|
|
|
|
*
|
|
|
|
|
* @param int $key the cache key, must be CACHE_KEY_CSS, CACHE_KEY_SELECTOR, CACHE_KEY_XPATH
|
|
|
|
|
* or CACHE_KEY_CSS_DECLARATION_BLOCK
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*
|
|
|
|
|
* @throws \InvalidArgumentException
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function clearCache($key)
|
|
|
|
|
{
|
2016-07-26 10:43:54 +00:00
|
|
|
$allowedCacheKeys = array(
|
2016-07-25 13:51:13 +00:00
|
|
|
self::CACHE_KEY_CSS,
|
|
|
|
|
self::CACHE_KEY_SELECTOR,
|
|
|
|
|
self::CACHE_KEY_XPATH,
|
|
|
|
|
self::CACHE_KEY_CSS_DECLARATIONS_BLOCK,
|
|
|
|
|
self::CACHE_KEY_COMBINED_STYLES,
|
2016-07-26 10:43:54 +00:00
|
|
|
);
|
2017-03-31 20:19:35 +00:00
|
|
|
if (!in_array($key, $allowedCacheKeys, true)) {
|
2016-07-26 15:50:23 +00:00
|
|
|
throw new InvalidArgumentException('Invalid cache key: ' . $key, 1391822035);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
$this->caches[$key] = array();
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Purges the visited nodes.
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function purgeVisitedNodes()
|
|
|
|
|
{
|
2016-07-26 10:43:54 +00:00
|
|
|
$this->visitedNodes = array();
|
|
|
|
|
$this->styleAttributesForNodes = array();
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Marks a tag for removal.
|
|
|
|
|
*
|
|
|
|
|
* There are some HTML tags that DOMDocument cannot process, and it will throw an error if it encounters them.
|
|
|
|
|
* In particular, DOMDocument will complain if you try to use HTML5 tags in an XHTML document.
|
|
|
|
|
*
|
|
|
|
|
* Note: The tags will not be removed if they have any content.
|
|
|
|
|
*
|
|
|
|
|
* @param string $tagName the tag name, e.g., "p"
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function addUnprocessableHtmlTag($tagName)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$this->unprocessableHtmlTags[] = $tagName;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Drops a tag from the removal list.
|
|
|
|
|
*
|
|
|
|
|
* @param string $tagName the tag name, e.g., "p"
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function removeUnprocessableHtmlTag($tagName)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$key = array_search($tagName, $this->unprocessableHtmlTags, true);
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($key !== false) {
|
|
|
|
|
unset($this->unprocessableHtmlTags[$key]);
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Marks a media query type to keep.
|
|
|
|
|
*
|
|
|
|
|
* @param string $mediaName the media type name, e.g., "braille"
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function addAllowedMediaType($mediaName)
|
|
|
|
|
{
|
|
|
|
|
$this->allowedMediaTypes[$mediaName] = true;
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Drops a media query type from the allowed list.
|
|
|
|
|
*
|
|
|
|
|
* @param string $mediaName the tag name, e.g., "braille"
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function removeAllowedMediaType($mediaName)
|
|
|
|
|
{
|
|
|
|
|
if (isset($this->allowedMediaTypes[$mediaName])) {
|
|
|
|
|
unset($this->allowedMediaTypes[$mediaName]);
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Adds a selector to exclude nodes from emogrification.
|
|
|
|
|
*
|
|
|
|
|
* Any nodes that match the selector will not have their style altered.
|
|
|
|
|
*
|
|
|
|
|
* @param string $selector the selector to exclude, e.g., ".editor"
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function addExcludedSelector($selector)
|
|
|
|
|
{
|
|
|
|
|
$this->excludedSelectors[$selector] = true;
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* No longer excludes the nodes matching this selector from emogrification.
|
|
|
|
|
*
|
|
|
|
|
* @param string $selector the selector to no longer exclude, e.g., ".editor"
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
public function removeExcludedSelector($selector)
|
|
|
|
|
{
|
|
|
|
|
if (isset($this->excludedSelectors[$selector])) {
|
|
|
|
|
unset($this->excludedSelectors[$selector]);
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* This removes styles from your email that contain display:none.
|
|
|
|
|
* We need to look for display:none, but we need to do a case-insensitive search. Since DOMDocument only
|
|
|
|
|
* supports XPath 1.0, lower-case() isn't available to us. We've thus far only set attributes to lowercase,
|
|
|
|
|
* not attribute values. Consequently, we need to translate() the letters that would be in 'NONE' ("NOE")
|
|
|
|
|
* to lowercase.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @param \DOMXPath $xPath
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function removeInvisibleNodes(DOMXPath $xPath)
|
|
|
|
|
{
|
|
|
|
|
$nodesWithStyleDisplayNone = $xPath->query(
|
2016-07-25 13:51:13 +00:00
|
|
|
'//*[contains(translate(translate(@style," ",""),"NOE","noe"),"display:none")]'
|
|
|
|
|
);
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($nodesWithStyleDisplayNone->length === 0) {
|
2016-07-25 13:51:13 +00:00
|
|
|
return;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// The checks on parentNode and is_callable below ensure that if we've deleted the parent node,
|
|
|
|
|
// we don't try to call removeChild on a nonexistent child node
|
2017-03-31 20:19:35 +00:00
|
|
|
/** @var \DOMNode $node */
|
|
|
|
|
foreach ($nodesWithStyleDisplayNone as $node) {
|
|
|
|
|
if ($node->parentNode && is_callable(array($node->parentNode, 'removeChild'))) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$node->parentNode->removeChild($node);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2016-07-26 15:34:17 +00:00
|
|
|
private function normalizeStyleAttributes_callback( $m ) {
|
|
|
|
|
return strtolower( $m[0] );
|
|
|
|
|
}
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
/**
|
|
|
|
|
* Normalizes the value of the "style" attribute and saves it.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @param \DOMElement $node
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function normalizeStyleAttributes(DOMElement $node)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$normalizedOriginalStyle = preg_replace_callback(
|
|
|
|
|
'/[A-z\\-]+(?=\\:)/S',
|
2016-07-26 15:34:17 +00:00
|
|
|
array( $this, 'normalizeStyleAttributes_callback' ),
|
2016-07-25 13:51:13 +00:00
|
|
|
$node->getAttribute('style')
|
|
|
|
|
);
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
// in order to not overwrite existing style attributes in the HTML, we
|
|
|
|
|
// have to save the original HTML styles
|
|
|
|
|
$nodePath = $node->getNodePath();
|
2017-03-31 20:19:35 +00:00
|
|
|
if (!isset($this->styleAttributesForNodes[$nodePath])) {
|
|
|
|
|
$this->styleAttributesForNodes[$nodePath] = $this->parseCssDeclarationsBlock($normalizedOriginalStyle);
|
|
|
|
|
$this->visitedNodes[$nodePath] = $node;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
$node->setAttribute('style', $normalizedOriginalStyle);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
/**
|
|
|
|
|
* Merges styles from styles attributes and style nodes and applies them to the attribute nodes
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function fillStyleAttributesWithMergedStyles()
|
|
|
|
|
{
|
|
|
|
|
foreach ($this->styleAttributesForNodes as $nodePath => $styleAttributesForNode) {
|
|
|
|
|
$node = $this->visitedNodes[$nodePath];
|
2016-07-25 13:51:13 +00:00
|
|
|
$currentStyleAttributes = $this->parseCssDeclarationsBlock($node->getAttribute('style'));
|
|
|
|
|
$node->setAttribute(
|
|
|
|
|
'style',
|
|
|
|
|
$this->generateStyleStringFromDeclarationsArrays(
|
|
|
|
|
$currentStyleAttributes,
|
|
|
|
|
$styleAttributesForNode
|
|
|
|
|
)
|
|
|
|
|
);
|
|
|
|
|
}
|
|
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* This method merges old or existing name/value array with new name/value array
|
2015-02-12 11:42:15 +00:00
|
|
|
* and then generates a string of the combined style suitable for placing inline.
|
2016-07-25 13:51:13 +00:00
|
|
|
* This becomes the single point for CSS string generation allowing for consistent
|
2015-02-12 11:42:15 +00:00
|
|
|
* CSS output no matter where the CSS originally came from.
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
|
|
|
|
* @param string[] $oldStyles
|
|
|
|
|
* @param string[] $newStyles
|
|
|
|
|
*
|
2015-02-12 11:42:15 +00:00
|
|
|
* @return string
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function generateStyleStringFromDeclarationsArrays(array $oldStyles, array $newStyles)
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
$combinedStyles = array_merge($oldStyles, $newStyles);
|
2017-03-31 20:19:35 +00:00
|
|
|
$cacheKey = serialize($combinedStyles);
|
|
|
|
|
if (isset($this->caches[self::CACHE_KEY_COMBINED_STYLES][$cacheKey])) {
|
|
|
|
|
return $this->caches[self::CACHE_KEY_COMBINED_STYLES][$cacheKey];
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
foreach ($oldStyles as $attributeName => $attributeValue) {
|
|
|
|
|
if (!isset($newStyles[$attributeName])) {
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
$newAttributeValue = $newStyles[$attributeName];
|
|
|
|
|
if ($this->attributeValueIsImportant($attributeValue)
|
|
|
|
|
&& !$this->attributeValueIsImportant($newAttributeValue)
|
|
|
|
|
) {
|
|
|
|
|
$combinedStyles[$attributeName] = $attributeValue;
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-02-12 11:42:15 +00:00
|
|
|
$style = '';
|
2017-03-31 20:19:35 +00:00
|
|
|
foreach ($combinedStyles as $attributeName => $attributeValue) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$style .= strtolower(trim($attributeName)) . ': ' . trim($attributeValue) . '; ';
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
$trimmedStyle = rtrim($style);
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
$this->caches[self::CACHE_KEY_COMBINED_STYLES][$cacheKey] = $trimmedStyle;
|
2016-07-25 13:51:13 +00:00
|
|
|
|
|
|
|
|
return $trimmedStyle;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
/**
|
|
|
|
|
* Checks whether $attributeValue is marked as !important.
|
|
|
|
|
*
|
|
|
|
|
* @param string $attributeValue
|
|
|
|
|
*
|
|
|
|
|
* @return bool
|
|
|
|
|
*/
|
|
|
|
|
private function attributeValueIsImportant($attributeValue)
|
|
|
|
|
{
|
|
|
|
|
return strtolower(substr(trim($attributeValue), -10)) === '!important';
|
|
|
|
|
}
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
/**
|
|
|
|
|
* Applies $css to $xmlDocument, limited to the media queries that actually apply to the document.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @param \DOMDocument $xmlDocument the document to match against
|
|
|
|
|
* @param \DOMXPath $xPath
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string $css a string of CSS
|
|
|
|
|
*
|
|
|
|
|
* @return void
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function copyCssWithMediaToStyleNode(DOMDocument $xmlDocument, DOMXPath $xPath, $css)
|
|
|
|
|
{
|
|
|
|
|
if ($css === '') {
|
2016-07-25 13:51:13 +00:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2016-07-26 10:43:54 +00:00
|
|
|
$mediaQueriesRelevantForDocument = array();
|
2016-07-25 13:51:13 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
foreach ($this->extractMediaQueriesFromCss($css) as $mediaQuery) {
|
|
|
|
|
foreach ($this->parseCssRules($mediaQuery['css']) as $selector) {
|
|
|
|
|
if ($this->existsMatchForCssSelector($xPath, $selector['selector'])) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$mediaQueriesRelevantForDocument[] = $mediaQuery['query'];
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
$this->addStyleElementToDocument($xmlDocument, implode($mediaQueriesRelevantForDocument));
|
|
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
|
|
|
|
/**
|
2017-03-31 20:19:35 +00:00
|
|
|
* Extracts the media queries from $css while skipping empty media queries.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string $css
|
|
|
|
|
*
|
|
|
|
|
* @return string[][] numeric array with string sub-arrays with the keys "css" and "query"
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function extractMediaQueriesFromCss($css)
|
|
|
|
|
{
|
|
|
|
|
preg_match_all('/@media\\b[^{]*({((?:[^{}]+|(?1))*)})/', $css, $rawMediaQueries, PREG_SET_ORDER);
|
|
|
|
|
$parsedQueries = array();
|
|
|
|
|
|
|
|
|
|
foreach ($rawMediaQueries as $mediaQuery) {
|
|
|
|
|
if ($mediaQuery[2] !== '') {
|
|
|
|
|
$parsedQueries[] = array(
|
|
|
|
|
'css' => $mediaQuery[2],
|
|
|
|
|
'query' => $mediaQuery[0],
|
|
|
|
|
);
|
|
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
2017-03-31 20:19:35 +00:00
|
|
|
|
|
|
|
|
return $parsedQueries;
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Checks whether there is at least one matching element for $cssSelector.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @param \DOMXPath $xPath
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string $cssSelector
|
|
|
|
|
*
|
|
|
|
|
* @return bool
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function existsMatchForCssSelector(DOMXPath $xPath, $cssSelector)
|
|
|
|
|
{
|
|
|
|
|
$nodesMatchingSelector = $xPath->query($this->translateCssToXpath($cssSelector));
|
2016-07-25 13:51:13 +00:00
|
|
|
|
|
|
|
|
return $nodesMatchingSelector !== false && $nodesMatchingSelector->length !== 0;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Returns CSS content.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @param \DOMXPath $xPath
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
2015-02-12 11:42:15 +00:00
|
|
|
* @return string
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function getCssFromAllStyleNodes(DOMXPath $xPath)
|
|
|
|
|
{
|
|
|
|
|
$styleNodes = $xPath->query('//style');
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($styleNodes === false) {
|
2015-02-12 11:42:15 +00:00
|
|
|
return '';
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
$css = '';
|
2017-03-31 20:19:35 +00:00
|
|
|
/** @var \DOMNode $styleNode */
|
|
|
|
|
foreach ($styleNodes as $styleNode) {
|
2015-02-12 11:42:15 +00:00
|
|
|
$css .= "\n\n" . $styleNode->nodeValue;
|
|
|
|
|
$styleNode->parentNode->removeChild($styleNode);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return $css;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Adds a style element with $css to $document.
|
|
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* This method is protected to allow overriding.
|
|
|
|
|
*
|
|
|
|
|
* @see https://github.com/jjriv/emogrifier/issues/103
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @param \DOMDocument $document
|
2015-02-12 11:42:15 +00:00
|
|
|
* @param string $css
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
|
|
|
|
* @return void
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
protected function addStyleElementToDocument(DOMDocument $document, $css)
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
$styleElement = $document->createElement('style', $css);
|
|
|
|
|
$styleAttribute = $document->createAttribute('type');
|
|
|
|
|
$styleAttribute->value = 'text/css';
|
|
|
|
|
$styleElement->appendChild($styleAttribute);
|
|
|
|
|
|
|
|
|
|
$head = $this->getOrCreateHeadElement($document);
|
|
|
|
|
$head->appendChild($styleElement);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Returns the existing or creates a new head element in $document.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @param \DOMDocument $document
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @return \DOMNode the head element
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function getOrCreateHeadElement(DOMDocument $document)
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
$head = $document->getElementsByTagName('head')->item(0);
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($head === null) {
|
2015-02-12 11:42:15 +00:00
|
|
|
$head = $document->createElement('head');
|
|
|
|
|
$html = $document->getElementsByTagName('html')->item(0);
|
|
|
|
|
$html->insertBefore($head, $document->getElementsByTagName('body')->item(0));
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return $head;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Splits input CSS code to an array where:
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* - key "css" will be contains clean CSS code
|
|
|
|
|
* - key "media" will be contains all valuable media queries
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* Example:
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* The CSS code
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* "@import "file.css"; h1 { color:red; } @media { h1 {}} @media tv { h1 {}}"
|
|
|
|
|
*
|
|
|
|
|
* will be parsed into the following array:
|
|
|
|
|
*
|
|
|
|
|
* "css" => "h1 { color:red; }"
|
|
|
|
|
* "media" => "@media { h1 {}}"
|
|
|
|
|
*
|
|
|
|
|
* @param string $css
|
2017-03-31 20:19:35 +00:00
|
|
|
*
|
|
|
|
|
* @return string[]
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function splitCssAndMediaQuery($css)
|
|
|
|
|
{
|
|
|
|
|
$cssWithoutComments = preg_replace('/\\/\\*.*\\*\\//sU', '', $css);
|
|
|
|
|
|
|
|
|
|
$mediaTypesExpression = '';
|
|
|
|
|
if (!empty($this->allowedMediaTypes)) {
|
|
|
|
|
$mediaTypesExpression = '|' . implode('|', array_keys($this->allowedMediaTypes));
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
$media = '';
|
|
|
|
|
$cssForAllowedMediaTypes = preg_replace_callback(
|
|
|
|
|
'#@media\\s+(?:only\\s)?(?:[\\s{\\(]' . $mediaTypesExpression . ')\\s?[^{]+{.*}\\s*}\\s*#misU',
|
|
|
|
|
array( $this, '_media_concat' ),
|
|
|
|
|
$cssWithoutComments
|
|
|
|
|
);
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
// filter the CSS
|
2016-07-26 10:43:54 +00:00
|
|
|
$search = array(
|
2017-03-31 20:19:35 +00:00
|
|
|
'import directives' => '/^\\s*@import\\s[^;]+;/misU',
|
|
|
|
|
'remaining media enclosures' => '/^\\s*@media\\s[^{]+{(.*)}\\s*}\\s/misU',
|
2016-07-26 10:43:54 +00:00
|
|
|
);
|
2017-03-31 20:19:35 +00:00
|
|
|
|
|
|
|
|
$cleanedCss = preg_replace($search, '', $cssForAllowedMediaTypes);
|
|
|
|
|
|
|
|
|
|
return array('css' => $cleanedCss, 'media' => self::$_media);
|
2016-07-26 15:50:23 +00:00
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2016-07-26 15:50:23 +00:00
|
|
|
private function _media_concat( $matches ) {
|
|
|
|
|
self::$_media .= $matches[0];
|
2015-01-23 12:41:15 +00:00
|
|
|
}
|
|
|
|
|
|
2015-02-12 11:42:15 +00:00
|
|
|
/**
|
|
|
|
|
* Creates a DOMDocument instance with the current HTML.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @return \DOMDocument
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function createXmlDocument()
|
|
|
|
|
{
|
|
|
|
|
$xmlDocument = new DOMDocument;
|
2016-07-25 13:51:13 +00:00
|
|
|
$xmlDocument->encoding = 'UTF-8';
|
2015-02-18 17:34:03 +00:00
|
|
|
$xmlDocument->strictErrorChecking = false;
|
|
|
|
|
$xmlDocument->formatOutput = true;
|
|
|
|
|
$libXmlState = libxml_use_internal_errors(true);
|
2015-02-12 11:42:15 +00:00
|
|
|
$xmlDocument->loadHTML($this->getUnifiedHtml());
|
|
|
|
|
libxml_clear_errors();
|
|
|
|
|
libxml_use_internal_errors($libXmlState);
|
|
|
|
|
$xmlDocument->normalizeDocument();
|
|
|
|
|
|
|
|
|
|
return $xmlDocument;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* Returns the HTML with the unprocessable HTML tags removed and
|
|
|
|
|
* with added document type and Content-Type meta tag if needed.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* @return string the unified HTML
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @throws \BadMethodCallException
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function getUnifiedHtml()
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$htmlWithoutUnprocessableTags = $this->removeUnprocessableTags($this->html);
|
|
|
|
|
$htmlWithDocumentType = $this->ensureDocumentType($htmlWithoutUnprocessableTags);
|
|
|
|
|
|
|
|
|
|
return $this->addContentTypeMetaTag($htmlWithDocumentType);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Removes the unprocessable tags from $html (if this feature is enabled).
|
|
|
|
|
*
|
|
|
|
|
* @param string $html
|
|
|
|
|
*
|
|
|
|
|
* @return string the reworked HTML with the unprocessable tags removed
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function removeUnprocessableTags($html)
|
|
|
|
|
{
|
|
|
|
|
if (empty($this->unprocessableHtmlTags)) {
|
2016-07-25 13:51:13 +00:00
|
|
|
return $html;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
$unprocessableHtmlTags = implode('|', $this->unprocessableHtmlTags);
|
|
|
|
|
|
|
|
|
|
return preg_replace(
|
|
|
|
|
'/<\\/?(' . $unprocessableHtmlTags . ')[^>]*>/i',
|
|
|
|
|
'',
|
|
|
|
|
$html
|
|
|
|
|
);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Makes sure that the passed HTML has a document type.
|
|
|
|
|
*
|
|
|
|
|
* @param string $html
|
|
|
|
|
*
|
|
|
|
|
* @return string HTML with document type
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function ensureDocumentType($html)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$hasDocumentType = stripos($html, '<!DOCTYPE') !== false;
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($hasDocumentType) {
|
2016-07-25 13:51:13 +00:00
|
|
|
return $html;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return self::DEFAULT_DOCUMENT_TYPE . $html;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Adds a Content-Type meta tag for the charset.
|
|
|
|
|
*
|
|
|
|
|
* @param string $html
|
|
|
|
|
*
|
|
|
|
|
* @return string the HTML with the meta tag added
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function addContentTypeMetaTag($html)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$hasContentTypeMetaTag = stristr($html, 'Content-Type') !== false;
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($hasContentTypeMetaTag) {
|
2016-07-25 13:51:13 +00:00
|
|
|
return $html;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// We are trying to insert the meta tag to the right spot in the DOM.
|
|
|
|
|
// If we just prepended it to the HTML, we would lose attributes set to the HTML tag.
|
|
|
|
|
$hasHeadTag = stripos($html, '<head') !== false;
|
|
|
|
|
$hasHtmlTag = stripos($html, '<html') !== false;
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($hasHeadTag) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$reworkedHtml = preg_replace('/<head(.*?)>/i', '<head$1>' . self::CONTENT_TYPE_META_TAG, $html);
|
2017-03-31 20:19:35 +00:00
|
|
|
} elseif ($hasHtmlTag) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$reworkedHtml = preg_replace(
|
|
|
|
|
'/<html(.*?)>/i',
|
|
|
|
|
'<html$1><head>' . self::CONTENT_TYPE_META_TAG . '</head>',
|
|
|
|
|
$html
|
|
|
|
|
);
|
2015-02-12 11:42:15 +00:00
|
|
|
} else {
|
2016-07-25 13:51:13 +00:00
|
|
|
$reworkedHtml = self::CONTENT_TYPE_META_TAG . $html;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
|
|
|
|
|
return $reworkedHtml;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string[] $a
|
|
|
|
|
* @param string[] $b
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @return int
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function sortBySelectorPrecedence(array $a, array $b)
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
$precedenceA = $this->getCssSelectorPrecedence($a['selector']);
|
|
|
|
|
$precedenceB = $this->getCssSelectorPrecedence($b['selector']);
|
|
|
|
|
|
|
|
|
|
// We want these sorted in ascending order so selectors with lesser precedence get processed first and
|
|
|
|
|
// selectors with greater precedence get sorted last.
|
2016-07-25 13:51:13 +00:00
|
|
|
$precedenceForEquals = ($a['line'] < $b['line'] ? -1 : 1);
|
|
|
|
|
$precedenceForNotEquals = ($precedenceA < $precedenceB ? -1 : 1);
|
2015-02-12 11:42:15 +00:00
|
|
|
return ($precedenceA === $precedenceB) ? $precedenceForEquals : $precedenceForNotEquals;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* @param string $selector
|
|
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @return int
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function getCssSelectorPrecedence($selector)
|
|
|
|
|
{
|
2015-02-12 11:42:15 +00:00
|
|
|
$selectorKey = md5($selector);
|
2017-03-31 20:19:35 +00:00
|
|
|
if (!isset($this->caches[self::CACHE_KEY_SELECTOR][$selectorKey])) {
|
2015-02-12 11:42:15 +00:00
|
|
|
$precedence = 0;
|
|
|
|
|
$value = 100;
|
|
|
|
|
// ids: worth 100, classes: worth 10, elements: worth 1
|
2017-03-31 20:19:35 +00:00
|
|
|
$search = array('\\#','\\.','');
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
foreach ($search as $s) {
|
|
|
|
|
if (trim($selector) === '') {
|
2015-02-12 11:42:15 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
$number = 0;
|
|
|
|
|
$selector = preg_replace('/' . $s . '\\w+/', '', $selector, -1, $number);
|
|
|
|
|
$precedence += ($value * $number);
|
|
|
|
|
$value /= 10;
|
|
|
|
|
}
|
2017-03-31 20:19:35 +00:00
|
|
|
$this->caches[self::CACHE_KEY_SELECTOR][$selectorKey] = $precedence;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
return $this->caches[self::CACHE_KEY_SELECTOR][$selectorKey];
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2016-07-26 15:50:23 +00:00
|
|
|
private function translateCssToXpath_callback( $matches ) {
|
|
|
|
|
return strtolower($matches[0]);
|
|
|
|
|
}
|
|
|
|
|
|
2015-02-12 11:42:15 +00:00
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* Maps a CSS selector to an XPath query string.
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* @see http://plasmasturm.org/log/444/
|
|
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string $cssSelector a CSS selector
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @return string the corresponding XPath selector
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function translateCssToXpath($cssSelector)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$paddedSelector = ' ' . $cssSelector . ' ';
|
|
|
|
|
$lowercasePaddedSelector = preg_replace_callback(
|
|
|
|
|
'/\\s+\\w+\\s+/',
|
2016-07-26 15:50:23 +00:00
|
|
|
array( $this, 'translateCssToXpath_callback' ),
|
2016-07-25 13:51:13 +00:00
|
|
|
$paddedSelector
|
|
|
|
|
);
|
2017-03-31 20:19:35 +00:00
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
$trimmedLowercaseSelector = trim($lowercasePaddedSelector);
|
2017-03-31 20:19:35 +00:00
|
|
|
$xPathKey = md5($trimmedLowercaseSelector);
|
|
|
|
|
if (!isset($this->caches[self::CACHE_KEY_XPATH][$xPathKey])) {
|
|
|
|
|
$roughXpath = '//' . preg_replace(
|
|
|
|
|
array_keys($this->xPathRules),
|
|
|
|
|
$this->xPathRules,
|
|
|
|
|
$trimmedLowercaseSelector
|
2016-07-26 10:43:54 +00:00
|
|
|
);
|
2017-03-31 20:19:35 +00:00
|
|
|
$xPathWithIdAttributeMatchers = preg_replace_callback(
|
2016-07-25 13:51:13 +00:00
|
|
|
self::ID_ATTRIBUTE_MATCHER,
|
2017-03-31 20:19:35 +00:00
|
|
|
array($this, 'matchIdAttributes'),
|
2016-07-25 13:51:13 +00:00
|
|
|
$roughXpath
|
2015-02-12 11:42:15 +00:00
|
|
|
);
|
2017-03-31 20:19:35 +00:00
|
|
|
$xPathWithIdAttributeAndClassMatchers = preg_replace_callback(
|
2016-07-25 13:51:13 +00:00
|
|
|
self::CLASS_ATTRIBUTE_MATCHER,
|
2017-03-31 20:19:35 +00:00
|
|
|
array($this, 'matchClassAttributes'),
|
|
|
|
|
$xPathWithIdAttributeMatchers
|
2015-02-12 11:42:15 +00:00
|
|
|
);
|
|
|
|
|
|
|
|
|
|
// Advanced selectors are going to require a bit more advanced emogrification.
|
|
|
|
|
// When we required PHP 5.3, we could do this with closures.
|
2017-03-31 20:19:35 +00:00
|
|
|
$xPathWithIdAttributeAndClassMatchers = preg_replace_callback(
|
2016-07-25 13:51:13 +00:00
|
|
|
'/([^\\/]+):nth-child\\(\\s*(odd|even|[+\\-]?\\d|[+\\-]?\\d?n(\\s*[+\\-]\\s*\\d)?)\\s*\\)/i',
|
2017-03-31 20:19:35 +00:00
|
|
|
array($this, 'translateNthChild'),
|
|
|
|
|
$xPathWithIdAttributeAndClassMatchers
|
2015-02-12 11:42:15 +00:00
|
|
|
);
|
2016-07-25 13:51:13 +00:00
|
|
|
$finalXpath = preg_replace_callback(
|
|
|
|
|
'/([^\\/]+):nth-of-type\\(\s*(odd|even|[+\\-]?\\d|[+\\-]?\\d?n(\\s*[+\\-]\\s*\\d)?)\\s*\\)/i',
|
2017-03-31 20:19:35 +00:00
|
|
|
array($this, 'translateNthOfType'),
|
|
|
|
|
$xPathWithIdAttributeAndClassMatchers
|
2015-02-12 11:42:15 +00:00
|
|
|
);
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
$this->caches[self::CACHE_KEY_SELECTOR][$xPathKey] = $finalXpath;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
2017-03-31 20:19:35 +00:00
|
|
|
return $this->caches[self::CACHE_KEY_SELECTOR][$xPathKey];
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string[] $match
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* @return string
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function matchIdAttributes(array $match)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
return ($match[1] !== '' ? $match[1] : '*') . '[@id="' . $match[2] . '"]';
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string[] $match
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* @return string
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function matchClassAttributes(array $match)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
return ($match[1] !== '' ? $match[1] : '*') . '[contains(concat(" ",@class," "),concat(" ","' .
|
2015-02-12 11:42:15 +00:00
|
|
|
implode(
|
|
|
|
|
'"," "))][contains(concat(" ",@class," "),concat(" ","',
|
|
|
|
|
explode('.', substr($match[2], 1))
|
|
|
|
|
) . '"," "))]';
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string[] $match
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* @return string
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function translateNthChild(array $match)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$parseResult = $this->parseNth($match);
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
if (isset($parseResult[self::MULTIPLIER])) {
|
|
|
|
|
if ($parseResult[self::MULTIPLIER] < 0) {
|
|
|
|
|
$parseResult[self::MULTIPLIER] = abs($parseResult[self::MULTIPLIER]);
|
2016-07-25 13:51:13 +00:00
|
|
|
$xPathExpression = sprintf(
|
|
|
|
|
'*[(last() - position()) mod %u = %u]/self::%s',
|
2017-03-31 20:19:35 +00:00
|
|
|
$parseResult[self::MULTIPLIER],
|
|
|
|
|
$parseResult[self::INDEX],
|
2016-07-25 13:51:13 +00:00
|
|
|
$match[1]
|
|
|
|
|
);
|
2015-02-12 11:42:15 +00:00
|
|
|
} else {
|
2016-07-25 13:51:13 +00:00
|
|
|
$xPathExpression = sprintf(
|
|
|
|
|
'*[position() mod %u = %u]/self::%s',
|
2017-03-31 20:19:35 +00:00
|
|
|
$parseResult[self::MULTIPLIER],
|
|
|
|
|
$parseResult[self::INDEX],
|
2016-07-25 13:51:13 +00:00
|
|
|
$match[1]
|
|
|
|
|
);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
} else {
|
2017-03-31 20:19:35 +00:00
|
|
|
$xPathExpression = sprintf('*[%u]/self::%s', $parseResult[self::INDEX], $match[1]);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
|
|
|
|
|
return $xPathExpression;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string[] $match
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* @return string
|
|
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function translateNthOfType(array $match)
|
|
|
|
|
{
|
2016-07-25 13:51:13 +00:00
|
|
|
$parseResult = $this->parseNth($match);
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
if (isset($parseResult[self::MULTIPLIER])) {
|
|
|
|
|
if ($parseResult[self::MULTIPLIER] < 0) {
|
|
|
|
|
$parseResult[self::MULTIPLIER] = abs($parseResult[self::MULTIPLIER]);
|
2016-07-25 13:51:13 +00:00
|
|
|
$xPathExpression = sprintf(
|
|
|
|
|
'%s[(last() - position()) mod %u = %u]',
|
|
|
|
|
$match[1],
|
2017-03-31 20:19:35 +00:00
|
|
|
$parseResult[self::MULTIPLIER],
|
|
|
|
|
$parseResult[self::INDEX]
|
2016-07-25 13:51:13 +00:00
|
|
|
);
|
2015-02-12 11:42:15 +00:00
|
|
|
} else {
|
2016-07-25 13:51:13 +00:00
|
|
|
$xPathExpression = sprintf(
|
|
|
|
|
'%s[position() mod %u = %u]',
|
|
|
|
|
$match[1],
|
2017-03-31 20:19:35 +00:00
|
|
|
$parseResult[self::MULTIPLIER],
|
|
|
|
|
$parseResult[self::INDEX]
|
2016-07-25 13:51:13 +00:00
|
|
|
);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
} else {
|
2017-03-31 20:19:35 +00:00
|
|
|
$xPathExpression = sprintf('%s[%u]', $match[1], $parseResult[self::INDEX]);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
|
|
|
|
|
return $xPathExpression;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string[] $match
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @return int[]
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function parseNth(array $match)
|
|
|
|
|
{
|
|
|
|
|
if (in_array(strtolower($match[2]), array('even', 'odd'), true)) {
|
2016-07-25 13:51:13 +00:00
|
|
|
// we have "even" or "odd"
|
|
|
|
|
$index = strtolower($match[2]) === 'even' ? 0 : 1;
|
2017-03-31 20:19:35 +00:00
|
|
|
return array(self::MULTIPLIER => 2, self::INDEX => $index);
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
2017-03-31 20:19:35 +00:00
|
|
|
if (stripos($match[2], 'n') === false) {
|
2015-02-12 11:42:15 +00:00
|
|
|
// if there is a multiplier
|
2016-07-25 13:51:13 +00:00
|
|
|
$index = (int) str_replace(' ', '', $match[2]);
|
2017-03-31 20:19:35 +00:00
|
|
|
return array(self::INDEX => $index);
|
2016-07-25 13:51:13 +00:00
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
if (isset($match[3])) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$multipleTerm = str_replace($match[3], '', $match[2]);
|
|
|
|
|
$index = (int) str_replace(' ', '', $match[3]);
|
|
|
|
|
} else {
|
|
|
|
|
$multipleTerm = $match[2];
|
|
|
|
|
$index = 0;
|
|
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2016-07-25 13:51:13 +00:00
|
|
|
$multiplier = str_ireplace('n', '', $multipleTerm);
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
if ($multiplier === '') {
|
2016-07-25 13:51:13 +00:00
|
|
|
$multiplier = 1;
|
2017-03-31 20:19:35 +00:00
|
|
|
} elseif ($multiplier === '0') {
|
|
|
|
|
return array(self::INDEX => $index);
|
2016-07-25 13:51:13 +00:00
|
|
|
} else {
|
|
|
|
|
$multiplier = (int) $multiplier;
|
|
|
|
|
}
|
2015-02-12 11:42:15 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
while ($index < 0) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$index += abs($multiplier);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
return array(self::MULTIPLIER => $multiplier, self::INDEX => $index);
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Parses a CSS declaration block into property name/value pairs.
|
|
|
|
|
*
|
|
|
|
|
* Example:
|
|
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* The declaration block
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* "color: #000; font-weight: bold;"
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
|
|
|
|
* will be parsed into the following array:
|
|
|
|
|
*
|
|
|
|
|
* "color" => "#000"
|
|
|
|
|
* "font-weight" => "bold"
|
|
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @param string $cssDeclarationsBlock the CSS declarations block without the curly braces, may be empty
|
2015-02-12 11:42:15 +00:00
|
|
|
*
|
2016-07-25 13:51:13 +00:00
|
|
|
* @return string[]
|
|
|
|
|
* the CSS declarations with the property names as array keys and the property values as array values
|
2015-02-12 11:42:15 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function parseCssDeclarationsBlock($cssDeclarationsBlock)
|
|
|
|
|
{
|
|
|
|
|
if (isset($this->caches[self::CACHE_KEY_CSS_DECLARATIONS_BLOCK][$cssDeclarationsBlock])) {
|
|
|
|
|
return $this->caches[self::CACHE_KEY_CSS_DECLARATIONS_BLOCK][$cssDeclarationsBlock];
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
|
|
|
|
|
2016-07-26 10:43:54 +00:00
|
|
|
$properties = array();
|
2016-07-25 13:51:13 +00:00
|
|
|
$declarations = preg_split('/;(?!base64|charset)/', $cssDeclarationsBlock);
|
|
|
|
|
|
2017-03-31 20:19:35 +00:00
|
|
|
foreach ($declarations as $declaration) {
|
2016-07-26 10:43:54 +00:00
|
|
|
$matches = array();
|
2017-03-31 20:19:35 +00:00
|
|
|
if (!preg_match('/^([A-Za-z\\-]+)\\s*:\\s*(.+)$/', trim($declaration), $matches)) {
|
2015-02-12 11:42:15 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
|
2015-02-12 11:42:15 +00:00
|
|
|
$propertyName = strtolower($matches[1]);
|
|
|
|
|
$propertyValue = $matches[2];
|
2017-03-31 20:19:35 +00:00
|
|
|
$properties[$propertyName] = $propertyValue;
|
2015-02-12 11:42:15 +00:00
|
|
|
}
|
2017-03-31 20:19:35 +00:00
|
|
|
$this->caches[self::CACHE_KEY_CSS_DECLARATIONS_BLOCK][$cssDeclarationsBlock] = $properties;
|
2015-02-12 11:42:15 +00:00
|
|
|
|
|
|
|
|
return $properties;
|
|
|
|
|
}
|
2016-07-25 13:51:13 +00:00
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Find the nodes that are not to be emogrified.
|
|
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @param \DOMXPath $xPath
|
2016-07-25 13:51:13 +00:00
|
|
|
*
|
2017-03-31 20:19:35 +00:00
|
|
|
* @return \DOMElement[]
|
2016-07-25 13:51:13 +00:00
|
|
|
*/
|
2017-03-31 20:19:35 +00:00
|
|
|
private function getNodesToExclude(DOMXPath $xPath)
|
|
|
|
|
{
|
2016-07-26 10:43:54 +00:00
|
|
|
$excludedNodes = array();
|
2017-03-31 20:19:35 +00:00
|
|
|
foreach (array_keys($this->excludedSelectors) as $selectorToExclude) {
|
|
|
|
|
foreach ($xPath->query($this->translateCssToXpath($selectorToExclude)) as $node) {
|
2016-07-25 13:51:13 +00:00
|
|
|
$excludedNodes[] = $node;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return $excludedNodes;
|
|
|
|
|
}
|
2017-03-31 20:19:35 +00:00
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Handles invalid xPath expression warnings, generated by process() method,
|
|
|
|
|
* during querying \DOMDocument and trigger \InvalidArgumentException
|
|
|
|
|
* with invalid selector.
|
|
|
|
|
*
|
|
|
|
|
* @param int $type
|
|
|
|
|
* @param string $message
|
|
|
|
|
* @param string $file
|
|
|
|
|
* @param int $line
|
|
|
|
|
* @param array $context
|
|
|
|
|
*
|
|
|
|
|
* @return bool always false
|
|
|
|
|
*
|
|
|
|
|
* @throws \InvalidArgumentException
|
|
|
|
|
*/
|
|
|
|
|
public function handleXpathError($type, $message, $file, $line, array $context)
|
|
|
|
|
{
|
|
|
|
|
if ($type === E_WARNING && isset($context['cssRule']['selector'])) {
|
|
|
|
|
throw new InvalidArgumentException(
|
|
|
|
|
sprintf(
|
|
|
|
|
'%s in selector >> %s << in %s on line %s',
|
|
|
|
|
$message,
|
|
|
|
|
$context['cssRule']['selector'],
|
|
|
|
|
$file,
|
|
|
|
|
$line
|
|
|
|
|
)
|
|
|
|
|
);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// the normal error handling continues when handler return false
|
|
|
|
|
return false;
|
|
|
|
|
}
|
2014-10-08 21:32:46 +00:00
|
|
|
}
|
