"HTMLText",
"BigSummary" => "HTMLText",
"ContextSummary" => "HTMLText",
"FirstParagraph" => "HTMLText",
"FirstSentence" => "HTMLText",
"LimitCharacters" => "HTMLText",
"LimitSentences" => "HTMLText",
"Lower" => "HTMLText",
"LowerCase" => "HTMLText",
"Summary" => "HTMLText",
"Upper" => "HTMLText",
"UpperCase" => "HTMLText",
'EscapeXML' => 'HTMLText',
'LimitWordCount' => 'HTMLText',
'LimitWordCountXML' => 'HTMLText',
'NoHTML' => 'Text',
);
protected $processShortcodes = true;
/**
* Check if shortcodes are enabled
*
* @return bool
*/
public function getProcessShortcodes() {
return $this->processShortcodes;
}
/**
* Set shortcodes on or off by default
*
* @param bool $process
* @return $this
*/
public function setProcessShortcodes($process) {
$this->processShortcodes = (bool)$process;
return $this;
}
protected $whitelist = false;
public function __construct($name = null, $options = array()) {
if(is_string($options)) {
$options = array('whitelist' => $options);
}
return parent::__construct($name, $options);
}
/**
* @param array $options
*
* Options accepted in addition to those provided by Text:
*
* - shortcodes: If true, shortcodes will be turned into the appropriate HTML.
* If false, shortcodes will not be processed.
*
* - whitelist: If provided, a comma-separated list of elements that will be allowed to be stored
* (be careful on relying on this for XSS protection - some seemingly-safe elements allow
* attributes that can be exploited, for instance )
* Text nodes outside of HTML tags are filtered out by default, but may be included by adding
* the text() directive. E.g. 'link,meta,text()' will allow only and text at
* the root level.
*/
public function setOptions(array $options = array()) {
parent::setOptions($options);
if(array_key_exists("shortcodes", $options)) {
$this->processShortcodes = !!$options["shortcodes"];
}
if(array_key_exists("whitelist", $options)) {
if(is_array($options['whitelist'])) {
$this->whitelist = $options['whitelist'];
}
else {
$this->whitelist = preg_split('/,\s*/', $options['whitelist']);
}
}
}
/**
* Create a summary of the content. This will be some section of the first paragraph, limited by
* $maxWords. All internal tags are stripped out - the return value is a string
*
* This is sort of the HTML aware equivilent to Text#Summary, although the logic for summarising is not exactly
* the same
*
* @param int $maxWords Maximum number of words to return - may return less, but never more. Pass -1 for no limit
* @param int $flex Number of words to search through when looking for a nice cut point
* @param string $add What to add to the end of the summary if we cut at a less-than-ideal cut point
* @return string A nice(ish) summary with no html tags (but possibly still some html entities)
*
* @see framework/core/model/fieldtypes/Text#Summary($maxWords)
*/
public function Summary($maxWords = 50, $flex = 15, $add = '...') {
$str = false;
/* First we need the text of the first paragraph, without tags. Try using SimpleXML first */
if (class_exists('SimpleXMLElement')) {
$doc = new DOMDocument();
// Catch warnings thrown by loadHTML and turn them into a failure boolean rather than a SilverStripe error
set_error_handler(create_function('$no, $str', 'throw new Exception("HTML Parse Error: ".$str);'), E_ALL);
// Nonbreaking spaces get converted into weird characters, so strip them
$value = str_replace(' ', ' ', $this->RAW());
try {
$res = $doc->loadHTML('' . $value);
}
catch (Exception $e) { $res = false; }
restore_error_handler();
if ($res) {
$xml = simplexml_import_dom($doc);
$res = $xml->xpath('//p');
if (!empty($res)) $str = strip_tags($res[0]->asXML());
}
}
/* If that failed, most likely the passed HTML is broken. use a simple regex + a custom more brutal strip_tags.
* We don't use strip_tags because that does very badly on broken HTML */
if (!$str) {
/* See if we can pull a paragraph out*/
// Strip out any images in case there's one at the beginning. Not doing this will return a blank paragraph
$str = preg_replace('{^\s*(<.+?>)*
]*>}', '', $this->value);
if (preg_match('{
]*)?>(.*[A-Za-z]+.*)
}', $str, $matches)) $str = $matches[2]; /* If _that_ failed, just use the whole text */ if (!$str) $str = $this->value; /* Now pull out all the html-alike stuff */ /* Take out anything that is obviously a tag */ $str = preg_replace('{?[a-zA-Z]+[^<>]*>}', '', $str); /* Strip out any left over looking bits. Textual < or > should already be encoded to < or > */ $str = preg_replace('{|<|>}', '', $str); } /* Now split into words. If we are under the maxWords limit, just return the whole string (re-implode for * whitespace normalization) */ $words = preg_split('/\s+/', $str); if ($maxWords == -1 || count($words) <= $maxWords) return implode(' ', $words); /* Otherwise work backwards for a looking for a sentence ending (we try to avoid abbreviations, but aren't * very good at it) */ for ($i = $maxWords; $i >= $maxWords - $flex && $i >= 0; $i--) { if (preg_match('/\.$/', $words[$i]) && !preg_match('/(Dr|Mr|Mrs|Ms|Miss|Sr|Jr|No)\.$/i', $words[$i])) { return implode(' ', array_slice($words, 0, $i+1)); } } // If we didn't find a sentence ending quickly enough, just cut at the maxWords point and add '...' to the end return implode(' ', array_slice($words, 0, $maxWords)) . $add; } /** * Returns the first sentence from the first paragraph. If it can't figure out what the first paragraph is (or * there isn't one), it returns the same as Summary() * * This is the HTML aware equivilent to Text#FirstSentence * * @see framework/core/model/fieldtypes/Text#FirstSentence() */ public function FirstSentence() { /* Use summary's html processing logic to get the first paragraph */ $paragraph = $this->Summary(-1); /* Then look for the first sentence ending. We could probably use a nice regex, but for now this will do */ $words = preg_split('/\s+/', $paragraph); foreach ($words as $i => $word) { if (preg_match('/(!|\?|\.)$/', $word) && !preg_match('/(Dr|Mr|Mrs|Ms|Miss|Sr|Jr|No)\.$/i', $word)) { return implode(' ', array_slice($words, 0, $i+1)); } } /* If we didn't find a sentence ending, use the summary. We re-call rather than using paragraph so that * Summary will limit the result this time */ return $this->Summary(); } public function RAW() { if ($this->processShortcodes) { return ShortcodeParser::get_active()->parse($this->value); } else { return $this->value; } } /** * Return the value of the field with relative links converted to absolute urls (with placeholders parsed). * @return string */ public function AbsoluteLinks() { return HTTP::absoluteURLs($this->forTemplate()); } public function forTemplate() { return $this->RAW(); } public function prepValueForDB($value) { return parent::prepValueForDB($this->whitelistContent($value)); } /** * Filter the given $value string through the whitelist filter * * @param string $value Input html content * @return string Value with all non-whitelisted content stripped (if applicable) */ public function whitelistContent($value) { if($this->whitelist) { $dom = Injector::inst()->create('HTMLValue', $value); $query = array(); $textFilter = ' | //body/text()'; foreach ($this->whitelist as $tag) { if($tag === 'text()') { $textFilter = ''; // Disable text filter if allowed } else { $query[] = 'not(self::'.$tag.')'; } } foreach($dom->query('//body//*['.implode(' and ', $query).']'.$textFilter) as $el) { if ($el->parentNode) $el->parentNode->removeChild($el); } $value = $dom->getContent(); } return $value; } /** * Returns true if the field has meaningful content. * Excludes null content like , ,etc * * @return boolean */ public function exists() { // If it's blank, it's blank if(!parent::exists()) { return false; } $value = $this->RAW(); // If it's got a content tag if(preg_match('/<(img|embed|object|iframe|meta|source|link)[^>]*>/i', $value)) { return true; } // If it's just one or two tags on its own (and not the above) it's empty. // This might be or or whatever. if(preg_match('/^[\\s]*(<[^>]+>[\\s]*){1,2}$/', $value)) { return false; } // Otherwise its content is genuine content return true; } public function scaffoldFormField($title = null, $params = null) { return new HtmlEditorField($this->name, $title); } public function scaffoldSearchField($title = null, $params = null) { return new TextField($this->name, $title); } }