1
0
mirror of https://github.com/moparisthebest/wallabag synced 2024-12-26 01:09:19 -05:00

fixes #963 and use our own readability.php file for mobiClass

This commit is contained in:
Thomas Citharel 2015-01-03 13:03:26 +01:00
parent 166ff0a093
commit 8ae45e7fe2
5 changed files with 7 additions and 1183 deletions

View File

@ -1,5 +1,5 @@
<?php <?php
require_once(dirname(__FILE__)."/readability/Readability.php"); require_once(dirname(__FILE__)."/../readability/Readability.php");
require_once(dirname(__FILE__).'/CharacterEntities.php'); require_once(dirname(__FILE__).'/CharacterEntities.php');
require_once(dirname(__FILE__).'/constants.php'); require_once(dirname(__FILE__).'/constants.php');
require_once(dirname(__FILE__).'/ContentProvider.php'); require_once(dirname(__FILE__).'/ContentProvider.php');

View File

@ -1,110 +0,0 @@
<?php
/**
* JavaScript-like HTML DOM Element
*
* This class extends PHP's DOMElement to allow
* users to get and set the innerHTML property of
* HTML elements in the same way it's done in
* JavaScript.
*
* Example usage:
* @code
* require_once 'JSLikeHTMLElement.php';
* header('Content-Type: text/plain');
* $doc = new DOMDocument();
* $doc->registerNodeClass('DOMElement', 'JSLikeHTMLElement');
* $doc->loadHTML('<div><p>Para 1</p><p>Para 2</p></div>');
* $elem = $doc->getElementsByTagName('div')->item(0);
*
* // print innerHTML
* echo $elem->innerHTML; // prints '<p>Para 1</p><p>Para 2</p>'
* echo "\n\n";
*
* // set innerHTML
* $elem->innerHTML = '<a href="http://fivefilters.org">FiveFilters.org</a>';
* echo $elem->innerHTML; // prints '<a href="http://fivefilters.org">FiveFilters.org</a>'
* echo "\n\n";
*
* // print document (with our changes)
* echo $doc->saveXML();
* @endcode
*
* @author Keyvan Minoukadeh - http://www.keyvan.net - keyvan@keyvan.net
* @see http://fivefilters.org (the project this was written for)
*/
class JSLikeHTMLElement extends DOMElement
{
/**
* Used for setting innerHTML like it's done in JavaScript:
* @code
* $div->innerHTML = '<h2>Chapter 2</h2><p>The story begins...</p>';
* @endcode
*/
public function __set($name, $value) {
if ($name == 'innerHTML') {
// first, empty the element
for ($x=$this->childNodes->length-1; $x>=0; $x--) {
$this->removeChild($this->childNodes->item($x));
}
// $value holds our new inner HTML
if ($value != '') {
$f = $this->ownerDocument->createDocumentFragment();
// appendXML() expects well-formed markup (XHTML)
$result = @$f->appendXML($value); // @ to suppress PHP warnings
if ($result) {
if ($f->hasChildNodes()) $this->appendChild($f);
} else {
// $value is probably ill-formed
$f = new DOMDocument();
$value = mb_convert_encoding($value, 'HTML-ENTITIES', 'UTF-8');
// Using <htmlfragment> will generate a warning, but so will bad HTML
// (and by this point, bad HTML is what we've got).
// We use it (and suppress the warning) because an HTML fragment will
// be wrapped around <html><body> tags which we don't really want to keep.
// Note: despite the warning, if loadHTML succeeds it will return true.
$result = @$f->loadHTML('<htmlfragment>'.$value.'</htmlfragment>');
if ($result) {
$import = $f->getElementsByTagName('htmlfragment')->item(0);
foreach ($import->childNodes as $child) {
$importedNode = $this->ownerDocument->importNode($child, true);
$this->appendChild($importedNode);
}
} else {
// oh well, we tried, we really did. :(
// this element is now empty
}
}
}
} else {
$trace = debug_backtrace();
trigger_error('Undefined property via __set(): '.$name.' in '.$trace[0]['file'].' on line '.$trace[0]['line'], E_USER_NOTICE);
}
}
/**
* Used for getting innerHTML like it's done in JavaScript:
* @code
* $string = $div->innerHTML;
* @endcode
*/
public function __get($name)
{
if ($name == 'innerHTML') {
$inner = '';
foreach ($this->childNodes as $child) {
$inner .= $this->ownerDocument->saveXML($child);
}
return $inner;
}
$trace = debug_backtrace();
trigger_error('Undefined property via __get(): '.$name.' in '.$trace[0]['file'].' on line '.$trace[0]['line'], E_USER_NOTICE);
return null;
}
public function __toString()
{
return '['.$this->tagName.']';
}
}
?>

File diff suppressed because it is too large Load Diff

View File

@ -46,6 +46,7 @@
// This class allows us to do JavaScript like assignements to innerHTML // This class allows us to do JavaScript like assignements to innerHTML
require_once(dirname(__FILE__).'/JSLikeHTMLElement.php'); require_once(dirname(__FILE__).'/JSLikeHTMLElement.php');
libxml_use_internal_errors(true);
// Alternative usage (for testing only!) // Alternative usage (for testing only!)
// uncomment the lines below and call Readability.php in your browser // uncomment the lines below and call Readability.php in your browser
@ -697,7 +698,7 @@ class Readability
$articleContent = $this->dom->createElement('div'); $articleContent = $this->dom->createElement('div');
$articleContent->setAttribute('id', 'readability-content'); $articleContent->setAttribute('id', 'readability-content');
$siblingScoreThreshold = max(10, ((int)$topCandidate->getAttribute('readability')) * 0.2); $siblingScoreThreshold = max(10, ((int)$topCandidate->getAttribute('readability')) * 0.2);
$siblingNodes = $topCandidate->parentNode->childNodes; $siblingNodes = @$topCandidate->parentNode->childNodes;
if (!isset($siblingNodes)) { if (!isset($siblingNodes)) {
$siblingNodes = new stdClass; $siblingNodes = new stdClass;
$siblingNodes->length = 0; $siblingNodes->length = 0;

View File

@ -31,6 +31,8 @@ along with this program. If not, see <http://www.gnu.org/licenses/>.
//error_reporting(E_ALL ^ E_NOTICE); //error_reporting(E_ALL ^ E_NOTICE);
ini_set("display_errors", 1); ini_set("display_errors", 1);
@set_time_limit(120); @set_time_limit(120);
libxml_use_internal_errors(true);
// Deal with magic quotes // Deal with magic quotes
if (get_magic_quotes_gpc()) { if (get_magic_quotes_gpc()) {