aboutsummaryrefslogtreecommitdiff
path: root/lib/SimplePie
diff options
context:
space:
mode:
Diffstat (limited to 'lib/SimplePie')
-rw-r--r--lib/SimplePie/SimplePie.php30
-rw-r--r--lib/SimplePie/SimplePie/File.php1
-rw-r--r--lib/SimplePie/SimplePie/Misc.php4
-rw-r--r--lib/SimplePie/SimplePie/Parser.php31
-rw-r--r--lib/SimplePie/SimplePie/Sanitize.php53
5 files changed, 95 insertions, 24 deletions
diff --git a/lib/SimplePie/SimplePie.php b/lib/SimplePie/SimplePie.php
index 9e532023a..f02037c10 100644
--- a/lib/SimplePie/SimplePie.php
+++ b/lib/SimplePie/SimplePie.php
@@ -602,6 +602,13 @@ class SimplePie
public $strip_attributes = array('bgsound', 'class', 'expr', 'id', 'style', 'onclick', 'onerror', 'onfinish', 'onmouseover', 'onmouseout', 'onfocus', 'onblur', 'lowsrc', 'dynsrc');
/**
+ * @var array Stores the default attributes to add to differet tags by add_attributes().
+ * @see SimplePie::add_attributes()
+ * @access private
+ */
+ public $add_attributes = array('audio' => array('preload' => 'none'), 'iframe' => array('sandbox' => 'allow-scripts allow-same-origin'), 'video' => array('preload' => 'none')); //FreshRSS
+
+ /**
* @var array Stores the default tags to be stripped by strip_htmltags().
* @see SimplePie::strip_htmltags()
* @access private
@@ -1073,6 +1080,7 @@ class SimplePie
$this->strip_comments(false);
$this->strip_htmltags(false);
$this->strip_attributes(false);
+ $this->add_attributes(false);
$this->set_image_handler(false);
}
}
@@ -1119,6 +1127,15 @@ class SimplePie
$this->sanitize->strip_attributes($attribs);
}
+ public function add_attributes($attribs = '')
+ {
+ if ($attribs === '')
+ {
+ $attribs = $this->add_attributes;
+ }
+ $this->sanitize->add_attributes($attribs);
+ }
+
/**
* Set the output encoding
*
@@ -1296,7 +1313,7 @@ class SimplePie
// First check to see if input has been overridden.
if ($this->input_encoding !== false)
{
- $encodings[] = $this->input_encoding;
+ $encodings[] = strtoupper($this->input_encoding);
}
$application_types = array('application/xml', 'application/xml-dtd', 'application/xml-external-parsed-entity');
@@ -1313,18 +1330,18 @@ class SimplePie
}
else
{
- $encodings[] = ''; //Let the DOM parser decide first
+ $encodings[] = ''; //FreshRSS: Let the DOM parser decide first
}
}
elseif (in_array($sniffed, $text_types) || substr($sniffed, 0, 5) === 'text/' && substr($sniffed, -4) === '+xml')
{
if (isset($headers['content-type']) && preg_match('/;\x20?charset=([^;]*)/i', $headers['content-type'], $charset))
{
- $encodings[] = $charset[1];
+ $encodings[] = strtoupper($charset[1]);
}
else
{
- $encodings[] = '';
+ $encodings[] = ''; //FreshRSS: Let the DOM parser decide first
}
$encodings[] = 'US-ASCII';
}
@@ -1347,13 +1364,14 @@ class SimplePie
foreach ($encodings as $encoding)
{
// Change the encoding to UTF-8 (as we always use UTF-8 internally)
- if ($utf8_data = (empty($encoding) || $encoding === 'UTF-8') ? $this->raw_data : $this->registry->call('Misc', 'change_encoding', array($this->raw_data, $encoding, 'UTF-8')))
+ if ($utf8_data = (empty($encoding) || $encoding === 'UTF-8') ? $this->raw_data : //FreshRSS
+ $this->registry->call('Misc', 'change_encoding', array($this->raw_data, $encoding, 'UTF-8')))
{
// Create new parser
$parser = $this->registry->create('Parser');
// If it's parsed fine
- if ($parser->parse($utf8_data, 'UTF-8'))
+ if ($parser->parse($utf8_data, empty($encoding) ? '' : 'UTF-8')) //FreshRSS
{
$this->data = $parser->get_data();
if (!($this->get_type() & ~SIMPLEPIE_TYPE_NONE))
diff --git a/lib/SimplePie/SimplePie/File.php b/lib/SimplePie/SimplePie/File.php
index 063ad955e..cf926cf5a 100644
--- a/lib/SimplePie/SimplePie/File.php
+++ b/lib/SimplePie/SimplePie/File.php
@@ -77,6 +77,7 @@ class SimplePie_File
$this->useragent = $useragent;
if (preg_match('/^http(s)?:\/\//i', $url))
{
+ syslog(LOG_INFO, 'SimplePie GET ' . $url); //FreshRSS
if ($useragent === null)
{
$useragent = ini_get('user_agent');
diff --git a/lib/SimplePie/SimplePie/Misc.php b/lib/SimplePie/SimplePie/Misc.php
index 621f2c062..347520303 100644
--- a/lib/SimplePie/SimplePie/Misc.php
+++ b/lib/SimplePie/SimplePie/Misc.php
@@ -79,6 +79,10 @@ class SimplePie_Misc
public static function absolutize_url($relative, $base)
{
+ if (substr($relative, 0, 2) === '//') //FreshRSS: disable absolutize_url for "//www.example.net" which will pick HTTP or HTTPS automatically
+ {
+ return $relative;
+ }
$iri = SimplePie_IRI::absolutize(new SimplePie_IRI($base), $relative);
if ($iri === false)
{
diff --git a/lib/SimplePie/SimplePie/Parser.php b/lib/SimplePie/SimplePie/Parser.php
index 72878c25a..bd6c4efd8 100644
--- a/lib/SimplePie/SimplePie/Parser.php
+++ b/lib/SimplePie/SimplePie/Parser.php
@@ -77,6 +77,8 @@ class SimplePie_Parser
public function parse(&$data, $encoding)
{
+ $xmlEncoding = '';
+
if (!empty($encoding))
{
// Use UTF-8 if we get passed US-ASCII, as every US-ASCII character is a UTF-8 character
@@ -121,6 +123,7 @@ class SimplePie_Parser
$declaration = $this->registry->create('XML_Declaration_Parser', array(substr($data, 5, $pos - 5)));
if ($declaration->parse())
{
+ $xmlEncoding = strtoupper($declaration->encoding); //FreshRSS
$data = substr($data, $pos + 2);
$data = '<?xml version="' . $declaration->version . '" encoding="' . $encoding . '" standalone="' . (($declaration->standalone) ? 'yes' : 'no') . '"?>' . $data;
}
@@ -132,18 +135,24 @@ class SimplePie_Parser
}
}
- try
- {
- $dom = new DOMDocument();
- $dom->recover = true;
- $dom->strictErrorChecking = false;
- $dom->loadXML($data);
- $this->encoding = $encoding = $dom->encoding = 'UTF-8';
- $data = $dom->saveXML();
- //file_put_contents('/home/alex/public_html/alexandre.alapetite.fr/prive/FreshRSS/log/parser.log', date('c') . ' ' . 'OK' . "\n", FILE_APPEND);
- }
- catch (Exception $e)
+ if ($xmlEncoding === '' || $xmlEncoding === 'UTF-8') //FreshRSS: case of no explicit HTTP encoding, and lax UTF-8
{
+ try
+ {
+ $dom = new DOMDocument();
+ $dom->recover = true;
+ $dom->strictErrorChecking = false;
+ $dom->loadXML($data);
+ $this->encoding = $encoding = $dom->encoding = 'UTF-8';
+ $data2 = $dom->saveXML();
+ if (strlen($data2) > (strlen($data) / 2.0))
+ {
+ $data = $data2;
+ }
+ }
+ catch (Exception $e)
+ {
+ }
}
$return = true;
diff --git a/lib/SimplePie/SimplePie/Sanitize.php b/lib/SimplePie/SimplePie/Sanitize.php
index 83a274ced..0974c150d 100644
--- a/lib/SimplePie/SimplePie/Sanitize.php
+++ b/lib/SimplePie/SimplePie/Sanitize.php
@@ -62,6 +62,7 @@ class SimplePie_Sanitize
var $strip_htmltags = array('base', 'blink', 'body', 'doctype', 'embed', 'font', 'form', 'frame', 'frameset', 'html', 'iframe', 'input', 'marquee', 'meta', 'noscript', 'object', 'param', 'script', 'style');
var $encode_instead_of_strip = false;
var $strip_attributes = array('bgsound', 'class', 'expr', 'id', 'style', 'onclick', 'onerror', 'onfinish', 'onmouseover', 'onmouseout', 'onfocus', 'onblur', 'lowsrc', 'dynsrc');
+ var $add_attributes = array('audio' => array('preload' => 'none'), 'iframe' => array('sandbox' => 'allow-scripts allow-same-origin'), 'video' => array('preload' => 'none')); //FreshRSS
var $strip_comments = false;
var $output_encoding = 'UTF-8';
var $enable_cache = true;
@@ -179,6 +180,25 @@ class SimplePie_Sanitize
}
}
+ public function add_attributes($attribs = array('audio' => array('preload' => 'none'), 'iframe' => array('sandbox' => 'allow-scripts allow-same-origin'), 'video' => array('preload' => 'none')))
+ {
+ if ($attribs)
+ {
+ if (is_array($attribs))
+ {
+ $this->add_attributes = $attribs;
+ }
+ else
+ {
+ $this->add_attributes = explode(',', $attribs);
+ }
+ }
+ else
+ {
+ $this->add_attributes = false;
+ }
+ }
+
public function strip_comments($strip = false)
{
$this->strip_comments = (bool) $strip;
@@ -255,10 +275,11 @@ class SimplePie_Sanitize
$document->loadHTML($data);
restore_error_handler();
+ $xpath = new DOMXPath($document); //FreshRSS
+
// Strip comments
if ($this->strip_comments)
{
- $xpath = new DOMXPath($document);
$comments = $xpath->query('//comment()');
foreach ($comments as $comment)
@@ -274,7 +295,7 @@ class SimplePie_Sanitize
{
foreach ($this->strip_htmltags as $tag)
{
- $this->strip_tag($tag, $document, $type);
+ $this->strip_tag($tag, $document, $xpath, $type);
}
}
@@ -282,7 +303,15 @@ class SimplePie_Sanitize
{
foreach ($this->strip_attributes as $attrib)
{
- $this->strip_attr($attrib, $document);
+ $this->strip_attr($attrib, $xpath);
+ }
+ }
+
+ if ($this->add_attributes)
+ {
+ foreach ($this->add_attributes as $tag => $valuePairs)
+ {
+ $this->add_attr($tag, $valuePairs, $document);
}
}
@@ -452,9 +481,8 @@ class SimplePie_Sanitize
}
}
- protected function strip_tag($tag, $document, $type)
+ protected function strip_tag($tag, $document, $xpath, $type)
{
- $xpath = new DOMXPath($document);
$elements = $xpath->query('body//' . $tag);
if ($this->encode_instead_of_strip)
{
@@ -537,9 +565,8 @@ class SimplePie_Sanitize
}
}
- protected function strip_attr($attrib, $document)
+ protected function strip_attr($attrib, $xpath)
{
- $xpath = new DOMXPath($document);
$elements = $xpath->query('//*[@' . $attrib . ']');
foreach ($elements as $element)
@@ -547,4 +574,16 @@ class SimplePie_Sanitize
$element->removeAttribute($attrib);
}
}
+
+ protected function add_attr($tag, $valuePairs, $document)
+ {
+ $elements = $document->getElementsByTagName($tag);
+ foreach ($elements as $element)
+ {
+ foreach ($valuePairs as $attrib => $value)
+ {
+ $element->setAttribute($attrib, $value);
+ }
+ }
+ }
}
generated by cgit v1.2.3 (git 2.39.1) at 2026-02-10 09:07:37 +0000