Browse Source

Added some more timestamp measuring

pull/1423/head
Michael Vogel 6 years ago
parent
commit
dbe8275ae5
4 changed files with 235 additions and 212 deletions
  1. +8
    -0
      include/bbcode.php
  2. +174
    -166
      include/network.php
  3. +0
    -1
      include/oembed.php
  4. +53
    -45
      mod/parse_url.php

+ 8
- 0
include/bbcode.php View File

@ -681,6 +681,8 @@ function bb_RemovePictureLinks($match) {
if(is_null($text)){
$a = get_app();
$stamp1 = microtime(true);
$ch = @curl_init($match[1]);
@curl_setopt($ch, CURLOPT_NOBODY, true);
@curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
@ -688,6 +690,8 @@ function bb_RemovePictureLinks($match) {
@curl_exec($ch);
$curl_info = @curl_getinfo($ch);
$a->save_timestamp($stamp1, "network");
if (substr($curl_info["content_type"], 0, 6) == "image/")
$text = "[url=".$match[1]."]".$match[1]."[/url]";
else {
@ -731,6 +735,8 @@ function bb_CleanPictureLinksSub($match) {
if(is_null($text)){
$a = get_app();
$stamp1 = microtime(true);
$ch = @curl_init($match[1]);
@curl_setopt($ch, CURLOPT_NOBODY, true);
@curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
@ -738,6 +744,8 @@ function bb_CleanPictureLinksSub($match) {
@curl_exec($ch);
$curl_info = @curl_getinfo($ch);
$a->save_timestamp($stamp1, "network");
// if its a link to a picture then embed this picture
if (substr($curl_info["content_type"], 0, 6) == "image/")
$text = "[img]".$match[1]."[/img]";


+ 174
- 166
include/network.php View File

@ -261,43 +261,43 @@ function http_status_exit($val) {
if(! function_exists('convert_xml_element_to_array')) {
function convert_xml_element_to_array($xml_element, &$recursion_depth=0) {
// If we're getting too deep, bail out
if ($recursion_depth > 512) {
return(null);
}
if (!is_string($xml_element) &&
!is_array($xml_element) &&
(get_class($xml_element) == 'SimpleXMLElement')) {
$xml_element_copy = $xml_element;
$xml_element = get_object_vars($xml_element);
}
if (is_array($xml_element)) {
$result_array = array();
if (count($xml_element) <= 0) {
return (trim(strval($xml_element_copy)));
}
foreach($xml_element as $key=>$value) {
$recursion_depth++;
$result_array[strtolower($key)] =
convert_xml_element_to_array($value, $recursion_depth);
$recursion_depth--;
}
if ($recursion_depth == 0) {
$temp_array = $result_array;
$result_array = array(
strtolower($xml_element_copy->getName()) => $temp_array,
);
}
return ($result_array);
} else {
return (trim(strval($xml_element)));
}
// If we're getting too deep, bail out
if ($recursion_depth > 512) {
return(null);
}
if (!is_string($xml_element) &&
!is_array($xml_element) &&
(get_class($xml_element) == 'SimpleXMLElement')) {
$xml_element_copy = $xml_element;
$xml_element = get_object_vars($xml_element);
}
if (is_array($xml_element)) {
$result_array = array();
if (count($xml_element) <= 0) {
return (trim(strval($xml_element_copy)));
}
foreach($xml_element as $key=>$value) {
$recursion_depth++;
$result_array[strtolower($key)] =
convert_xml_element_to_array($value, $recursion_depth);
$recursion_depth--;
}
if ($recursion_depth == 0) {
$temp_array = $result_array;
$result_array = array(
strtolower($xml_element_copy->getName()) => $temp_array,
);
}
return ($result_array);
} else {
return (trim(strval($xml_element)));
}
}}
// Given an email style address, perform webfinger lookup and
@ -964,8 +964,8 @@ function xml2array($contents, $namespaces = true, $get_attributes=1, $priority =
if(!$contents) return array();
if(!function_exists('xml_parser_create')) {
logger('xml2array: parser function missing');
return array();
logger('xml2array: parser function missing');
return array();
}
@ -1008,29 +1008,29 @@ function xml2array($contents, $namespaces = true, $get_attributes=1, $priority =
// Go through the tags.
$repeated_tag_index = array(); // Multiple tags with same name will be turned into an array
foreach($xml_values as $data) {
unset($attributes,$value); // Remove existing values, or there will be trouble
unset($attributes,$value); // Remove existing values, or there will be trouble
// This command will extract these variables into the foreach scope
// tag(string), type(string), level(int), attributes(array).
extract($data); // We could use the array by itself, but this cooler.
// This command will extract these variables into the foreach scope
// tag(string), type(string), level(int), attributes(array).
extract($data); // We could use the array by itself, but this cooler.
$result = array();
$attributes_data = array();
$result = array();
$attributes_data = array();
if(isset($value)) {
if($priority == 'tag') $result = $value;
else $result['value'] = $value; // Put the value in a assoc array if we are in the 'Attribute' mode
}
if(isset($value)) {
if($priority == 'tag') $result = $value;
else $result['value'] = $value; // Put the value in a assoc array if we are in the 'Attribute' mode
}
//Set the attributes too.
if(isset($attributes) and $get_attributes) {
foreach($attributes as $attr => $val) {
if($priority == 'tag') $attributes_data[$attr] = $val;
else $result['@attributes'][$attr] = $val; // Set all the attributes in a array called 'attr'
}
}
//Set the attributes too.
if(isset($attributes) and $get_attributes) {
foreach($attributes as $attr => $val) {
if($priority == 'tag') $attributes_data[$attr] = $val;
else $result['@attributes'][$attr] = $val; // Set all the attributes in a array called 'attr'
}
}
// See tag status and do the needed.
// See tag status and do the needed.
if($namespaces && strpos($tag,':')) {
$namespc = substr($tag,0,strrpos($tag,':'));
$tag = strtolower(substr($tag,strlen($namespc)+1));
@ -1039,72 +1039,72 @@ function xml2array($contents, $namespaces = true, $get_attributes=1, $priority =
$tag = strtolower($tag);
if($type == "open") { // The starting of the tag '<tag>'
$parent[$level-1] = &$current;
if(!is_array($current) or (!in_array($tag, array_keys($current)))) { // Insert New tag
$current[$tag] = $result;
if($attributes_data) $current[$tag. '_attr'] = $attributes_data;
$repeated_tag_index[$tag.'_'.$level] = 1;
$current = &$current[$tag];
} else { // There was another element with the same tag name
if(isset($current[$tag][0])) { // If there is a 0th element it is already an array
$current[$tag][$repeated_tag_index[$tag.'_'.$level]] = $result;
$repeated_tag_index[$tag.'_'.$level]++;
} else { // This section will make the value an array if multiple tags with the same name appear together
$current[$tag] = array($current[$tag],$result); // This will combine the existing item and the new item together to make an array
$repeated_tag_index[$tag.'_'.$level] = 2;
if(isset($current[$tag.'_attr'])) { // The attribute of the last(0th) tag must be moved as well
$current[$tag]['0_attr'] = $current[$tag.'_attr'];
unset($current[$tag.'_attr']);
}
}
$last_item_index = $repeated_tag_index[$tag.'_'.$level]-1;
$current = &$current[$tag][$last_item_index];
}
} elseif($type == "complete") { // Tags that ends in 1 line '<tag />'
//See if the key is already taken.
if(!isset($current[$tag])) { //New Key
$current[$tag] = $result;
$repeated_tag_index[$tag.'_'.$level] = 1;
if($priority == 'tag' and $attributes_data) $current[$tag. '_attr'] = $attributes_data;
} else { // If taken, put all things inside a list(array)
if(isset($current[$tag][0]) and is_array($current[$tag])) { // If it is already an array...
// ...push the new element into that array.
$current[$tag][$repeated_tag_index[$tag.'_'.$level]] = $result;
if($priority == 'tag' and $get_attributes and $attributes_data) {
$current[$tag][$repeated_tag_index[$tag.'_'.$level] . '_attr'] = $attributes_data;
}
$repeated_tag_index[$tag.'_'.$level]++;
} else { // If it is not an array...
$current[$tag] = array($current[$tag],$result); //...Make it an array using using the existing value and the new value
$repeated_tag_index[$tag.'_'.$level] = 1;
if($priority == 'tag' and $get_attributes) {
if(isset($current[$tag.'_attr'])) { // The attribute of the last(0th) tag must be moved as well
$current[$tag]['0_attr'] = $current[$tag.'_attr'];
unset($current[$tag.'_attr']);
}
if($attributes_data) {
$current[$tag][$repeated_tag_index[$tag.'_'.$level] . '_attr'] = $attributes_data;
}
}
$repeated_tag_index[$tag.'_'.$level]++; // 0 and 1 indexes are already taken
}
}
} elseif($type == 'close') { // End of tag '</tag>'
$current = &$parent[$level-1];
}
$parent[$level-1] = &$current;
if(!is_array($current) or (!in_array($tag, array_keys($current)))) { // Insert New tag
$current[$tag] = $result;
if($attributes_data) $current[$tag. '_attr'] = $attributes_data;
$repeated_tag_index[$tag.'_'.$level] = 1;
$current = &$current[$tag];
} else { // There was another element with the same tag name
if(isset($current[$tag][0])) { // If there is a 0th element it is already an array
$current[$tag][$repeated_tag_index[$tag.'_'.$level]] = $result;
$repeated_tag_index[$tag.'_'.$level]++;
} else { // This section will make the value an array if multiple tags with the same name appear together
$current[$tag] = array($current[$tag],$result); // This will combine the existing item and the new item together to make an array
$repeated_tag_index[$tag.'_'.$level] = 2;
if(isset($current[$tag.'_attr'])) { // The attribute of the last(0th) tag must be moved as well
$current[$tag]['0_attr'] = $current[$tag.'_attr'];
unset($current[$tag.'_attr']);
}
}
$last_item_index = $repeated_tag_index[$tag.'_'.$level]-1;
$current = &$current[$tag][$last_item_index];
}
} elseif($type == "complete") { // Tags that ends in 1 line '<tag />'
//See if the key is already taken.
if(!isset($current[$tag])) { //New Key
$current[$tag] = $result;
$repeated_tag_index[$tag.'_'.$level] = 1;
if($priority == 'tag' and $attributes_data) $current[$tag. '_attr'] = $attributes_data;
} else { // If taken, put all things inside a list(array)
if(isset($current[$tag][0]) and is_array($current[$tag])) { // If it is already an array...
// ...push the new element into that array.
$current[$tag][$repeated_tag_index[$tag.'_'.$level]] = $result;
if($priority == 'tag' and $get_attributes and $attributes_data) {
$current[$tag][$repeated_tag_index[$tag.'_'.$level] . '_attr'] = $attributes_data;
}
$repeated_tag_index[$tag.'_'.$level]++;
} else { // If it is not an array...
$current[$tag] = array($current[$tag],$result); //...Make it an array using using the existing value and the new value
$repeated_tag_index[$tag.'_'.$level] = 1;
if($priority == 'tag' and $get_attributes) {
if(isset($current[$tag.'_attr'])) { // The attribute of the last(0th) tag must be moved as well
$current[$tag]['0_attr'] = $current[$tag.'_attr'];
unset($current[$tag.'_attr']);
}
if($attributes_data) {
$current[$tag][$repeated_tag_index[$tag.'_'.$level] . '_attr'] = $attributes_data;
}
}
$repeated_tag_index[$tag.'_'.$level]++; // 0 and 1 indexes are already taken
}
}
} elseif($type == 'close') { // End of tag '</tag>'
$current = &$parent[$level-1];
}
}
return($xml_array);
@ -1146,32 +1146,36 @@ function original_url($url, $depth=1, $fetchbody = false) {
$url = substr($url, 0, -1);
}
if ($depth > 10)
return($url);
if ($depth > 10)
return($url);
$url = trim($url, "'");
$url = trim($url, "'");
$stamp1 = microtime(true);
$siteinfo = array();
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
$siteinfo = array();
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
curl_setopt($ch, CURLOPT_HEADER, 1);
curl_setopt($ch, CURLOPT_NOBODY, 1);
curl_setopt($ch, CURLOPT_TIMEOUT, 10);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_setopt($ch, CURLOPT_TIMEOUT, 10);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
$header = curl_exec($ch);
$curl_info = @curl_getinfo($ch);
$http_code = $curl_info['http_code'];
curl_close($ch);
$header = curl_exec($ch);
$curl_info = @curl_getinfo($ch);
$http_code = $curl_info['http_code'];
curl_close($ch);
$a->save_timestamp($stamp1, "network");
if ((($curl_info['http_code'] == "301") OR ($curl_info['http_code'] == "302"))
AND (($curl_info['redirect_url'] != "") OR ($curl_info['location'] != ""))) {
if ($curl_info['redirect_url'] != "")
return(original_url($curl_info['redirect_url'], ++$depth, $fetchbody));
else
return(original_url($curl_info['location'], ++$depth, $fetchbody));
}
if ((($curl_info['http_code'] == "301") OR ($curl_info['http_code'] == "302"))
AND (($curl_info['redirect_url'] != "") OR ($curl_info['location'] != ""))) {
if ($curl_info['redirect_url'] != "")
return(original_url($curl_info['redirect_url'], ++$depth, $fetchbody));
else
return(original_url($curl_info['location'], ++$depth, $fetchbody));
}
// Check for redirects in the meta elements of the body if there are no redirects in the header.
if (!$fetchbody)
@ -1185,6 +1189,8 @@ function original_url($url, $depth=1, $fetchbody = false) {
if (($curl_info["content_type"] != "") AND !strstr(strtolower($curl_info["content_type"]),"html"))
return($url);
$stamp1 = microtime(true);
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
curl_setopt($ch, CURLOPT_HEADER, 0);
@ -1196,33 +1202,35 @@ function original_url($url, $depth=1, $fetchbody = false) {
$body = curl_exec($ch);
curl_close($ch);
if (trim($body) == "")
$a->save_timestamp($stamp1, "network");
if (trim($body) == "")
return($url);
// Check for redirect in meta elements
$doc = new DOMDocument();
@$doc->loadHTML($body);
$xpath = new DomXPath($doc);
$list = $xpath->query("//meta[@content]");
foreach ($list as $node) {
$attr = array();
if ($node->attributes->length)
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
if (@$attr["http-equiv"] == 'refresh') {
$path = $attr["content"];
$pathinfo = explode(";", $path);
$content = "";
foreach ($pathinfo AS $value)
if (substr(strtolower($value), 0, 4) == "url=")
return(original_url(substr($value, 4), ++$depth));
}
}
return($url);
$doc = new DOMDocument();
@$doc->loadHTML($body);
$xpath = new DomXPath($doc);
$list = $xpath->query("//meta[@content]");
foreach ($list as $node) {
$attr = array();
if ($node->attributes->length)
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
if (@$attr["http-equiv"] == 'refresh') {
$path = $attr["content"];
$pathinfo = explode(";", $path);
$content = "";
foreach ($pathinfo AS $value)
if (substr(strtolower($value), 0, 4) == "url=")
return(original_url(substr($value, 4), ++$depth));
}
}
return($url);
}
if (!function_exists('short_link')) {


+ 0
- 1
include/oembed.php View File

@ -11,7 +11,6 @@ function oembed_replacecb($matches){
function oembed_fetch_url($embedurl, $no_rich_type = false){
$embedurl = trim($embedurl, "'");
$embedurl = trim($embedurl, '"');


+ 53
- 45
mod/parse_url.php View File

@ -24,30 +24,30 @@ if(!function_exists('deletenode')) {
}
function completeurl($url, $scheme) {
$urlarr = parse_url($url);
$urlarr = parse_url($url);
if (isset($urlarr["scheme"]))
return($url);
if (isset($urlarr["scheme"]))
return($url);
$schemearr = parse_url($scheme);
$schemearr = parse_url($scheme);
$complete = $schemearr["scheme"]."://".$schemearr["host"];
$complete = $schemearr["scheme"]."://".$schemearr["host"];
if (@$schemearr["port"] != "")
$complete .= ":".$schemearr["port"];
if (@$schemearr["port"] != "")
$complete .= ":".$schemearr["port"];
if(strpos($urlarr['path'],'/') !== 0)
$complete .= '/';
$complete .= $urlarr["path"];
$complete .= $urlarr["path"];
if (@$urlarr["query"] != "")
$complete .= "?".$urlarr["query"];
if (@$urlarr["query"] != "")
$complete .= "?".$urlarr["query"];
if (@$urlarr["fragment"] != "")
$complete .= "#".$urlarr["fragment"];
if (@$urlarr["fragment"] != "")
$complete .= "#".$urlarr["fragment"];
return($complete);
return($complete);
}
function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1) {
@ -70,6 +70,8 @@ function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $co
$siteinfo["url"] = $url;
$siteinfo["type"] = "link";
$stamp1 = microtime(true);
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
curl_setopt($ch, CURLOPT_HEADER, 1);
@ -81,9 +83,11 @@ function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $co
$header = curl_exec($ch);
$curl_info = @curl_getinfo($ch);
$http_code = $curl_info['http_code'];
$http_code = $curl_info['http_code'];
curl_close($ch);
$a->save_timestamp($stamp1, "network");
if ((($curl_info['http_code'] == "301") OR ($curl_info['http_code'] == "302") OR ($curl_info['http_code'] == "303") OR ($curl_info['http_code'] == "307"))
AND (($curl_info['redirect_url'] != "") OR ($curl_info['location'] != ""))) {
if ($curl_info['redirect_url'] != "")
@ -110,6 +114,8 @@ function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $co
if (($curl_info["content_type"] != "") AND !strstr(strtolower($curl_info["content_type"]),"html"))
return($siteinfo);
$stamp1 = microtime(true);
// Now fetch the body as well
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
@ -121,9 +127,11 @@ function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $co
$header = curl_exec($ch);
$curl_info = @curl_getinfo($ch);
$http_code = $curl_info['http_code'];
$http_code = $curl_info['http_code'];
curl_close($ch);
$a->save_timestamp($stamp1, "network");
// Fetch the first mentioned charset. Can be in body or header
$charset = "";
if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches))
@ -165,25 +173,25 @@ function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $co
$xpath = new DomXPath($doc);
$list = $xpath->query("//meta[@content]");
foreach ($list as $node) {
$attr = array();
if ($node->attributes->length)
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
if (@$attr["http-equiv"] == 'refresh') {
$path = $attr["content"];
$pathinfo = explode(";", $path);
$content = "";
foreach ($pathinfo AS $value) {
if (substr(strtolower($value), 0, 4) == "url=")
$content = substr($value, 4);
}
if ($content != "") {
$siteinfo = parseurl_getsiteinfo($content, $no_guessing, $do_oembed, ++$count);
return($siteinfo);
}
}
foreach ($list as $node) {
$attr = array();
if ($node->attributes->length)
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
if (@$attr["http-equiv"] == 'refresh') {
$path = $attr["content"];
$pathinfo = explode(";", $path);
$content = "";
foreach ($pathinfo AS $value) {
if (substr(strtolower($value), 0, 4) == "url=")
$content = substr($value, 4);
}
if ($content != "") {
$siteinfo = parseurl_getsiteinfo($content, $no_guessing, $do_oembed, ++$count);
return($siteinfo);
}
}
}
//$list = $xpath->query("head/title");
@ -196,8 +204,8 @@ function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $co
foreach ($list as $node) {
$attr = array();
if ($node->attributes->length)
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
$attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
@ -256,8 +264,8 @@ function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $co
foreach ($list as $node) {
$attr = array();
if ($node->attributes->length)
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
$attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
@ -285,12 +293,12 @@ function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $co
}
if ((@$siteinfo["image"] == "") AND !$no_guessing) {
$list = $xpath->query("//img[@src]");
foreach ($list as $node) {
$attr = array();
if ($node->attributes->length)
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
$list = $xpath->query("//img[@src]");
foreach ($list as $node) {
$attr = array();
if ($node->attributes->length)
foreach ($node->attributes as $attribute)
$attr[$attribute->name] = $attribute->value;
$src = completeurl($attr["src"], $url);
$photodata = @getimagesize($src);
@ -309,7 +317,7 @@ function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $co
"height"=>$photodata[1]);
}
}
}
} else {
$src = completeurl($siteinfo["image"], $url);


Loading…
Cancel
Save