Unused OEmbed functionality is removed

This commit is contained in:
Michael 2024-02-24 10:58:18 +00:00
parent 0ff37c0075
commit 821a135033
5 changed files with 48 additions and 274 deletions

View File

@ -58,7 +58,7 @@ class OEmbed
public static function replaceCallback(array $matches): string public static function replaceCallback(array $matches): string
{ {
$embedurl = $matches[1]; $embedurl = $matches[1];
$j = self::fetchURL($embedurl, !self::isAllowedURL($embedurl)); $j = self::fetchURL($embedurl);
$s = self::formatObject($j); $s = self::formatObject($j);
return $s; return $s;
@ -67,14 +67,12 @@ class OEmbed
/** /**
* Get data from an URL to embed its content. * Get data from an URL to embed its content.
* *
* @param string $embedurl The URL from which the data should be fetched. * @param string $embedurl The URL from which the data should be fetched.
* @param bool $no_rich_type If set to true rich type content won't be fetched.
* @param bool $use_parseurl Use the "ParseUrl" functionality to add additional data
* *
* @return \Friendica\Object\OEmbed * @return \Friendica\Object\OEmbed
* @throws \Friendica\Network\HTTPException\InternalServerErrorException * @throws \Friendica\Network\HTTPException\InternalServerErrorException
*/ */
public static function fetchURL(string $embedurl, bool $no_rich_type = false, bool $use_parseurl = true): \Friendica\Object\OEmbed private static function fetchURL(string $embedurl): \Friendica\Object\OEmbed
{ {
$embedurl = trim($embedurl, '\'"'); $embedurl = trim($embedurl, '\'"');
@ -119,7 +117,7 @@ class OEmbed
$href = str_replace(['http://www.youtube.com/', 'http://player.vimeo.com/'], $href = str_replace(['http://www.youtube.com/', 'http://player.vimeo.com/'],
['https://www.youtube.com/', 'https://player.vimeo.com/'], $href); ['https://www.youtube.com/', 'https://player.vimeo.com/'], $href);
$result = DI::httpClient()->fetchFull($href . '&maxwidth=' . $a->getThemeInfoValue('videowidth')); $result = DI::httpClient()->fetchFull($href . '&maxwidth=' . $a->getThemeInfoValue('videowidth'));
if ($result->getReturnCode() === 200) { if ($result->isSuccess()) {
$json_string = $result->getBodyString(); $json_string = $result->getBodyString();
break; break;
} }
@ -157,57 +155,55 @@ class OEmbed
} }
// Improve the OEmbed data with data from OpenGraph, Twitter cards and other sources // Improve the OEmbed data with data from OpenGraph, Twitter cards and other sources
if ($use_parseurl) { $data = ParseUrl::getSiteinfoCached($embedurl);
$data = ParseUrl::getSiteinfoCached($embedurl, false);
if (($oembed->type == 'error') && empty($data['title']) && empty($data['text'])) { if (($oembed->type == 'error') && empty($data['title']) && empty($data['text'])) {
return $oembed; return $oembed;
} }
if ($no_rich_type || ($oembed->type == 'error')) { if (!self::isAllowedURL($oembed->url) || ($oembed->type == 'error')) {
$oembed->html = ''; $oembed->html = '';
$oembed->type = $data['type']; $oembed->type = $data['type'];
if ($oembed->type == 'photo') { if ($oembed->type == 'photo') {
if (!empty($data['images'])) { if (!empty($data['images'])) {
$oembed->url = $data['images'][0]['src']; $oembed->url = $data['images'][0]['src'];
$oembed->width = $data['images'][0]['width']; $oembed->width = $data['images'][0]['width'];
$oembed->height = $data['images'][0]['height']; $oembed->height = $data['images'][0]['height'];
} else { } else {
$oembed->type = 'link'; $oembed->type = 'link';
}
} }
} }
}
if (!empty($data['title'])) { if (!empty($data['title'])) {
$oembed->title = $data['title']; $oembed->title = $data['title'];
} }
if (!empty($data['text'])) { if (!empty($data['text'])) {
$oembed->description = $data['text']; $oembed->description = $data['text'];
} }
if (!empty($data['publisher_name'])) { if (!empty($data['publisher_name'])) {
$oembed->provider_name = $data['publisher_name']; $oembed->provider_name = $data['publisher_name'];
} }
if (!empty($data['publisher_url'])) { if (!empty($data['publisher_url'])) {
$oembed->provider_url = $data['publisher_url']; $oembed->provider_url = $data['publisher_url'];
} }
if (!empty($data['author_name'])) { if (!empty($data['author_name'])) {
$oembed->author_name = $data['author_name']; $oembed->author_name = $data['author_name'];
} }
if (!empty($data['author_url'])) { if (!empty($data['author_url'])) {
$oembed->author_url = $data['author_url']; $oembed->author_url = $data['author_url'];
} }
if (!empty($data['images']) && ($oembed->type != 'photo')) { if (!empty($data['images']) && ($oembed->type != 'photo')) {
$oembed->thumbnail_url = $data['images'][0]['src']; $oembed->thumbnail_url = $data['images'][0]['src'];
$oembed->thumbnail_width = $data['images'][0]['width']; $oembed->thumbnail_width = $data['images'][0]['width'];
$oembed->thumbnail_height = $data['images'][0]['height']; $oembed->thumbnail_height = $data['images'][0]['height'];
}
} }
Hook::callAll('oembed_fetch_url', $embedurl, $oembed); Hook::callAll('oembed_fetch_url', $embedurl, $oembed);
@ -318,43 +314,6 @@ class OEmbed
return preg_replace_callback("/\[embed\](.+?)\[\/embed\]/is", [self::class, 'replaceCallback'], $text); return preg_replace_callback("/\[embed\](.+?)\[\/embed\]/is", [self::class, 'replaceCallback'], $text);
} }
/**
* Find <span class='oembed'>..<a href='url' rel='oembed'>..</a></span>
* and replace it with [embed]url[/embed]
*
* @param string $text
* @return string
*/
public static function HTML2BBCode(string $text): string
{
// start parser only if 'oembed' is in text
if (strpos($text, 'oembed')) {
// convert non ascii chars to html entities
$html_text = mb_convert_encoding($text, 'HTML-ENTITIES', mb_detect_encoding($text));
// If it doesn't parse at all, just return the text.
$dom = new DOMDocument();
if (!@$dom->loadHTML($html_text)) {
return $text;
}
$xpath = new DOMXPath($dom);
$xattr = self::buildXPath('class', 'oembed');
$entries = $xpath->query("//div[$xattr]");
$xattr = "@rel='oembed'"; //oe_build_xpath("rel","oembed");
foreach ($entries as $e) {
$href = $xpath->evaluate("a[$xattr]/@href", $e)->item(0)->nodeValue;
if (!is_null($href)) {
$e->parentNode->replaceChild(new DOMText('[embed]' . $href . '[/embed]'), $e);
}
}
return self::getInnerHTML($dom->getElementsByTagName('body')->item(0));
} else {
return $text;
}
}
/** /**
* Determines if rich content OEmbed is allowed for the provided URL * Determines if rich content OEmbed is allowed for the provided URL
* *
@ -392,7 +351,7 @@ class OEmbed
*/ */
public static function getHTML(string $url, string $title = ''): string public static function getHTML(string $url, string $title = ''): string
{ {
$o = self::fetchURL($url, !self::isAllowedURL($url)); $o = self::fetchURL($url);
if (!is_object($o) || property_exists($o, 'type') && $o->type == 'error') { if (!is_object($o) || property_exists($o, 'type') && $o->type == 'error') {
throw new Exception('OEmbed failed for URL: ' . $url); throw new Exception('OEmbed failed for URL: ' . $url);
@ -406,70 +365,4 @@ class OEmbed
return $html; return $html;
} }
/**
* Generates the iframe HTML for an oembed attachment.
*
* Width and height are given by the remote, and are regularly too small for
* the generated iframe.
*
* The width is entirely discarded for the actual width of the post, while fixed
* height is used as a starting point before the inevitable resizing.
*
* Since the iframe is automatically resized on load, there are no need for ugly
* and impractical scrollbars.
*
* @todo This function is currently unused until someone™ adds support for a separate OEmbed domain
*
* @param string $src Original remote URL to embed
* @param string $width
* @param string $height
* @return string Formatted HTML
*
* @throws \Friendica\Network\HTTPException\InternalServerErrorException
* @see oembed_format_object()
*/
private static function iframe(string $src, string $width, string $height): string
{
if (!$height || strstr($height, '%')) {
$height = '200';
}
$width = '100%';
$src = DI::baseUrl() . '/oembed/' . Strings::base64UrlEncode($src);
return '<iframe onload="resizeIframe(this);" class="embed_rich" height="' . $height . '" width="' . $width . '" src="' . $src . '" allowfullscreen scrolling="no" frameborder="no">' . DI::l10n()->t('Embedded content') . '</iframe>';
}
/**
* Generates attribute search XPath string
*
* Generates an XPath query to select elements whose provided attribute contains
* the provided value in a space-separated list.
*
* @param string $attr Name of the attribute to search
* @param string $value Value to search in a space-separated list
* @return string
*/
private static function buildXPath(string $attr, $value): string
{
// https://www.westhoffswelt.de/blog/2009/6/9/select-html-elements-with-more-than-one-css-class-using-xpath
return "contains(normalize-space(@$attr), ' $value ') or substring(normalize-space(@$attr), 1, string-length('$value') + 1) = '$value ' or substring(normalize-space(@$attr), string-length(@$attr) - string-length('$value')) = ' $value' or @$attr = '$value'";
}
/**
* Returns the inner XML string of a provided DOMNode
*
* @param DOMNode $node
* @return string
*/
private static function getInnerHTML(DOMNode $node): string
{
$innerHTML = '';
$children = $node->childNodes;
foreach ($children as $child) {
$innerHTML .= $child->ownerDocument->saveXML($child);
}
return $innerHTML;
}
} }

View File

@ -2058,9 +2058,6 @@ class BBCode
// Default iframe allowed domains/path // Default iframe allowed domains/path
$allowedIframeDomains = [ $allowedIframeDomains = [
DI::baseUrl()->getHost()
. (DI::baseUrl()->getPath() ? '/' . DI::baseUrl()->getPath() : '')
. '/oembed/', # The path part has to change with the source in Content\Oembed::iframe
'www.youtube.com/embed/', 'www.youtube.com/embed/',
'player.vimeo.com/video/', 'player.vimeo.com/video/',
]; ];

View File

@ -1,74 +0,0 @@
<?php
/**
* @copyright Copyright (C) 2010-2024, the Friendica project
*
* @license GNU AGPL version 3 or any later version
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as
* published by the Free Software Foundation, either version 3 of the
* License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <https://www.gnu.org/licenses/>.
*
*/
namespace Friendica\Module;
use Friendica\BaseModule;
use Friendica\Content;
use Friendica\Core\System;
use Friendica\DI;
use Friendica\Util\Strings;
/**
* Oembed module
*
* Displays stored embed content based on a base64 hash of a remote URL
*
* Example: /oembed/aHR0cHM6Ly9...
*
* @author Hypolite Petovan <hypolite@mrpetovan.com>
*/
class Oembed extends BaseModule
{
protected function content(array $request = []): string
{
// Unused form: /oembed/b2h?url=...
if (DI::args()->getArgv()[1] == 'b2h') {
$url = ["", trim(hex2bin($_REQUEST['url']))];
echo Content\OEmbed::replaceCallback($url);
System::exit();
}
// Unused form: /oembed/h2b?text=...
if (DI::args()->getArgv()[1] == 'h2b') {
$text = trim(hex2bin($_REQUEST['text']));
echo Content\OEmbed::HTML2BBCode($text);
System::exit();
}
// @TODO: Replace with parameter from router
if (DI::args()->getArgc() == 2) {
echo '<html><body>';
$url = Strings::base64UrlDecode(DI::args()->getArgv()[1]);
$j = Content\OEmbed::fetchURL($url);
// workaround for media.ccc.de (and any other endpoint that return size 0)
if (substr($j->html, 0, 7) == "<iframe" && strstr($j->html, 'width="0"')) {
$j->html = '<style>html,body{margin:0;padding:0;} iframe{width:100%;height:100%;}</style>' . $j->html;
$j->html = str_replace('width="0"', '', $j->html);
$j->html = str_replace('height="0"', '', $j->html);
}
echo $j->html;
echo '</body></html>';
}
System::exit();
}
}

View File

@ -99,8 +99,6 @@ class ParseUrl
* Search for cached embeddable data of an url otherwise fetch it * Search for cached embeddable data of an url otherwise fetch it
* *
* @param string $url The url of the page which should be scraped * @param string $url The url of the page which should be scraped
* @param bool $do_oembed The false option is used by the function fetch_oembed()
* to avoid endless loops
* *
* @return array which contains needed data for embedding * @return array which contains needed data for embedding
* string 'url' => The url of the parsed page * string 'url' => The url of the parsed page
@ -115,7 +113,7 @@ class ParseUrl
* @see ParseUrl::getSiteinfo() for more information about scraping * @see ParseUrl::getSiteinfo() for more information about scraping
* embeddable content * embeddable content
*/ */
public static function getSiteinfoCached(string $url, bool $do_oembed = true): array public static function getSiteinfoCached(string $url): array
{ {
if (empty($url)) { if (empty($url)) {
return [ return [
@ -127,14 +125,14 @@ class ParseUrl
$urlHash = hash('sha256', $url); $urlHash = hash('sha256', $url);
$parsed_url = DBA::selectFirst('parsed_url', ['content'], $parsed_url = DBA::selectFirst('parsed_url', ['content'],
['url_hash' => $urlHash, 'oembed' => $do_oembed] ['url_hash' => $urlHash, 'oembed' => false]
); );
if (!empty($parsed_url['content'])) { if (!empty($parsed_url['content'])) {
$data = unserialize($parsed_url['content']); $data = unserialize($parsed_url['content']);
return $data; return $data;
} }
$data = self::getSiteinfo($url, $do_oembed); $data = self::getSiteinfo($url);
$expires = $data['expires']; $expires = $data['expires'];
@ -144,7 +142,7 @@ class ParseUrl
'parsed_url', 'parsed_url',
[ [
'url_hash' => $urlHash, 'url_hash' => $urlHash,
'oembed' => $do_oembed, 'oembed' => false,
'url' => $url, 'url' => $url,
'content' => serialize($data), 'content' => serialize($data),
'created' => DateTimeFormat::utcNow(), 'created' => DateTimeFormat::utcNow(),
@ -194,7 +192,7 @@ class ParseUrl
* </body> * </body>
* @endverbatim * @endverbatim
*/ */
public static function getSiteinfo(string $url, bool $do_oembed = true, int $count = 1): array public static function getSiteinfo(string $url, int $count = 1): array
{ {
if (empty($url)) { if (empty($url)) {
return [ return [
@ -254,41 +252,6 @@ class ParseUrl
$body = $curlResult->getBodyString(); $body = $curlResult->getBodyString();
if ($do_oembed) {
$oembed_data = OEmbed::fetchURL($url, false, false);
if (!empty($oembed_data->type)) {
if (!in_array($oembed_data->type, ['error', 'rich', 'image', 'video', 'audio', ''])) {
$siteinfo['type'] = $oembed_data->type;
}
// See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
if ($siteinfo['type'] != 'photo') {
if (!empty($oembed_data->title)) {
$siteinfo['title'] = trim($oembed_data->title);
}
if (!empty($oembed_data->description)) {
$siteinfo['text'] = trim($oembed_data->description);
}
if (!empty($oembed_data->author_name)) {
$siteinfo['author_name'] = trim($oembed_data->author_name);
}
if (!empty($oembed_data->author_url)) {
$siteinfo['author_url'] = Network::sanitizeUrl($oembed_data->author_url);
}
if (!empty($oembed_data->provider_name)) {
$siteinfo['publisher_name'] = trim($oembed_data->provider_name);
}
if (!empty($oembed_data->provider_url)) {
$siteinfo['publisher_url'] = Network::sanitizeUrl($oembed_data->provider_url);
}
if (!empty($oembed_data->thumbnail_url)) {
$siteinfo['image'] = $oembed_data->thumbnail_url;
}
}
}
}
$charset = ''; $charset = '';
try { try {
// Look for a charset, first in headers // Look for a charset, first in headers
@ -351,7 +314,7 @@ class ParseUrl
} }
} }
if ($content != '') { if ($content != '') {
$siteinfo = self::getSiteinfo($content, $do_oembed, ++$count); $siteinfo = self::getSiteinfo($content, ++$count);
return $siteinfo; return $siteinfo;
} }
} }

View File

@ -557,11 +557,6 @@ return [
'/objects/{guid}[/{activity}]' => [Module\ActivityPub\Objects::class, [R::GET]], '/objects/{guid}[/{activity}]' => [Module\ActivityPub\Objects::class, [R::GET]],
'/oembed' => [
'/b2h' => [Module\Oembed::class, [R::GET]],
'/h2b' => [Module\Oembed::class, [R::GET]],
'/{hash}' => [Module\Oembed::class, [R::GET]],
],
'/outbox/{nickname}' => [Module\ActivityPub\Outbox::class, [R::GET, R::POST]], '/outbox/{nickname}' => [Module\ActivityPub\Outbox::class, [R::GET, R::POST]],
'/owa' => [Module\Owa::class, [R::GET]], '/owa' => [Module\Owa::class, [R::GET]],
'/openid' => [Module\Security\OpenID::class, [R::GET]], '/openid' => [Module\Security\OpenID::class, [R::GET]],