Move /parse_url module to /parseurl

- Update oexchange module to use BBCode::embedURL instead of a self-request
- Remove mod/parse_url.php file
- Restrict ParseUrl module to authenticated users

mod/oexchange.php

@@ -47,16 +47,12 @@ function oexchange_content(App $a) {
 		return;
 	}
 
-	$url = ((!empty($_REQUEST['url']))
+	$url         = !empty($_REQUEST['url'])         ? trim($_REQUEST['url'])         : '';
-		? urlencode(Strings::escapeTags(trim($_REQUEST['url']))) : '');
+	$title       = !empty($_REQUEST['title'])       ? trim($_REQUEST['title'])       : '';
-	$title = ((!empty($_REQUEST['title']))
+	$description = !empty($_REQUEST['description']) ? trim($_REQUEST['description']) : '';
-		? '&title=' . urlencode(Strings::escapeTags(trim($_REQUEST['title']))) : '');
+	$tags        = !empty($_REQUEST['tags'])        ? trim($_REQUEST['tags'])        : '';
-	$description = ((!empty($_REQUEST['description']))
-		? '&description=' . urlencode(Strings::escapeTags(trim($_REQUEST['description']))) : '');
-	$tags = ((!empty($_REQUEST['tags']))
-		? '&tags=' . urlencode(Strings::escapeTags(trim($_REQUEST['tags']))) : '');
 
-	$s = DI::httpRequest()->fetch(DI::baseUrl() . '/parse_url?url=' . $url . $title . $description . $tags);
+	$s = \Friendica\Content\Text\BBCode::embedURL($url, true, $title, $description, $tags);
 
 	if (!strlen($s)) {
 		return;

mod/parse_url.php

@@ -1,182 +0,0 @@
-<?php
-/**
- * @copyright Copyright (C) 2020, Friendica
- *
- * @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/>.
- *
- * This module does parse an url for embeddable content (audio, video, image files or link)
- * information and does format this information to BBCode
- *
- * @see ParseUrl::getSiteinfo() for more information about scraping embeddable content
- */
-
-use Friendica\App;
-use Friendica\Content\PageInfo;
-use Friendica\Core\Hook;
-use Friendica\Core\Logger;
-use Friendica\Core\System;
-use Friendica\DI;
-use Friendica\Util\ParseUrl;
-use Friendica\Util\Strings;
-
-function parse_url_content(App $a)
-{
-	$text = null;
-	$str_tags = '';
-	$format = '';
-	$ret= ['success' => false, 'contentType' => ''];
-
-	$br = "\n";
-
-	if (!empty($_GET['binurl']) && Strings::isHex($_GET['binurl'])) {
-		$url = trim(hex2bin($_GET['binurl']));
-	} elseif (!empty($_GET['url'])) {
-		$url = trim($_GET['url']);
-	// fallback in case no url is valid
-	} else {
-		Logger::info('No url given');
-		exit();
-	}
-
-	if (!empty($_GET['title'])) {
-		$title = strip_tags(trim($_GET['title']));
-	}
-
-	if (!empty($_GET['description'])) {
-		$text = strip_tags(trim($_GET['description']));
-	}
-
-	if (!empty($_GET['tags'])) {
-		$arr_tags = ParseUrl::convertTagsToArray($_GET['tags']);
-		if (count($arr_tags)) {
-			$str_tags = $br . implode(' ', $arr_tags) . $br;
-		}
-	}
-
-	if (isset($_GET['format']) && $_GET['format'] == 'json') {
-		$format = 'json';
-	}
-
-	// Add url scheme if it is missing
-	$arrurl = parse_url($url);
-	if (empty($arrurl['scheme'])) {
-		if (!empty($arrurl['host'])) {
-			$url = 'http:' . $url;
-		} else {
-			$url = 'http://' . $url;
-		}
-	}
-
-	Logger::log($url);
-
-	// Check if the URL is an image, video or audio file. If so format
-	// the URL with the corresponding BBCode media tag
-	// Fetch the header of the URL
-	$curlResponse = DI::httpRequest()->head($url);
-
-	if ($curlResponse->isSuccess()) {
-		$hdrs = $curlResponse->getHeaderArray();
-
-		$type = null;
-		$content_type = '';
-		$bbcode = '';
-		if (array_key_exists('Content-Type', $hdrs)) {
-			$type = $hdrs['Content-Type'];
-		}
-		if ($type) {
-			if (stripos($type, 'image/') !== false) {
-				$content_type = 'image';
-				$bbcode = $br . '[img]' . $url . '[/img]' . $br;
-			}
-			if (stripos($type, 'video/') !== false) {
-				$content_type = 'video';
-				$bbcode = $br . '[video]' . $url . '[/video]' . $br;
-			}
-			if (stripos($type, 'audio/') !== false) {
-				$content_type = 'audio';
-				$bbcode = $br . '[audio]' . $url . '[/audio]' . $br;
-			}
-		}
-		if (!empty($content_type)) {
-			if ($format == 'json') {
-				$ret['contentType'] = $content_type;
-				$ret['data'] = ['url' => $url];
-				$ret['success'] = true;
-				System::jsonExit($ret);
-			}
-
-			echo $bbcode;
-			exit();
-		}
-	}
-
-
-	$template = '[bookmark=%s]%s[/bookmark]%s';
-
-	$arr = ['url' => $url, 'format' => $format, 'text' => null];
-
-	Hook::callAll('parse_link', $arr);
-
-	if ($arr['text']) {
-		if ($format == 'json') {
-			System::jsonExit($arr['text']);
-		} else {
-			echo $arr['text'];
-			exit();
-		}
-	}
-
-	// If there is already some content information submitted we don't
-	// need to parse the url for content.
-	if (!empty($url) && !empty($title) && !empty($text)) {
-		$title = str_replace(["\r", "\n"], ['', ''], $title);
-
-		$text = '[quote]' . trim($text) . '[/quote]' . $br;
-
-		$result = sprintf($template, $url, ($title) ? $title : $url, $text) . $str_tags;
-
-		Logger::log('(unparsed): returns: ' . $result);
-
-		echo $result;
-		exit();
-	}
-
-	// Fetch the information directly from the webpage
-	$siteinfo = ParseUrl::getSiteinfo($url);
-
-	unset($siteinfo['keywords']);
-
-	// Bypass attachment if parse url for a comment
-	if (!empty($_GET['noAttachment'])) {
-		echo $br . '[url=' . $url . ']' . $siteinfo['title'] . '[/url]';
-		exit();
-	}
-
-	if ($format == 'json') {
-		$ret['data'] = $siteinfo;
-		$ret['contentType'] = 'attachment';
-		$ret['success'] = true;
-
-		System::jsonExit($ret);
-	}
-
-	// Format it as BBCode attachment
-	$info = "\n" . PageInfo::getFooterFromData($siteinfo);
-
-	echo $info;
-
-	exit();
-}

src/Module/ParseUrl.php

@@ -0,0 +1,129 @@
+<?php
+/**
+ * @copyright Copyright (C) 2020, Friendica
+ *
+ * @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\Text\BBCode;
+use Friendica\Core\Hook;
+use Friendica\Core\Session;
+use Friendica\Core\System;
+use Friendica\Network\HTTPException\BadRequestException;
+use Friendica\Util;
+
+class ParseUrl extends BaseModule
+{
+	public static function rawContent(array $parameters = [])
+	{
+		if (!Session::isAuthenticated()) {
+			throw new \Friendica\Network\HTTPException\ForbiddenException();
+		}
+
+		$format = '';
+		$title = '';
+		$description = '';
+		$ret = ['success' => false, 'contentType' => ''];
+
+		if (!empty($_GET['binurl']) && Util\Strings::isHex($_GET['binurl'])) {
+			$url = trim(hex2bin($_GET['binurl']));
+		} elseif (!empty($_GET['url'])) {
+			$url = trim($_GET['url']);
+			// fallback in case no url is valid
+		} else {
+			throw new BadRequestException('No url given');
+		}
+
+		if (!empty($_GET['title'])) {
+			$title = strip_tags(trim($_GET['title']));
+		}
+
+		if (!empty($_GET['description'])) {
+			$description = strip_tags(trim($_GET['description']));
+		}
+
+		if (!empty($_GET['tags'])) {
+			$arr_tags = Util\ParseUrl::convertTagsToArray($_GET['tags']);
+			if (count($arr_tags)) {
+				$str_tags = "\n" . implode(' ', $arr_tags) . "\n";
+			}
+		}
+
+		if (isset($_GET['format']) && $_GET['format'] == 'json') {
+			$format = 'json';
+		}
+
+		// Add url scheme if it is missing
+		$arrurl = parse_url($url);
+		if (empty($arrurl['scheme'])) {
+			if (!empty($arrurl['host'])) {
+				$url = 'http:' . $url;
+			} else {
+				$url = 'http://' . $url;
+			}
+		}
+
+		$arr = ['url' => $url, 'format' => $format, 'text' => null];
+
+		Hook::callAll('parse_link', $arr);
+
+		if ($arr['text']) {
+			if ($format == 'json') {
+				System::jsonExit($arr['text']);
+			} else {
+				echo $arr['text'];
+				exit();
+			}
+		}
+
+		if ($format == 'json') {
+			$siteinfo = Util\ParseUrl::getSiteinfoCached($url);
+
+			if (in_array($siteinfo['type'], ['image', 'video', 'audio'])) {
+				switch ($siteinfo['type']) {
+					case 'video':
+						$content_type = 'video';
+						break;
+					case 'audio':
+						$content_type = 'audio';
+						break;
+					default:
+						$content_type = 'image';
+						break;
+				}
+
+				$ret['contentType'] = $content_type;
+				$ret['data'] = ['url' => $url];
+				$ret['success'] = true;
+			} else {
+				unset($siteinfo['keywords']);
+
+				$ret['data'] = $siteinfo;
+				$ret['contentType'] = 'attachment';
+				$ret['success'] = true;
+			}
+
+			System::jsonExit($ret);
+		} else {
+			echo BBCode::embedURL($url, empty($_GET['noAttachment']), $title, $description, $_GET['tags'] ?? '');
+			exit();
+		}
+	}
+}

static/routes.config.php

@@ -346,6 +346,7 @@ return [
 	'/openid'         => [Module\Security\OpenID::class, [R::GET]],
 	'/opensearch'     => [Module\OpenSearch::class,      [R::GET]],
 
+	'/parseurl'                           => [Module\ParseUrl::class,          [R::GET]],
 	'/permission/tooltip/{type}/{id:\d+}' => [Module\PermissionTooltip::class, [R::GET]],
 
 	'/photo' => [

view/js/linkPreview.js

@@ -178,7 +178,7 @@
 		 * @returns {void}
 		 */
 		var getContentData = function(binurl, callback) {
-			$.get('parse_url?binurl='+ binurl + '&format=json', function (answer) {
+			$.get('parseurl?binurl='+ binurl + '&format=json', function (answer) {
 				obj = sanitizeInputData(answer);
 
 				// Put the data into a cache

view/templates/jot-header.tpl

@@ -103,7 +103,7 @@ function enableOnUser(){
 		if(reply && reply.length) {
 			reply = bin2hex(reply);
 			$('#profile-rotator').show();
-			$.get('parse_url?binurl=' + reply, function(data) {
+			$.get('parseurl?binurl=' + reply, function(data) {
 				addeditortext(data);
 				$('#profile-rotator').hide();
 			});
@@ -160,7 +160,7 @@ function enableOnUser(){
 		if(reply && reply.length) {
 			reply = bin2hex(reply);
 			$('#profile-rotator').show();
-			$.get('parse_url?binurl=' + reply, function(data) {
+			$.get('parseurl?binurl=' + reply, function(data) {
 				if (!editor) $("#profile-jot-text").val("");
 				initEditor(function(){
 					addeditortext(data);

view/templates/msg-header.tpl

@@ -23,7 +23,7 @@
 		reply = prompt("{{$linkurl}}");
 		if(reply && reply.length) {
 			$('#profile-rotator').show();
-			$.get('parse_url?url=' + reply, function(data) {
+			$.get('parseurl?url=' + reply, function(data) {
 				addeditortext(data);
 				$('#profile-rotator').hide();
 			});
@@ -42,7 +42,7 @@
 		event.preventDefault();
 		if(reply && reply.length) {
 			$('#profile-rotator').show();
-			$.get('parse_url?url=' + reply, function(data) {
+			$.get('parseurl?url=' + reply, function(data) {
 				addeditortext(data);
 				$('#profile-rotator').hide();
 			});

view/templates/wallmsg-header.tpl

@@ -7,7 +7,7 @@
 		reply = prompt("{{$linkurl}}");
 		if(reply && reply.length) {
 			$('#profile-rotator').show();
-			$.get('parse_url?url=' + reply, function(data) {
+			$.get('parseurl?url=' + reply, function(data) {
 				addeditortext(data);
 				$('#profile-rotator').hide();
 			});
@@ -26,7 +26,7 @@
 		event.preventDefault();
 		if(reply && reply.length) {
 			$('#profile-rotator').show();
-			$.get('parse_url?url=' + reply, function(data) {
+			$.get('parseurl?url=' + reply, function(data) {
 				addeditortext(data);
 				$('#profile-rotator').hide();
 			});

view/theme/frio/js/jot.js

@@ -30,7 +30,7 @@ function jotGetLink() {
 			// Fallback: insert the attachment bbcode directly into the textarea
 			// if the attachment live preview isn't available
 		} else {
-			$.get("parse_url?binurl=" + bin2hex(reply) + noAttachment, function (data) {
+			$.get("parseurl?binurl=" + bin2hex(reply) + noAttachment, function (data) {
 				addeditortext(data);
 				$("#profile-rotator").hide();
 			});

view/theme/frio/js/textedit.js

@@ -40,7 +40,7 @@ function commentGetLink(id, prompttext) {
 	reply = prompt(prompttext);
 	if (reply && reply.length) {
 		reply = bin2hex(reply);
-		$.get("parse_url?noAttachment=1&binurl=" + reply, function (data) {
+		$.get("parseurl?noAttachment=1&binurl=" + reply, function (data) {
 			addCommentText(data, id);
 		});
 	}
@@ -64,7 +64,7 @@ function commentLinkDrop(event, id) {
 	event.preventDefault();
 	if (reply && reply.length) {
 		reply = bin2hex(reply);
-		$.get("parse_url?noAttachment=1&binurl=" + reply, function (data) {
+		$.get("parseurl?noAttachment=1&binurl=" + reply, function (data) {
 			addCommentText(data, id);
 		});
 	}

view/theme/frio/templates/jot-header.tpl

@@ -223,7 +223,7 @@
 			if (currentText.includes("[attachment") && currentText.includes("[/attachment]")) {
 				noAttachment = '&noAttachment=1';
 			}
-			$.get('parse_url?binurl=' + reply + noAttachment, function(data) {
+			$.get('parseurl?binurl=' + reply + noAttachment, function(data) {
 				if (!editor) $("#profile-jot-text").val("");
 				initEditor(function(){
 					addeditortext(data);

view/theme/smoothly/templates/jot-header.tpl

@@ -129,7 +129,7 @@ function enableOnUser(){
 		if(reply && reply.length) {
 			reply = bin2hex(reply);
 			$('#profile-rotator').show();
-			$.get('parse_url?binurl=' + reply, function(data) {
+			$.get('parseurl?binurl=' + reply, function(data) {
 				addeditortext(data);
 				$('#profile-rotator').hide();
 			});
@@ -190,7 +190,7 @@ function enableOnUser(){
 		if(reply && reply.length) {
 			reply = bin2hex(reply);
 			$('#profile-rotator').show();
-			$.get('parse_url?binurl=' + reply, function(data) {
+			$.get('parseurl?binurl=' + reply, function(data) {
 				if (!editor) $("#profile-jot-text").val("");
 				initEditor(function(){
 					addeditortext(data);