From 3395f5603699aa5a76cadbbea64db4ca8c681f4e Mon Sep 17 00:00:00 2001
From: Hypolite Petovan <hypolite@mrpetovan.com>
Date: Sun, 19 May 2019 08:45:29 -0400
Subject: [PATCH 1/3] Add Unicode support to autolink regular expression

- Explicitly exclude non-breaking spaces from URLs as \s doesn't include them
---
 src/Util/Strings.php | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/src/Util/Strings.php b/src/Util/Strings.php
index b2b710d96f..3f8990d6c1 100644
--- a/src/Util/Strings.php
+++ b/src/Util/Strings.php
@@ -355,25 +355,25 @@ class Strings
 	 */
 	public static function autoLinkRegEx()
 	{
-		return '@(?xi)
+		return '@
 (?<![=\'\]"/])          # Not preceded by [, =, \', ], ", /
 \b
 (                              # Capture 1: entire matched URL
   https?://                            # http or https protocol
   (?:
-    [^/\s`!()\[\]{};:\'",<>?«»“”‘’.]    # Domain can\'t start with a . 
-    [^/\s`!()\[\]{};:\'",<>?«»“”‘’]+    # Domain can\'t end with a .
+    [^/\s\xA0`!()\[\]{};:\'",<>?«»“”‘’.]    # Domain can\'t start with a . 
+    [^/\s\xA0`!()\[\]{};:\'",<>?«»“”‘’]+    # Domain can\'t end with a .
     \.
-    [^/\s`!()\[\]{};:\'".,<>?«»“”‘’]+/? # Followed by a slash
+    [^/\s\xA0`!()\[\]{};:\'".,<>?«»“”‘’]+/? # Followed by a slash
   )
   (?:                                  # One or more:
-    [^\s()<>]+                         # Run of non-space, non-()<>
+    [^\s\xA0()<>]+                         # Run of non-space, non-()<>
     |                                  #   or
-    \(([^\s()<>]+|(\([^\s()<>]+\)))*\) # balanced parens, up to 2 levels
+    \(([^\s\xA0()<>]+|(\([^\s()<>]+\)))*\) # balanced parens, up to 2 levels
     |                                  #   or
-    [^\s`!()\[\]{};:\'".,<>?«»“”‘’]    # not a space or one of these punct chars
+    [^\s\xA0`!()\[\]{};:\'".,<>?«»“”‘’]    # not a space or one of these punct chars
   )*
-)@';
+)@xiu';
 	}
 
 	/**

From 325ba20141dfa4a85411cf24ea692ea0174cf5b2 Mon Sep 17 00:00:00 2001
From: Hypolite Petovan <hypolite@mrpetovan.com>
Date: Sun, 19 May 2019 08:45:54 -0400
Subject: [PATCH 2/3] Improve Diaspora raw text support in /babel module

---
 src/Module/Debug/Babel.php | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/src/Module/Debug/Babel.php b/src/Module/Debug/Babel.php
index 899c8033c4..1ce1ac4c6e 100644
--- a/src/Module/Debug/Babel.php
+++ b/src/Module/Debug/Babel.php
@@ -7,6 +7,7 @@ use Friendica\Content\Text;
 use Friendica\Core\L10n;
 use Friendica\Core\Renderer;
 use Friendica\Model\Item;
+use Friendica\Util\XML;
 
 /**
  * Translates input text into different formats (HTML, BBCode, Markdown)
@@ -98,10 +99,10 @@ class Babel extends BaseModule
 					$markdown = trim($_REQUEST['text']);
 					$results[] = [
 						'title'   => L10n::t('Source input (Diaspora format)'),
-						'content' => '<pre>' . $markdown . '</pre>'
+						'content' => '<pre>' . htmlspecialchars($markdown) . '</pre>'
 					];
 
-					$html = Text\Markdown::convert($markdown);
+					$html = Text\Markdown::convert(html_entity_decode($markdown,ENT_COMPAT, 'UTF-8'));
 					$results[] = [
 						'title'   => L10n::t('Markdown::convert (raw HTML)'),
 						'content' => visible_whitespace(htmlspecialchars($html))
@@ -112,7 +113,7 @@ class Babel extends BaseModule
 						'content' => $html
 					];
 
-					$bbcode = Text\Markdown::toBBCode($markdown);
+					$bbcode = Text\Markdown::toBBCode(XML::unescape($markdown));
 					$results[] = [
 						'title'   => L10n::t('Markdown::toBBCode'),
 						'content' => '<pre>' . $bbcode . '</pre>'

From c21a227d0a9509c85cafd7316f631570b12a53c9 Mon Sep 17 00:00:00 2001
From: Hypolite Petovan <hypolite@mrpetovan.com>
Date: Sun, 19 May 2019 14:57:28 -0400
Subject: [PATCH 3/3] Add test for bug 7150

---
 tests/src/Content/Text/BBCodeTest.php | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/tests/src/Content/Text/BBCodeTest.php b/tests/src/Content/Text/BBCodeTest.php
index 802a90278b..864a3794d3 100644
--- a/tests/src/Content/Text/BBCodeTest.php
+++ b/tests/src/Content/Text/BBCodeTest.php
@@ -109,6 +109,10 @@ class BBCodeTest extends MockedTest
 				'data' => "http://example.com<ul>",
 				'assertHTML' => false
 			],
+			'bug-7150' => [
+				'data' => html_entity_decode('http://example.com&nbsp;', ENT_QUOTES, 'UTF-8'),
+				'assertHTML' => false
+			],
 		];
 	}