From 74a61d559f6270fd98c1ec0003db65e33a297b98 Mon Sep 17 00:00:00 2001
From: Laurent Destailleur <eldy@destailleur.fr>
Date: Sun, 14 Mar 2021 15:38:10 +0100
Subject: [PATCH] FIX sanitizing with GETPOST(alphanohtml) #yogosha5629

---
 htdocs/core/lib/functions.lib.php | 5 +++--
 test/phpunit/SecurityTest.php     | 5 +++++
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/htdocs/core/lib/functions.lib.php b/htdocs/core/lib/functions.lib.php
index 4fc283b8082..e6547c8c5f1 100644
--- a/htdocs/core/lib/functions.lib.php
+++ b/htdocs/core/lib/functions.lib.php
@@ -6167,12 +6167,13 @@ function dol_string_nohtmltag($stringtoclean, $removelinefeed = 1, $pagecodeto =
 	if ($strip_tags) {
 		$temp = strip_tags($temp);
 	} else {
+		$temp = str_replace('<>', '', $temp);	// No reason to have this into a text, except if value is to try bypass the next html cleaning
 		$pattern = "/<[^<>]+>/";
 		// Example of $temp: <a href="/myurl" title="<u>A title</u>">0000-021</a>
 		$temp = preg_replace($pattern, "", $temp); // pass 1 - $temp after pass 1: <a href="/myurl" title="A title">0000-021
 		$temp = preg_replace($pattern, "", $temp); // pass 2 - $temp after pass 2: 0000-021
-		// Remove '<' into remainging, so non closing html tags like '<abc'. Note: '<123abc' is not a html tag (can be kept), but '<abc123' is (must be removed).
-		$temp = preg_replace('/<([a-z]+)/i', '\1', $temp);
+		// Remove '<' into remainging, so remove non closing html tags like '<abc' or '<<abc'. Note: '<123abc' is not a html tag (can be kept), but '<abc123' is (must be removed).
+		$temp = preg_replace('/<+([a-z]+)/i', '\1', $temp);
 	}
 
 	$temp = dol_html_entity_decode($temp, ENT_COMPAT, $pagecodeto);
diff --git a/test/phpunit/SecurityTest.php b/test/phpunit/SecurityTest.php
index 1c29c2ab3e5..9aaf1349f14 100644
--- a/test/phpunit/SecurityTest.php
+++ b/test/phpunit/SecurityTest.php
@@ -322,6 +322,7 @@ class SecurityTest extends PHPUnit\Framework\TestCase
 		$_POST['param8c']='< with space after is ok';
 		$_POST['param8d']='<abc123 is html to clean';
 		$_POST['param8e']='<123abc is not html to clean';
+		$_POST['param8f']='abc<<svg <><<animate onbegin=alert(document.domain) a';
 		$_POST["param9"]='is_object($object) ? ($object->id < 10 ? round($object->id / 2, 2) : (2 * $user->id) * (int) substr($mysoc->zip, 1, 2)) : \'objnotdefined\'';
 		$_POST["param10"]='is_object($object) ? ($object->id < 10 ? round($object->id / 2, 2) : (2 * $user->id) * (int) substr($mysoc->zip, 1, 2)) : \'<abc>objnotdefined\'';
 		$_POST["param11"]=' Name <email@email.com> ';
@@ -411,6 +412,10 @@ class SecurityTest extends PHPUnit\Framework\TestCase
 		print __METHOD__." result=".$result."\n";
 		$this->assertEquals($_POST['param8e'], $result, 'Test a string with non closing html tag with alphanohtml');
 
+		$result=GETPOST("param8f", 'alphanohtml');
+		print __METHOD__." result=".$result."\n";
+		$this->assertEquals('abcsvg animate onbegin=alert(document.domain) a', $result, 'Test a string with html tag open with several <');
+
 		$result=GETPOST("param9", 'alphanohtml');
 		print __METHOD__." result=".$result."\n";
 		$this->assertEquals($_POST["param9"], $result);