From 70351389f710e34eb7767f688e607f6738bfce03 Mon Sep 17 00:00:00 2001
From: Tim Hunt <T.J.Hunt@open.ac.uk>
Date: Fri, 26 Apr 2013 15:59:53 +0100
Subject: [PATCH] MDL-39378 Improve s() performance.

These changes give about a 10% speed-up in this function. The significant changes are:
1. Simplify the if logic to remove unnecssary cases.
2. Dont pass default argument values to htmlspecialchars, just using the
   defaults is faster.
3. I can confirm that /i regex is faster than the equivalent regex without the i.

I also added more unit tests to test the edge cases.
---
 lib/tests/weblib_test.php | 22 ++++++++++++++++++++--
 lib/weblib.php            | 12 ++++++------
 2 files changed, 26 insertions(+), 8 deletions(-)

diff --git a/lib/tests/weblib_test.php b/lib/tests/weblib_test.php
index 2b9eadc822d..6f94b08ea4b 100644
--- a/lib/tests/weblib_test.php
+++ b/lib/tests/weblib_test.php
@@ -70,9 +70,27 @@ class web_testcase extends advanced_testcase {
     }
 
     function test_s() {
-        $this->assertEquals(s("This Breaks \" Strict"), "This Breaks &quot; Strict");
-        $this->assertEquals(s("This Breaks <a>\" Strict</a>"), "This Breaks &lt;a&gt;&quot; Strict&lt;/a&gt;");
+        // Special cases.
+        $this->assertSame('0', s(0));
+        $this->assertSame('0', s('0'));
+        $this->assertSame('0', s(false));
+        $this->assertSame('', s(null));
+
+        // Normal cases.
+        $this->assertEquals('This Breaks &quot; Strict', s('This Breaks " Strict'));
+        $this->assertEquals('This Breaks &lt;a&gt;&quot; Strict&lt;/a&gt;', s('This Breaks <a>" Strict</a>'));
+
+        // Unicode characters.
+        $this->assertEquals('Café', s('Café'));
+        $this->assertEquals('一, 二, 三', s('一, 二, 三'));
+
+        // Don't escape already-escaped numeric entities. (Note, this behaviour
+        // may not be desirable. Perhaps we should remove these tests and that
+        // functionality, but we can only do that if we understand why it was added.)
         $this->assertEquals('An entity: &#x09ff;.', s('An entity: &#x09ff;.'));
+        $this->assertEquals('An entity: &#1073;.', s('An entity: &#1073;.'));
+        $this->assertEquals('An entity: &amp;amp;.', s('An entity: &amp;.'));
+        $this->assertEquals('Not an entity: &amp;amp;#x09ff;.', s('Not an entity: &amp;#x09ff;.'));
     }
 
     function test_format_text_email() {
diff --git a/lib/weblib.php b/lib/weblib.php
index d8c9d567a21..647504a19cd 100644
--- a/lib/weblib.php
+++ b/lib/weblib.php
@@ -85,19 +85,19 @@ define('URL_MATCH_EXACT', 2);
  * Returns $var with HTML characters (like "<", ">", etc.) properly quoted.
  * This function is very similar to {@link p()}
  *
- * @todo Remove obsolete param $obsolete if not used anywhere
- *
  * @param string $var the string potentially containing HTML characters
- * @param boolean $obsolete no longer used.
  * @return string
  */
-function s($var, $obsolete = false) {
+function s($var) {
 
-    if ($var === '0' or $var === false or $var === 0) {
+    if ($var === false) {
         return '0';
     }
 
-    return preg_replace("/&amp;#(\d+|x[0-9a-f]+);/i", "&#$1;", htmlspecialchars($var, ENT_QUOTES, 'UTF-8', true));
+    // When we move to PHP 5.4 as a minimum version, change ENT_QUOTES on the
+    // next line to ENT_QUOTES | ENT_HTML5 | ENT_SUBSTITUTE, and remove the
+    // 'UTF-8' argument. Both bring a speed-increase.
+    return preg_replace('/&amp;#(\d+|x[0-9a-f]+);/i', '&#$1;', htmlspecialchars($var, ENT_QUOTES, 'UTF-8'));
 }
 
 /**