From b71e34d890415ed06db6c059cacdda4d0f8ff617 Mon Sep 17 00:00:00 2001
From: "Eloy Lafuente (stronk7)" <stronk7@moodle.org>
Date: Mon, 4 Jun 2012 12:40:37 +0200
Subject: [PATCH] MDL-33353 make fix_utf8() to work better under some systems.

This is a partial backport of MDL-32586 and MDL-33007 in order
to get some systems handling utf-8 strings better:

- debian/ubuntu. Avoid some glibc iconv() implementations to
  throw one PHP notice each time the //IGNORE option is used.
- redhat. Avoid some implementations to return false
  each time the //IGNORE option is used.

Note this does not include any backport of the admin notification
not the configmin stuff.
---
 lib/moodlelib.php                | 39 +++++++++++++++++++++++++++++++-
 lib/simpletest/testmoodlelib.php |  3 ++-
 2 files changed, 40 insertions(+), 2 deletions(-)

diff --git a/lib/moodlelib.php b/lib/moodlelib.php
index 38bf0fec0ec..ae868b86afd 100644
--- a/lib/moodlelib.php
+++ b/lib/moodlelib.php
@@ -1127,7 +1127,44 @@ function fix_utf8($value) {
             // shortcut
             return $value;
         }
-        return iconv('UTF-8', 'UTF-8//IGNORE', $value);
+
+        // Note: This is a partial backport of MDL-32586 and MDL-33007 to stable branches.
+        // Lower error reporting because glibc throws bogus notices.
+        $olderror = error_reporting();
+        if ($olderror & E_NOTICE) {
+            error_reporting($olderror ^ E_NOTICE);
+        }
+
+        // Detect buggy iconv implementations borking results.
+        static $buggyiconv = null;
+        if ($buggyiconv === null) {
+            $buggyiconv = (!function_exists('iconv') or iconv('UTF-8', 'UTF-8//IGNORE', '100'.chr(130).'€') !== '100€');
+        }
+
+        if ($buggyiconv) {
+            if (function_exists('mb_convert_encoding')) {
+                // Fallback to mbstring if available.
+                $subst = mb_substitute_character();
+                mb_substitute_character('');
+                $result = mb_convert_encoding($value, 'utf-8', 'utf-8');
+                mb_substitute_character($subst);
+
+            } else {
+                // Return unmodified text, mbstring not available.
+                $result = $value;
+            }
+
+        } else {
+            // Working iconv, use it normally (with PHP notices disabled)
+            $result = iconv('UTF-8', 'UTF-8//IGNORE', $value);
+        }
+
+        // Back to original reporting level
+        if ($olderror & E_NOTICE) {
+            error_reporting($olderror);
+        }
+
+        return $result;
 
     } else if (is_array($value)) {
         foreach ($value as $k=>$v) {
diff --git a/lib/simpletest/testmoodlelib.php b/lib/simpletest/testmoodlelib.php
index 0fed885b0ff..a17fc1697a3 100644
--- a/lib/simpletest/testmoodlelib.php
+++ b/lib/simpletest/testmoodlelib.php
@@ -311,6 +311,7 @@ class moodlelib_test extends UnitTestCase {
         $this->assertidentical(1, fix_utf8(1));
         $this->assertidentical(1.1, fix_utf8(1.1));
         $this->assertidentical(true, fix_utf8(true));
+        $this->assertidentical('abc', fix_utf8('abc'));
         $this->assertidentical('', fix_utf8(''));
         $array = array('do', 're', 'mi');
         $this->assertidentical($array, fix_utf8($array));
@@ -323,7 +324,7 @@ class moodlelib_test extends UnitTestCase {
         $this->assertidentical("žlutý koníček přeskočil potůček \n\t\r\0", fix_utf8("žlutý koníček přeskočil potůček \n\t\r\0"));
 
         // invalid utf8 string
-        $this->assertidentical('aaabbb', fix_utf8('aaa'.chr(130).'bbb'));
+        $this->assertidentical('aš', fix_utf8('a'.chr(130).'š'), 'This fails with buggy iconv() when mbstring extenstion is not available as fallback.');
     }
 
     function test_optional_param() {