From c31c2d10b972b9efc750fcec28acdd90da81a4cd Mon Sep 17 00:00:00 2001
From: Mikael Nordfeldth <mmn@hethane.se>
Date: Tue, 17 Feb 2015 20:11:35 +0100
Subject: [PATCH] PHP>=5.4.0 lets us use Transliterator, tags now asciified!
MIME-Version: 1.0
Content-Type: text/plain; charset=utf8
Content-Transfer-Encoding: 8bit

For example: #REVOLUCIÓN becomes #revolucion instead of #revolución
---
 lib/util.php | 22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)

diff --git a/lib/util.php b/lib/util.php
index a32c35395e..b81aa5ef35 100644
--- a/lib/util.php
+++ b/lib/util.php
@@ -1116,6 +1116,20 @@ function common_xml_safe_str($str)
     return preg_replace('/[\p{Cc}\p{Cs}]/u', '*', $str);
 }
 
+function common_slugify($str)
+{
+    $str = transliterator_transliterate(
+                        'Any-Latin;' .      // any charset to latin compatible
+                            'NFD;' .        // decompose
+                            '[:Nonspacing Mark:] Remove;' . // remove nonspacing marks (accents etc.)
+                            'NFC;' .        // composite again
+                            '[:Punctuation:] Remove;' . // remove punctuation (.,¿? etc.)
+                            'Lower();' .    // turn into lowercase
+                            'Latin-ASCII;',  // get ASCII equivalents (ð to d for example)
+                        $str);
+    return preg_replace('/[^\pL\pN]/', '', $str);
+}
+
 function common_tag_link($tag)
 {
     $canonical = common_canonical_tag($tag);
@@ -1139,11 +1153,9 @@ function common_tag_link($tag)
 
 function common_canonical_tag($tag)
 {
-  // only alphanum
-  $tag = preg_replace('/[^\pL\pN]/u', '', $tag);
-  $tag = mb_convert_case($tag, MB_CASE_LOWER, "UTF-8");
-  $tag = substr($tag, 0, 64);
-  return $tag;
+    $tag = common_slugify($tag);
+    $tag = substr($tag, 0, 64);
+    return $tag;
 }
 
 function common_valid_profile_tag($str)
-- 
2.39.5