[LON-CAPA-cvs] cvs: loncom /localize transliterate.pm

[raeburn]

raeburn		Sat Mar  2 23:08:51 2019 EDT

  Modified files:              
    /loncom/localize	transliterate.pm 
  Log:
  - Bug 6792
    Normalize unicode representations before transliteration.
  
  
Index: loncom/localize/transliterate.pm
diff -u loncom/localize/transliterate.pm:1.1 loncom/localize/transliterate.pm:1.2
--- loncom/localize/transliterate.pm:1.1	Tue Feb 26 14:42:22 2019
+++ loncom/localize/transliterate.pm	Sat Mar  2 23:08:51 2019
@@ -1,7 +1,7 @@
 # The LearningOnline Network with CAPA
 # Transliteration to ascii
 #
-# $Id: transliterate.pm,v 1.1 2019/02/26 14:42:22 raeburn Exp $
+# $Id: transliterate.pm,v 1.2 2019/03/02 23:08:51 raeburn Exp $
 #
 # Copyright Michigan State University Board of Trustees
 #
@@ -60,6 +60,7 @@
 use utf8;
 use Text::Unidecode qw(unidecode);
 use Encode qw(decode_utf8 encode_utf8);
+use Unicode::Normalize qw(normalize);
 
 =pod
 
@@ -90,12 +91,14 @@
 sub fname_to_ascii {
     my ($fname,$language) = @_;
     if ($fname =~ /([^\x{00}-\x{7f}])/) {
-        $fname=&decode_utf8($fname);
+        $fname = decode_utf8($fname);
+        $fname = normalize('D',$fname);
+        $fname = normalize('C',$fname);
         if ($language eq 'de') {
             $fname = &german_to_ascii($fname);
         }
         $fname = unidecode($fname);
-        $fname=&encode_utf8($fname);
+        $fname = encode_utf8($fname);
     }
     return $fname;
 }