Bug fixing charsets Using Unicode::Normalize

[koha.git] / C4 / Charset.pm

diff --git a/C4/Charset.pm b/C4/Charset.pm
index 001cf01..541ebce 100644 (file)
--- a/C4/Charset.pm
+++ b/C4/Charset.pm
@@ -18,8 +18,11 @@ package C4::Charset;
 # Suite 330, Boston, MA  02111-1307 USA
 
 use strict;
+use warnings;
+
 use MARC::Charset qw/marc8_to_utf8/;
 use Text::Iconv;
+use Unicode::Normalize;
 
 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS);
 
@@ -33,6 +36,7 @@ BEGIN {
         MarcToUTF8Record
         SetMarcUnicodeFlag
         StripNonXmlChars
+               Normalize_String
     );
 }
 
@@ -108,6 +112,30 @@ sub IsStringUTF8ish {
     return utf8::decode($str);
 }
 
+
+=head2 Normalize_String
+
+=over 4
+
+my $$string_normalized = Normalize_String($string);
+
+=back
+
+Returns normalized string C<$string> in C Form 
+
+
+=cut
+
+sub Normalize_String {
+    my $string = shift;
+       if (IsStringUTF8ish($string)){
+               return NFC($string);
+       }
+       else {
+               return $string;
+       }
+}
+
 =head2 MarcToUTF8Record
 
 =over 4
@@ -153,7 +181,23 @@ sub MarcToUTF8Record {
         $marc =~ s/^\s+//;
         $marc =~ s/\s+$//;
         $marc_blob_is_utf8 = IsStringUTF8ish($marc);
-        $marc_record = MARC::Record->new_from_usmarc($marc);
+        eval {
+            $marc_record = MARC::Record->new_from_usmarc($marc);
+        };
+        if ($@) {
+            # if we fail the first time, one likely problem
+            # is that we have a MARC21 record that says that it's
+            # UTF-8 (Leader/09 = 'a') but contains non-UTF-8 characters.
+            # We'll try parsing it again.
+            substr($marc, 9, 1) = ' ';
+            eval {
+                $marc_record = MARC::Record->new_from_usmarc($marc);
+            };
+            if ($@) {
+                # it's hopeless; return an empty MARC::Record
+                return MARC::Record->new(), 'failed', ['could not parse MARC blob'];
+            }
+        }
     }
 
     # If we do not know the source encoding, try some guesses
@@ -197,7 +241,7 @@ sub MarcToUTF8Record {
             @errors = _marc_iso5426_to_utf8($marc_record, $marc_flavour);
         } else {
             # assume any other character encoding is for Text::Iconv
-            @errors = _marc_to_utf8_via_text_iconv($marc_record, $marc_flavour, 'iso-8859-1');
+            @errors = _marc_to_utf8_via_text_iconv($marc_record, $marc_flavour, $source_encoding);
         }
 
         if (@errors) {
@@ -437,6 +481,7 @@ sub _marc_marc8_to_utf8 {
                     # that the resulting string is UTF-8.
                     utf8::upgrade($utf8sf);
                 }
+                           $utf8sf=NFC($utf8sf);
                 push @converted_subfields, $subfield->[0], $utf8sf;
             }
 
@@ -485,6 +530,7 @@ sub _marc_iso5426_to_utf8 {
             my @converted_subfields;
             foreach my $subfield ($field->subfields()) {
                 my $utf8sf = char_decode5426($subfield->[1]);
+                           $utf8sf=NFC($utf8sf);
                 push @converted_subfields, $subfield->[0], $utf8sf;
             }
 
@@ -555,6 +601,7 @@ sub _marc_to_utf8_via_text_iconv {
                     push @converted_subfields, $subfield->[0], $converted_value;
                 } else {
                     $converted_value = $subfield->[1];
+                               $converted_value=NFC($converted_value);
                     $converted_value =~ s/[\200-\377]/\xef\xbf\xbd/g;
                     push @converted_subfields, $subfield->[0], $converted_value;
                 }
@@ -602,6 +649,7 @@ sub _marc_to_utf8_replacement_char {
             my @converted_subfields;
             foreach my $subfield ($field->subfields()) {
                 my $value = $subfield->[1];
+                               $value=NFC($value);
                 $value =~ s/[\200-\377]/\xef\xbf\xbd/g;
                 push @converted_subfields, $subfield->[0], $value;
             }