4d4582566e
After Jonathan said this was the wrong way to correct the issue, www.utf8-chartable.de made it clear to me that the \c3\a9 were missing x's. TEST PLAN --------- 1) prove t/Charset.t -- noise 2) apply patch 3) prove t/Charset.t -- no noise 4) koha qa test tools Signed-off-by: Aleisha <aleishaamohia@hotmail.com> Signed-off-by: Jonathan Druart <jonathan.druart@koha-community.org> Signed-off-by: Tomas Cohen Arazi <tomascohen@gmail.com>
74 lines
2.7 KiB
Perl
Executable file
74 lines
2.7 KiB
Perl
Executable file
#!/usr/bin/perl
|
||
|
||
# This file is part of Koha.
|
||
#
|
||
# Koha is free software; you can redistribute it and/or modify it
|
||
# under the terms of the GNU General Public License as published by
|
||
# the Free Software Foundation; either version 3 of the License, or
|
||
# (at your option) any later version.
|
||
#
|
||
# Koha is distributed in the hope that it will be useful, but
|
||
# WITHOUT ANY WARRANTY; without even the implied warranty of
|
||
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
||
# GNU General Public License for more details.
|
||
#
|
||
# You should have received a copy of the GNU General Public License
|
||
# along with Koha; if not, see <http://www.gnu.org/licenses>.
|
||
|
||
use Modern::Perl;
|
||
|
||
use Test::More tests => 16;
|
||
use Encode qw( is_utf8 );
|
||
|
||
use MARC::Record;
|
||
|
||
use utf8;
|
||
use open ':std', ':encoding(utf8)';
|
||
|
||
BEGIN {
|
||
use_ok('C4::Charset');
|
||
}
|
||
|
||
my $string;
|
||
ok(!defined(NormalizeString($string,undef,1)),'Uninitialized string case 1 normalizes to uninitialized string.');
|
||
|
||
$string = 'Sample';
|
||
ok(defined(NormalizeString($string,undef,0)), 'Initialized string case 1 normalizes to some string.');
|
||
ok(defined(NormalizeString($string,undef,1)), 'Initialized string case 2 normalizes to some string.');
|
||
ok(defined(NormalizeString($string,1,0)), 'Initialized string case 3 normalizes to some string.');
|
||
ok(defined(NormalizeString($string,1,1)), 'Initialized string case 4 normalizes to some string.');
|
||
|
||
my $octets = "abc";
|
||
ok(IsStringUTF8ish($octets), "verify octets are valid UTF-8 (ASCII)");
|
||
|
||
$octets = "flamb\xc3\xa9";
|
||
ok(!Encode::is_utf8($octets), "verify that string does not have Perl UTF-8 flag on");
|
||
ok(IsStringUTF8ish($octets), "verify octets are valid UTF-8 (LATIN SMALL LETTER E WITH ACUTE)");
|
||
ok(!Encode::is_utf8($octets), "verify that IsStringUTF8ish does not magically turn Perl UTF-8 flag on");
|
||
|
||
$octets = "a\xc2" . "c";
|
||
ok(!IsStringUTF8ish($octets), "verify octets are not valid UTF-8");
|
||
|
||
ok( !SetUTF8Flag(), 'SetUTF8Flag returns undef if no record passed' );
|
||
|
||
my $record = MARC::Record->new();
|
||
ok( !SetUTF8Flag($record), 'SetUTF8Flag returns undef if the record has no subfields' );
|
||
# Add some fields/subfields
|
||
$record->append_fields(
|
||
MARC::Field->new('100', ' ', ' ', a => 'Julio Cortazar'),
|
||
MARC::Field->new('245', ' ', ' ', a => 'Rayuela'),
|
||
);
|
||
# Verify our data serves its purpose
|
||
ok( !Encode::is_utf8($record->subfield('100','a')) &&
|
||
!Encode::is_utf8($record->subfield('245','a')),
|
||
'Verify that the subfields are NOT set the UTF-8 flag yet' );
|
||
|
||
SetUTF8Flag($record);
|
||
|
||
ok( Encode::is_utf8($record->subfield('100','a')) &&
|
||
Encode::is_utf8($record->subfield('245','a')),
|
||
'SetUTF8Flag sets the UTF-8 flag to all subfields' );
|
||
|
||
is( nsb_clean("Le Moyen Âge"), "Le Moyen Âge", "nsb_clean removes and " );
|
||
|
||
1;
|