Commit | Line | Data |
5dd9c59c |
1 | package DBIx::Class::UTF8Columns; |
2 | use strict; |
3 | use warnings; |
4 | use base qw/DBIx::Class/; |
5dd9c59c |
5 | |
4e8964d5 |
6 | __PACKAGE__->mk_classdata( '_utf8_columns' ); |
5dd9c59c |
7 | |
8 | =head1 NAME |
9 | |
3c2a505c |
10 | DBIx::Class::UTF8Columns - Force UTF8 (Unicode) flag on columns (DEPRECATED) |
7c14c3cf |
11 | |
5dd9c59c |
12 | =head1 SYNOPSIS |
13 | |
14 | package Artist; |
d88ecca6 |
15 | use base 'DBIx::Class::Core'; |
16 | |
17 | __PACKAGE__->load_components(qw/UTF8Columns/); |
5dd9c59c |
18 | __PACKAGE__->utf8_columns(qw/name description/); |
d4daee7b |
19 | |
5dd9c59c |
20 | # then belows return strings with utf8 flag |
21 | $artist->name; |
22 | $artist->get_column('description'); |
23 | |
24 | =head1 DESCRIPTION |
25 | |
7c14c3cf |
26 | This module allows you to get and store utf8 (unicode) column data |
27 | in a database that does not natively support unicode. It ensures |
28 | that column data is correctly serialised as a byte stream when |
29 | stored and de-serialised to unicode strings on retrieval. |
30 | |
3c2a505c |
31 | THE USE OF THIS MODULE (AND ITS COUSIN DBIx::Class::ForceUTF8) IS VERY |
32 | STRONGLY DISCOURAGED, PLEASE READ THE WARNINGS BELOW FOR AN EXPLANATION. |
33 | |
aa9bcbfb |
34 | If you want to continue using this module and do not want to receive |
4d1e63f4 |
35 | further warnings set the environment variable C<DBIC_UTF8COLUMNS_OK> |
3c2a505c |
36 | to a true value. |
37 | |
38 | =head2 Warning - Module does not function properly on create/insert |
39 | |
40 | Recently (April 2010) a bug was found deep in the core of L<DBIx::Class> |
41 | which affects any component attempting to perform encoding/decoding by |
42 | overloading L<store_column|DBIx::Class::Row/store_column> and |
43 | L<get_columns|DBIx::Class::Row/get_columns>. As a result of this problem |
44 | L<create|DBIx::Class::ResultSet/create> sends the original column values |
45 | to the database, while L<update|DBIx::Class::ResultSet/update> sends the |
46 | encoded values. L<DBIx::Class::UTF8Columns> and L<DBIx::Class::ForceUTF8> |
4a0eed52 |
47 | are both affected by this bug. |
3c2a505c |
48 | |
49 | It is unclear how this bug went undetected for so long (it was |
50 | introduced in March 2006), No attempts to fix it will be made while the |
51 | implications of changing such a fundamental behavior of DBIx::Class are |
52 | being evaluated. However in this day and age you should not be using |
53 | this module anyway as Unicode is properly supported by all major |
54 | database engines, as explained below. |
55 | |
56 | If you have specific questions about the integrity of your data in light |
8273e845 |
57 | of this development - please |
3c2a505c |
58 | L<join us on IRC or the mailing list|DBIx::Class/GETTING HELP/SUPPORT> |
59 | to further discuss your concerns with the team. |
60 | |
7c14c3cf |
61 | =head2 Warning - Native Database Unicode Support |
62 | |
63 | If your database natively supports Unicode (as does SQLite with the |
64 | C<sqlite_unicode> connect flag, MySQL with C<mysql_enable_utf8> |
65 | connect flag or Postgres with the C<pg_enable_utf8> connect flag), |
66 | then this component should B<not> be used, and will corrupt unicode |
67 | data in a subtle and unexpected manner. |
68 | |
69 | It is far better to do Unicode support within the database if |
3c2a505c |
70 | possible rather than converting data to and from raw bytes on every |
71 | database round trip. |
5dd9c59c |
72 | |
7c14c3cf |
73 | =head2 Warning - Component Overloading |
d38cd95c |
74 | |
75 | Note that this module overloads L<DBIx::Class::Row/store_column> in a way |
76 | that may prevent other components overloading the same method from working |
77 | correctly. This component must be the last one before L<DBIx::Class::Row> |
78 | (which is provided by L<DBIx::Class::Core>). DBIx::Class will detect such |
79 | incorrect component order and issue an appropriate warning, advising which |
80 | components need to be loaded differently. |
81 | |
5dd9c59c |
82 | =head1 SEE ALSO |
83 | |
84 | L<Template::Stash::ForceUTF8>, L<DBIx::Class::UUIDColumns>. |
85 | |
86 | =head1 METHODS |
87 | |
88 | =head2 utf8_columns |
89 | |
90 | =cut |
91 | |
92 | sub utf8_columns { |
93 | my $self = shift; |
4e8964d5 |
94 | if (@_) { |
95 | foreach my $col (@_) { |
96 | $self->throw_exception("column $col doesn't exist") |
97 | unless $self->has_column($col); |
d38cd95c |
98 | } |
4e8964d5 |
99 | return $self->_utf8_columns({ map { $_ => 1 } @_ }); |
100 | } else { |
101 | return $self->_utf8_columns; |
5dd9c59c |
102 | } |
5dd9c59c |
103 | } |
104 | |
105 | =head1 EXTENDED METHODS |
106 | |
107 | =head2 get_column |
108 | |
109 | =cut |
110 | |
111 | sub get_column { |
112 | my ( $self, $column ) = @_; |
113 | my $value = $self->next::method($column); |
114 | |
d38cd95c |
115 | utf8::decode($value) if ( |
116 | defined $value and $self->_is_utf8_column($column) and ! utf8::is_utf8($value) |
117 | ); |
5dd9c59c |
118 | |
55087b99 |
119 | return $value; |
5dd9c59c |
120 | } |
121 | |
e063fe2c |
122 | =head2 get_columns |
123 | |
124 | =cut |
125 | |
126 | sub get_columns { |
127 | my $self = shift; |
128 | my %data = $self->next::method(@_); |
129 | |
d38cd95c |
130 | foreach my $col (keys %data) { |
131 | utf8::decode($data{$col}) if ( |
132 | exists $data{$col} and defined $data{$col} and $self->_is_utf8_column($col) and ! utf8::is_utf8($data{$col}) |
133 | ); |
e063fe2c |
134 | } |
135 | |
55087b99 |
136 | return %data; |
e063fe2c |
137 | } |
138 | |
5dd9c59c |
139 | =head2 store_column |
140 | |
141 | =cut |
142 | |
143 | sub store_column { |
144 | my ( $self, $column, $value ) = @_; |
145 | |
4a0eed52 |
146 | # the dirtiness comparison must happen on the non-encoded value |
d38cd95c |
147 | my $copy; |
148 | |
149 | if ( defined $value and $self->_is_utf8_column($column) and utf8::is_utf8($value) ) { |
150 | $copy = $value; |
151 | utf8::encode($value); |
5dd9c59c |
152 | } |
153 | |
154 | $self->next::method( $column, $value ); |
d38cd95c |
155 | |
156 | return $copy || $value; |
5dd9c59c |
157 | } |
158 | |
d38cd95c |
159 | # override this if you want to force everything to be encoded/decoded |
160 | sub _is_utf8_column { |
2ba92e45 |
161 | # my ($self, $col) = @_; |
162 | return ($_[0]->utf8_columns || {})->{$_[1]}; |
d38cd95c |
163 | } |
5dd9c59c |
164 | |
d38cd95c |
165 | =head1 AUTHORS |
5dd9c59c |
166 | |
d38cd95c |
167 | See L<DBIx::Class/CONTRIBUTORS>. |
5dd9c59c |
168 | |
d38cd95c |
169 | =head1 LICENSE |
5dd9c59c |
170 | |
d38cd95c |
171 | You may distribute this code under the same terms as Perl itself. |
5dd9c59c |
172 | |
173 | =cut |
174 | |
175 | 1; |