Revision history for Perl extension DBIx::Class::Schema::Loader
- support for DBD::Firebird
+ - support for unicode Firebird data types
0.07010 2011-03-04 08:26:31
- add result_component_map option
my $type_num = $colinfo->{data_type};
my $type_name;
if (defined $type_num && $type_num =~ /^-?\d+\z/ && $dbh->can('type_info')) {
- my $type_info = $self->_dbh_type_info($type_num);
- $type_name = $type_info->{TYPE_NAME} if $type_info;
+ my $type_name = $self->_dbh_type_info_type_name($type_num);
$colinfo->{data_type} = lc $type_name if $type_name;
}
}
}
# Need to override this for the buggy Firebird ODBC driver.
-sub _dbh_type_info {
+sub _dbh_type_info_type_name {
my ($self, $type_num) = @_;
my $dbh = $self->schema->storage->dbh;
- return $dbh->type_info($type_num);
+ my $type_info = $dbh->type_info($type_num);
+
+ return $type_info ? $type_info->{TYPE_NAME} : undef;
}
# do not use this, override _columns_info_for instead
# fix up types
$sth = $dbh->prepare(<<'EOF');
-SELECT f.rdb$field_precision, f.rdb$field_scale, f.rdb$field_type, f.rdb$field_sub_type, t.rdb$type_name, st.rdb$type_name
+SELECT f.rdb$field_precision, f.rdb$field_scale, f.rdb$field_type, f.rdb$field_sub_type, f.rdb$character_set_id, f.rdb$character_length, t.rdb$type_name, st.rdb$type_name
FROM rdb$fields f
JOIN rdb$relation_fields rf ON rf.rdb$field_source = f.rdb$field_name
LEFT JOIN rdb$types t ON f.rdb$field_type = t.rdb$type AND t.rdb$field_name = 'RDB$FIELD_TYPE'
AND rf.rdb$field_name = ?
EOF
$sth->execute($table, $self->_uc($column));
- my ($precision, $scale, $type_num, $sub_type_num, $type_name, $sub_type_name) = $sth->fetchrow_array;
+ my ($precision, $scale, $type_num, $sub_type_num, $char_set_id, $char_length, $type_name, $sub_type_name) = $sth->fetchrow_array;
$scale = -$scale if $scale && $scale < 0;
if ($type_name && $sub_type_name) {
$info->{data_type} = 'blob';
}
elsif ($sub_type_name eq 'TEXT') {
- $info->{data_type} = 'blob sub_type text';
+ if ($char_set_id == 3) {
+ $info->{data_type} = 'blob sub_type text character set unicode_fss';
+ }
+ else {
+ $info->{data_type} = 'blob sub_type text';
+ }
}
}
}
$info->{data_type} = 'bigint';
}
- # DBD::InterBase sets scale to '0' for some reason for char types
- if ($info->{data_type} =~ /^(?:char|varchar)\z/ && ref($info->{size}) eq 'ARRAY') {
- $info->{size} = $info->{size}[0];
+ if ($info->{data_type} =~ /^(?:char|varchar)\z/) {
+ $info->{size} = $char_length;
+
+ if ($char_set_id == 3) {
+ $info->{data_type} .= '(x) character set unicode_fss';
+ }
}
- elsif ($info->{data_type} !~ /^(?:char|varchar|numeric|decimal)\z/) {
+ elsif ($info->{data_type} !~ /^(?:numeric|decimal)\z/) {
delete $info->{size};
}
# Some (current) versions of the ODBC driver have a bug where ->type_info breaks
# with "data truncated". This "fixes" it, but some type names are truncated.
-sub _dbh_type_info {
+sub _dbh_type_info_type_name {
my ($self, $type_num) = @_;
my $dbh = $self->schema->storage->dbh;
local $dbh->{LongReadLen} = 100_000;
local $dbh->{LongTruncOk} = 1;
- return $dbh->type_info($type_num);
+ my $type_info = $dbh->type_info($type_num);
+
+ return undef if not $type_info;
+
+ my $type_name = $type_info->{TYPE_NAME};
+
+ # fix up truncated type names
+ if ($type_name eq "VARCHAR(x) CHARACTER SET UNICODE_\0") {
+ return 'VARCHAR(x) CHARACTER SET UNICODE_FSS';
+ }
+ elsif ($type_name eq "BLOB SUB_TYPE TEXT CHARACTER SET \0") {
+ return 'BLOB SUB_TYPE TEXT CHARACTER SET UNICODE_FSS';
+ }
+
+ return $type_name;
}
=head1 SEE ALSO
'char' => { data_type => 'char', size => 1 },
'char(11)' => { data_type => 'char', size => 11 },
'varchar(20)' => { data_type => 'varchar', size => 20 },
+ 'char(22) character set unicode_fss' =>
+ => { data_type => 'char(x) character set unicode_fss', size => 22 },
+ 'varchar(33) character set unicode_fss' =>
+ => { data_type => 'varchar(x) character set unicode_fss', size => 33 },
+
# Blob types
'blob' => { data_type => 'blob' },
'blob sub_type text'
=> { data_type => 'blob sub_type text' },
+ 'blob sub_type text character set unicode_fss'
+ => { data_type => 'blob sub_type text character set unicode_fss' },
},
extra => {
- count => 6,
+ count => 9,
run => sub {
$schema = shift;
my ($monikers, $classes, $self) = @_;
is $col_info->{sequence}, 'Gen_Firebird_Loader_Test1_Id', 'correct mixed case sequence name';
is eval { $rsrc->column_info('Foo')->{default_value} }, 42, 'default_value detected for mixed case column';
+
+ # test the fixed up ->_dbh_type_info_type_name for the ODBC driver
+ if ($schema->storage->_dbi_connect_info->[0] =~ /:ODBC:/i) {
+ my %truncated_types = (
+ 4 => 'INTEGER',
+ -9 => 'VARCHAR(x) CHARACTER SET UNICODE_FSS',
+ -10 => 'BLOB SUB_TYPE TEXT CHARACTER SET UNICODE_FSS',
+ );
+
+ for my $type_num (keys %truncated_types) {
+ is $schema->_loader->_dbh_type_info_type_name($type_num),
+ $truncated_types{$type_num},
+ "ODBC ->_dbh_type_info_type_name correct for '$truncated_types{$type_num}'";
+ }
+ }
+ else {
+ my $tb = Test::More->builder;
+ $tb->skip('not testing _dbh_type_info_type_name on DBD::InterBase') for 1..3;
+ }
},
},
);