byterock

ts fix the other bug day here in the Moose-Pen.

The second of the bugs that I am after squishing today was in Driver::DBI and in test case '15_alias.t'
It was dropping the elements/fields from the update command like this

Expected->UPDATE people SET first_name = ?

Generated -> UPDATE people SET 

Looking at the code in Driver::DBI the line in question is the '_update' sub;

    my (@field_sql) =

        $self->_insert_update_container( Database::Accessor::Constants::UPDATE,

        $container );

The first thing I checked was what was in the '$container' I was playing with. A line of two or waring and a run later it that turned out it was

container=$VAR1 = {};

an empty hash. Now that is a little concerning as you should not be able to do an update with an empty container. I moved the debugging back in the code to where the '_update' sub is called in the 'execute' sub

sub execute {

    my $self = shift;

    my ( $result, $action, $dbh, $container, $opt ) = @_;

    warn("execute container=".Dumper($container));

and it was empty there as well. I guess it is into the Database::Accessor and see what the value is before create the DAD. Now I am not going to boar you with the six or so warings I inserted into Database::Accessor but I did find my $container was going awry with this element;

Database::Accessor::Element= bless( {

                 'view' => 'sys_users',

                 'name' => 'first_name',

                 'alias' => 'First Name'

               }, 'Database::Accessor::Element' );

on this line

    next

          if ( ($field->view) and ($field->view ne $self->view()->name()));

so the view 'alias' view was buggering me up. The fix was simple enough;

    next

         if (($field->view) 

              and ($field->view ne $self->view()->name() 

                      and $field->view ne $self->view()->alias()));

Though it did take me some 45 mins to get the logic just right forgot that I should only 'next' when neither the view name and view alias match not when just one matches;

After a few mins of removing way to many lines of warings I finally get a full pass;

1..18

ok 1 - Basic table alias create SQL correct

...

ok 18 - Field alias with spaces delete SQL correct

Onto other things; even though for just a postette it still took me over an hour to debug.