Re: [svn:dbi] r14361 - dbi/trunk/t

2010/8/31 Tim Bunce <[email protected]>:
> On Mon, Aug 30, 2010 at 01:49:10PM -0700, [email protected] wrote:
>> Log:
>> Fix for failing t/50 when old SQL::Statement available (e.g. 1.16)
>
> Does this fix http://www.cpantesters.org/cpan/report/8392757 ?

No, because this fixes t/50dbm_simple.t and the error is in t/51dbm_file.t

>>  my $dbi_sql_nano = $ENV{DBI_SQL_NANO};
>>  unless( $dbi_sql_nano ) {
>> -    $@ = undef;
>> +    my $haveSS = 0;
>>      eval {
>>       require SQL::Statement;
>> +     $haveSS = DBD::DBM::Statement->isa('SQL::Statement');
>>      };
>> -    $@ and $dbi_sql_nano = 1;
>> +    $dbi_sql_nano = !$haveSS;
>>  }
>
> Could this:
>
>>  my $dbi_sql_nano = $ENV{DBI_SQL_NANO};
>>  unless( $dbi_sql_nano ) {
>>      my $haveSS = 0;
>>      eval {
>>        require SQL::Statement;
>>        $haveSS = DBD::DBM::Statement->isa('SQL::Statement');
>>      };
>>      $dbi_sql_nano = !$haveSS;
>>  }
>
> be expressed as:
>
>   my $dbi_sql_nano = $ENV{DBI_SQL_NANO};
>   unless( $dbi_sql_nano ) {
>       $dbi_sql_nano = not eval {
>         require SQL::Statement;
>         DBD::DBM::Statement->isa('SQL::Statement');
>       };
>   }

I made a quick fix and send it to Tux for testing - and he committed instantly.
I'll handle that together with the win32 issue
(http://www.cpantesters.org/cpan/report/8392757) before Sunday.

> or even
>
>   # use Nano if requested or if SQL::Statement isn't usable for DBD::DBM
>   my $dbi_sql_nano = $ENV{DBI_SQL_NANO} || not eval {
>        require SQL::Statement; DBD::DBM::Statement->isa('SQL::Statement')
>    };

If we shorten that all, we don't need to require SQL::Statement. This
will be done
in DBI::SQL::Nano. Following would be enough:

# use Nano if requested or if SQL::Statement isn't usable for DBD::DBM
my $dbi_sql_nano = not eval {
     DBD::DBM::Statement->isa('SQL::Statement');
};

Or, if we can assume that DBD::DBM can always successfully loaded:

my $dbi_sql_nano = not DBD::DBM::Statement->isa('SQL::Statement');

/Jens