view release on metacpan or  search on metacpan

lib/DBI/PurePerl.pm  view on Meta::CPAN

    my $h = shift;
    $h->{'_fbav'} = undef;
    $h->{'Active'} = 0;
    return 1;
}
sub rows {
    my $h = shift;
    my $rows = $h->{'_rows'};
    return -1 unless defined $rows;
    return $rows;
}

1;
__END__

=pod

=head1 NAME

DBI::PurePerl -- a DBI emulation using pure perl (no C/XS compilation required)

=head1 SYNOPSIS

 BEGIN { $ENV{DBI_PUREPERL} = 2 }
 use DBI;

=head1 DESCRIPTION

This is a pure perl emulation of the DBI internals.  In almost all
cases you will be better off using standard DBI since the portions
of the standard version written in C make it *much* faster.

However, if you are in a situation where it isn't possible to install
a compiled version of standard DBI, and you're using pure-perl DBD
drivers, then this module allows you to use most common features
of DBI without needing any changes in your scripts.

=head1 EXPERIMENTAL STATUS

DBI::PurePerl is new so please treat it as experimental pending
more extensive testing.  So far it has passed all tests with DBD::CSV,
DBD::AnyData, DBD::XBase, DBD::Sprite, DBD::mysqlPP.  Please send
bug reports to Jeff Zucker at <jeff@vpservices.com> with a cc to
<dbi-dev@perl.org>.

=head1 USAGE

The usage is the same as for standard DBI with the exception
that you need to set the environment variable DBI_PUREPERL if
you want to use the PurePerl version.

 DBI_PUREPERL == 0 (the default) Always use compiled DBI, die
                   if it isn't properly compiled & installed

 DBI_PUREPERL == 1 Use compiled DBI if it is properly compiled
                   & installed, otherwise use PurePerl

 DBI_PUREPERL == 2 Always use PurePerl

You may set the environment variable in your shell (e.g. with
set or setenv or export, etc) or else set it in your script like
this:

 BEGIN { $ENV{DBI_PUREPERL}=2 }

before you C<use DBI;>.

=head1 INSTALLATION

In most situations simply install DBI (see the DBI pod for details).

In the situation in which you can not install DBI itself, you
may manually copy DBI.pm and PurePerl.pm into the appropriate
directories.

For example:

 cp DBI.pm      /usr/jdoe/mylibs/.
 cp PurePerl.pm /usr/jdoe/mylibs/DBI/.

Then add this to the top of scripts:

 BEGIN {
   $ENV{DBI_PUREPERL} = 1;	# or =2
   unshift @INC, '/usr/jdoe/mylibs';
 }

(Or should we perhaps patch Makefile.PL so that if DBI_PUREPERL
is set to 2 prior to make, the normal compile process is skipped
and the files are installed automatically?)

=head1 DIFFERENCES BETWEEN DBI AND DBI::PurePerl

=head2 Attributes

Boolean attributes still return boolean values but the actual values
used may be different, i.e., 0 or undef instead of an empty string.

Some handle attributes are either not supported or have very limited
functionality:

  ActiveKids
  InactiveDestroy
  AutoInactiveDestroy
  Kids
  Taint
  TaintIn
  TaintOut

(and probably others)

=head2 Tracing

Trace functionality is more limited and the code to handle tracing is
only embedded into DBI:PurePerl if the DBI_TRACE environment variable
is defined.  To enable total tracing you can set the DBI_TRACE
environment variable as usual.  But to enable individual handle
tracing using the trace() method you also need to set the DBI_TRACE
environment variable, but set it to 0.

=head2 Parameter Usage Checking