You've not given me much to go on, but I'd guess it's related to the timing of when perl invokes the DESTROY method (which has changed between perl versions). In which case it may be mostly beyond the control of the DBI.
A small self-contained example that behaves differently between different DBI/Perl versions will buy you more of my time :) Tim. On Fri, May 02, 2008 at 01:23:33PM -0700, David E. Wheeler wrote: > Howdy, > > I started getting test failures for Exception::Class::DBI on Perl 5.6 > recently. The failures seem to be due to this test: > > # For some reason, under perl < 5.8.0, $dbh->{Kids} returns a different > value > # inside the HandleError scope than it does outside that scope. So we're > # checking for the perl version here to cover our butts on this test. This > may > # be fixed in the DBI soon. > is( $err->kids, ($] < 5.008 ? 1 : 0), "Check kids" ); > > Was this fixed in the DBI? If so, in what version (I saw a change to > CachedKids in 1.55 that looked close, but that's not the same as {kids}, is > it?)? > > Thanks, > > david