-----BEGIN PGP SIGNED MESSAGE----- Hash: RIPEMD160
I noticed something odd with NAME and NAME_lc / NAME_uc: if they are called, and then called *again* after $sth->finish() is called, both NAME_lc and NAME_uc return the previous values instead of an undef. I'm not sure if this is a DBI or DBD::Pg issue, but I thought I'd see if people can perhaps duplicate this on other DBDs: $sth = $dbh->prepare(q{SELECT 123 AS "Goldfish"}); $sth->execute(); print Dumper $sth->{NAME}; print Dumper $sth->{NAME_lc}; $sth->finish(); print Dumper $sth->{NAME}; print Dumper $sth->{NAME_lc}; print Dumper $sth->{NAME_uc}; For DBD::Pg, it prints: $VAR1 = [ 'Goldfish' ]; $VAR1 = [ 'goldfish' ]; $VAR1 = undef; $VAR1 = [ 'goldfish' ]; $VAR1 = undef; I also got a CPAN tester report today that is almost certainly related. It can be seen here: http://www.nntp.perl.org/group/perl.cpan.testers/2008/08/msg1997411.html The relevant info is: Assertion i == (((((SV *) (name_av))->sv_flags & 0x00800000)) ? Perl_mg_size(my_perl, (SV *) name_av) : ((XPVAV*) (name_av)->sv_any)->xav_fill)+1 failed: file "DBI.xs", line 2001 at /usr/lib/perl5/5.10.0/Test/More.pm line 821. # Looks like you planned 250 tests but only ran 95. Test 96 is this on line 623 of 02attribs.t: $t='Statement handle attribute "NAME_uc" returns empty arrayref for updates'; is_deeply ($sth->{'NAME_uc'}, [], $t); The file is here to see it in context: http://search.cpan.org/src/TURNSTEP/DBD-Pg-2.9.0/t/02attribs.t - -- Greg Sabino Mullane [EMAIL PROTECTED] End Point Corporation PGP Key: 0x14964AC8 200808071224 http://biglumber.com/x/web?pk=2529DF6AB8F79407E94445B4BC9B906714964AC8 -----BEGIN PGP SIGNATURE----- iEYEAREDAAYFAkibIqkACgkQvJuQZxSWSsgNsACg3VoEuo0AffJo8hHOC5xizqA1 H10AoIAi7UzE++v+7/8peJfzfgdTdJ82 =CoBV -----END PGP SIGNATURE-----