Re: DBI->disconnect is not a DBI method. Read the DBI manual. at ff.pl line 224

On 5/23/05, Hemanth Kumar <HKumar@xxxxxxxxxxx> wrote:
> What happens when you replace $dbh->disconnect(); with $dbh->disconnect;
>
> hth
> Hemanth
>

Son-of-a-monkey's-uncle, that worked!

Okay, now I'm lost. Does anyone want to explain this behavior? The
best that I can come up with is that:
1) there is some deep magic in DBI
2) the () changes the signature of the disconnect call in such a way
that it does not match the prototype given in DBI and that since
MDISS::DB is a subclass of DBI and not a DBI, we've confused perl in a
bizzare way.

>
> -----Original Message-----
> From: Matthew Persico [mailto:matthew.persico@xxxxxxxxx]
> Sent: Monday, May 23, 2005 10:27 AM
> To: dbi-users@xxxxxxxx
> Subject: Re: DBI->disconnect is not a DBI method. Read the DBI manual.
> at ff.pl line 224
>
> On 5/20/05, Tim Bunce <Tim.Bunce@xxxxxxxxx> wrote:
> > On Fri, May 20, 2005 at 10:08:36AM -0400, Matthew Persico wrote:
> > > Has anyone ever seen this message? I have a valid handle - all of my
>
> > > database commands work. The command is simply
> > >
> > > $dbh->disconnect();
> >
> > I'd guess that $dbh contains the string "DBI".
> >
> > Tim.
> >
>
> That was a rather sparse message I posted Friday. A few more details are
> in order. If anyone has any ideas after reading this, thanks in advance.
>
> Perl 5.6.1
> DBI 1.37
> DBD::Sybase 1.00
>
> Yes, I am stuck with these versions (esp. Perl) for the foreseable
> future.
>
> The connection is created in a module I have inherited (consultant long
> gone...) called MDISS::DB that contains, among other things, this
> code:
>
> BEGIN { # setup as a subclass of DBI
> @MDISS2::DB::EXPORT = (@DBI::EXPORT);
> @MDISS2::DB::EXPORT_OK = (@DBI::EXPORT_OK);
> @MDISS2::DB::EXPORT_TAGS = (@DBI::EXPORT_TAGS);
> @MDISS2::DB::ISA = qw(DBI DBI::db Exporter);
> @MDISS2::DB::db::ISA = qw(DBI::db MDISS2::DB Exporter);
> @MDISS2::DB::st::ISA = qw(DBI::st MDISS2::DB Exporter);
> DBI::init_rootclass('MDISS2::DB');
> }
>
> and this code:
>
> sub new {
>
> my %args = ();
> my $proto = undef;
>
> if (scalar(@_)) {
> ($proto, %args) = @_;
> }
>
> my %clone = ();
> my $class = undef;
> my $self = undef;
> my $private = undef;
> my %login = ();
> my $n = undef;
>
> $n = 'MDISS2::DB->new()';
>
> =item legacy code
>
> my($proto, %args, %clone, $class, $self, $private, %login, $n) = @_;
>
> =cut
>
> %clone = ();
> if(ref($proto)) {
> if($proto->can('_private')) {
> $private = $proto->_private();
> if(exists($private->{'login'})) {
> %clone = %{$private->{'login'}};
> if(exists($clone{'driver'}) &&
> exists($args{'driver'})) {
> croak("$n - cannot clone
> '$args{'driver'}' from '$clone{'driver'}'")
>
> unless(uc($args{'driver'}) eq uc($clone{'driver'}));
> }
> }
> }
> } #endif
>
>
> $args{'driver'} = ucfirst(lc($args{'driver'})) || 'Sybase';
> if($args{'driver'} eq 'Oracle') {
> %login = mdiss_oracle_login(%clone, %args);
> } else {
> %login = mdiss_sybase_login(%args);
> }
> # print Dumper(%login);
>
> ### This connectDBI is a wrapper that mashals args and makes the
> ### typical DBI->connect call.
> $self = connectDBI(%login) || croak("$n - failed to login to
> database: $DBI::errstr, aborting");
> # $class = ref($proto) || $proto;
> $class = 'MDISS2::DB';
> bless($self, $class);
>
> $private = $self->_private();
> $private->{'login'} = {%login};
> $private->{'cachedQueries'} = [];
> $private->{'cachedStatements'} = {};
> $private->{'cachedColumns'} = {};
> $private->{'cachedColumnTypes'} = {};
> $private->{'portiaDb'} = $ENV{'AM_PORTIA_DB_APO'};
> if($DEBUG > 1) {
> my $pri = $self->_private();
> print STDERR "$n - DBH Private Data:\n";, Dumper($pri);
> }
>
> $ALL_DBH{"$self"} = $self;
> return $self;
> } # end new()
>
> All of the database calls we are making using handles (such as do)
> created by MDISS::DB (do) and calls made by sths created from these
> handles all work.
>
> The only one that doesn't work is the disconnect. When I debug, I get to
> line 654
>
> sub disconnect { # a regular beginners bug
> Carp::croak("DBI->disconnect is not a DBI method. Read the DBI
> manual."); }
>
> Now, I've read the DBI manual and I seem to have missed the discussion
> of the regular beginners bug.
>
> Here is some DBI_TRACE=1 ouput. The first section is the conenction and
> the first prepare statement
>
> ->
> DBI->connect(dbi:Sybase:server=A2S1_402;appname=MDISS2::DB;hostname=nydu
> x-ast402,
> AMGDEV, ****)
> -> DBI->install_driver(Sybase) for solaris perl=5.006001 pid=27583
> ruid=8030 euid=8030
> install_driver: DBD::Sybase version 1.00 loaded from
> /opt/perl/lib/site_perl/5.6.1/sun4-solaris/DBD/Sybase.pm
> <- install_driver= DBI::dr=HASH(0x52c8e8)
> <-
> connect('server=A2S1_402;appname=MDISS2::DB;hostname=nydux-ast402'
> 'AMGDEV' ...)= DBI::db=HASH(0x1243d0) at DBI.pm line 582
> <- STORE('PrintError' 1)= 1 at DBI.pm line 622
> <- STORE('AutoCommit' 1)= 1 at DBI.pm line 622
> <- STORE('Username' 'AMGDEV')= 1 at DBI.pm line 625
> <- connect= DBI::db=HASH(0x1243d0)
> <- STORE('dbi_connect_closure' CODE(0xa0472c))= 1 at DBI.pm line 639
> <- err= undef at DBI.pm line 936
> <- STORE('RaiseError' 1)= 1 at DBI.pm line 950
> <- STORE('PrintError' 0)= 1 at DBI.pm line 951
> <- STORE('AutoCommit' 0)= 1 at DBI.pm line 952
> <- STORE('syb_show_sql' 1)= 1 at DBI.pm line 954
> <- STORE('syb_show_eed' 1)= 1 at DBI.pm line 955
> 1 <- prepare('use mdissdb' undef)= DBI::st=HASH(0xa09364) at
> Sybase.pm line 150
> <- execute= -1 at Sybase.pm line 151
> <- err= undef at Sybase.pm line 152
> <- rows= -1 at Sybase.pm line 153
> <- FETCH('syb_more_results')= undef at Sybase.pm line 154
> <- do('use mdissdb')= -1 at DBI.pm line 971
> <- DESTROY= undef at DBI.pm line 973
> <- err= undef at DBI.pm line 973
> <- STORE('private_AMG_DBI_data' HASH(0xa120c4))= 1 at DBI.pm line
> 988
> 1 <- FETCH('private_mdiss')= undef at DBI.pm line 1168
> <- EXISTS('private_mdiss')= '' at DB.pm line 196
> <- STORE('private_mdiss' HASH(0x10f220))= 1 at DB.pm line 196
> <- FETCH('private_mdiss')= HASH(0x10f220)0keys ('private_mdiss'
> from cache) at DB.pm line 199
> 1 <- FETCH('private_mdiss')= HASH(0x10f220)6keys ('private_mdiss'
> from cache) at DBI.pm line 1168
> <- EXISTS('private_mdiss')= 1 at DB.pm line 196
> <- FETCH('private_mdiss')= HASH(0x10f220)6keys ('private_mdiss'
> from cache) at DB.pm line 196
> <- FETCH('private_mdiss')= HASH(0x10f220)6keys ('private_mdiss'
> from cache) at DB.pm line 199
> <- prepare('
> select am_md_rule_id
> from am_md_rule
> where name = ?
> ' undef)= DBI::st=HASH(0xa12088) at DB.pm line
> 282
>
> I'l skip the rest of the trace - propriatary stuff that just works. At
> the end:
>
> code:
>
> print "dbh Active is $dbh->{Active}
> Kids is $dbh->{Kids}
> ActiveKids is $dbh->{ActiveKids}
> just before disconnect
> ";
> $dbh->disconnect();
>
> trace:
> dbh Active is 1
> Kids is 1
> ActiveKids is 0
> just before disconnect
> DBI->disconnect is not a DBI method. Read the DBI manual. at ff.pl line
> DBI->223
> <- DESTROY= undef
> <- DESTROY= undef
> <- DESTROY= undef
>
>
>
>
> --
> Matthew O. Persico
>


--
Matthew O. Persico
.

Relevant Pages

  • Re: DBI->disconnect is not a DBI method. Read the DBI manual. at ff.pl line 224
    ... > 1) there is some deep magic in DBI ... >> The only one that doesn't work is the disconnect. ... >> from cache) at DBI.pm line 1168 ... >> Matthew O. Persico ...
    (perl.dbi.users)
  • Re: mod_perl2 DBI handle freshining problem solved "once and for all"...
    ... > able to verify that Apache::DBI wasn't loaded yet at that point, and I was issuing a disconnectbefore the fork> took place. ... It has to load before DBI in order to work. ... disconnect() would fail because the handle> was "already disconnected", and the handle did *not* disappear out of> CachedKids. ... At this point, I think we don't know what the problem was with your database connections, and I'm not convinced it was a result of an Apache::DBI bug. ...
    (perl.dbi.users)
  • Re: mod_perl2 DBI handle freshining problem solved "once and for all"...
    ... The point here was that I *wanted* the disconnect() to take place. ... >> server that disconnects *every* DBI handle active before apache forks? ... that means you essentially have this TCP connection to the ...
    (perl.dbi.users)