Skip Menu |

This queue is for tickets about the DBD-SQLite CPAN distribution.

Report information
The Basics
Id: 21406
Status: resolved
Priority: 0/
Queue: DBD-SQLite

People
Owner: Nobody in particular
Requestors: ATOURBIN [...] cpan.org
Cc:
AdminCc:

Bug Information
Severity: Important
Broken in:
  • 1.13
  • 1.14
Fixed in: (no value)



Subject: DBD-SQLite 1.13 broke Class-DBI

Message body is not shown because it is too large.

From: sabol [...] alderaan.gsfc.nasa.gov
On Sat Sep 09 13:35:28 2006, ATOURBIN wrote: Show quoted text
> After I upgraded from 1.12 to 1.13, I found some applications issuing > "statement handle DBIx::ContextualFetch::st=HASH(0x8e86ee4) still > Active at /usr/lib/perl5/vendor_perl/Ima/DBI.pm line 391" warnings. > Moreover, some test suites got broken, most notable that of Class- > DBI. After I rolled back to 1.12, things work fine again.
This continues to be a widespread problem. Matt, any fix (besides downgrading to DBD:: SQLite 1.12) in sight? Do you need a failing test case to resolve it?
From: ATOURBIN [...] cpan.org
On Mon Jan 22 13:13:55 2007, sabol@alderaan.gsfc.nasa.gov wrote: Show quoted text
> This continues to be a widespread problem. Matt, any fix (besides > downgrading to DBD:: > SQLite 1.12) in sight? Do you need a failing test case to resolve it?
Hello, that's me again. I managed to hack around the problem. I had to patch Class-DBI. Some details are available here: http://git.altlinux.org/people/at/packages/?p=perl-Class-DBI.git
On Mon Jan 22 13:28:45 2007, ATOURBIN wrote: Show quoted text
> On Mon Jan 22 13:13:55 2007, sabol@alderaan.gsfc.nasa.gov wrote:
> > This continues to be a widespread problem. Matt, any fix (besides > > downgrading to DBD:: > > SQLite 1.12) in sight? Do you need a failing test case to resolve it?
> > Hello, that's me again. I managed to hack around the problem. > I had to patch Class-DBI. Some details are available here: > http://git.altlinux.org/people/at/packages/?p=perl-Class-DBI.git
Looking at DBI's documentation, I think you have pointed to the solution that is needed. "If execute() is called on a statement handle that's still active ($sth->{Active} is true) then it should effectively call finish() to tidy up the previous execution results before starting this new execution." Obviously, it seems that the finish is not being called on the DBI/DBD side to make the statement handle inactive.
This appears to be resolved now.