Thread: Use lo_creat() when possible instead of lo_create()
Hi. I just created a ticket against psycopg2, and Daniele asked me to bring the issue on the list to open it to a wider audience. The ticket is here: http://psycopg.lighthouseapp.com/projects/62710/tickets/88-use-lo_creat-when-possible-instead-of-lo_create For your convenience, I copy the description here: --- I'm using psycopg2 against a PostgreSQL database replicated using pgpool-II. On this database I also need large objects support. Apparently pgpool-II is able to correctly handle lo_creat() calls, but not lo_create() (see http://pgpool.projects.postgresql.org/pgpool-II/doc/pgpool-en.html#lobj_lock_table). Unfortunately, psycopg2 only create large objects with lo_create(), even when no new_oid is specified. Could you please consider applying the attached patch, that modifies this behaviour? --- Please, Cc: me as I'm not subscribed to the list. Thanks, Giovanni. -- Giovanni Mascellani <mascellani@poisson.phc.unipi.it> Pisa, Italy Web: http://poisson.phc.unipi.it/~mascellani Jabber: g.mascellani@jabber.org / giovanni@elabor.homelinux.org
Attachment
On 10/01/12 09:51, Giovanni Mascellani wrote: > I just created a ticket against psycopg2, and Daniele asked me to bring > the issue on the list to open it to a wider audience. The ticket is here: > > http://psycopg.lighthouseapp.com/projects/62710/tickets/88-use-lo_creat-when-possible-instead-of-lo_create The patch seems fine to me. Doesn't add any overhead and improves compatibility with an existing product. I vote +1 [that means "unless Daniele votes -1 I'll include it" ;)] federico
On Tue, Jan 10, 2012 at 10:06 AM, Federico Di Gregorio <fog@dndg.it> wrote: > On 10/01/12 09:51, Giovanni Mascellani wrote: >> >> I just created a ticket against psycopg2, and Daniele asked me to bring >> the issue on the list to open it to a wider audience. The ticket is here: >> >> >> http://psycopg.lighthouseapp.com/projects/62710/tickets/88-use-lo_creat-when-possible-instead-of-lo_create > > > The patch seems fine to me. Doesn't add any overhead and improves > compatibility with an existing product. I vote +1 [that means "unless > Daniele votes -1 I'll include it" ;)] I don't know much about the difference between the two functions: if anybody has any interesting story about them I'd be happy to know. Meanwhile I'll make sure we have tests for both the code paths in the patch. -- Daniele
On Tue, Jan 10, 2012 at 12:18 PM, Daniele Varrazzo <daniele.varrazzo@gmail.com> wrote: > On Tue, Jan 10, 2012 at 10:06 AM, Federico Di Gregorio <fog@dndg.it> wrote: >> On 10/01/12 09:51, Giovanni Mascellani wrote: >>> >>> I just created a ticket against psycopg2, and Daniele asked me to bring >>> the issue on the list to open it to a wider audience. The ticket is here: >>> >>> >>> http://psycopg.lighthouseapp.com/projects/62710/tickets/88-use-lo_creat-when-possible-instead-of-lo_create >> >> >> The patch seems fine to me. Doesn't add any overhead and improves >> compatibility with an existing product. I vote +1 [that means "unless >> Daniele votes -1 I'll include it" ;)] > > I don't know much about the difference between the two functions: if > anybody has any interesting story about them I'd be happy to know. > Meanwhile I'll make sure we have tests for both the code paths in the > patch. I've made some tests and everything seems fine. I've integrated the patch in my branch. -- Daniele