Re: psycopg2 vs psycopg1: the _v_database_connection problem in Zope 2

Il giorno mer, 24/09/2008 alle 11.33 +0200, Marco Bizzarri ha scritto:
> On Wed, Sep 24, 2008 at 10:14 AM, Federico Di Gregorio <fog-NGVKUo/i/[email protected]> wrote:
> 
> >> But I do not understand what you're meaning about the persistent
> >> attribute. Can you explain it?
> >
> > An attribute that does not start with _v_.
> >
> > federico
> 
> Ehm... I know what a persisten attribute is ;)
> 
> The question was more in how you wanted to use it.

Let suppose that when you ask the pool for a connection the pool keeps
the connection in a hash table indexed by a serial number and you get

	(serial, conn)

then assign

	self.conn_serial = serial
	self._v_connection = conn

Then, before trying to use the connection you do:

	if not hasattr(self, "_v_connection") or not self._v_connection:
		self._v_connection = pool.getconn(self.conn_serial)

and you get exactly the living connection you had before. This way even
if the object is frozen and lose its _v_ attributes you have a way to
get back a living connection. Then, after the commit/rollback you do:

	pool.putconn(serial)

to "put back" the connection and make it available to other threads from
the pool.

federico

-- 
Federico Di Gregorio                         http://people.initd.org/fog
Debian GNU/Linux Developer                                [email protected]
INIT.D Developer                                           fog-NGVKUo/i/[email protected]
 Questa lista non pretende di essere il riferimento torinese per il
  software libero o altre cazzate del genere.       -- Silvano, su torte

_______________________________________________
Psycopg mailing list
Psycopg-IAPFreCvJWPBWskQ1e/[email protected]
http://lists.initd.org/mailman/listinfo/psycopg

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEABECAAYFAkjaG4AACgkQvcCgrgZGjetTjwCfVdwriRv9dWM+Jk+nxIyXUNPI
Nc4AnRtwOtLiu6576p9RVFaJpRn+IkYc
=4kym
-----END PGP SIGNATURE-----