Re: Automatic transactions in psql

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

> I believe Fernando Nasser at Red Hat is currently working on backend
> changes to do this properly; so I recommend we not apply the psql hack

Sounds good to me. If he is not, I'll volunteer if someone will point
me in the right direction.

> The notion of customizing the psql prompt based on
> in-an-xact-block-or-not seems cool; but I do not see how to do it
> reliably without a protocol change, and it's not worth that.

<devil's advocate>
Surely it's being done at least as reliably as the hack in large_obj.c? :)
</da>
I don't think it's that unreliable myself: as long as the backend outputs
the standard confirmation message (e.g. "BEGIN"), we should always be
able to keep track. And a dropped connection means we switch to "not in a
transaction"

-----BEGIN PGP SIGNATURE-----
Comment: http://www.turnstep.com/pgp.html

iD8DBQE8dnuTvJuQZxSWSsgRAnnFAJ9vz2o1bYiDB1jfnBE2cWohDAw0IwCglMX5
LUgc2CcQAsdueSpFWH+mwxE=
=e0qZ
-----END PGP SIGNATURE-----