[Stackless] announcing Concurrence Framework

Bob Ippolito bob at redivi.com
Fri Jan 16 20:55:42 CET 2009


On Fri, Jan 16, 2009 at 11:28 AM, Andreas Kostyrka <andreas at kostyrka.org> wrote:
> Am Fri, 16 Jan 2009 10:49:27 -0800
> schrieb "Bob Ippolito" <bob at redivi.com>:
>
>> pg8000 is probably a better base for building a PostgreSQL binding,
>> it's pure python. psycopg2 does wrap libpq's async functionality, but
>> it's ugly to use.
>
> pg8000 looks nice, although I'm slightly wary of Python-only
> reimplementations (I once had to use a Twisted-python-only PG
> interface, don't remember it's name, and it was awful buggy, that left
> me with an instinctive wariness when it comes to libpq
> reimplementations).

I've tried pg8000 and was able to use it to dump a few hundred gigs
out of an existing PostgreSQL DB with COPY with zero issues. I haven't
done much in the way of other SQL, and I contributed the (small) patch
to add COPY support, but it was on top of existing code that looked
good. In contrast, I have had very bad experiences with psycopg2's
COPY support in the past.

Just because one person failed to properly implement the PostgreSQL
protocol doesn't mean that all people have failed :) There are
actually maybe four or five different ones out there, but pg8000 is
the best of breed. I'm a lot more likely to trust a pure Python
implementation than something in C; if there's a problem it's a heck
of a lot easier to fix in Python.

-bob




More information about the Stackless mailing list