On Sat, Feb 04, 2006 at 12:57:02AM -0500, Colin F. wrote:
} > From my experience… it’s worth the few extra steps to get the C
} > bindings working.
} Robby, why do you say that?
} Blair and Amr seem to think that the -pr versions are fine… but, in
} slightly misdirected posting to the Ruby mailing list some time ago, I
} got a response very similar to yours.
} But, that’s why I asked the question of which is better… some
} examples, knowledge of what causes differences in behavior, and maybe
} some pointers to locations in the code that cause that behavior would
} a long way towards helping us all figure out what the real differences
} here are, and how they’d affect systems under development.
} Just to throw my $0.02 in, I’m using the postgres-pr version, with
} postgres 8, and it’s working fine… if there are compelling reasons
} switch though, which seem to be darkly hinted at, I’d rather do it now
} than later.
So it seems to me that it very much depends on your development platform
and your deployment platform. For me, both are Debian and it’s dead easy
set up the C bindings (apt-get install libdbd-pg-ruby); it doesn’t even
involve a gem install. For platforms with less complete package
systems (cough Windows cough), it’s easier to use gem for package
management, and the -pr version doesn’t require any work other than the
When it comes to deployment, you probably want the extra performance of
C bindings, so you may want to get that working on your deployment
even if you don’t bother for your development platform. Then again, the
only good reason to optimize (and the C bindings are an optimization,
no mistake) is because you have a performance bottleneck; if it ain’t
broke, don’t fix it, and the only way to tell is by monitoring real-life
(or simulated) performance.