On Sat, Feb 26, 2011 at 4:19 PM, Josh Berkus <josh@agliodbs.com> wrote:
> Anyway, if it's ASCII-only, that's a guaranteed way to make sure it
> isn't taken seriously.
Pre-9.1 levenshtein is ASCII-only, and I think some of the other stuff
in contrib/fuzzystrmatch still is. We had to work pretty hard to
avoid a massive performance loss when we made it multi-byte aware, and
IIRC there is still a pretty significant performance loss if any
multibyte characters are actually present. But at least now it
doesn't return totally bogus answers.
So I have some sympathy with the OP's desire not to burden himself
with the non-ASCII case if he doesn't need it for his application, but
I also agree with your point that we probably wouldn't accept code
into contrib that doesn't.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company