On Fri, Sep 23, 2011 at 12:37 PM, Josh Berkus <josh@agliodbs.com> wrote:
>> I'm OK with the proposed behavior change and I agree that it's
>> probably what people want, but I am awfully suspicious that those
>> extra casts are going to break something you haven't thought about.
>> It might be worth posting a rough version first just to see if I (or
>> someone else) can break it before you spend a lot of time on it.
>
> Additional breakage confirmed (hash functions, etc.) Looks like I need
> to add a lot more support functions and test. This is still worth
> doing, but don't expect it for the next commitfest.
I would also be looking carefully at whether you can construct a
scenario where the operator resolution code can't decide between
=(text,citext) or =(text,text) - you probably need a third type like
varchar or bpchar in the mix to trigger a failure, if there's one to
be found. Or you might have a problem with citext = bpchar being
ambiguous between =(text,citext) and =(varchar,text), or some such
thing.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company