Re: Planner hints in Postgresql

On 03/17/2014 11:15 PM, Josh Berkus wrote:
> 2. they eventually become a barrier to upgrading, once the performance
> of the DB engine changes in a way that makes older query decorators
> crippling and/or erroneous.  Because they are scattered all around the
> code, it then becomes a major refactoring effort to fix them.

My idea of associating them with major versions and issuing warnings
would help with this.

> 5. Application developers will add them without telling the DBAs, and
> vice-versa.  Hilarity ensues.

And this would be unmasked at next major upgrade when the logs fill up.

-- 
Vik