On Wed, Feb 14, 2018 at 2:19 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Andres Freund <andres@anarazel.de> writes: > On 2018-02-13 09:13:09 -0800, Shay Rojansky wrote: >> Was wondering if anyone has a reaction to my email below about statement >> preparation, was it too long? :)
> Well, the issue is that implementing this is a major piece of work. This > post doesn't offer either resources nor a simpler way to do so. There's > no huge debate about the benefit of having a global plan cache, so I'm > not that surprised there's not a huge debate about a post arguing that > we should have one.
Actually, I'm pretty darn skeptical about the value of such a cache for most use-cases. But as long as it can be turned off and doesn't leave residual overhead nor massive added code cruft, I won't stand in the way of someone else investing their time in it.
In any case, as you say, it's moot until somebody steps up to do it.