Re: ANY_VALUE aggregate

On 09.02.23 10:42, Vik Fearing wrote:
> v4 attached.  I am putting this back to Needs Review in the commitfest 
> app, but these changes were editorial so it is probably RfC like Peter 
> had set it.  I will let you be the judge of that.

I have committed this.

I made a few small last-minute tweaks:

- Changed "non-deterministic" to "arbitrary", as suggested by Maciek 
Sakrejda nearby.  This seemed like a handier and less jargony term.

- Removed trailing whitespace in misc.c.

- Changed the function description in pg_proc.dat.  Apparently, we are 
using 'aggregate transition function' there for all aggregate functions 
(instead of 'any_value transition function' etc.).

- Made the tests a bit more interested by feeding in more rows and a mix 
of null and nonnull values.