Re: [HACKERS] Arrays of domains

Andrew Dunstan <andrew.dunstan@2ndquadrant.com> writes:
> On 08/11/2017 01:17 PM, Tom Lane wrote:
>> Attached is a patch series that allows us to create arrays of domain
>> types.

> I've reviewed and tested the updated versions of these patches. The
> patches apply but there's an apparent typo in arrayfuncs.c -
> DatumGetAnyArray instead of DatumGetAnyArrayP

Thanks for reviewing!  The DatumGetAnyArrayP thing is another artifact
of 4bd199465 --- sorry for missing that.

> Some of the line breaking in argument lists for some of the code
> affected by these patches is a bit bizarre. It hasn't been made worse by
> these patches but it hasn't been made better either. That's especially
> true of patch 1.

Yeah, perhaps.  A lot of these argument lists are long enough that I'm
not especially thrilled with the idea of making them one-arg-per-line;
that seems like it would consume a lot of vertical space and make it
harder to see context in a finite-size window.  I think there's been
some attempt at grouping the arguments into related groups on single
lines, though I concede it's probably not very obvious nor 100%
consistent.

> I wonder if we need to do any benchmarking to assure ourselves that the
> changes to ArrayCoerceExpr don't have a significant performance impact?

That would likely be a good idea, though I'm not very sure what or
how to benchmark.
        regards, tom lane


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers