Hallo Peter,
>>>> Would it make sense to accumulate in the other direction, older to newer,
>>>> so that new attributes are added at the end of the select?
>>> I think that could make sense, but the current style was introduced by
>>> daa9fe8a5264a3f192efa5ddee8fb011ad9da365. Should we revise that?
>> It seems to me more logical to do it while you're at it, but you are the
>> one writing the patches:-)
>
> What do you think about this patch to reorganize the existing code from that
> old commit?
I think it is a definite further improvement.
Patch applies cleanly, compiles, global & pg_dump tap test ok, looks ok to
me.
--
Fabien.