"David G. Johnston" <david.g.johnston@gmail.com> writes:
> On Mon, Nov 13, 2023 at 12:36 PM Laurenz Albe <laurenz.albe@cybertec.at>
> wrote:
>> On Mon, 2023-11-13 at 11:27 +0100, Erik Wienhold wrote:
>>> On 2023-11-09 20:19 +0100, Tom Lane wrote:
>>>> So, just to clarify, we're settling on your v4 from [1]?
>>> Yes from my side.
>> +1
> +0.5 for the reasons already stated; but I get and accept the argument for
> NULL.
Patch pushed with minor adjustments, mainly rewriting some comments.
One notable change is that I dropped the newline whitespace printed
by printACLColumn. That was contrary to the policy expressed in the
function's comment, and IMO it made -E output look worse not better.
The problem is that the calling code determines the indentation
that this targetlist item should have, and we don't want to outdent
from that. I think it's better to make it one line, even though
that will run a bit over 80 columns.
I also got rid of the use of a created superuser in the test case.
The test seems pretty duplicative to me anyway, so let's just not
test the object types that need superuser.
> I will reiterate my preference for writing an explicit IS NULL branch in
> the case expression instead of relying upon the strict-ness of
> array_to_string.
Meh. We were relying on that already, and it wasn't a problem.
I might have done it, except that it'd have made the one line
even longer and harder to read (and slower to execute, probably).
regards, tom lane