On Tue, Aug 10, 2021 at 09:31:37AM +0200, Michael Meskes wrote:
>> that it could be a good thing. declare.pgc seems to rely on that
>> already but the tests are incorrect as I mentioned in [2]. For
>> DESCRIBE, that provides data about a result set, I find the
>> assignment
>> of a connection a bit strange, and even if this would allow the use
>> of
>> the same statement name for multiple connections, it seems to me that
>> there is a risk of breaking existing applications. There should not
>> be that many, so perhaps that's fine anyway.
>
> I don't think we'd break anything given that DECLARE STATEMENT is new.
Sure, DECLARE does not matter as it is new. However, please note that
the specific point I was trying to make with my link [2] from upthread
is related to the use of cached connection names with DEALLOCATE, as
of this line in the new test declare.pgc:
EXEC SQL DEALLOCATE PREPARE stmt_2;
And DEALLOCATE is far from being new.
> Also please keep in mind that you can use EXEC SQL AT ... DESCRIBE ...;
> already anyway. Again, not very meaningful but why should we accept a
> connection one way but not the other?
No objections to that.
--
Michael