Thomas Munro <[email protected]> writes:
> On Tue, Mar 16, 2021 at 3:28 AM Tom Lane <[email protected]> wrote:
>> Usually, if we issue a SET in the regression tests, we explicitly RESET
>> as soon thereafter as practical, so as to have a well-defined scope
>> where the script is running under unusual conditions.
> Oh, of course. Thanks.
> I was wrong to blame that commit, and there are many other tests that
> fail in the back branches. But since we were down to just one, I went
> ahead and fixed this in the master branch only.
Makes sense to me. Committed patch looks good.
regards, tom lane
