"Kevin Grittner" <Kevin.Grittner@wicourts.gov> writes:
> Robert Haas <robertmhaas@gmail.com> wrote:
>> This patch makes me a little nervous, because the existing
>> behavior seems to have been coded for quite deliberately.
> It does, although I'm not clear *why* it was. I suspect it may have
> been based on an assumption that whatever value is in the reset_val
> field had to have been already determined to be good, so it was a
> waste of cycles to check it again -- without considering that the
> validity of making a change might depend on context.
Yes, I'm inclined to think the same, although obviously we need to
review the patch carefully. The GUC code is a bit ticklish.
The main thing I would be worried about is whether you're sure that
you have separated the RESET-as-a-command case from the cases where
we actually are rolling back to a previous state.
regards, tom lane