Bruce Momjian <bruce@momjian.us> writes:
> I have created the attached patch to mention
> suppress_redundant_updates_trigger() in this case. I don't think having
> an actual example is warranted.
I don't like this patch, because in fact
suppress_redundant_updates_trigger is entirely unrelated to the stated
purpose of that example (namely, to log something if the table was
changed). The way you've written it makes it sound like
suppress_redundant_updates_trigger could be used as an alternative
implementation of that requirement.
If you want to mention suppress_redundant_updates_trigger somewhere
in this area, that's fine, but it should be treated as an independent
topic rather than being wedged into the middle of an unrelated example.
Maybe a distinct para saying something like "To suppress no-op updates
of a table, see suppress_redundant_updates_trigger." (and making that
an actual link would be a good idea).
regards, tom lane