Re: per user/database connections limit again

Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> I have "worked over" your patch and I think it is ready for application.

I've made another pass over this and should be able to commit tomorrow
(I'm about to knock off for today, and ran out of time to test
pg_dumpall). One thing I changed was that it didn't make sense to me
for CREATE DATABASE to copy the template database's datconnlimit.
We don't copy its datallowconn or datconfig, so why datconnlimit?

BTW I disagree with removing datallowconn; that is different from
datconnlimit = 0 because it is enforced even against superusers.
(Similar remarks apply for rolcanlogin vs rolconnlimit.)

regards, tom lane