Brent Verner <brent@rcfile.org> writes:
> Thanks for fixing that problem. Out of curiousity, did you put that
> code in the renameatt() function for any reason, or is it just a
> style thing?
Just to avoid closing and reopening the target relation and pg_attribute
relation. Releasing and then reacquiring the lock on pg_attribute
seemed like a not-so-good idea, although it probably shouldn't make any
difference.
regards, tom lane