Rod Taylor <rbt@rbt.ca> writes:
> I'd like ALTER SEQUENCE to be transaction safe.
I think that's inherently impossible without breaking the existing
behavior of setval/nextval, which is something we will not accept.
ALTER SEQUENCE would be better thought of as a form of setval with
even more parameters, but not fundamentally different semantics.
> Can I assume that
> a standard simple_heap_update() is valid against the sequence,
You can be certain that it is not. There can be only one tuple in a
sequence table.
regards, tom lane