On Wed, Nov 12, 2014 at 9:49 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> On Tue, Nov 11, 2014 at 7:48 PM, Kouhei Kaigai <kaigai(at)ak(dot)jp(dot)nec(dot)com> wrote:
>> Isn't provolatile = PROVOLATILE_IMMUTABLE sufficient?
>
> There are certainly things that are parallel-safe that are not
> immutable. It might be the case that everything immutable is
> parallel-safe.
FWIW, when working on the concept of expression and clause
shippability for Postgres-XC (aka the possibility to pass it safely to
another backend, but in another PG node in this case), we discussed
similar things and if I recall correctly we even discussed about
adding a flag to pg_proc to define if a function was shippable or not.
Finally what we finished with was not adding a new flag and use as
rule that all the immutable functions can be safely shipped, and
others not, even some stable functions that *could* be safe. Maybe
Ashutosh has more comments on that, my memory may be failing.
In the end, I think that you would finish with something similar.
My 2c.
--
Michael