Pedro Gimeno <pgsql-004(at)personal(dot)formauri(dot)es> writes:
> Tom Lane wrote, On 2015-01-21 03:23:
>> Done.
> Thank you. As I said I consider that a poor solution in the long term
> (though probably a necessary one for current branches, for the reasons
> you stated). Would a patch in this area have any chance?
Personally I see no reason whatsoever to replace random(), and multiple
reasons not to. If you want a random-number generator with different
properties from what libc provides, write an extension.
regards, tom lane