On Fri, 2012-11-09 at 10:18 -0500, Robert Haas wrote:
> Sure, I agree. I don't think it should stay that way forever, but
> removing the burden of dealing with this issue from the initial commit
> would likely allow that commit to happen this release cycle, perhaps
> even in the next CommitFest. And then we'd have half a loaf, which is
> better than none, and we could deal with the issues of switching it on
> and off as a further enhancement.
Just after sending the last email, I realized that it can be separated
into separate commits fairly naturally, I think. So, I agree with you
that we should focus on an initdb setting for the next commitfest and
try for at least an offline migration tool (if not online) later.
Regards,Jeff Davis