On Thu, Mar 11, 2021 at 8:27 AM Tom Lane <[email protected]> wrote:
> Strongly agree.  I hadn't realized that the patch actually *removed*
> the index reloption, rather than just turning it into a no-op.  I think
> that's a complete nonstarter.  It will break pg_upgrade scenarios where
> users have indexes with that option set.  It's an especially bad idea
> to have done it in a stable branch, but I doubt we can get away with
> it even in HEAD.

But I haven't removed the index reloption (or the GUC of the same
name) on the 13 branch. I just made it a no-op there, and removed the
documentation.

Evidently there is a consensus that the reloption should not have been
removed on HEAD. I'll discuss that on the other -committers thread
now.

--
Peter Geoghegan


Reply via email to