On 2018/11/08 0:04, Alvaro Herrera wrote:
>> On Wed, Nov 07, 2018 at 03:34:38PM +0900, Amit Langote wrote:
>>> I think we can design the interface of partition_bounds_create such that
>>> it returns information needed to re-arrange OIDs to be in the canonical
>>> partition bound order, so the actual re-ordering of OIDs can be done by
>>> the caller itself. (It may not always be OIDs, could be something else
>>> like a struct to represent fake partitions.)
>
> This is interesting. I don't think the current interface is so bad that
> we can't make a few tweaks later; IOW let's focus on the current patch
> for now, and we can improve later. You (and David, and maybe someone
> else) already have half a dozen patches touching this code, and I bet
> some of these will have to be rebased over this one.
The patch on ATTACH/DETACH concurrently thread is perhaps touching close
to here, but I'm not sure if any of it should be concerned about how we
generate the PartitionBoundInfo which the patch here trying to make into
its own function?
Thanks,
Amit