Andrew Dunstan <andrew@dunslane.net> writes:
> On 2022-08-23 Tu 15:32, Jonathan S. Katz wrote:
>> On 8/23/22 1:26 PM, Andres Freund wrote:
>>> We could decide to revert this for 15, but leave it in tree for HEAD.
>> If it comes to that, I think that is a reasonable suggestion so long
>> as we're committed to making the requisite changes.
I'm not particularly on board with that. In the first place, I'm
unconvinced that very much of the current code will survive, and
I don't want people contorting the rewrite in order to salvage
committed code that would be better off junked. In the second
place, if we still don't have a shippable feature in a year, then
undoing it again is going to be just that much harder.
> One good reason for this is that way we're not fighting against the node
> changes, which complicate any reversion significantly.
Having said that, I'm prepared to believe that a lot of the node
infrastructure won't change because it's dictated by the SQL-spec
grammar. So we could leave that part alone in HEAD; at worst
it adds some dead code in backend/nodes.
regards, tom lane