On Fri, Jan 12, 2024 at 12:28 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> After sleeping on it, I'm inclined to the opinion that unconditionally
> materializing newslot here is a good idea.
I'm glad you came to that conclusion, and I agree. I think it's really
bad that we're passing around un-materialized slots across many layers
of the call stack; having the behavior of the inner code know in
detail which outer code paths materialize the slot and which don't
seems like a recipe for infinite future bugs. The only concern is
overhead, IMV.
> 75e03eabe
> should be sufficient evidence that there's a lot of hazard here
> (and I especially like being able to revert that wart).
I haven't analyzed whether this allow reverting that, but it seems
nice if it does.
Do you want to take this forward from here?
--
Robert Haas
EDB: http://www.enterprisedb.com