On Fri, Apr 05, 2019 at 09:10:31AM -0400, Andrew Dunstan wrote:
> Well, that would be a bit sad. ISTM if we conclude that the current
> behaviour is a bug we could commit the current patch and backpatch a
> fix to honor a lower toast_tuple_threshold. But yes, time is tight.
48 hours remain, which is very tight. Let's see but the chances are
small :(
If we think that lowering toast_tuple_threshold should be supported
then the patch on the other thread should be used first, perhaps
back-patched (it lacks pieces with ALTER TABLE as pointed out as
well). If we don't use the other patch, then what's proposed on this
thread is actually wrong and should be reworked. In any case,
something is wrong.
--
Michael