Kyotaro HORIGUCHI wrote:
> 1. some messages are missing partitioned table/index
> The first attached. I'm not sure how the orering ought to be
> but I arranged them in mainly in the appearance order in if()
> conditions, or the order of case label in switch()
> construct. One exception is ATExecChangeOwner, in which the
> order is the others regardless of the case label order just
> This is backpatchable to 10 but 10 doesn't have partitioned
> index so needs a different fix (Second attached).
> # But, I'm not sure if the list may be that long...
I *think* the idea here is that a partitioned table is a table, so there
is no need to say "foo is not a table or partitioned table". We only
mention partitioned tables when we want to make a distinction between
those that are partitioned and those that aren't.
Same with indexes.
> 2. GUC comment for autovacuum_work_mem has a typo, "max num of
> parallel workers *than* can be..". (Third attached)
Yeah, pushed, though it's parallel_max_workers not autovacuum_work_mem.
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services