Re: BUG #17872: Dropping an attribute of a composite type breaks indexes over the type silently

Alexander Lakhin <exclusion(at)gmail(dot)com> writes:
> Yes, I think something like that can resolve the issue.
> But I would also note that the problem is not with indexes only, but also
> with "... partition by list(comp_type_value)", for example.

Hmm ... really? I'd just concluded that a partitioned table is okay
as long as it doesn't yet have any partitions. Even if the modified
type is a partitioning column, there's no structure yet that could
depend on the contents of the type. (If it does have partitions,
we'll fail when we get to one of those.)

[ thinks some more... ] I guess there's the corner case where we
replace, say, a hashable type with a non-hashable one and thereby
break decisions about whether PARTITION BY HASH is allowable.
That's kind of a stretch. But find_composite_type_dependencies
currently rejects partitioned tables, so we're not taking away any
functionality if we continue to forbid that.

regards, tom lane