Re: tweak to a few index tests to hits ambuildempty() routine.

I still wonder, if assert doesn't catch why that place is marked as
covered here?
https://coverage.postgresql.org/src/backend/access/gin/ginvacuum.c.gcov.html

a(dot)kozhemyakin(at)postgrespro(dot)ru писал 2022-09-12 15:47:
> Hi,
>
> The commit 4fb5c794e5 eliminates the ginbulkdelete() test coverage
> provided by the commit 4c51a2d1e4 two years ago.
> With the following Assert added:
> @@ -571,7 +571,7 @@ ginbulkdelete(IndexVacuumInfo *info,
> IndexBulkDeleteResult *stats,
> Buffer buffer;
> BlockNumber rootOfPostingTree[BLCKSZ / (sizeof(IndexTupleData)
> + sizeof(ItemId))];
> uint32 nRoot;
> -
> + Assert(0);
> gvs.tmpCxt = AllocSetContextCreate(CurrentMemoryContext,
>
> "Gin vacuum temporary context",
>
> ALLOCSET_DEFAULT_SIZES);
> I have check-world passed successfully.
>
> Amul Sul писал 2021-11-29 12:04:
>> Hi,
>>
>> Attached patch is doing small changes to brin, gin & gist index tests
>> to use an unlogged table without changing the original intention of
>> those tests and that is able to hit ambuildempty() routing which is
>> otherwise not reachable by the current tests.