>> I don't think that's a bug. You may not intend ever to cluster on that
>> index again, and if you try it will tell you about the problem.
>
>
> Except it breaks the 'cluster everything' case:
>
> test=# cluster;
> ERROR: cannot cluster when index access method does not handle null values
> HINT: You may be able to work around this by marking column "a" NOT NULL.
Any further comments on this? I still think it's a bug, ie. an
invariant that's not being maintained...
Chris