Re: [PATCH] bms_prev_member() can read beyond the end of the array of allocated words

Greg Burd <greg@burd.me> writes:
> Spoiler alert, I have larger plans[2] for Bitmapset which is why I
> started down this road to begin with as I wanted to capture the existing
> API/contract for it before I proposed changes to it.

FWIW, I'd be seriously resistant to simply replacing Bitmapset with
something like that, as I think the present implementation is about
right for the planner's uses and complicating it would just slow
things down.  There may be specific places where a compressed
implementation could win, though.

Anyway, that's off-topic for the present thread.  I believe it's
middle-of-the-night in Rowley's time zone, so I was waiting for
further comment from him before taking any action.

            regards, tom lane